commit 4442389b0020e4446a406b39e584c61613acbf38 Author: ModelHub XC Date: Sat May 23 08:16:13 2026 +0800 初始化项目,由ModelHub XC社区提供模型 Model: laion/exp_tas_timeout_multiplier_1_0_traces Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..e90a0f8 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,56 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text + + +*.bz2 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 +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack 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 +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +saved_model/**/* 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 +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zstandard filter=lfs diff=lfs merge=lfs -text +*.tfevents* filter=lfs diff=lfs merge=lfs -text +*.db* filter=lfs diff=lfs merge=lfs -text +*.ark* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text + +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.gguf* filter=lfs diff=lfs merge=lfs -text +*.ggml filter=lfs diff=lfs merge=lfs -text +*.llamafile* filter=lfs diff=lfs merge=lfs -text +*.pt2 filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text + +model-00003-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +model-00002-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text +vocab.json filter=lfs diff=lfs merge=lfs -text +training_args.bin filter=lfs diff=lfs merge=lfs -text +merges.txt filter=lfs diff=lfs merge=lfs -text +model-00001-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +model-00004-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text \ No newline at end of file diff --git a/README.md b/README.md new file mode 100644 index 0000000..e049d15 --- /dev/null +++ b/README.md @@ -0,0 +1,61 @@ +--- +library_name: transformers +license: apache-2.0 +base_model: Qwen/Qwen3-8B +tags: +- llama-factory +- full +- generated_from_trainer +model-index: +- name: exp_tas_timeout_multiplier_1_0_traces + results: [] +--- + + + +# exp_tas_timeout_multiplier_1_0_traces + +This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on the DCAgent/exp_tas_timeout_multiplier_1.0_traces dataset. + +## Model description + +More information needed + +## Intended uses & limitations + +More information needed + +## Training and evaluation data + +More information needed + +## Training procedure + +### Training hyperparameters + +The following hyperparameters were used during training: +- learning_rate: 4e-05 +- train_batch_size: 1 +- eval_batch_size: 8 +- seed: 42 +- distributed_type: multi-GPU +- num_devices: 8 +- gradient_accumulation_steps: 2 +- total_train_batch_size: 16 +- total_eval_batch_size: 64 +- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.98) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments +- lr_scheduler_type: cosine +- lr_scheduler_warmup_ratio: 0.1 +- num_epochs: 7.0 + +### Training results + + + +### Framework versions + +- Transformers 4.57.3 +- Pytorch 2.9.0+cu128 +- Datasets 4.4.1 +- Tokenizers 0.22.2 diff --git a/added_tokens.json b/added_tokens.json new file mode 100644 index 0000000..b54f913 --- /dev/null +++ b/added_tokens.json @@ -0,0 +1,28 @@ +{ + "": 151668, + "": 151658, + "": 151666, + "": 151667, + "": 151657, + "": 151665, + "<|box_end|>": 151649, + "<|box_start|>": 151648, + "<|endoftext|>": 151643, + "<|file_sep|>": 151664, + "<|fim_middle|>": 151660, + "<|fim_pad|>": 151662, + "<|fim_prefix|>": 151659, + "<|fim_suffix|>": 151661, + "<|im_end|>": 151645, + "<|im_start|>": 151644, + "<|image_pad|>": 151655, + "<|object_ref_end|>": 151647, + "<|object_ref_start|>": 151646, + "<|quad_end|>": 151651, + "<|quad_start|>": 151650, + "<|repo_name|>": 151663, + "<|video_pad|>": 151656, + "<|vision_end|>": 151653, + "<|vision_pad|>": 151654, + "<|vision_start|>": 151652 +} diff --git a/all_results.json b/all_results.json new file mode 100644 index 0000000..997573a --- /dev/null +++ b/all_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 4.491016057964447, + "achieved_tflops_per_gpu_theoretical": 504.797779860652, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18683922290802002, + "mfu_percent": 0.4540966691571736, + "mfu_percent_theoretical": 51.041231532927405, + "total_flos": 1.0544849440651346e+18, + "train_loss": 0.4722494138848214, + "train_runtime": 29349.8434, + "train_samples_per_second": 2.381, + "train_steps_per_second": 0.149, + "valid_targets_mean": 3776.8, + "valid_targets_min": 2389 +} \ No newline at end of file diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000..01be9b3 --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,89 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0].role == 'system' %} + {{- messages[0].content + '\n\n' }} + {%- endif %} + {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0].role == 'system' %} + {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} +{%- for message in messages[::-1] %} + {%- set index = (messages|length - 1) - loop.index0 %} + {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('') and message.content.endswith('')) %} + {%- set ns.multi_step_tool = false %} + {%- set ns.last_query_index = index %} + {%- endif %} +{%- endfor %} +{%- for message in messages %} + {%- if message.content is string %} + {%- set content = message.content %} + {%- else %} + {%- set content = '' %} + {%- endif %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) %} + {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {%- set reasoning_content = '' %} + {%- if message.reasoning_content is string %} + {%- set reasoning_content = message.reasoning_content %} + {%- else %} + {%- if '' in content %} + {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} + {%- set content = content.split('')[-1].lstrip('\n') %} + {%- endif %} + {%- endif %} + {%- if loop.index0 > ns.last_query_index %} + {%- if loop.last or (not loop.last and reasoning_content) %} + {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- else %} + {{- '<|im_start|>' + message.role + '\n' + content }} + {%- endif %} + {%- if message.tool_calls %} + {%- for tool_call in message.tool_calls %} + {%- if (loop.first and content) or (not loop.first) %} + {{- '\n' }} + {%- endif %} + {%- if tool_call.function %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {%- if tool_call.arguments is string %} + {{- tool_call.arguments }} + {%- else %} + {{- tool_call.arguments | tojson }} + {%- endif %} + {{- '}\n' }} + {%- endfor %} + {%- endif %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} + {%- if enable_thinking is defined and enable_thinking is false %} + {{- '\n\n\n\n' }} + {%- endif %} +{%- endif %} \ No newline at end of file diff --git a/config.json b/config.json new file mode 100644 index 0000000..f8a98d4 --- /dev/null +++ b/config.json @@ -0,0 +1,68 @@ +{ + "architectures": [ + "Qwen3ForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "dtype": "bfloat16", + "eos_token_id": 151645, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 12288, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 40960, + "max_window_layers": 36, + "model_type": "qwen3", + "num_attention_heads": 32, + "num_hidden_layers": 36, + "num_key_value_heads": 8, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_scaling": null, + "rope_theta": 1000000, + "sliding_window": null, + "tie_word_embeddings": false, + "transformers_version": "4.57.3", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/configuration.json b/configuration.json new file mode 100644 index 0000000..159097f --- /dev/null +++ b/configuration.json @@ -0,0 +1 @@ +{"framework": "pytorch", "task": "others", "allow_remote": true} \ No newline at end of file diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..9047f32 --- /dev/null +++ b/generation_config.json @@ -0,0 +1,12 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "temperature": 0.6, + "top_k": 20, + "top_p": 0.95, + "transformers_version": "4.57.3" +} diff --git a/merges.txt b/merges.txt new file mode 100644 index 0000000..80c1a19 --- /dev/null +++ b/merges.txt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5 +size 1671853 diff --git a/model-00001-of-00004.safetensors b/model-00001-of-00004.safetensors new file mode 100644 index 0000000..f590953 --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ef4837fbcbdcf285e83be8028a7361d814f4f10e141b2578daef79faf78729a +size 4902257696 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..eb3c25e --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a2f5595da2900bda971378598d0ec1166b6fd7afcf602ab5f5828ea0d0d0370 +size 4915960368 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..fc5b068 --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:32bfdf43cb869137dbff43e0df37e6ff01ab9afc63c048cc6897e206651305c3 +size 4983068496 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..3257e4d --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5d0ac799e6ce34cdb22b2c7c017cf920b6176b06afee53e762b3666974a2892d +size 1580230264 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000..ba886c0 --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,407 @@ +{ + "metadata": { + "total_parameters": 308224, + "total_size": 16381470720 + }, + "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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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-00002-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_norm.weight": "model-00002-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_norm.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-00002-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-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_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.q_norm.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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-00003-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-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-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_norm.weight": "model-00003-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_norm.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.32.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.k_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.q_norm.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.35.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.gate_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.35.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_norm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.35.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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_norm.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-00001-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_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_norm.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/run_summary.json b/run_summary.json new file mode 100644 index 0000000..3ec7f95 --- /dev/null +++ b/run_summary.json @@ -0,0 +1,12 @@ +{ + "agent_name": "terminus-2", + "training_start": null, + "training_end": null, + "created_by": "DCAgent", + "base_model_name": "Qwen/Qwen3-8B", + "dataset_name": "DCAgent/exp_tas_timeout_multiplier_1.0_traces", + "training_type": "SFT", + "training_parameters": "https://huggingface.co/laion/exp_tas_timeout_multiplier_1_0_traces/blob/main/config.json", + "wandb_link": "https://wandb.ai/dogml/OpenThoughts-Agent/runs/sft_exp_tas_timeout_multiplier_1-0_traces_Qwen3-8B", + "traces_location_s3": null +} \ No newline at end of file diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000..ac23c0a --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,31 @@ +{ + "additional_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "eos_token": { + "content": "<|im_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..cd71f61 --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4 +size 11422654 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..e9dc937 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,240 @@ +{ + "add_bos_token": false, + "add_prefix_space": false, + "added_tokens_decoder": { + "151643": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151644": { + "content": "<|im_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151645": { + "content": "<|im_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151646": { + "content": "<|object_ref_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151647": { + "content": "<|object_ref_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151648": { + "content": "<|box_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151649": { + "content": "<|box_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151650": { + "content": "<|quad_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151651": { + "content": "<|quad_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151652": { + "content": "<|vision_start|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151653": { + "content": "<|vision_end|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151654": { + "content": "<|vision_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151655": { + "content": "<|image_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151656": { + "content": "<|video_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151657": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151658": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151659": { + "content": "<|fim_prefix|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151660": { + "content": "<|fim_middle|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151661": { + "content": "<|fim_suffix|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151662": { + "content": "<|fim_pad|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151663": { + "content": "<|repo_name|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151664": { + "content": "<|file_sep|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151665": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151666": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151667": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151668": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": {}, + "model_max_length": 32768, + "pad_token": "<|endoftext|>", + "padding_side": "right", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/train_results.json b/train_results.json new file mode 100644 index 0000000..997573a --- /dev/null +++ b/train_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 4.491016057964447, + "achieved_tflops_per_gpu_theoretical": 504.797779860652, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18683922290802002, + "mfu_percent": 0.4540966691571736, + "mfu_percent_theoretical": 51.041231532927405, + "total_flos": 1.0544849440651346e+18, + "train_loss": 0.4722494138848214, + "train_runtime": 29349.8434, + "train_samples_per_second": 2.381, + "train_steps_per_second": 0.149, + "valid_targets_mean": 3776.8, + "valid_targets_min": 2389 +} \ No newline at end of file diff --git a/trainer_log.jsonl b/trainer_log.jsonl new file mode 100644 index 0000000..2f12d46 --- /dev/null +++ b/trainer_log.jsonl @@ -0,0 +1,874 @@ +{"current_steps": 5, "total_steps": 4368, "loss": 0.9655, "lr": 3.661327231121282e-07, "epoch": 0.008012820512820512, "percentage": 0.11, "elapsed_time": "0:00:43", "remaining_time": "10:33:09"} +{"current_steps": 10, "total_steps": 4368, "loss": 0.9567, "lr": 8.237986270022884e-07, "epoch": 0.016025641025641024, "percentage": 0.23, "elapsed_time": "0:01:19", "remaining_time": "9:35:01"} +{"current_steps": 15, "total_steps": 4368, "loss": 0.9797, "lr": 1.2814645308924487e-06, "epoch": 0.02403846153846154, "percentage": 0.34, "elapsed_time": "0:01:56", "remaining_time": "9:24:31"} +{"current_steps": 20, "total_steps": 4368, "loss": 0.9819, "lr": 1.7391304347826088e-06, "epoch": 0.03205128205128205, "percentage": 0.46, "elapsed_time": "0:02:34", "remaining_time": "9:21:28"} +{"current_steps": 25, "total_steps": 4368, "loss": 0.943, "lr": 2.196796338672769e-06, "epoch": 0.04006410256410257, "percentage": 0.57, "elapsed_time": "0:03:06", "remaining_time": "8:58:40"} +{"current_steps": 30, "total_steps": 4368, "loss": 0.894, "lr": 2.654462242562929e-06, "epoch": 0.04807692307692308, "percentage": 0.69, "elapsed_time": "0:03:37", "remaining_time": "8:45:05"} +{"current_steps": 35, "total_steps": 4368, "loss": 0.898, "lr": 3.1121281464530894e-06, "epoch": 0.05608974358974359, "percentage": 0.8, "elapsed_time": "0:04:14", "remaining_time": "8:44:15"} +{"current_steps": 40, "total_steps": 4368, "loss": 0.8348, "lr": 3.56979405034325e-06, "epoch": 0.0641025641025641, "percentage": 0.92, "elapsed_time": "0:04:48", "remaining_time": "8:40:55"} +{"current_steps": 45, "total_steps": 4368, "loss": 0.8442, "lr": 4.0274599542334094e-06, "epoch": 0.07211538461538461, "percentage": 1.03, "elapsed_time": "0:05:36", "remaining_time": "8:58:35"} +{"current_steps": 50, "total_steps": 4368, "loss": 0.8109, "lr": 4.48512585812357e-06, "epoch": 0.08012820512820513, "percentage": 1.14, "elapsed_time": "0:06:09", "remaining_time": "8:51:42"} +{"current_steps": 55, "total_steps": 4368, "loss": 0.8115, "lr": 4.94279176201373e-06, "epoch": 0.08814102564102565, "percentage": 1.26, "elapsed_time": "0:06:49", "remaining_time": "8:55:32"} +{"current_steps": 60, "total_steps": 4368, "loss": 0.7536, "lr": 5.400457665903891e-06, "epoch": 0.09615384615384616, "percentage": 1.37, "elapsed_time": "0:07:29", "remaining_time": "8:57:54"} +{"current_steps": 65, "total_steps": 4368, "loss": 0.7537, "lr": 5.858123569794051e-06, "epoch": 0.10416666666666667, "percentage": 1.49, "elapsed_time": "0:08:05", "remaining_time": "8:55:56"} +{"current_steps": 70, "total_steps": 4368, "loss": 0.7417, "lr": 6.31578947368421e-06, "epoch": 0.11217948717948718, "percentage": 1.6, "elapsed_time": "0:08:45", "remaining_time": "8:57:34"} +{"current_steps": 75, "total_steps": 4368, "loss": 0.7748, "lr": 6.773455377574372e-06, "epoch": 0.1201923076923077, "percentage": 1.72, "elapsed_time": "0:09:24", "remaining_time": "8:58:47"} +{"current_steps": 80, "total_steps": 4368, "loss": 0.7666, "lr": 7.231121281464531e-06, "epoch": 0.1282051282051282, "percentage": 1.83, "elapsed_time": "0:10:03", "remaining_time": "8:59:23"} +{"current_steps": 85, "total_steps": 4368, "loss": 0.6723, "lr": 7.688787185354691e-06, "epoch": 0.1362179487179487, "percentage": 1.95, "elapsed_time": "0:10:46", "remaining_time": "9:02:57"} +{"current_steps": 90, "total_steps": 4368, "loss": 0.7063, "lr": 8.146453089244852e-06, "epoch": 0.14423076923076922, "percentage": 2.06, "elapsed_time": "0:11:19", "remaining_time": "8:58:35"} +{"current_steps": 95, "total_steps": 4368, "loss": 0.7005, "lr": 8.604118993135013e-06, "epoch": 0.15224358974358973, "percentage": 2.17, "elapsed_time": "0:11:50", "remaining_time": "8:52:22"} +{"current_steps": 100, "total_steps": 4368, "loss": 0.6994, "lr": 9.061784897025172e-06, "epoch": 0.16025641025641027, "percentage": 2.29, "elapsed_time": "0:12:23", "remaining_time": "8:49:08"} +{"current_steps": 105, "total_steps": 4368, "loss": 0.7064, "lr": 9.519450800915333e-06, "epoch": 0.16826923076923078, "percentage": 2.4, "elapsed_time": "0:12:58", "remaining_time": "8:46:28"} +{"current_steps": 110, "total_steps": 4368, "loss": 0.6694, "lr": 9.977116704805492e-06, "epoch": 0.1762820512820513, "percentage": 2.52, "elapsed_time": "0:13:42", "remaining_time": "8:50:29"} +{"current_steps": 115, "total_steps": 4368, "loss": 0.6781, "lr": 1.0434782608695653e-05, "epoch": 0.1842948717948718, "percentage": 2.63, "elapsed_time": "0:14:19", "remaining_time": "8:49:29"} +{"current_steps": 120, "total_steps": 4368, "loss": 0.6593, "lr": 1.0892448512585814e-05, "epoch": 0.19230769230769232, "percentage": 2.75, "elapsed_time": "0:14:58", "remaining_time": "8:49:53"} +{"current_steps": 125, "total_steps": 4368, "loss": 0.6411, "lr": 1.1350114416475973e-05, "epoch": 0.20032051282051283, "percentage": 2.86, "elapsed_time": "0:15:32", "remaining_time": "8:47:49"} +{"current_steps": 130, "total_steps": 4368, "loss": 0.6848, "lr": 1.1807780320366134e-05, "epoch": 0.20833333333333334, "percentage": 2.98, "elapsed_time": "0:16:11", "remaining_time": "8:47:46"} +{"current_steps": 135, "total_steps": 4368, "loss": 0.6421, "lr": 1.2265446224256295e-05, "epoch": 0.21634615384615385, "percentage": 3.09, "elapsed_time": "0:16:40", "remaining_time": "8:42:51"} +{"current_steps": 140, "total_steps": 4368, "loss": 0.6348, "lr": 1.2723112128146454e-05, "epoch": 0.22435897435897437, "percentage": 3.21, "elapsed_time": "0:17:10", "remaining_time": "8:38:29"} +{"current_steps": 145, "total_steps": 4368, "loss": 0.683, "lr": 1.3180778032036615e-05, "epoch": 0.23237179487179488, "percentage": 3.32, "elapsed_time": "0:17:42", "remaining_time": "8:35:44"} +{"current_steps": 150, "total_steps": 4368, "loss": 0.656, "lr": 1.3638443935926776e-05, "epoch": 0.2403846153846154, "percentage": 3.43, "elapsed_time": "0:18:25", "remaining_time": "8:38:08"} +{"current_steps": 155, "total_steps": 4368, "loss": 0.6734, "lr": 1.4096109839816933e-05, "epoch": 0.2483974358974359, "percentage": 3.55, "elapsed_time": "0:18:57", "remaining_time": "8:35:29"} +{"current_steps": 160, "total_steps": 4368, "loss": 0.6344, "lr": 1.4553775743707096e-05, "epoch": 0.2564102564102564, "percentage": 3.66, "elapsed_time": "0:19:33", "remaining_time": "8:34:31"} +{"current_steps": 165, "total_steps": 4368, "loss": 0.6547, "lr": 1.5011441647597256e-05, "epoch": 0.2644230769230769, "percentage": 3.78, "elapsed_time": "0:20:07", "remaining_time": "8:32:44"} +{"current_steps": 170, "total_steps": 4368, "loss": 0.6144, "lr": 1.5469107551487414e-05, "epoch": 0.2724358974358974, "percentage": 3.89, "elapsed_time": "0:20:43", "remaining_time": "8:31:37"} +{"current_steps": 175, "total_steps": 4368, "loss": 0.6454, "lr": 1.5926773455377575e-05, "epoch": 0.28044871794871795, "percentage": 4.01, "elapsed_time": "0:21:24", "remaining_time": "8:32:45"} +{"current_steps": 180, "total_steps": 4368, "loss": 0.6317, "lr": 1.6384439359267736e-05, "epoch": 0.28846153846153844, "percentage": 4.12, "elapsed_time": "0:21:52", "remaining_time": "8:28:59"} +{"current_steps": 185, "total_steps": 4368, "loss": 0.6377, "lr": 1.6842105263157896e-05, "epoch": 0.296474358974359, "percentage": 4.24, "elapsed_time": "0:22:31", "remaining_time": "8:29:12"} +{"current_steps": 190, "total_steps": 4368, "loss": 0.6429, "lr": 1.7299771167048057e-05, "epoch": 0.30448717948717946, "percentage": 4.35, "elapsed_time": "0:23:04", "remaining_time": "8:27:33"} +{"current_steps": 195, "total_steps": 4368, "loss": 0.6447, "lr": 1.7757437070938218e-05, "epoch": 0.3125, "percentage": 4.46, "elapsed_time": "0:23:41", "remaining_time": "8:27:05"} +{"current_steps": 200, "total_steps": 4368, "loss": 0.6205, "lr": 1.8215102974828376e-05, "epoch": 0.32051282051282054, "percentage": 4.58, "elapsed_time": "0:24:19", "remaining_time": "8:27:02"} +{"current_steps": 205, "total_steps": 4368, "loss": 0.611, "lr": 1.8672768878718537e-05, "epoch": 0.328525641025641, "percentage": 4.69, "elapsed_time": "0:24:53", "remaining_time": "8:25:35"} +{"current_steps": 210, "total_steps": 4368, "loss": 0.6301, "lr": 1.9130434782608697e-05, "epoch": 0.33653846153846156, "percentage": 4.81, "elapsed_time": "0:25:31", "remaining_time": "8:25:32"} +{"current_steps": 215, "total_steps": 4368, "loss": 0.5953, "lr": 1.9588100686498858e-05, "epoch": 0.34455128205128205, "percentage": 4.92, "elapsed_time": "0:26:07", "remaining_time": "8:24:35"} +{"current_steps": 220, "total_steps": 4368, "loss": 0.656, "lr": 2.004576659038902e-05, "epoch": 0.3525641025641026, "percentage": 5.04, "elapsed_time": "0:26:36", "remaining_time": "8:21:40"} +{"current_steps": 225, "total_steps": 4368, "loss": 0.6255, "lr": 2.050343249427918e-05, "epoch": 0.3605769230769231, "percentage": 5.15, "elapsed_time": "0:27:07", "remaining_time": "8:19:24"} +{"current_steps": 230, "total_steps": 4368, "loss": 0.614, "lr": 2.0961098398169337e-05, "epoch": 0.3685897435897436, "percentage": 5.27, "elapsed_time": "0:27:45", "remaining_time": "8:19:19"} +{"current_steps": 235, "total_steps": 4368, "loss": 0.6089, "lr": 2.14187643020595e-05, "epoch": 0.3766025641025641, "percentage": 5.38, "elapsed_time": "0:28:12", "remaining_time": "8:16:09"} +{"current_steps": 240, "total_steps": 4368, "loss": 0.6354, "lr": 2.187643020594966e-05, "epoch": 0.38461538461538464, "percentage": 5.49, "elapsed_time": "0:28:43", "remaining_time": "8:14:01"} +{"current_steps": 245, "total_steps": 4368, "loss": 0.6026, "lr": 2.2334096109839817e-05, "epoch": 0.3926282051282051, "percentage": 5.61, "elapsed_time": "0:29:22", "remaining_time": "8:14:17"} +{"current_steps": 250, "total_steps": 4368, "loss": 0.6119, "lr": 2.279176201372998e-05, "epoch": 0.40064102564102566, "percentage": 5.72, "elapsed_time": "0:30:06", "remaining_time": "8:15:50"} +{"current_steps": 255, "total_steps": 4368, "loss": 0.5854, "lr": 2.3249427917620138e-05, "epoch": 0.40865384615384615, "percentage": 5.84, "elapsed_time": "0:30:34", "remaining_time": "8:13:15"} +{"current_steps": 260, "total_steps": 4368, "loss": 0.5917, "lr": 2.37070938215103e-05, "epoch": 0.4166666666666667, "percentage": 5.95, "elapsed_time": "0:31:07", "remaining_time": "8:11:42"} +{"current_steps": 265, "total_steps": 4368, "loss": 0.6211, "lr": 2.4164759725400463e-05, "epoch": 0.42467948717948717, "percentage": 6.07, "elapsed_time": "0:31:33", "remaining_time": "8:08:32"} +{"current_steps": 270, "total_steps": 4368, "loss": 0.5775, "lr": 2.462242562929062e-05, "epoch": 0.4326923076923077, "percentage": 6.18, "elapsed_time": "0:32:12", "remaining_time": "8:08:47"} +{"current_steps": 275, "total_steps": 4368, "loss": 0.5995, "lr": 2.508009153318078e-05, "epoch": 0.4407051282051282, "percentage": 6.3, "elapsed_time": "0:32:42", "remaining_time": "8:06:50"} +{"current_steps": 280, "total_steps": 4368, "loss": 0.5977, "lr": 2.5537757437070943e-05, "epoch": 0.44871794871794873, "percentage": 6.41, "elapsed_time": "0:33:14", "remaining_time": "8:05:17"} +{"current_steps": 285, "total_steps": 4368, "loss": 0.6142, "lr": 2.59954233409611e-05, "epoch": 0.4567307692307692, "percentage": 6.52, "elapsed_time": "0:33:58", "remaining_time": "8:06:49"} +{"current_steps": 290, "total_steps": 4368, "loss": 0.5716, "lr": 2.645308924485126e-05, "epoch": 0.46474358974358976, "percentage": 6.64, "elapsed_time": "0:34:32", "remaining_time": "8:05:44"} +{"current_steps": 295, "total_steps": 4368, "loss": 0.5892, "lr": 2.6910755148741422e-05, "epoch": 0.47275641025641024, "percentage": 6.75, "elapsed_time": "0:35:05", "remaining_time": "8:04:25"} +{"current_steps": 300, "total_steps": 4368, "loss": 0.6112, "lr": 2.7368421052631583e-05, "epoch": 0.4807692307692308, "percentage": 6.87, "elapsed_time": "0:35:40", "remaining_time": "8:03:47"} +{"current_steps": 305, "total_steps": 4368, "loss": 0.5587, "lr": 2.782608695652174e-05, "epoch": 0.48878205128205127, "percentage": 6.98, "elapsed_time": "0:36:12", "remaining_time": "8:02:18"} +{"current_steps": 310, "total_steps": 4368, "loss": 0.6255, "lr": 2.8283752860411904e-05, "epoch": 0.4967948717948718, "percentage": 7.1, "elapsed_time": "0:36:44", "remaining_time": "8:00:54"} +{"current_steps": 315, "total_steps": 4368, "loss": 0.6187, "lr": 2.8741418764302062e-05, "epoch": 0.5048076923076923, "percentage": 7.21, "elapsed_time": "0:37:12", "remaining_time": "7:58:48"} +{"current_steps": 320, "total_steps": 4368, "loss": 0.5639, "lr": 2.9199084668192223e-05, "epoch": 0.5128205128205128, "percentage": 7.33, "elapsed_time": "0:37:50", "remaining_time": "7:58:37"} +{"current_steps": 325, "total_steps": 4368, "loss": 0.5994, "lr": 2.9656750572082384e-05, "epoch": 0.5208333333333334, "percentage": 7.44, "elapsed_time": "0:38:22", "remaining_time": "7:57:23"} +{"current_steps": 330, "total_steps": 4368, "loss": 0.5855, "lr": 3.0114416475972544e-05, "epoch": 0.5288461538461539, "percentage": 7.55, "elapsed_time": "0:38:55", "remaining_time": "7:56:15"} +{"current_steps": 335, "total_steps": 4368, "loss": 0.5956, "lr": 3.05720823798627e-05, "epoch": 0.5368589743589743, "percentage": 7.67, "elapsed_time": "0:39:31", "remaining_time": "7:55:48"} +{"current_steps": 340, "total_steps": 4368, "loss": 0.5798, "lr": 3.102974828375286e-05, "epoch": 0.5448717948717948, "percentage": 7.78, "elapsed_time": "0:40:04", "remaining_time": "7:54:45"} +{"current_steps": 345, "total_steps": 4368, "loss": 0.5892, "lr": 3.1487414187643024e-05, "epoch": 0.5528846153846154, "percentage": 7.9, "elapsed_time": "0:40:33", "remaining_time": "7:53:01"} +{"current_steps": 350, "total_steps": 4368, "loss": 0.5648, "lr": 3.1945080091533184e-05, "epoch": 0.5608974358974359, "percentage": 8.01, "elapsed_time": "0:41:14", "remaining_time": "7:53:23"} +{"current_steps": 355, "total_steps": 4368, "loss": 0.567, "lr": 3.240274599542334e-05, "epoch": 0.5689102564102564, "percentage": 8.13, "elapsed_time": "0:41:48", "remaining_time": "7:52:34"} +{"current_steps": 360, "total_steps": 4368, "loss": 0.6071, "lr": 3.2860411899313506e-05, "epoch": 0.5769230769230769, "percentage": 8.24, "elapsed_time": "0:42:23", "remaining_time": "7:51:57"} +{"current_steps": 365, "total_steps": 4368, "loss": 0.6091, "lr": 3.331807780320366e-05, "epoch": 0.5849358974358975, "percentage": 8.36, "elapsed_time": "0:42:51", "remaining_time": "7:50:04"} +{"current_steps": 370, "total_steps": 4368, "loss": 0.5993, "lr": 3.377574370709382e-05, "epoch": 0.592948717948718, "percentage": 8.47, "elapsed_time": "0:43:23", "remaining_time": "7:48:50"} +{"current_steps": 375, "total_steps": 4368, "loss": 0.582, "lr": 3.423340961098399e-05, "epoch": 0.6009615384615384, "percentage": 8.59, "elapsed_time": "0:44:05", "remaining_time": "7:49:29"} +{"current_steps": 380, "total_steps": 4368, "loss": 0.5572, "lr": 3.469107551487414e-05, "epoch": 0.6089743589743589, "percentage": 8.7, "elapsed_time": "0:44:42", "remaining_time": "7:49:09"} +{"current_steps": 385, "total_steps": 4368, "loss": 0.5954, "lr": 3.5148741418764304e-05, "epoch": 0.6169871794871795, "percentage": 8.81, "elapsed_time": "0:45:14", "remaining_time": "7:48:02"} +{"current_steps": 390, "total_steps": 4368, "loss": 0.5986, "lr": 3.5606407322654464e-05, "epoch": 0.625, "percentage": 8.93, "elapsed_time": "0:45:47", "remaining_time": "7:47:00"} +{"current_steps": 395, "total_steps": 4368, "loss": 0.5691, "lr": 3.6064073226544625e-05, "epoch": 0.6330128205128205, "percentage": 9.04, "elapsed_time": "0:46:21", "remaining_time": "7:46:18"} +{"current_steps": 400, "total_steps": 4368, "loss": 0.5678, "lr": 3.6521739130434786e-05, "epoch": 0.6410256410256411, "percentage": 9.16, "elapsed_time": "0:46:59", "remaining_time": "7:46:06"} +{"current_steps": 405, "total_steps": 4368, "loss": 0.5637, "lr": 3.697940503432495e-05, "epoch": 0.6490384615384616, "percentage": 9.27, "elapsed_time": "0:47:33", "remaining_time": "7:45:23"} +{"current_steps": 410, "total_steps": 4368, "loss": 0.5706, "lr": 3.743707093821511e-05, "epoch": 0.657051282051282, "percentage": 9.39, "elapsed_time": "0:48:12", "remaining_time": "7:45:23"} +{"current_steps": 415, "total_steps": 4368, "loss": 0.6069, "lr": 3.789473684210526e-05, "epoch": 0.6650641025641025, "percentage": 9.5, "elapsed_time": "0:48:49", "remaining_time": "7:45:04"} +{"current_steps": 420, "total_steps": 4368, "loss": 0.5823, "lr": 3.835240274599543e-05, "epoch": 0.6730769230769231, "percentage": 9.62, "elapsed_time": "0:49:30", "remaining_time": "7:45:22"} +{"current_steps": 425, "total_steps": 4368, "loss": 0.5661, "lr": 3.8810068649885584e-05, "epoch": 0.6810897435897436, "percentage": 9.73, "elapsed_time": "0:49:58", "remaining_time": "7:43:38"} +{"current_steps": 430, "total_steps": 4368, "loss": 0.5573, "lr": 3.9267734553775745e-05, "epoch": 0.6891025641025641, "percentage": 9.84, "elapsed_time": "0:50:34", "remaining_time": "7:43:10"} +{"current_steps": 435, "total_steps": 4368, "loss": 0.5766, "lr": 3.9725400457665905e-05, "epoch": 0.6971153846153846, "percentage": 9.96, "elapsed_time": "0:51:04", "remaining_time": "7:41:43"} +{"current_steps": 440, "total_steps": 4368, "loss": 0.5625, "lr": 3.999997445219712e-05, "epoch": 0.7051282051282052, "percentage": 10.07, "elapsed_time": "0:51:39", "remaining_time": "7:41:10"} +{"current_steps": 445, "total_steps": 4368, "loss": 0.5622, "lr": 3.999968704016428e-05, "epoch": 0.7131410256410257, "percentage": 10.19, "elapsed_time": "0:52:10", "remaining_time": "7:39:58"} +{"current_steps": 450, "total_steps": 4368, "loss": 0.5867, "lr": 3.9999080285949514e-05, "epoch": 0.7211538461538461, "percentage": 10.3, "elapsed_time": "0:52:42", "remaining_time": "7:38:56"} +{"current_steps": 455, "total_steps": 4368, "loss": 0.545, "lr": 3.999815419924108e-05, "epoch": 0.7291666666666666, "percentage": 10.42, "elapsed_time": "0:53:16", "remaining_time": "7:38:06"} +{"current_steps": 460, "total_steps": 4368, "loss": 0.6004, "lr": 3.999690879482614e-05, "epoch": 0.7371794871794872, "percentage": 10.53, "elapsed_time": "0:53:44", "remaining_time": "7:36:33"} +{"current_steps": 465, "total_steps": 4368, "loss": 0.5831, "lr": 3.9995344092590506e-05, "epoch": 0.7451923076923077, "percentage": 10.65, "elapsed_time": "0:54:22", "remaining_time": "7:36:21"} +{"current_steps": 470, "total_steps": 4368, "loss": 0.5489, "lr": 3.999346011751835e-05, "epoch": 0.7532051282051282, "percentage": 10.76, "elapsed_time": "0:54:49", "remaining_time": "7:34:42"} +{"current_steps": 475, "total_steps": 4368, "loss": 0.5706, "lr": 3.999125689969176e-05, "epoch": 0.7612179487179487, "percentage": 10.87, "elapsed_time": "0:55:25", "remaining_time": "7:34:17"} +{"current_steps": 480, "total_steps": 4368, "loss": 0.5563, "lr": 3.9988734474290324e-05, "epoch": 0.7692307692307693, "percentage": 10.99, "elapsed_time": "0:55:55", "remaining_time": "7:33:01"} +{"current_steps": 485, "total_steps": 4368, "loss": 0.5372, "lr": 3.9985892881590513e-05, "epoch": 0.7772435897435898, "percentage": 11.1, "elapsed_time": "0:56:26", "remaining_time": "7:31:49"} +{"current_steps": 490, "total_steps": 4368, "loss": 0.6105, "lr": 3.9982732166965054e-05, "epoch": 0.7852564102564102, "percentage": 11.22, "elapsed_time": "0:56:59", "remaining_time": "7:31:00"} +{"current_steps": 495, "total_steps": 4368, "loss": 0.6127, "lr": 3.997925238088221e-05, "epoch": 0.7932692307692307, "percentage": 11.33, "elapsed_time": "0:57:32", "remaining_time": "7:30:12"} +{"current_steps": 500, "total_steps": 4368, "loss": 0.5466, "lr": 3.9975453578904975e-05, "epoch": 0.8012820512820513, "percentage": 11.45, "elapsed_time": "0:58:07", "remaining_time": "7:29:37"} +{"current_steps": 505, "total_steps": 4368, "loss": 0.5801, "lr": 3.997133582169018e-05, "epoch": 0.8092948717948718, "percentage": 11.56, "elapsed_time": "0:58:43", "remaining_time": "7:29:14"} +{"current_steps": 510, "total_steps": 4368, "loss": 0.5661, "lr": 3.996689917498754e-05, "epoch": 0.8173076923076923, "percentage": 11.68, "elapsed_time": "0:59:13", "remaining_time": "7:28:03"} +{"current_steps": 515, "total_steps": 4368, "loss": 0.5698, "lr": 3.9962143709638585e-05, "epoch": 0.8253205128205128, "percentage": 11.79, "elapsed_time": "0:59:48", "remaining_time": "7:27:25"} +{"current_steps": 520, "total_steps": 4368, "loss": 0.5813, "lr": 3.995706950157554e-05, "epoch": 0.8333333333333334, "percentage": 11.9, "elapsed_time": "1:00:15", "remaining_time": "7:25:55"} +{"current_steps": 525, "total_steps": 4368, "loss": 0.561, "lr": 3.995167663182008e-05, "epoch": 0.8413461538461539, "percentage": 12.02, "elapsed_time": "1:00:47", "remaining_time": "7:25:00"} +{"current_steps": 530, "total_steps": 4368, "loss": 0.5973, "lr": 3.994596518648214e-05, "epoch": 0.8493589743589743, "percentage": 12.13, "elapsed_time": "1:01:25", "remaining_time": "7:24:47"} +{"current_steps": 535, "total_steps": 4368, "loss": 0.5672, "lr": 3.993993525675838e-05, "epoch": 0.8573717948717948, "percentage": 12.25, "elapsed_time": "1:02:01", "remaining_time": "7:24:23"} +{"current_steps": 540, "total_steps": 4368, "loss": 0.5876, "lr": 3.993358693893086e-05, "epoch": 0.8653846153846154, "percentage": 12.36, "elapsed_time": "1:02:38", "remaining_time": "7:24:05"} +{"current_steps": 545, "total_steps": 4368, "loss": 0.5817, "lr": 3.9926920334365457e-05, "epoch": 0.8733974358974359, "percentage": 12.48, "elapsed_time": "1:03:12", "remaining_time": "7:23:25"} +{"current_steps": 550, "total_steps": 4368, "loss": 0.5339, "lr": 3.991993554951023e-05, "epoch": 0.8814102564102564, "percentage": 12.59, "elapsed_time": "1:03:50", "remaining_time": "7:23:08"} +{"current_steps": 555, "total_steps": 4368, "loss": 0.5681, "lr": 3.991263269589376e-05, "epoch": 0.8894230769230769, "percentage": 12.71, "elapsed_time": "1:04:21", "remaining_time": "7:22:11"} +{"current_steps": 560, "total_steps": 4368, "loss": 0.5348, "lr": 3.990501189012332e-05, "epoch": 0.8974358974358975, "percentage": 12.82, "elapsed_time": "1:05:01", "remaining_time": "7:22:11"} +{"current_steps": 565, "total_steps": 4368, "loss": 0.5521, "lr": 3.989707325388305e-05, "epoch": 0.905448717948718, "percentage": 12.93, "elapsed_time": "1:05:45", "remaining_time": "7:22:36"} +{"current_steps": 570, "total_steps": 4368, "loss": 0.5455, "lr": 3.9888816913932016e-05, "epoch": 0.9134615384615384, "percentage": 13.05, "elapsed_time": "1:06:16", "remaining_time": "7:21:37"} +{"current_steps": 575, "total_steps": 4368, "loss": 0.5792, "lr": 3.988024300210215e-05, "epoch": 0.9214743589743589, "percentage": 13.16, "elapsed_time": "1:06:47", "remaining_time": "7:20:35"} +{"current_steps": 580, "total_steps": 4368, "loss": 0.5451, "lr": 3.987135165529618e-05, "epoch": 0.9294871794871795, "percentage": 13.28, "elapsed_time": "1:07:20", "remaining_time": "7:19:47"} +{"current_steps": 585, "total_steps": 4368, "loss": 0.5535, "lr": 3.9862143015485446e-05, "epoch": 0.9375, "percentage": 13.39, "elapsed_time": "1:07:53", "remaining_time": "7:19:04"} +{"current_steps": 590, "total_steps": 4368, "loss": 0.5085, "lr": 3.985261722970759e-05, "epoch": 0.9455128205128205, "percentage": 13.51, "elapsed_time": "1:08:28", "remaining_time": "7:18:25"} +{"current_steps": 595, "total_steps": 4368, "loss": 0.5729, "lr": 3.984277445006426e-05, "epoch": 0.9535256410256411, "percentage": 13.62, "elapsed_time": "1:09:01", "remaining_time": "7:17:39"} +{"current_steps": 600, "total_steps": 4368, "loss": 0.5466, "lr": 3.9832614833718654e-05, "epoch": 0.9615384615384616, "percentage": 13.74, "elapsed_time": "1:09:40", "remaining_time": "7:17:31"} +{"current_steps": 605, "total_steps": 4368, "loss": 0.5446, "lr": 3.9822138542893005e-05, "epoch": 0.969551282051282, "percentage": 13.85, "elapsed_time": "1:10:15", "remaining_time": "7:16:57"} +{"current_steps": 610, "total_steps": 4368, "loss": 0.552, "lr": 3.9811345744866014e-05, "epoch": 0.9775641025641025, "percentage": 13.97, "elapsed_time": "1:10:43", "remaining_time": "7:15:40"} +{"current_steps": 615, "total_steps": 4368, "loss": 0.6033, "lr": 3.980023661197016e-05, "epoch": 0.9855769230769231, "percentage": 14.08, "elapsed_time": "1:11:22", "remaining_time": "7:15:33"} +{"current_steps": 620, "total_steps": 4368, "loss": 0.5286, "lr": 3.978881132158896e-05, "epoch": 0.9935897435897436, "percentage": 14.19, "elapsed_time": "1:11:54", "remaining_time": "7:14:41"} +{"current_steps": 625, "total_steps": 4368, "loss": 0.5658, "lr": 3.9777070056154124e-05, "epoch": 1.001602564102564, "percentage": 14.31, "elapsed_time": "1:12:27", "remaining_time": "7:13:57"} +{"current_steps": 630, "total_steps": 4368, "loss": 0.5828, "lr": 3.976501300314264e-05, "epoch": 1.0096153846153846, "percentage": 14.42, "elapsed_time": "1:12:59", "remaining_time": "7:13:02"} +{"current_steps": 635, "total_steps": 4368, "loss": 0.541, "lr": 3.9752640355073825e-05, "epoch": 1.017628205128205, "percentage": 14.54, "elapsed_time": "1:13:28", "remaining_time": "7:11:54"} +{"current_steps": 640, "total_steps": 4368, "loss": 0.5157, "lr": 3.9739952309506175e-05, "epoch": 1.0256410256410255, "percentage": 14.65, "elapsed_time": "1:14:04", "remaining_time": "7:11:29"} +{"current_steps": 645, "total_steps": 4368, "loss": 0.489, "lr": 3.972694906903427e-05, "epoch": 1.0336538461538463, "percentage": 14.77, "elapsed_time": "1:14:36", "remaining_time": "7:10:35"} +{"current_steps": 650, "total_steps": 4368, "loss": 0.5247, "lr": 3.971363084128552e-05, "epoch": 1.0416666666666667, "percentage": 14.88, "elapsed_time": "1:15:06", "remaining_time": "7:09:36"} +{"current_steps": 655, "total_steps": 4368, "loss": 0.5068, "lr": 3.969999783891685e-05, "epoch": 1.0496794871794872, "percentage": 15.0, "elapsed_time": "1:15:41", "remaining_time": "7:09:06"} +{"current_steps": 660, "total_steps": 4368, "loss": 0.5414, "lr": 3.96860502796113e-05, "epoch": 1.0576923076923077, "percentage": 15.11, "elapsed_time": "1:16:08", "remaining_time": "7:07:47"} +{"current_steps": 665, "total_steps": 4368, "loss": 0.5552, "lr": 3.967178838607456e-05, "epoch": 1.0657051282051282, "percentage": 15.22, "elapsed_time": "1:16:41", "remaining_time": "7:07:03"} +{"current_steps": 670, "total_steps": 4368, "loss": 0.5292, "lr": 3.965721238603139e-05, "epoch": 1.0737179487179487, "percentage": 15.34, "elapsed_time": "1:17:13", "remaining_time": "7:06:11"} +{"current_steps": 675, "total_steps": 4368, "loss": 0.5445, "lr": 3.964232251222203e-05, "epoch": 1.0817307692307692, "percentage": 15.45, "elapsed_time": "1:17:45", "remaining_time": "7:05:27"} +{"current_steps": 680, "total_steps": 4368, "loss": 0.5318, "lr": 3.962711900239844e-05, "epoch": 1.0897435897435896, "percentage": 15.57, "elapsed_time": "1:18:18", "remaining_time": "7:04:43"} +{"current_steps": 685, "total_steps": 4368, "loss": 0.5339, "lr": 3.961160209932051e-05, "epoch": 1.0977564102564104, "percentage": 15.68, "elapsed_time": "1:18:49", "remaining_time": "7:03:47"} +{"current_steps": 690, "total_steps": 4368, "loss": 0.5215, "lr": 3.95957720507522e-05, "epoch": 1.1057692307692308, "percentage": 15.8, "elapsed_time": "1:19:24", "remaining_time": "7:03:15"} +{"current_steps": 695, "total_steps": 4368, "loss": 0.526, "lr": 3.957962910945759e-05, "epoch": 1.1137820512820513, "percentage": 15.91, "elapsed_time": "1:19:56", "remaining_time": "7:02:26"} +{"current_steps": 700, "total_steps": 4368, "loss": 0.5335, "lr": 3.9563173533196805e-05, "epoch": 1.1217948717948718, "percentage": 16.03, "elapsed_time": "1:20:33", "remaining_time": "7:02:06"} +{"current_steps": 705, "total_steps": 4368, "loss": 0.5364, "lr": 3.954640558472195e-05, "epoch": 1.1298076923076923, "percentage": 16.14, "elapsed_time": "1:21:12", "remaining_time": "7:01:54"} +{"current_steps": 710, "total_steps": 4368, "loss": 0.5172, "lr": 3.952932553177287e-05, "epoch": 1.1378205128205128, "percentage": 16.25, "elapsed_time": "1:21:49", "remaining_time": "7:01:35"} +{"current_steps": 715, "total_steps": 4368, "loss": 0.5399, "lr": 3.95119336470729e-05, "epoch": 1.1458333333333333, "percentage": 16.37, "elapsed_time": "1:22:19", "remaining_time": "7:00:34"} +{"current_steps": 720, "total_steps": 4368, "loss": 0.5593, "lr": 3.949423020832451e-05, "epoch": 1.1538461538461537, "percentage": 16.48, "elapsed_time": "1:22:57", "remaining_time": "7:00:19"} +{"current_steps": 725, "total_steps": 4368, "loss": 0.5088, "lr": 3.947621549820485e-05, "epoch": 1.1618589743589745, "percentage": 16.6, "elapsed_time": "1:23:31", "remaining_time": "6:59:42"} +{"current_steps": 730, "total_steps": 4368, "loss": 0.5143, "lr": 3.945788980436129e-05, "epoch": 1.169871794871795, "percentage": 16.71, "elapsed_time": "1:24:04", "remaining_time": "6:58:57"} +{"current_steps": 735, "total_steps": 4368, "loss": 0.5198, "lr": 3.943925341940673e-05, "epoch": 1.1778846153846154, "percentage": 16.83, "elapsed_time": "1:24:34", "remaining_time": "6:58:00"} +{"current_steps": 740, "total_steps": 4368, "loss": 0.5195, "lr": 3.942030664091503e-05, "epoch": 1.185897435897436, "percentage": 16.94, "elapsed_time": "1:25:05", "remaining_time": "6:57:10"} +{"current_steps": 745, "total_steps": 4368, "loss": 0.5314, "lr": 3.9401049771416214e-05, "epoch": 1.1939102564102564, "percentage": 17.06, "elapsed_time": "1:25:39", "remaining_time": "6:56:35"} +{"current_steps": 750, "total_steps": 4368, "loss": 0.5278, "lr": 3.938148311839162e-05, "epoch": 1.2019230769230769, "percentage": 17.17, "elapsed_time": "1:26:13", "remaining_time": "6:55:58"} +{"current_steps": 755, "total_steps": 4368, "loss": 0.517, "lr": 3.9361606994269014e-05, "epoch": 1.2099358974358974, "percentage": 17.28, "elapsed_time": "1:26:43", "remaining_time": "6:54:59"} +{"current_steps": 760, "total_steps": 4368, "loss": 0.5336, "lr": 3.934142171641763e-05, "epoch": 1.217948717948718, "percentage": 17.4, "elapsed_time": "1:27:20", "remaining_time": "6:54:38"} +{"current_steps": 765, "total_steps": 4368, "loss": 0.544, "lr": 3.9320927607143003e-05, "epoch": 1.2259615384615385, "percentage": 17.51, "elapsed_time": "1:27:52", "remaining_time": "6:53:53"} +{"current_steps": 770, "total_steps": 4368, "loss": 0.5579, "lr": 3.9300124993681976e-05, "epoch": 1.233974358974359, "percentage": 17.63, "elapsed_time": "1:28:25", "remaining_time": "6:53:10"} +{"current_steps": 775, "total_steps": 4368, "loss": 0.5541, "lr": 3.9279014208197317e-05, "epoch": 1.2419871794871795, "percentage": 17.74, "elapsed_time": "1:28:53", "remaining_time": "6:52:08"} +{"current_steps": 780, "total_steps": 4368, "loss": 0.5336, "lr": 3.925759558777252e-05, "epoch": 1.25, "percentage": 17.86, "elapsed_time": "1:29:28", "remaining_time": "6:51:33"} +{"current_steps": 785, "total_steps": 4368, "loss": 0.5268, "lr": 3.923586947440639e-05, "epoch": 1.2580128205128205, "percentage": 17.97, "elapsed_time": "1:30:06", "remaining_time": "6:51:18"} +{"current_steps": 790, "total_steps": 4368, "loss": 0.5041, "lr": 3.921383621500758e-05, "epoch": 1.266025641025641, "percentage": 18.09, "elapsed_time": "1:30:42", "remaining_time": "6:50:49"} +{"current_steps": 795, "total_steps": 4368, "loss": 0.5446, "lr": 3.919149616138906e-05, "epoch": 1.2740384615384617, "percentage": 18.2, "elapsed_time": "1:31:11", "remaining_time": "6:49:52"} +{"current_steps": 800, "total_steps": 4368, "loss": 0.5421, "lr": 3.916884967026246e-05, "epoch": 1.282051282051282, "percentage": 18.32, "elapsed_time": "1:31:47", "remaining_time": "6:49:22"} +{"current_steps": 805, "total_steps": 4368, "loss": 0.4931, "lr": 3.914589710323245e-05, "epoch": 1.2900641025641026, "percentage": 18.43, "elapsed_time": "1:32:18", "remaining_time": "6:48:35"} +{"current_steps": 810, "total_steps": 4368, "loss": 0.5118, "lr": 3.912263882679091e-05, "epoch": 1.2980769230769231, "percentage": 18.54, "elapsed_time": "1:32:49", "remaining_time": "6:47:44"} +{"current_steps": 815, "total_steps": 4368, "loss": 0.5049, "lr": 3.9099075212311076e-05, "epoch": 1.3060897435897436, "percentage": 18.66, "elapsed_time": "1:33:21", "remaining_time": "6:46:59"} +{"current_steps": 820, "total_steps": 4368, "loss": 0.5442, "lr": 3.9075206636041646e-05, "epoch": 1.314102564102564, "percentage": 18.77, "elapsed_time": "1:33:53", "remaining_time": "6:46:16"} +{"current_steps": 825, "total_steps": 4368, "loss": 0.5011, "lr": 3.905103347910075e-05, "epoch": 1.3221153846153846, "percentage": 18.89, "elapsed_time": "1:34:33", "remaining_time": "6:46:03"} +{"current_steps": 830, "total_steps": 4368, "loss": 0.5498, "lr": 3.902655612746985e-05, "epoch": 1.330128205128205, "percentage": 19.0, "elapsed_time": "1:35:04", "remaining_time": "6:45:18"} +{"current_steps": 835, "total_steps": 4368, "loss": 0.5102, "lr": 3.900177497198761e-05, "epoch": 1.3381410256410255, "percentage": 19.12, "elapsed_time": "1:35:43", "remaining_time": "6:45:03"} +{"current_steps": 840, "total_steps": 4368, "loss": 0.5582, "lr": 3.8976690408343635e-05, "epoch": 1.3461538461538463, "percentage": 19.23, "elapsed_time": "1:36:20", "remaining_time": "6:44:37"} +{"current_steps": 845, "total_steps": 4368, "loss": 0.5099, "lr": 3.8951302837072165e-05, "epoch": 1.3541666666666667, "percentage": 19.35, "elapsed_time": "1:37:00", "remaining_time": "6:44:26"} +{"current_steps": 850, "total_steps": 4368, "loss": 0.5142, "lr": 3.892561266354566e-05, "epoch": 1.3621794871794872, "percentage": 19.46, "elapsed_time": "1:37:42", "remaining_time": "6:44:24"} +{"current_steps": 855, "total_steps": 4368, "loss": 0.5306, "lr": 3.889962029796833e-05, "epoch": 1.3701923076923077, "percentage": 19.57, "elapsed_time": "1:38:13", "remaining_time": "6:43:35"} +{"current_steps": 860, "total_steps": 4368, "loss": 0.5239, "lr": 3.887332615536962e-05, "epoch": 1.3782051282051282, "percentage": 19.69, "elapsed_time": "1:38:43", "remaining_time": "6:42:43"} +{"current_steps": 865, "total_steps": 4368, "loss": 0.5186, "lr": 3.8846730655597535e-05, "epoch": 1.3862179487179487, "percentage": 19.8, "elapsed_time": "1:39:15", "remaining_time": "6:41:59"} +{"current_steps": 870, "total_steps": 4368, "loss": 0.5468, "lr": 3.881983422331198e-05, "epoch": 1.3942307692307692, "percentage": 19.92, "elapsed_time": "1:39:49", "remaining_time": "6:41:23"} +{"current_steps": 875, "total_steps": 4368, "loss": 0.5015, "lr": 3.879263728797792e-05, "epoch": 1.4022435897435899, "percentage": 20.03, "elapsed_time": "1:40:27", "remaining_time": "6:41:02"} +{"current_steps": 880, "total_steps": 4368, "loss": 0.4915, "lr": 3.876514028385861e-05, "epoch": 1.4102564102564101, "percentage": 20.15, "elapsed_time": "1:41:02", "remaining_time": "6:40:28"} +{"current_steps": 885, "total_steps": 4368, "loss": 0.5515, "lr": 3.873734365000857e-05, "epoch": 1.4182692307692308, "percentage": 20.26, "elapsed_time": "1:41:28", "remaining_time": "6:39:23"} +{"current_steps": 890, "total_steps": 4368, "loss": 0.515, "lr": 3.870924783026663e-05, "epoch": 1.4262820512820513, "percentage": 20.38, "elapsed_time": "1:41:56", "remaining_time": "6:38:24"} +{"current_steps": 895, "total_steps": 4368, "loss": 0.5177, "lr": 3.8680853273248826e-05, "epoch": 1.4342948717948718, "percentage": 20.49, "elapsed_time": "1:42:34", "remaining_time": "6:38:01"} +{"current_steps": 900, "total_steps": 4368, "loss": 0.4936, "lr": 3.865216043234126e-05, "epoch": 1.4423076923076923, "percentage": 20.6, "elapsed_time": "1:43:09", "remaining_time": "6:37:29"} +{"current_steps": 905, "total_steps": 4368, "loss": 0.5145, "lr": 3.862316976569281e-05, "epoch": 1.4503205128205128, "percentage": 20.72, "elapsed_time": "1:43:38", "remaining_time": "6:36:35"} +{"current_steps": 910, "total_steps": 4368, "loss": 0.4967, "lr": 3.859388173620785e-05, "epoch": 1.4583333333333333, "percentage": 20.83, "elapsed_time": "1:44:11", "remaining_time": "6:35:55"} +{"current_steps": 915, "total_steps": 4368, "loss": 0.5083, "lr": 3.8564296811538874e-05, "epoch": 1.4663461538461537, "percentage": 20.95, "elapsed_time": "1:44:44", "remaining_time": "6:35:17"} +{"current_steps": 920, "total_steps": 4368, "loss": 0.5424, "lr": 3.853441546407898e-05, "epoch": 1.4743589743589745, "percentage": 21.06, "elapsed_time": "1:45:11", "remaining_time": "6:34:13"} +{"current_steps": 925, "total_steps": 4368, "loss": 0.5221, "lr": 3.850423817095438e-05, "epoch": 1.482371794871795, "percentage": 21.18, "elapsed_time": "1:45:50", "remaining_time": "6:33:55"} +{"current_steps": 930, "total_steps": 4368, "loss": 0.5446, "lr": 3.847376541401674e-05, "epoch": 1.4903846153846154, "percentage": 21.29, "elapsed_time": "1:46:22", "remaining_time": "6:33:15"} +{"current_steps": 935, "total_steps": 4368, "loss": 0.5334, "lr": 3.844299767983551e-05, "epoch": 1.498397435897436, "percentage": 21.41, "elapsed_time": "1:46:53", "remaining_time": "6:32:28"} +{"current_steps": 940, "total_steps": 4368, "loss": 0.5161, "lr": 3.841193545969015e-05, "epoch": 1.5064102564102564, "percentage": 21.52, "elapsed_time": "1:47:23", "remaining_time": "6:31:37"} +{"current_steps": 945, "total_steps": 4368, "loss": 0.5105, "lr": 3.8380579249562265e-05, "epoch": 1.5144230769230769, "percentage": 21.63, "elapsed_time": "1:47:58", "remaining_time": "6:31:05"} +{"current_steps": 950, "total_steps": 4368, "loss": 0.4847, "lr": 3.8348929550127734e-05, "epoch": 1.5224358974358974, "percentage": 21.75, "elapsed_time": "1:48:30", "remaining_time": "6:30:23"} +{"current_steps": 955, "total_steps": 4368, "loss": 0.5687, "lr": 3.831698686674866e-05, "epoch": 1.530448717948718, "percentage": 21.86, "elapsed_time": "1:48:55", "remaining_time": "6:29:16"} +{"current_steps": 960, "total_steps": 4368, "loss": 0.5203, "lr": 3.828475170946534e-05, "epoch": 1.5384615384615383, "percentage": 21.98, "elapsed_time": "1:49:27", "remaining_time": "6:28:34"} +{"current_steps": 965, "total_steps": 4368, "loss": 0.5172, "lr": 3.8252224592988087e-05, "epoch": 1.546474358974359, "percentage": 22.09, "elapsed_time": "1:50:04", "remaining_time": "6:28:11"} +{"current_steps": 970, "total_steps": 4368, "loss": 0.529, "lr": 3.821940603668906e-05, "epoch": 1.5544871794871795, "percentage": 22.21, "elapsed_time": "1:50:34", "remaining_time": "6:27:21"} +{"current_steps": 975, "total_steps": 4368, "loss": 0.5386, "lr": 3.8186296564593924e-05, "epoch": 1.5625, "percentage": 22.32, "elapsed_time": "1:50:58", "remaining_time": "6:26:11"} +{"current_steps": 980, "total_steps": 4368, "loss": 0.4983, "lr": 3.815289670537351e-05, "epoch": 1.5705128205128205, "percentage": 22.44, "elapsed_time": "1:51:33", "remaining_time": "6:25:40"} +{"current_steps": 985, "total_steps": 4368, "loss": 0.5514, "lr": 3.811920699233535e-05, "epoch": 1.578525641025641, "percentage": 22.55, "elapsed_time": "1:52:08", "remaining_time": "6:25:09"} +{"current_steps": 990, "total_steps": 4368, "loss": 0.5246, "lr": 3.8085227963415186e-05, "epoch": 1.5865384615384617, "percentage": 22.66, "elapsed_time": "1:52:50", "remaining_time": "6:25:00"} +{"current_steps": 995, "total_steps": 4368, "loss": 0.5319, "lr": 3.805096016116838e-05, "epoch": 1.594551282051282, "percentage": 22.78, "elapsed_time": "1:53:27", "remaining_time": "6:24:36"} +{"current_steps": 1000, "total_steps": 4368, "loss": 0.5422, "lr": 3.801640413276121e-05, "epoch": 1.6025641025641026, "percentage": 22.89, "elapsed_time": "1:53:57", "remaining_time": "6:23:50"} +{"current_steps": 1005, "total_steps": 4368, "loss": 0.487, "lr": 3.7981560429962204e-05, "epoch": 1.6105769230769231, "percentage": 23.01, "elapsed_time": "1:54:34", "remaining_time": "6:23:24"} +{"current_steps": 1010, "total_steps": 4368, "loss": 0.5211, "lr": 3.7946429609133274e-05, "epoch": 1.6185897435897436, "percentage": 23.12, "elapsed_time": "1:55:13", "remaining_time": "6:23:06"} +{"current_steps": 1015, "total_steps": 4368, "loss": 0.5188, "lr": 3.791101223122084e-05, "epoch": 1.626602564102564, "percentage": 23.24, "elapsed_time": "1:55:40", "remaining_time": "6:22:07"} +{"current_steps": 1020, "total_steps": 4368, "loss": 0.5088, "lr": 3.787530886174688e-05, "epoch": 1.6346153846153846, "percentage": 23.35, "elapsed_time": "1:56:13", "remaining_time": "6:21:28"} +{"current_steps": 1025, "total_steps": 4368, "loss": 0.5486, "lr": 3.783932007079992e-05, "epoch": 1.6426282051282053, "percentage": 23.47, "elapsed_time": "1:56:50", "remaining_time": "6:21:04"} +{"current_steps": 1030, "total_steps": 4368, "loss": 0.4929, "lr": 3.7803046433025905e-05, "epoch": 1.6506410256410255, "percentage": 23.58, "elapsed_time": "1:57:27", "remaining_time": "6:20:40"} +{"current_steps": 1035, "total_steps": 4368, "loss": 0.52, "lr": 3.7766488527619024e-05, "epoch": 1.6586538461538463, "percentage": 23.7, "elapsed_time": "1:58:00", "remaining_time": "6:20:01"} +{"current_steps": 1040, "total_steps": 4368, "loss": 0.4955, "lr": 3.772964693831247e-05, "epoch": 1.6666666666666665, "percentage": 23.81, "elapsed_time": "1:58:31", "remaining_time": "6:19:17"} +{"current_steps": 1045, "total_steps": 4368, "loss": 0.5276, "lr": 3.7692522253369136e-05, "epoch": 1.6746794871794872, "percentage": 23.92, "elapsed_time": "1:59:08", "remaining_time": "6:18:52"} +{"current_steps": 1050, "total_steps": 4368, "loss": 0.5139, "lr": 3.7655115065572194e-05, "epoch": 1.6826923076923077, "percentage": 24.04, "elapsed_time": "1:59:31", "remaining_time": "6:17:40"} +{"current_steps": 1055, "total_steps": 4368, "loss": 0.4682, "lr": 3.7617425972215626e-05, "epoch": 1.6907051282051282, "percentage": 24.15, "elapsed_time": "1:59:58", "remaining_time": "6:16:44"} +{"current_steps": 1060, "total_steps": 4368, "loss": 0.5348, "lr": 3.757945557509472e-05, "epoch": 1.6987179487179487, "percentage": 24.27, "elapsed_time": "2:00:30", "remaining_time": "6:16:05"} +{"current_steps": 1065, "total_steps": 4368, "loss": 0.5192, "lr": 3.7541204480496444e-05, "epoch": 1.7067307692307692, "percentage": 24.38, "elapsed_time": "2:01:04", "remaining_time": "6:15:30"} +{"current_steps": 1070, "total_steps": 4368, "loss": 0.5074, "lr": 3.7502673299189745e-05, "epoch": 1.7147435897435899, "percentage": 24.5, "elapsed_time": "2:01:41", "remaining_time": "6:15:04"} +{"current_steps": 1075, "total_steps": 4368, "loss": 0.5368, "lr": 3.746386264641583e-05, "epoch": 1.7227564102564101, "percentage": 24.61, "elapsed_time": "2:02:14", "remaining_time": "6:14:27"} +{"current_steps": 1080, "total_steps": 4368, "loss": 0.5334, "lr": 3.7424773141878324e-05, "epoch": 1.7307692307692308, "percentage": 24.73, "elapsed_time": "2:02:48", "remaining_time": "6:13:52"} +{"current_steps": 1085, "total_steps": 4368, "loss": 0.5129, "lr": 3.738540540973338e-05, "epoch": 1.7387820512820513, "percentage": 24.84, "elapsed_time": "2:03:27", "remaining_time": "6:13:34"} +{"current_steps": 1090, "total_steps": 4368, "loss": 0.5256, "lr": 3.7345760078579695e-05, "epoch": 1.7467948717948718, "percentage": 24.95, "elapsed_time": "2:04:02", "remaining_time": "6:13:02"} +{"current_steps": 1095, "total_steps": 4368, "loss": 0.5195, "lr": 3.730583778144852e-05, "epoch": 1.7548076923076923, "percentage": 25.07, "elapsed_time": "2:04:39", "remaining_time": "6:12:36"} +{"current_steps": 1100, "total_steps": 4368, "loss": 0.522, "lr": 3.7265639155793494e-05, "epoch": 1.7628205128205128, "percentage": 25.18, "elapsed_time": "2:05:10", "remaining_time": "6:11:53"} +{"current_steps": 1105, "total_steps": 4368, "loss": 0.528, "lr": 3.7225164843480503e-05, "epoch": 1.7708333333333335, "percentage": 25.3, "elapsed_time": "2:05:46", "remaining_time": "6:11:24"} +{"current_steps": 1110, "total_steps": 4368, "loss": 0.5171, "lr": 3.7184415490777426e-05, "epoch": 1.7788461538461537, "percentage": 25.41, "elapsed_time": "2:06:22", "remaining_time": "6:10:56"} +{"current_steps": 1115, "total_steps": 4368, "loss": 0.5287, "lr": 3.714339174834379e-05, "epoch": 1.7868589743589745, "percentage": 25.53, "elapsed_time": "2:07:03", "remaining_time": "6:10:42"} +{"current_steps": 1120, "total_steps": 4368, "loss": 0.493, "lr": 3.710209427122044e-05, "epoch": 1.7948717948717947, "percentage": 25.64, "elapsed_time": "2:07:32", "remaining_time": "6:09:52"} +{"current_steps": 1125, "total_steps": 4368, "loss": 0.5003, "lr": 3.7060523718819e-05, "epoch": 1.8028846153846154, "percentage": 25.76, "elapsed_time": "2:08:00", "remaining_time": "6:09:01"} +{"current_steps": 1130, "total_steps": 4368, "loss": 0.5375, "lr": 3.701868075491139e-05, "epoch": 1.810897435897436, "percentage": 25.87, "elapsed_time": "2:08:28", "remaining_time": "6:08:08"} +{"current_steps": 1135, "total_steps": 4368, "loss": 0.5086, "lr": 3.697656604761926e-05, "epoch": 1.8189102564102564, "percentage": 25.98, "elapsed_time": "2:09:03", "remaining_time": "6:07:36"} +{"current_steps": 1140, "total_steps": 4368, "loss": 0.5056, "lr": 3.693418026940325e-05, "epoch": 1.8269230769230769, "percentage": 26.1, "elapsed_time": "2:09:41", "remaining_time": "6:07:14"} +{"current_steps": 1145, "total_steps": 4368, "loss": 0.4896, "lr": 3.689152409705229e-05, "epoch": 1.8349358974358974, "percentage": 26.21, "elapsed_time": "2:10:09", "remaining_time": "6:06:23"} +{"current_steps": 1150, "total_steps": 4368, "loss": 0.4998, "lr": 3.6848598211672794e-05, "epoch": 1.842948717948718, "percentage": 26.33, "elapsed_time": "2:10:36", "remaining_time": "6:05:28"} +{"current_steps": 1155, "total_steps": 4368, "loss": 0.4874, "lr": 3.6805403298677797e-05, "epoch": 1.8509615384615383, "percentage": 26.44, "elapsed_time": "2:11:12", "remaining_time": "6:04:58"} +{"current_steps": 1160, "total_steps": 4368, "loss": 0.5301, "lr": 3.6761940047775966e-05, "epoch": 1.858974358974359, "percentage": 26.56, "elapsed_time": "2:11:44", "remaining_time": "6:04:20"} +{"current_steps": 1165, "total_steps": 4368, "loss": 0.5226, "lr": 3.671820915296063e-05, "epoch": 1.8669871794871795, "percentage": 26.67, "elapsed_time": "2:12:16", "remaining_time": "6:03:39"} +{"current_steps": 1170, "total_steps": 4368, "loss": 0.5085, "lr": 3.667421131249869e-05, "epoch": 1.875, "percentage": 26.79, "elapsed_time": "2:12:43", "remaining_time": "6:02:47"} +{"current_steps": 1175, "total_steps": 4368, "loss": 0.5034, "lr": 3.662994722891946e-05, "epoch": 1.8830128205128205, "percentage": 26.9, "elapsed_time": "2:13:12", "remaining_time": "6:01:58"} +{"current_steps": 1180, "total_steps": 4368, "loss": 0.5253, "lr": 3.658541760900344e-05, "epoch": 1.891025641025641, "percentage": 27.01, "elapsed_time": "2:13:52", "remaining_time": "6:01:41"} +{"current_steps": 1185, "total_steps": 4368, "loss": 0.4923, "lr": 3.654062316377106e-05, "epoch": 1.8990384615384617, "percentage": 27.13, "elapsed_time": "2:14:23", "remaining_time": "6:00:59"} +{"current_steps": 1190, "total_steps": 4368, "loss": 0.4968, "lr": 3.649556460847131e-05, "epoch": 1.907051282051282, "percentage": 27.24, "elapsed_time": "2:14:52", "remaining_time": "6:00:12"} +{"current_steps": 1195, "total_steps": 4368, "loss": 0.5464, "lr": 3.6450242662570314e-05, "epoch": 1.9150641025641026, "percentage": 27.36, "elapsed_time": "2:15:28", "remaining_time": "5:59:41"} +{"current_steps": 1200, "total_steps": 4368, "loss": 0.5017, "lr": 3.6404658049739854e-05, "epoch": 1.9230769230769231, "percentage": 27.47, "elapsed_time": "2:16:05", "remaining_time": "5:59:16"} +{"current_steps": 1205, "total_steps": 4368, "loss": 0.5117, "lr": 3.63588114978458e-05, "epoch": 1.9310897435897436, "percentage": 27.59, "elapsed_time": "2:16:42", "remaining_time": "5:58:51"} +{"current_steps": 1210, "total_steps": 4368, "loss": 0.4967, "lr": 3.6312703738936504e-05, "epoch": 1.939102564102564, "percentage": 27.7, "elapsed_time": "2:17:19", "remaining_time": "5:58:23"} +{"current_steps": 1215, "total_steps": 4368, "loss": 0.4842, "lr": 3.626633550923111e-05, "epoch": 1.9471153846153846, "percentage": 27.82, "elapsed_time": "2:17:58", "remaining_time": "5:58:02"} +{"current_steps": 1220, "total_steps": 4368, "loss": 0.4889, "lr": 3.621970754910778e-05, "epoch": 1.9551282051282053, "percentage": 27.93, "elapsed_time": "2:18:36", "remaining_time": "5:57:40"} +{"current_steps": 1225, "total_steps": 4368, "loss": 0.4942, "lr": 3.6172820603091885e-05, "epoch": 1.9631410256410255, "percentage": 28.04, "elapsed_time": "2:19:08", "remaining_time": "5:56:58"} +{"current_steps": 1230, "total_steps": 4368, "loss": 0.4815, "lr": 3.612567541984413e-05, "epoch": 1.9711538461538463, "percentage": 28.16, "elapsed_time": "2:19:44", "remaining_time": "5:56:30"} +{"current_steps": 1235, "total_steps": 4368, "loss": 0.5211, "lr": 3.6078272752148574e-05, "epoch": 1.9791666666666665, "percentage": 28.27, "elapsed_time": "2:20:24", "remaining_time": "5:56:11"} +{"current_steps": 1240, "total_steps": 4368, "loss": 0.5182, "lr": 3.6030613356900635e-05, "epoch": 1.9871794871794872, "percentage": 28.39, "elapsed_time": "2:20:57", "remaining_time": "5:55:34"} +{"current_steps": 1245, "total_steps": 4368, "loss": 0.4999, "lr": 3.598269799509498e-05, "epoch": 1.9951923076923077, "percentage": 28.5, "elapsed_time": "2:21:27", "remaining_time": "5:54:50"} +{"current_steps": 1250, "total_steps": 4368, "loss": 0.5314, "lr": 3.5934527431813385e-05, "epoch": 2.003205128205128, "percentage": 28.62, "elapsed_time": "2:22:06", "remaining_time": "5:54:29"} +{"current_steps": 1255, "total_steps": 4368, "loss": 0.4891, "lr": 3.5886102436212536e-05, "epoch": 2.011217948717949, "percentage": 28.73, "elapsed_time": "2:22:34", "remaining_time": "5:53:39"} +{"current_steps": 1260, "total_steps": 4368, "loss": 0.4868, "lr": 3.583742378151171e-05, "epoch": 2.019230769230769, "percentage": 28.85, "elapsed_time": "2:23:09", "remaining_time": "5:53:07"} +{"current_steps": 1265, "total_steps": 4368, "loss": 0.4944, "lr": 3.5788492244980464e-05, "epoch": 2.02724358974359, "percentage": 28.96, "elapsed_time": "2:23:42", "remaining_time": "5:52:29"} +{"current_steps": 1270, "total_steps": 4368, "loss": 0.4826, "lr": 3.573930860792621e-05, "epoch": 2.03525641025641, "percentage": 29.08, "elapsed_time": "2:24:13", "remaining_time": "5:51:48"} +{"current_steps": 1275, "total_steps": 4368, "loss": 0.4955, "lr": 3.568987365568173e-05, "epoch": 2.043269230769231, "percentage": 29.19, "elapsed_time": "2:24:45", "remaining_time": "5:51:11"} +{"current_steps": 1280, "total_steps": 4368, "loss": 0.4912, "lr": 3.564018817759266e-05, "epoch": 2.051282051282051, "percentage": 29.3, "elapsed_time": "2:25:17", "remaining_time": "5:50:30"} +{"current_steps": 1285, "total_steps": 4368, "loss": 0.4789, "lr": 3.559025296700484e-05, "epoch": 2.059294871794872, "percentage": 29.42, "elapsed_time": "2:25:48", "remaining_time": "5:49:49"} +{"current_steps": 1290, "total_steps": 4368, "loss": 0.4888, "lr": 3.554006882125173e-05, "epoch": 2.0673076923076925, "percentage": 29.53, "elapsed_time": "2:26:20", "remaining_time": "5:49:10"} +{"current_steps": 1295, "total_steps": 4368, "loss": 0.4921, "lr": 3.5489636541641586e-05, "epoch": 2.0753205128205128, "percentage": 29.65, "elapsed_time": "2:26:47", "remaining_time": "5:48:19"} +{"current_steps": 1300, "total_steps": 4368, "loss": 0.478, "lr": 3.543895693344472e-05, "epoch": 2.0833333333333335, "percentage": 29.76, "elapsed_time": "2:27:20", "remaining_time": "5:47:44"} +{"current_steps": 1305, "total_steps": 4368, "loss": 0.4908, "lr": 3.538803080588063e-05, "epoch": 2.0913461538461537, "percentage": 29.88, "elapsed_time": "2:27:52", "remaining_time": "5:47:04"} +{"current_steps": 1310, "total_steps": 4368, "loss": 0.4789, "lr": 3.5336858972105076e-05, "epoch": 2.0993589743589745, "percentage": 29.99, "elapsed_time": "2:28:22", "remaining_time": "5:46:20"} +{"current_steps": 1315, "total_steps": 4368, "loss": 0.4858, "lr": 3.528544224919708e-05, "epoch": 2.1073717948717947, "percentage": 30.11, "elapsed_time": "2:28:55", "remaining_time": "5:45:44"} +{"current_steps": 1320, "total_steps": 4368, "loss": 0.4942, "lr": 3.5233781458145934e-05, "epoch": 2.1153846153846154, "percentage": 30.22, "elapsed_time": "2:29:32", "remaining_time": "5:45:17"} +{"current_steps": 1325, "total_steps": 4368, "loss": 0.4767, "lr": 3.5181877423838034e-05, "epoch": 2.123397435897436, "percentage": 30.33, "elapsed_time": "2:30:08", "remaining_time": "5:44:49"} +{"current_steps": 1330, "total_steps": 4368, "loss": 0.4903, "lr": 3.512973097504371e-05, "epoch": 2.1314102564102564, "percentage": 30.45, "elapsed_time": "2:30:39", "remaining_time": "5:44:08"} +{"current_steps": 1335, "total_steps": 4368, "loss": 0.4905, "lr": 3.507734294440403e-05, "epoch": 2.139423076923077, "percentage": 30.56, "elapsed_time": "2:31:14", "remaining_time": "5:43:35"} +{"current_steps": 1340, "total_steps": 4368, "loss": 0.4654, "lr": 3.50247141684175e-05, "epoch": 2.1474358974358974, "percentage": 30.68, "elapsed_time": "2:31:41", "remaining_time": "5:42:47"} +{"current_steps": 1345, "total_steps": 4368, "loss": 0.4741, "lr": 3.497184548742667e-05, "epoch": 2.155448717948718, "percentage": 30.79, "elapsed_time": "2:32:18", "remaining_time": "5:42:19"} +{"current_steps": 1350, "total_steps": 4368, "loss": 0.4819, "lr": 3.491873774560473e-05, "epoch": 2.1634615384615383, "percentage": 30.91, "elapsed_time": "2:32:48", "remaining_time": "5:41:36"} +{"current_steps": 1355, "total_steps": 4368, "loss": 0.4889, "lr": 3.486539179094208e-05, "epoch": 2.171474358974359, "percentage": 31.02, "elapsed_time": "2:33:20", "remaining_time": "5:40:57"} +{"current_steps": 1360, "total_steps": 4368, "loss": 0.496, "lr": 3.481180847523272e-05, "epoch": 2.1794871794871793, "percentage": 31.14, "elapsed_time": "2:33:51", "remaining_time": "5:40:17"} +{"current_steps": 1365, "total_steps": 4368, "loss": 0.4668, "lr": 3.4757988654060684e-05, "epoch": 2.1875, "percentage": 31.25, "elapsed_time": "2:34:24", "remaining_time": "5:39:40"} +{"current_steps": 1370, "total_steps": 4368, "loss": 0.4848, "lr": 3.470393318678637e-05, "epoch": 2.1955128205128207, "percentage": 31.36, "elapsed_time": "2:34:53", "remaining_time": "5:38:58"} +{"current_steps": 1375, "total_steps": 4368, "loss": 0.4529, "lr": 3.4649642936532836e-05, "epoch": 2.203525641025641, "percentage": 31.48, "elapsed_time": "2:35:25", "remaining_time": "5:38:19"} +{"current_steps": 1380, "total_steps": 4368, "loss": 0.4754, "lr": 3.4595118770171984e-05, "epoch": 2.2115384615384617, "percentage": 31.59, "elapsed_time": "2:35:57", "remaining_time": "5:37:41"} +{"current_steps": 1385, "total_steps": 4368, "loss": 0.4613, "lr": 3.454036155831077e-05, "epoch": 2.219551282051282, "percentage": 31.71, "elapsed_time": "2:36:39", "remaining_time": "5:37:25"} +{"current_steps": 1390, "total_steps": 4368, "loss": 0.4866, "lr": 3.4485372175277236e-05, "epoch": 2.2275641025641026, "percentage": 31.82, "elapsed_time": "2:37:11", "remaining_time": "5:36:47"} +{"current_steps": 1395, "total_steps": 4368, "loss": 0.4799, "lr": 3.44301514991066e-05, "epoch": 2.235576923076923, "percentage": 31.94, "elapsed_time": "2:37:43", "remaining_time": "5:36:09"} +{"current_steps": 1400, "total_steps": 4368, "loss": 0.4898, "lr": 3.4374700411527225e-05, "epoch": 2.2435897435897436, "percentage": 32.05, "elapsed_time": "2:38:22", "remaining_time": "5:35:44"} +{"current_steps": 1405, "total_steps": 4368, "loss": 0.5002, "lr": 3.431901979794653e-05, "epoch": 2.251602564102564, "percentage": 32.17, "elapsed_time": "2:38:54", "remaining_time": "5:35:07"} +{"current_steps": 1410, "total_steps": 4368, "loss": 0.4801, "lr": 3.426311054743685e-05, "epoch": 2.2596153846153846, "percentage": 32.28, "elapsed_time": "2:39:20", "remaining_time": "5:34:16"} +{"current_steps": 1415, "total_steps": 4368, "loss": 0.4954, "lr": 3.420697355272127e-05, "epoch": 2.2676282051282053, "percentage": 32.39, "elapsed_time": "2:39:57", "remaining_time": "5:33:49"} +{"current_steps": 1420, "total_steps": 4368, "loss": 0.4855, "lr": 3.415060971015933e-05, "epoch": 2.2756410256410255, "percentage": 32.51, "elapsed_time": "2:40:29", "remaining_time": "5:33:11"} +{"current_steps": 1425, "total_steps": 4368, "loss": 0.4701, "lr": 3.4094019919732736e-05, "epoch": 2.2836538461538463, "percentage": 32.62, "elapsed_time": "2:40:57", "remaining_time": "5:32:25"} +{"current_steps": 1430, "total_steps": 4368, "loss": 0.4814, "lr": 3.403720508503098e-05, "epoch": 2.2916666666666665, "percentage": 32.74, "elapsed_time": "2:41:33", "remaining_time": "5:31:56"} +{"current_steps": 1435, "total_steps": 4368, "loss": 0.4613, "lr": 3.398016611323693e-05, "epoch": 2.2996794871794872, "percentage": 32.85, "elapsed_time": "2:42:03", "remaining_time": "5:31:14"} +{"current_steps": 1440, "total_steps": 4368, "loss": 0.4826, "lr": 3.392290391511232e-05, "epoch": 2.3076923076923075, "percentage": 32.97, "elapsed_time": "2:42:40", "remaining_time": "5:30:46"} +{"current_steps": 1445, "total_steps": 4368, "loss": 0.4692, "lr": 3.386541940498322e-05, "epoch": 2.315705128205128, "percentage": 33.08, "elapsed_time": "2:43:14", "remaining_time": "5:30:12"} +{"current_steps": 1450, "total_steps": 4368, "loss": 0.4883, "lr": 3.380771350072543e-05, "epoch": 2.323717948717949, "percentage": 33.2, "elapsed_time": "2:43:43", "remaining_time": "5:29:28"} +{"current_steps": 1455, "total_steps": 4368, "loss": 0.4807, "lr": 3.374978712374986e-05, "epoch": 2.331730769230769, "percentage": 33.31, "elapsed_time": "2:44:15", "remaining_time": "5:28:50"} +{"current_steps": 1460, "total_steps": 4368, "loss": 0.4669, "lr": 3.369164119898774e-05, "epoch": 2.33974358974359, "percentage": 33.42, "elapsed_time": "2:44:51", "remaining_time": "5:28:22"} +{"current_steps": 1465, "total_steps": 4368, "loss": 0.5001, "lr": 3.363327665487593e-05, "epoch": 2.34775641025641, "percentage": 33.54, "elapsed_time": "2:45:31", "remaining_time": "5:28:00"} +{"current_steps": 1470, "total_steps": 4368, "loss": 0.5013, "lr": 3.357469442334206e-05, "epoch": 2.355769230769231, "percentage": 33.65, "elapsed_time": "2:46:07", "remaining_time": "5:27:31"} +{"current_steps": 1475, "total_steps": 4368, "loss": 0.4724, "lr": 3.351589543978965e-05, "epoch": 2.363782051282051, "percentage": 33.77, "elapsed_time": "2:46:34", "remaining_time": "5:26:43"} +{"current_steps": 1480, "total_steps": 4368, "loss": 0.4442, "lr": 3.345688064308317e-05, "epoch": 2.371794871794872, "percentage": 33.88, "elapsed_time": "2:47:03", "remaining_time": "5:25:59"} +{"current_steps": 1485, "total_steps": 4368, "loss": 0.4715, "lr": 3.339765097553307e-05, "epoch": 2.3798076923076925, "percentage": 34.0, "elapsed_time": "2:47:39", "remaining_time": "5:25:29"} +{"current_steps": 1490, "total_steps": 4368, "loss": 0.4965, "lr": 3.33382073828807e-05, "epoch": 2.3878205128205128, "percentage": 34.11, "elapsed_time": "2:48:24", "remaining_time": "5:25:17"} +{"current_steps": 1495, "total_steps": 4368, "loss": 0.4658, "lr": 3.327855081428326e-05, "epoch": 2.3958333333333335, "percentage": 34.23, "elapsed_time": "2:48:56", "remaining_time": "5:24:39"} +{"current_steps": 1500, "total_steps": 4368, "loss": 0.4668, "lr": 3.3218682222298584e-05, "epoch": 2.4038461538461537, "percentage": 34.34, "elapsed_time": "2:49:28", "remaining_time": "5:24:01"} +{"current_steps": 1505, "total_steps": 4368, "loss": 0.4531, "lr": 3.315860256286996e-05, "epoch": 2.4118589743589745, "percentage": 34.46, "elapsed_time": "2:50:38", "remaining_time": "5:24:36"} +{"current_steps": 1510, "total_steps": 4368, "loss": 0.4799, "lr": 3.3098312795310894e-05, "epoch": 2.4198717948717947, "percentage": 34.57, "elapsed_time": "2:51:11", "remaining_time": "5:24:00"} +{"current_steps": 1515, "total_steps": 4368, "loss": 0.4576, "lr": 3.303781388228974e-05, "epoch": 2.4278846153846154, "percentage": 34.68, "elapsed_time": "2:51:53", "remaining_time": "5:23:42"} +{"current_steps": 1520, "total_steps": 4368, "loss": 0.4892, "lr": 3.297710678981435e-05, "epoch": 2.435897435897436, "percentage": 34.8, "elapsed_time": "2:52:27", "remaining_time": "5:23:07"} +{"current_steps": 1525, "total_steps": 4368, "loss": 0.503, "lr": 3.291619248721667e-05, "epoch": 2.4439102564102564, "percentage": 34.91, "elapsed_time": "2:52:58", "remaining_time": "5:22:29"} +{"current_steps": 1530, "total_steps": 4368, "loss": 0.4559, "lr": 3.285507194713724e-05, "epoch": 2.451923076923077, "percentage": 35.03, "elapsed_time": "2:53:31", "remaining_time": "5:21:52"} +{"current_steps": 1535, "total_steps": 4368, "loss": 0.4832, "lr": 3.279374614550966e-05, "epoch": 2.4599358974358974, "percentage": 35.14, "elapsed_time": "2:54:05", "remaining_time": "5:21:18"} +{"current_steps": 1540, "total_steps": 4368, "loss": 0.4852, "lr": 3.2732216061545e-05, "epoch": 2.467948717948718, "percentage": 35.26, "elapsed_time": "2:54:36", "remaining_time": "5:20:37"} +{"current_steps": 1545, "total_steps": 4368, "loss": 0.5175, "lr": 3.2670482677716214e-05, "epoch": 2.4759615384615383, "percentage": 35.37, "elapsed_time": "2:55:11", "remaining_time": "5:20:07"} +{"current_steps": 1550, "total_steps": 4368, "loss": 0.4546, "lr": 3.2608546979742394e-05, "epoch": 2.483974358974359, "percentage": 35.49, "elapsed_time": "2:55:47", "remaining_time": "5:19:35"} +{"current_steps": 1555, "total_steps": 4368, "loss": 0.498, "lr": 3.254640995657307e-05, "epoch": 2.4919871794871793, "percentage": 35.6, "elapsed_time": "2:56:24", "remaining_time": "5:19:06"} +{"current_steps": 1560, "total_steps": 4368, "loss": 0.4668, "lr": 3.248407260037239e-05, "epoch": 2.5, "percentage": 35.71, "elapsed_time": "2:56:54", "remaining_time": "5:18:25"} +{"current_steps": 1565, "total_steps": 4368, "loss": 0.4543, "lr": 3.24215359065033e-05, "epoch": 2.5080128205128203, "percentage": 35.83, "elapsed_time": "2:57:31", "remaining_time": "5:17:57"} +{"current_steps": 1570, "total_steps": 4368, "loss": 0.4411, "lr": 3.235880087351164e-05, "epoch": 2.516025641025641, "percentage": 35.94, "elapsed_time": "2:58:07", "remaining_time": "5:17:27"} +{"current_steps": 1575, "total_steps": 4368, "loss": 0.5194, "lr": 3.2295868503110184e-05, "epoch": 2.5240384615384617, "percentage": 36.06, "elapsed_time": "2:58:41", "remaining_time": "5:16:52"} +{"current_steps": 1580, "total_steps": 4368, "loss": 0.4734, "lr": 3.22327398001627e-05, "epoch": 2.532051282051282, "percentage": 36.17, "elapsed_time": "2:59:19", "remaining_time": "5:16:25"} +{"current_steps": 1585, "total_steps": 4368, "loss": 0.4595, "lr": 3.216941577266783e-05, "epoch": 2.5400641025641026, "percentage": 36.29, "elapsed_time": "2:59:51", "remaining_time": "5:15:48"} +{"current_steps": 1590, "total_steps": 4368, "loss": 0.4874, "lr": 3.210589743174308e-05, "epoch": 2.5480769230769234, "percentage": 36.4, "elapsed_time": "3:00:21", "remaining_time": "5:15:07"} +{"current_steps": 1595, "total_steps": 4368, "loss": 0.4805, "lr": 3.204218579160857e-05, "epoch": 2.5560897435897436, "percentage": 36.52, "elapsed_time": "3:00:52", "remaining_time": "5:14:28"} +{"current_steps": 1600, "total_steps": 4368, "loss": 0.4735, "lr": 3.197828186957094e-05, "epoch": 2.564102564102564, "percentage": 36.63, "elapsed_time": "3:01:26", "remaining_time": "5:13:52"} +{"current_steps": 1605, "total_steps": 4368, "loss": 0.4776, "lr": 3.191418668600705e-05, "epoch": 2.5721153846153846, "percentage": 36.74, "elapsed_time": "3:01:59", "remaining_time": "5:13:18"} +{"current_steps": 1610, "total_steps": 4368, "loss": 0.523, "lr": 3.184990126434771e-05, "epoch": 2.5801282051282053, "percentage": 36.86, "elapsed_time": "3:02:31", "remaining_time": "5:12:39"} +{"current_steps": 1615, "total_steps": 4368, "loss": 0.4761, "lr": 3.178542663106131e-05, "epoch": 2.5881410256410255, "percentage": 36.97, "elapsed_time": "3:03:01", "remaining_time": "5:11:59"} +{"current_steps": 1620, "total_steps": 4368, "loss": 0.521, "lr": 3.172076381563748e-05, "epoch": 2.5961538461538463, "percentage": 37.09, "elapsed_time": "3:03:34", "remaining_time": "5:11:24"} +{"current_steps": 1625, "total_steps": 4368, "loss": 0.4565, "lr": 3.165591385057058e-05, "epoch": 2.6041666666666665, "percentage": 37.2, "elapsed_time": "3:04:10", "remaining_time": "5:10:53"} +{"current_steps": 1630, "total_steps": 4368, "loss": 0.4784, "lr": 3.1590877771343316e-05, "epoch": 2.6121794871794872, "percentage": 37.32, "elapsed_time": "3:04:44", "remaining_time": "5:10:19"} +{"current_steps": 1635, "total_steps": 4368, "loss": 0.4735, "lr": 3.152565661641008e-05, "epoch": 2.6201923076923075, "percentage": 37.43, "elapsed_time": "3:05:17", "remaining_time": "5:09:43"} +{"current_steps": 1640, "total_steps": 4368, "loss": 0.4606, "lr": 3.1460251427180474e-05, "epoch": 2.628205128205128, "percentage": 37.55, "elapsed_time": "3:05:50", "remaining_time": "5:09:07"} +{"current_steps": 1645, "total_steps": 4368, "loss": 0.4749, "lr": 3.139466324800263e-05, "epoch": 2.636217948717949, "percentage": 37.66, "elapsed_time": "3:06:23", "remaining_time": "5:08:32"} +{"current_steps": 1650, "total_steps": 4368, "loss": 0.4964, "lr": 3.132889312614655e-05, "epoch": 2.644230769230769, "percentage": 37.77, "elapsed_time": "3:06:55", "remaining_time": "5:07:54"} +{"current_steps": 1655, "total_steps": 4368, "loss": 0.4728, "lr": 3.126294211178737e-05, "epoch": 2.65224358974359, "percentage": 37.89, "elapsed_time": "3:07:31", "remaining_time": "5:07:25"} +{"current_steps": 1660, "total_steps": 4368, "loss": 0.4895, "lr": 3.1196811257988634e-05, "epoch": 2.66025641025641, "percentage": 38.0, "elapsed_time": "3:07:58", "remaining_time": "5:06:38"} +{"current_steps": 1665, "total_steps": 4368, "loss": 0.4868, "lr": 3.1130501620685394e-05, "epoch": 2.668269230769231, "percentage": 38.12, "elapsed_time": "3:08:32", "remaining_time": "5:06:05"} +{"current_steps": 1670, "total_steps": 4368, "loss": 0.4694, "lr": 3.106401425866745e-05, "epoch": 2.676282051282051, "percentage": 38.23, "elapsed_time": "3:09:11", "remaining_time": "5:05:39"} +{"current_steps": 1675, "total_steps": 4368, "loss": 0.4814, "lr": 3.099735023356236e-05, "epoch": 2.684294871794872, "percentage": 38.35, "elapsed_time": "3:09:45", "remaining_time": "5:05:05"} +{"current_steps": 1680, "total_steps": 4368, "loss": 0.4999, "lr": 3.0930510609818564e-05, "epoch": 2.6923076923076925, "percentage": 38.46, "elapsed_time": "3:10:12", "remaining_time": "5:04:20"} +{"current_steps": 1685, "total_steps": 4368, "loss": 0.48, "lr": 3.086349645468831e-05, "epoch": 2.7003205128205128, "percentage": 38.58, "elapsed_time": "3:10:42", "remaining_time": "5:03:39"} +{"current_steps": 1690, "total_steps": 4368, "loss": 0.4785, "lr": 3.079630883821067e-05, "epoch": 2.7083333333333335, "percentage": 38.69, "elapsed_time": "3:11:14", "remaining_time": "5:03:03"} +{"current_steps": 1695, "total_steps": 4368, "loss": 0.4662, "lr": 3.0728948833194436e-05, "epoch": 2.7163461538461537, "percentage": 38.8, "elapsed_time": "3:11:45", "remaining_time": "5:02:24"} +{"current_steps": 1700, "total_steps": 4368, "loss": 0.4703, "lr": 3.066141751520099e-05, "epoch": 2.7243589743589745, "percentage": 38.92, "elapsed_time": "3:12:21", "remaining_time": "5:01:53"} +{"current_steps": 1705, "total_steps": 4368, "loss": 0.4707, "lr": 3.059371596252712e-05, "epoch": 2.7323717948717947, "percentage": 39.03, "elapsed_time": "3:12:54", "remaining_time": "5:01:17"} +{"current_steps": 1710, "total_steps": 4368, "loss": 0.4602, "lr": 3.0525845256187834e-05, "epoch": 2.7403846153846154, "percentage": 39.15, "elapsed_time": "3:13:33", "remaining_time": "5:00:51"} +{"current_steps": 1715, "total_steps": 4368, "loss": 0.4771, "lr": 3.0457806479899044e-05, "epoch": 2.748397435897436, "percentage": 39.26, "elapsed_time": "3:14:01", "remaining_time": "5:00:09"} +{"current_steps": 1720, "total_steps": 4368, "loss": 0.4698, "lr": 3.0389600720060318e-05, "epoch": 2.7564102564102564, "percentage": 39.38, "elapsed_time": "3:14:38", "remaining_time": "4:59:39"} +{"current_steps": 1725, "total_steps": 4368, "loss": 0.4479, "lr": 3.0321229065737522e-05, "epoch": 2.7644230769230766, "percentage": 39.49, "elapsed_time": "3:15:13", "remaining_time": "4:59:06"} +{"current_steps": 1730, "total_steps": 4368, "loss": 0.4466, "lr": 3.0252692608645384e-05, "epoch": 2.7724358974358974, "percentage": 39.61, "elapsed_time": "3:15:44", "remaining_time": "4:58:28"} +{"current_steps": 1735, "total_steps": 4368, "loss": 0.4691, "lr": 3.0183992443130127e-05, "epoch": 2.780448717948718, "percentage": 39.72, "elapsed_time": "3:16:27", "remaining_time": "4:58:09"} +{"current_steps": 1740, "total_steps": 4368, "loss": 0.4869, "lr": 3.011512966615195e-05, "epoch": 2.7884615384615383, "percentage": 39.84, "elapsed_time": "3:17:04", "remaining_time": "4:57:39"} +{"current_steps": 1745, "total_steps": 4368, "loss": 0.4615, "lr": 3.0046105377267523e-05, "epoch": 2.796474358974359, "percentage": 39.95, "elapsed_time": "3:17:37", "remaining_time": "4:57:03"} +{"current_steps": 1750, "total_steps": 4368, "loss": 0.4963, "lr": 2.9976920678612456e-05, "epoch": 2.8044871794871797, "percentage": 40.06, "elapsed_time": "3:18:05", "remaining_time": "4:56:21"} +{"current_steps": 1755, "total_steps": 4368, "loss": 0.4788, "lr": 2.9907576674883664e-05, "epoch": 2.8125, "percentage": 40.18, "elapsed_time": "3:18:38", "remaining_time": "4:55:44"} +{"current_steps": 1760, "total_steps": 4368, "loss": 0.4996, "lr": 2.983807447332174e-05, "epoch": 2.8205128205128203, "percentage": 40.29, "elapsed_time": "3:19:13", "remaining_time": "4:55:13"} +{"current_steps": 1765, "total_steps": 4368, "loss": 0.509, "lr": 2.9768415183693293e-05, "epoch": 2.828525641025641, "percentage": 40.41, "elapsed_time": "3:19:41", "remaining_time": "4:54:30"} +{"current_steps": 1770, "total_steps": 4368, "loss": 0.4845, "lr": 2.9698599918273197e-05, "epoch": 2.8365384615384617, "percentage": 40.52, "elapsed_time": "3:20:16", "remaining_time": "4:53:58"} +{"current_steps": 1775, "total_steps": 4368, "loss": 0.4682, "lr": 2.962862979182686e-05, "epoch": 2.844551282051282, "percentage": 40.64, "elapsed_time": "3:20:50", "remaining_time": "4:53:24"} +{"current_steps": 1780, "total_steps": 4368, "loss": 0.4888, "lr": 2.95585059215924e-05, "epoch": 2.8525641025641026, "percentage": 40.75, "elapsed_time": "3:21:22", "remaining_time": "4:52:47"} +{"current_steps": 1785, "total_steps": 4368, "loss": 0.4737, "lr": 2.948822942726284e-05, "epoch": 2.8605769230769234, "percentage": 40.87, "elapsed_time": "3:21:56", "remaining_time": "4:52:13"} +{"current_steps": 1790, "total_steps": 4368, "loss": 0.4671, "lr": 2.941780143096817e-05, "epoch": 2.8685897435897436, "percentage": 40.98, "elapsed_time": "3:22:30", "remaining_time": "4:51:39"} +{"current_steps": 1795, "total_steps": 4368, "loss": 0.4892, "lr": 2.9347223057257505e-05, "epoch": 2.876602564102564, "percentage": 41.09, "elapsed_time": "3:23:03", "remaining_time": "4:51:04"} +{"current_steps": 1800, "total_steps": 4368, "loss": 0.4479, "lr": 2.927649543308106e-05, "epoch": 2.8846153846153846, "percentage": 41.21, "elapsed_time": "3:23:32", "remaining_time": "4:50:22"} +{"current_steps": 1805, "total_steps": 4368, "loss": 0.5043, "lr": 2.9205619687772212e-05, "epoch": 2.8926282051282053, "percentage": 41.32, "elapsed_time": "3:24:05", "remaining_time": "4:49:48"} +{"current_steps": 1810, "total_steps": 4368, "loss": 0.4703, "lr": 2.9134596953029413e-05, "epoch": 2.9006410256410255, "percentage": 41.44, "elapsed_time": "3:24:36", "remaining_time": "4:49:10"} +{"current_steps": 1815, "total_steps": 4368, "loss": 0.4819, "lr": 2.9063428362898168e-05, "epoch": 2.9086538461538463, "percentage": 41.55, "elapsed_time": "3:25:14", "remaining_time": "4:48:42"} +{"current_steps": 1820, "total_steps": 4368, "loss": 0.4456, "lr": 2.8992115053752905e-05, "epoch": 2.9166666666666665, "percentage": 41.67, "elapsed_time": "3:25:49", "remaining_time": "4:48:09"} +{"current_steps": 1825, "total_steps": 4368, "loss": 0.4694, "lr": 2.8920658164278816e-05, "epoch": 2.9246794871794872, "percentage": 41.78, "elapsed_time": "3:26:28", "remaining_time": "4:47:42"} +{"current_steps": 1830, "total_steps": 4368, "loss": 0.4882, "lr": 2.884905883545373e-05, "epoch": 2.9326923076923075, "percentage": 41.9, "elapsed_time": "3:26:56", "remaining_time": "4:47:00"} +{"current_steps": 1835, "total_steps": 4368, "loss": 0.4582, "lr": 2.877731821052981e-05, "epoch": 2.940705128205128, "percentage": 42.01, "elapsed_time": "3:27:33", "remaining_time": "4:46:30"} +{"current_steps": 1840, "total_steps": 4368, "loss": 0.4697, "lr": 2.8705437435015375e-05, "epoch": 2.948717948717949, "percentage": 42.12, "elapsed_time": "3:28:12", "remaining_time": "4:46:04"} +{"current_steps": 1845, "total_steps": 4368, "loss": 0.4801, "lr": 2.8633417656656566e-05, "epoch": 2.956730769230769, "percentage": 42.24, "elapsed_time": "3:28:44", "remaining_time": "4:45:26"} +{"current_steps": 1850, "total_steps": 4368, "loss": 0.5233, "lr": 2.8561260025419036e-05, "epoch": 2.96474358974359, "percentage": 42.35, "elapsed_time": "3:29:16", "remaining_time": "4:44:50"} +{"current_steps": 1855, "total_steps": 4368, "loss": 0.49, "lr": 2.8488965693469583e-05, "epoch": 2.97275641025641, "percentage": 42.47, "elapsed_time": "3:29:54", "remaining_time": "4:44:22"} +{"current_steps": 1860, "total_steps": 4368, "loss": 0.4719, "lr": 2.8416535815157763e-05, "epoch": 2.980769230769231, "percentage": 42.58, "elapsed_time": "3:30:30", "remaining_time": "4:43:50"} +{"current_steps": 1865, "total_steps": 4368, "loss": 0.4551, "lr": 2.8343971546997434e-05, "epoch": 2.988782051282051, "percentage": 42.7, "elapsed_time": "3:31:06", "remaining_time": "4:43:19"} +{"current_steps": 1870, "total_steps": 4368, "loss": 0.473, "lr": 2.827127404764831e-05, "epoch": 2.996794871794872, "percentage": 42.81, "elapsed_time": "3:31:44", "remaining_time": "4:42:51"} +{"current_steps": 1875, "total_steps": 4368, "loss": 0.4595, "lr": 2.8198444477897467e-05, "epoch": 3.0048076923076925, "percentage": 42.93, "elapsed_time": "3:32:14", "remaining_time": "4:42:11"} +{"current_steps": 1880, "total_steps": 4368, "loss": 0.4477, "lr": 2.8125484000640787e-05, "epoch": 3.0128205128205128, "percentage": 43.04, "elapsed_time": "3:32:46", "remaining_time": "4:41:34"} +{"current_steps": 1885, "total_steps": 4368, "loss": 0.4509, "lr": 2.8052393780864394e-05, "epoch": 3.0208333333333335, "percentage": 43.15, "elapsed_time": "3:33:25", "remaining_time": "4:41:08"} +{"current_steps": 1890, "total_steps": 4368, "loss": 0.4363, "lr": 2.797917498562607e-05, "epoch": 3.0288461538461537, "percentage": 43.27, "elapsed_time": "3:33:58", "remaining_time": "4:40:32"} +{"current_steps": 1895, "total_steps": 4368, "loss": 0.4465, "lr": 2.7905828784036596e-05, "epoch": 3.0368589743589745, "percentage": 43.38, "elapsed_time": "3:34:30", "remaining_time": "4:39:56"} +{"current_steps": 1900, "total_steps": 4368, "loss": 0.4506, "lr": 2.78323563472411e-05, "epoch": 3.0448717948717947, "percentage": 43.5, "elapsed_time": "3:35:08", "remaining_time": "4:39:27"} +{"current_steps": 1905, "total_steps": 4368, "loss": 0.4415, "lr": 2.7758758848400354e-05, "epoch": 3.0528846153846154, "percentage": 43.61, "elapsed_time": "3:35:42", "remaining_time": "4:38:53"} +{"current_steps": 1910, "total_steps": 4368, "loss": 0.45, "lr": 2.7685037462672043e-05, "epoch": 3.0608974358974357, "percentage": 43.73, "elapsed_time": "3:36:16", "remaining_time": "4:38:19"} +{"current_steps": 1915, "total_steps": 4368, "loss": 0.4363, "lr": 2.7611193367191993e-05, "epoch": 3.0689102564102564, "percentage": 43.84, "elapsed_time": "3:36:50", "remaining_time": "4:37:46"} +{"current_steps": 1920, "total_steps": 4368, "loss": 0.4282, "lr": 2.7537227741055378e-05, "epoch": 3.076923076923077, "percentage": 43.96, "elapsed_time": "3:37:29", "remaining_time": "4:37:18"} +{"current_steps": 1925, "total_steps": 4368, "loss": 0.4557, "lr": 2.746314176529791e-05, "epoch": 3.0849358974358974, "percentage": 44.07, "elapsed_time": "3:38:04", "remaining_time": "4:36:45"} +{"current_steps": 1930, "total_steps": 4368, "loss": 0.4514, "lr": 2.7388936622876957e-05, "epoch": 3.092948717948718, "percentage": 44.18, "elapsed_time": "3:38:35", "remaining_time": "4:36:07"} +{"current_steps": 1935, "total_steps": 4368, "loss": 0.4517, "lr": 2.7314613498652663e-05, "epoch": 3.1009615384615383, "percentage": 44.3, "elapsed_time": "3:39:13", "remaining_time": "4:35:38"} +{"current_steps": 1940, "total_steps": 4368, "loss": 0.4373, "lr": 2.7240173579369025e-05, "epoch": 3.108974358974359, "percentage": 44.41, "elapsed_time": "3:39:50", "remaining_time": "4:35:08"} +{"current_steps": 1945, "total_steps": 4368, "loss": 0.4448, "lr": 2.7165618053634962e-05, "epoch": 3.1169871794871793, "percentage": 44.53, "elapsed_time": "3:40:24", "remaining_time": "4:34:34"} +{"current_steps": 1950, "total_steps": 4368, "loss": 0.4531, "lr": 2.7090948111905304e-05, "epoch": 3.125, "percentage": 44.64, "elapsed_time": "3:40:57", "remaining_time": "4:33:59"} +{"current_steps": 1955, "total_steps": 4368, "loss": 0.4368, "lr": 2.701616494646183e-05, "epoch": 3.1330128205128207, "percentage": 44.76, "elapsed_time": "3:41:27", "remaining_time": "4:33:20"} +{"current_steps": 1960, "total_steps": 4368, "loss": 0.4155, "lr": 2.6941269751394174e-05, "epoch": 3.141025641025641, "percentage": 44.87, "elapsed_time": "3:41:59", "remaining_time": "4:32:44"} +{"current_steps": 1965, "total_steps": 4368, "loss": 0.429, "lr": 2.686626372258081e-05, "epoch": 3.1490384615384617, "percentage": 44.99, "elapsed_time": "3:42:31", "remaining_time": "4:32:08"} +{"current_steps": 1970, "total_steps": 4368, "loss": 0.4498, "lr": 2.6791148057669913e-05, "epoch": 3.157051282051282, "percentage": 45.1, "elapsed_time": "3:43:04", "remaining_time": "4:31:32"} +{"current_steps": 1975, "total_steps": 4368, "loss": 0.4441, "lr": 2.671592395606027e-05, "epoch": 3.1650641025641026, "percentage": 45.22, "elapsed_time": "3:43:34", "remaining_time": "4:30:54"} +{"current_steps": 1980, "total_steps": 4368, "loss": 0.4213, "lr": 2.6640592618882114e-05, "epoch": 3.173076923076923, "percentage": 45.33, "elapsed_time": "3:44:07", "remaining_time": "4:30:18"} +{"current_steps": 1985, "total_steps": 4368, "loss": 0.4915, "lr": 2.656515524897795e-05, "epoch": 3.1810897435897436, "percentage": 45.44, "elapsed_time": "3:44:37", "remaining_time": "4:29:40"} +{"current_steps": 1990, "total_steps": 4368, "loss": 0.4379, "lr": 2.6489613050883343e-05, "epoch": 3.189102564102564, "percentage": 45.56, "elapsed_time": "3:45:09", "remaining_time": "4:29:03"} +{"current_steps": 1995, "total_steps": 4368, "loss": 0.4356, "lr": 2.6413967230807677e-05, "epoch": 3.1971153846153846, "percentage": 45.67, "elapsed_time": "3:45:38", "remaining_time": "4:28:23"} +{"current_steps": 2000, "total_steps": 4368, "loss": 0.4592, "lr": 2.6338218996614924e-05, "epoch": 3.2051282051282053, "percentage": 45.79, "elapsed_time": "3:46:15", "remaining_time": "4:27:53"} +{"current_steps": 2005, "total_steps": 4368, "loss": 0.4566, "lr": 2.6262369557804325e-05, "epoch": 3.2131410256410255, "percentage": 45.9, "elapsed_time": "3:46:56", "remaining_time": "4:27:28"} +{"current_steps": 2010, "total_steps": 4368, "loss": 0.4709, "lr": 2.6186420125491094e-05, "epoch": 3.2211538461538463, "percentage": 46.02, "elapsed_time": "3:47:30", "remaining_time": "4:26:54"} +{"current_steps": 2015, "total_steps": 4368, "loss": 0.4659, "lr": 2.6110371912387083e-05, "epoch": 3.2291666666666665, "percentage": 46.13, "elapsed_time": "3:47:57", "remaining_time": "4:26:11"} +{"current_steps": 2020, "total_steps": 4368, "loss": 0.4449, "lr": 2.6034226132781407e-05, "epoch": 3.2371794871794872, "percentage": 46.25, "elapsed_time": "3:48:25", "remaining_time": "4:25:31"} +{"current_steps": 2025, "total_steps": 4368, "loss": 0.4912, "lr": 2.5957984002521066e-05, "epoch": 3.2451923076923075, "percentage": 46.36, "elapsed_time": "3:49:02", "remaining_time": "4:25:00"} +{"current_steps": 2030, "total_steps": 4368, "loss": 0.4263, "lr": 2.588164673899151e-05, "epoch": 3.253205128205128, "percentage": 46.47, "elapsed_time": "3:49:37", "remaining_time": "4:24:27"} +{"current_steps": 2035, "total_steps": 4368, "loss": 0.4229, "lr": 2.580521556109724e-05, "epoch": 3.261217948717949, "percentage": 46.59, "elapsed_time": "3:50:10", "remaining_time": "4:23:53"} +{"current_steps": 2040, "total_steps": 4368, "loss": 0.4726, "lr": 2.57286916892423e-05, "epoch": 3.269230769230769, "percentage": 46.7, "elapsed_time": "3:50:42", "remaining_time": "4:23:16"} +{"current_steps": 2045, "total_steps": 4368, "loss": 0.425, "lr": 2.5652076345310822e-05, "epoch": 3.27724358974359, "percentage": 46.82, "elapsed_time": "3:51:13", "remaining_time": "4:22:39"} +{"current_steps": 2050, "total_steps": 4368, "loss": 0.4533, "lr": 2.5575370752647507e-05, "epoch": 3.28525641025641, "percentage": 46.93, "elapsed_time": "3:51:49", "remaining_time": "4:22:07"} +{"current_steps": 2055, "total_steps": 4368, "loss": 0.4375, "lr": 2.5498576136038077e-05, "epoch": 3.293269230769231, "percentage": 47.05, "elapsed_time": "3:52:16", "remaining_time": "4:21:26"} +{"current_steps": 2060, "total_steps": 4368, "loss": 0.4532, "lr": 2.542169372168976e-05, "epoch": 3.301282051282051, "percentage": 47.16, "elapsed_time": "3:52:48", "remaining_time": "4:20:49"} +{"current_steps": 2065, "total_steps": 4368, "loss": 0.4632, "lr": 2.5344724737211646e-05, "epoch": 3.309294871794872, "percentage": 47.28, "elapsed_time": "3:53:25", "remaining_time": "4:20:20"} +{"current_steps": 2070, "total_steps": 4368, "loss": 0.4304, "lr": 2.5267670411595152e-05, "epoch": 3.3173076923076925, "percentage": 47.39, "elapsed_time": "3:53:54", "remaining_time": "4:19:40"} +{"current_steps": 2075, "total_steps": 4368, "loss": 0.4502, "lr": 2.5190531975194345e-05, "epoch": 3.3253205128205128, "percentage": 47.5, "elapsed_time": "3:54:21", "remaining_time": "4:18:58"} +{"current_steps": 2080, "total_steps": 4368, "loss": 0.4721, "lr": 2.5113310659706322e-05, "epoch": 3.3333333333333335, "percentage": 47.62, "elapsed_time": "3:54:57", "remaining_time": "4:18:26"} +{"current_steps": 2085, "total_steps": 4368, "loss": 0.4301, "lr": 2.5036007698151553e-05, "epoch": 3.3413461538461537, "percentage": 47.73, "elapsed_time": "3:55:31", "remaining_time": "4:17:52"} +{"current_steps": 2090, "total_steps": 4368, "loss": 0.4438, "lr": 2.4958624324854185e-05, "epoch": 3.3493589743589745, "percentage": 47.85, "elapsed_time": "3:56:04", "remaining_time": "4:17:18"} +{"current_steps": 2095, "total_steps": 4368, "loss": 0.4625, "lr": 2.4881161775422303e-05, "epoch": 3.3573717948717947, "percentage": 47.96, "elapsed_time": "3:56:41", "remaining_time": "4:16:47"} +{"current_steps": 2100, "total_steps": 4368, "loss": 0.5032, "lr": 2.480362128672824e-05, "epoch": 3.3653846153846154, "percentage": 48.08, "elapsed_time": "3:57:06", "remaining_time": "4:16:04"} +{"current_steps": 2105, "total_steps": 4368, "loss": 0.4448, "lr": 2.4726004096888817e-05, "epoch": 3.373397435897436, "percentage": 48.19, "elapsed_time": "3:57:39", "remaining_time": "4:15:29"} +{"current_steps": 2110, "total_steps": 4368, "loss": 0.4615, "lr": 2.4648311445245558e-05, "epoch": 3.3814102564102564, "percentage": 48.31, "elapsed_time": "3:58:12", "remaining_time": "4:14:55"} +{"current_steps": 2115, "total_steps": 4368, "loss": 0.4455, "lr": 2.457054457234493e-05, "epoch": 3.389423076923077, "percentage": 48.42, "elapsed_time": "3:58:47", "remaining_time": "4:14:22"} +{"current_steps": 2120, "total_steps": 4368, "loss": 0.4104, "lr": 2.4492704719918497e-05, "epoch": 3.3974358974358974, "percentage": 48.53, "elapsed_time": "3:59:26", "remaining_time": "4:13:53"} +{"current_steps": 2125, "total_steps": 4368, "loss": 0.4627, "lr": 2.4414793130863134e-05, "epoch": 3.405448717948718, "percentage": 48.65, "elapsed_time": "3:59:59", "remaining_time": "4:13:18"} +{"current_steps": 2130, "total_steps": 4368, "loss": 0.4218, "lr": 2.433681104922114e-05, "epoch": 3.4134615384615383, "percentage": 48.76, "elapsed_time": "4:00:30", "remaining_time": "4:12:42"} +{"current_steps": 2135, "total_steps": 4368, "loss": 0.4261, "lr": 2.4258759720160412e-05, "epoch": 3.421474358974359, "percentage": 48.88, "elapsed_time": "4:01:00", "remaining_time": "4:12:04"} +{"current_steps": 2140, "total_steps": 4368, "loss": 0.4442, "lr": 2.4180640389954534e-05, "epoch": 3.4294871794871793, "percentage": 48.99, "elapsed_time": "4:01:32", "remaining_time": "4:11:28"} +{"current_steps": 2145, "total_steps": 4368, "loss": 0.475, "lr": 2.4102454305962892e-05, "epoch": 3.4375, "percentage": 49.11, "elapsed_time": "4:02:03", "remaining_time": "4:10:51"} +{"current_steps": 2150, "total_steps": 4368, "loss": 0.4877, "lr": 2.402420271661076e-05, "epoch": 3.4455128205128207, "percentage": 49.22, "elapsed_time": "4:02:31", "remaining_time": "4:10:11"} +{"current_steps": 2155, "total_steps": 4368, "loss": 0.4468, "lr": 2.3945886871369338e-05, "epoch": 3.453525641025641, "percentage": 49.34, "elapsed_time": "4:03:07", "remaining_time": "4:09:39"} +{"current_steps": 2160, "total_steps": 4368, "loss": 0.4411, "lr": 2.3867508020735865e-05, "epoch": 3.4615384615384617, "percentage": 49.45, "elapsed_time": "4:03:41", "remaining_time": "4:09:06"} +{"current_steps": 2165, "total_steps": 4368, "loss": 0.45, "lr": 2.3789067416213568e-05, "epoch": 3.469551282051282, "percentage": 49.57, "elapsed_time": "4:04:12", "remaining_time": "4:08:29"} +{"current_steps": 2170, "total_steps": 4368, "loss": 0.4333, "lr": 2.3710566310291733e-05, "epoch": 3.4775641025641026, "percentage": 49.68, "elapsed_time": "4:04:47", "remaining_time": "4:07:56"} +{"current_steps": 2175, "total_steps": 4368, "loss": 0.4327, "lr": 2.36320059564257e-05, "epoch": 3.485576923076923, "percentage": 49.79, "elapsed_time": "4:05:21", "remaining_time": "4:07:23"} +{"current_steps": 2180, "total_steps": 4368, "loss": 0.4424, "lr": 2.3553387609016833e-05, "epoch": 3.4935897435897436, "percentage": 49.91, "elapsed_time": "4:05:52", "remaining_time": "4:06:46"} +{"current_steps": 2185, "total_steps": 4368, "loss": 0.4338, "lr": 2.347471252339252e-05, "epoch": 3.501602564102564, "percentage": 50.02, "elapsed_time": "4:06:28", "remaining_time": "4:06:15"} +{"current_steps": 2190, "total_steps": 4368, "loss": 0.4342, "lr": 2.339598195578608e-05, "epoch": 3.5096153846153846, "percentage": 50.14, "elapsed_time": "4:07:02", "remaining_time": "4:05:41"} +{"current_steps": 2195, "total_steps": 4368, "loss": 0.461, "lr": 2.3317197163316757e-05, "epoch": 3.5176282051282053, "percentage": 50.25, "elapsed_time": "4:07:33", "remaining_time": "4:05:04"} +{"current_steps": 2200, "total_steps": 4368, "loss": 0.4543, "lr": 2.3238359403969608e-05, "epoch": 3.5256410256410255, "percentage": 50.37, "elapsed_time": "4:08:08", "remaining_time": "4:04:32"} +{"current_steps": 2205, "total_steps": 4368, "loss": 0.442, "lr": 2.315946993657543e-05, "epoch": 3.5336538461538463, "percentage": 50.48, "elapsed_time": "4:08:42", "remaining_time": "4:03:58"} +{"current_steps": 2210, "total_steps": 4368, "loss": 0.4852, "lr": 2.3080530020790673e-05, "epoch": 3.5416666666666665, "percentage": 50.6, "elapsed_time": "4:09:23", "remaining_time": "4:03:30"} +{"current_steps": 2215, "total_steps": 4368, "loss": 0.4376, "lr": 2.300154091707731e-05, "epoch": 3.5496794871794872, "percentage": 50.71, "elapsed_time": "4:09:59", "remaining_time": "4:02:59"} +{"current_steps": 2220, "total_steps": 4368, "loss": 0.4586, "lr": 2.2922503886682706e-05, "epoch": 3.5576923076923075, "percentage": 50.82, "elapsed_time": "4:10:34", "remaining_time": "4:02:26"} +{"current_steps": 2225, "total_steps": 4368, "loss": 0.4556, "lr": 2.28434201916195e-05, "epoch": 3.565705128205128, "percentage": 50.94, "elapsed_time": "4:11:07", "remaining_time": "4:01:52"} +{"current_steps": 2230, "total_steps": 4368, "loss": 0.4144, "lr": 2.2764291094645446e-05, "epoch": 3.573717948717949, "percentage": 51.05, "elapsed_time": "4:11:41", "remaining_time": "4:01:18"} +{"current_steps": 2235, "total_steps": 4368, "loss": 0.4595, "lr": 2.2685117859243223e-05, "epoch": 3.581730769230769, "percentage": 51.17, "elapsed_time": "4:12:12", "remaining_time": "4:00:41"} +{"current_steps": 2240, "total_steps": 4368, "loss": 0.4554, "lr": 2.2605901749600312e-05, "epoch": 3.58974358974359, "percentage": 51.28, "elapsed_time": "4:12:42", "remaining_time": "4:00:04"} +{"current_steps": 2245, "total_steps": 4368, "loss": 0.473, "lr": 2.2526644030588764e-05, "epoch": 3.59775641025641, "percentage": 51.4, "elapsed_time": "4:13:12", "remaining_time": "3:59:27"} +{"current_steps": 2250, "total_steps": 4368, "loss": 0.4373, "lr": 2.2447345967745036e-05, "epoch": 3.605769230769231, "percentage": 51.51, "elapsed_time": "4:13:52", "remaining_time": "3:58:58"} +{"current_steps": 2255, "total_steps": 4368, "loss": 0.4463, "lr": 2.2368008827249756e-05, "epoch": 3.613782051282051, "percentage": 51.63, "elapsed_time": "4:14:22", "remaining_time": "3:58:21"} +{"current_steps": 2260, "total_steps": 4368, "loss": 0.4369, "lr": 2.228863387590752e-05, "epoch": 3.621794871794872, "percentage": 51.74, "elapsed_time": "4:14:53", "remaining_time": "3:57:45"} +{"current_steps": 2265, "total_steps": 4368, "loss": 0.4472, "lr": 2.2209222381126687e-05, "epoch": 3.6298076923076925, "percentage": 51.85, "elapsed_time": "4:15:25", "remaining_time": "3:57:09"} +{"current_steps": 2270, "total_steps": 4368, "loss": 0.4591, "lr": 2.212977561089908e-05, "epoch": 3.6378205128205128, "percentage": 51.97, "elapsed_time": "4:16:01", "remaining_time": "3:56:37"} +{"current_steps": 2275, "total_steps": 4368, "loss": 0.4355, "lr": 2.20502948337798e-05, "epoch": 3.6458333333333335, "percentage": 52.08, "elapsed_time": "4:16:40", "remaining_time": "3:56:08"} +{"current_steps": 2280, "total_steps": 4368, "loss": 0.4439, "lr": 2.1970781318866953e-05, "epoch": 3.6538461538461537, "percentage": 52.2, "elapsed_time": "4:17:19", "remaining_time": "3:55:39"} +{"current_steps": 2285, "total_steps": 4368, "loss": 0.4445, "lr": 2.1891236335781363e-05, "epoch": 3.6618589743589745, "percentage": 52.31, "elapsed_time": "4:17:48", "remaining_time": "3:55:00"} +{"current_steps": 2290, "total_steps": 4368, "loss": 0.4735, "lr": 2.1811661154646332e-05, "epoch": 3.6698717948717947, "percentage": 52.43, "elapsed_time": "4:18:15", "remaining_time": "3:54:21"} +{"current_steps": 2295, "total_steps": 4368, "loss": 0.4986, "lr": 2.173205704606735e-05, "epoch": 3.6778846153846154, "percentage": 52.54, "elapsed_time": "4:18:43", "remaining_time": "3:53:41"} +{"current_steps": 2300, "total_steps": 4368, "loss": 0.4557, "lr": 2.1652425281111785e-05, "epoch": 3.685897435897436, "percentage": 52.66, "elapsed_time": "4:19:21", "remaining_time": "3:53:11"} +{"current_steps": 2305, "total_steps": 4368, "loss": 0.4787, "lr": 2.1572767131288607e-05, "epoch": 3.6939102564102564, "percentage": 52.77, "elapsed_time": "4:19:56", "remaining_time": "3:52:39"} +{"current_steps": 2310, "total_steps": 4368, "loss": 0.4328, "lr": 2.1493083868528095e-05, "epoch": 3.7019230769230766, "percentage": 52.88, "elapsed_time": "4:20:31", "remaining_time": "3:52:06"} +{"current_steps": 2315, "total_steps": 4368, "loss": 0.4616, "lr": 2.141337676516151e-05, "epoch": 3.7099358974358974, "percentage": 53.0, "elapsed_time": "4:21:07", "remaining_time": "3:51:34"} +{"current_steps": 2320, "total_steps": 4368, "loss": 0.4168, "lr": 2.1333647093900772e-05, "epoch": 3.717948717948718, "percentage": 53.11, "elapsed_time": "4:21:43", "remaining_time": "3:51:02"} +{"current_steps": 2325, "total_steps": 4368, "loss": 0.431, "lr": 2.1253896127818175e-05, "epoch": 3.7259615384615383, "percentage": 53.23, "elapsed_time": "4:22:21", "remaining_time": "3:50:32"} +{"current_steps": 2330, "total_steps": 4368, "loss": 0.4593, "lr": 2.1174125140326013e-05, "epoch": 3.733974358974359, "percentage": 53.34, "elapsed_time": "4:22:58", "remaining_time": "3:50:01"} +{"current_steps": 2335, "total_steps": 4368, "loss": 0.4249, "lr": 2.1094335405156277e-05, "epoch": 3.7419871794871797, "percentage": 53.46, "elapsed_time": "4:23:33", "remaining_time": "3:49:28"} +{"current_steps": 2340, "total_steps": 4368, "loss": 0.4456, "lr": 2.1014528196340316e-05, "epoch": 3.75, "percentage": 53.57, "elapsed_time": "4:24:09", "remaining_time": "3:48:56"} +{"current_steps": 2345, "total_steps": 4368, "loss": 0.4293, "lr": 2.093470478818847e-05, "epoch": 3.7580128205128203, "percentage": 53.69, "elapsed_time": "4:24:42", "remaining_time": "3:48:21"} +{"current_steps": 2350, "total_steps": 4368, "loss": 0.4573, "lr": 2.0854866455269756e-05, "epoch": 3.766025641025641, "percentage": 53.8, "elapsed_time": "4:25:15", "remaining_time": "3:47:46"} +{"current_steps": 2355, "total_steps": 4368, "loss": 0.4343, "lr": 2.0775014472391496e-05, "epoch": 3.7740384615384617, "percentage": 53.91, "elapsed_time": "4:25:41", "remaining_time": "3:47:06"} +{"current_steps": 2360, "total_steps": 4368, "loss": 0.4277, "lr": 2.0695150114578958e-05, "epoch": 3.782051282051282, "percentage": 54.03, "elapsed_time": "4:26:16", "remaining_time": "3:46:33"} +{"current_steps": 2365, "total_steps": 4368, "loss": 0.4593, "lr": 2.061527465705502e-05, "epoch": 3.7900641025641026, "percentage": 54.14, "elapsed_time": "4:26:47", "remaining_time": "3:45:57"} +{"current_steps": 2370, "total_steps": 4368, "loss": 0.4198, "lr": 2.0535389375219773e-05, "epoch": 3.7980769230769234, "percentage": 54.26, "elapsed_time": "4:27:20", "remaining_time": "3:45:22"} +{"current_steps": 2375, "total_steps": 4368, "loss": 0.4433, "lr": 2.045549554463019e-05, "epoch": 3.8060897435897436, "percentage": 54.37, "elapsed_time": "4:27:54", "remaining_time": "3:44:49"} +{"current_steps": 2380, "total_steps": 4368, "loss": 0.4185, "lr": 2.0375594440979744e-05, "epoch": 3.814102564102564, "percentage": 54.49, "elapsed_time": "4:28:25", "remaining_time": "3:44:12"} +{"current_steps": 2385, "total_steps": 4368, "loss": 0.4298, "lr": 2.0295687340078037e-05, "epoch": 3.8221153846153846, "percentage": 54.6, "elapsed_time": "4:28:56", "remaining_time": "3:43:36"} +{"current_steps": 2390, "total_steps": 4368, "loss": 0.4531, "lr": 2.0215775517830437e-05, "epoch": 3.8301282051282053, "percentage": 54.72, "elapsed_time": "4:29:27", "remaining_time": "3:43:00"} +{"current_steps": 2395, "total_steps": 4368, "loss": 0.4517, "lr": 2.013586025021769e-05, "epoch": 3.8381410256410255, "percentage": 54.83, "elapsed_time": "4:29:59", "remaining_time": "3:42:25"} +{"current_steps": 2400, "total_steps": 4368, "loss": 0.4619, "lr": 2.0055942813275564e-05, "epoch": 3.8461538461538463, "percentage": 54.95, "elapsed_time": "4:30:36", "remaining_time": "3:41:53"} +{"current_steps": 2405, "total_steps": 4368, "loss": 0.4574, "lr": 1.9976024483074456e-05, "epoch": 3.8541666666666665, "percentage": 55.06, "elapsed_time": "4:31:06", "remaining_time": "3:41:17"} +{"current_steps": 2410, "total_steps": 4368, "loss": 0.4382, "lr": 1.9896106535699025e-05, "epoch": 3.8621794871794872, "percentage": 55.17, "elapsed_time": "4:31:37", "remaining_time": "3:40:41"} +{"current_steps": 2415, "total_steps": 4368, "loss": 0.451, "lr": 1.9816190247227834e-05, "epoch": 3.8701923076923075, "percentage": 55.29, "elapsed_time": "4:32:12", "remaining_time": "3:40:08"} +{"current_steps": 2420, "total_steps": 4368, "loss": 0.4264, "lr": 1.9736276893712954e-05, "epoch": 3.878205128205128, "percentage": 55.4, "elapsed_time": "4:32:46", "remaining_time": "3:39:34"} +{"current_steps": 2425, "total_steps": 4368, "loss": 0.4344, "lr": 1.9656367751159565e-05, "epoch": 3.886217948717949, "percentage": 55.52, "elapsed_time": "4:33:25", "remaining_time": "3:39:04"} +{"current_steps": 2430, "total_steps": 4368, "loss": 0.4489, "lr": 1.957646409550565e-05, "epoch": 3.894230769230769, "percentage": 55.63, "elapsed_time": "4:34:01", "remaining_time": "3:38:32"} +{"current_steps": 2435, "total_steps": 4368, "loss": 0.4735, "lr": 1.9496567202601545e-05, "epoch": 3.90224358974359, "percentage": 55.75, "elapsed_time": "4:34:39", "remaining_time": "3:38:01"} +{"current_steps": 2440, "total_steps": 4368, "loss": 0.4378, "lr": 1.9416678348189627e-05, "epoch": 3.91025641025641, "percentage": 55.86, "elapsed_time": "4:35:06", "remaining_time": "3:37:22"} +{"current_steps": 2445, "total_steps": 4368, "loss": 0.4167, "lr": 1.9336798807883907e-05, "epoch": 3.918269230769231, "percentage": 55.98, "elapsed_time": "4:35:40", "remaining_time": "3:36:48"} +{"current_steps": 2450, "total_steps": 4368, "loss": 0.4526, "lr": 1.9256929857149686e-05, "epoch": 3.926282051282051, "percentage": 56.09, "elapsed_time": "4:36:08", "remaining_time": "3:36:10"} +{"current_steps": 2455, "total_steps": 4368, "loss": 0.4541, "lr": 1.9177072771283167e-05, "epoch": 3.934294871794872, "percentage": 56.2, "elapsed_time": "4:36:43", "remaining_time": "3:35:37"} +{"current_steps": 2460, "total_steps": 4368, "loss": 0.4309, "lr": 1.9097228825391087e-05, "epoch": 3.9423076923076925, "percentage": 56.32, "elapsed_time": "4:37:21", "remaining_time": "3:35:07"} +{"current_steps": 2465, "total_steps": 4368, "loss": 0.4793, "lr": 1.9017399294370413e-05, "epoch": 3.9503205128205128, "percentage": 56.43, "elapsed_time": "4:37:58", "remaining_time": "3:34:36"} +{"current_steps": 2470, "total_steps": 4368, "loss": 0.4545, "lr": 1.893758545288791e-05, "epoch": 3.9583333333333335, "percentage": 56.55, "elapsed_time": "4:38:29", "remaining_time": "3:34:00"} +{"current_steps": 2475, "total_steps": 4368, "loss": 0.4811, "lr": 1.8857788575359847e-05, "epoch": 3.9663461538461537, "percentage": 56.66, "elapsed_time": "4:38:59", "remaining_time": "3:33:23"} +{"current_steps": 2480, "total_steps": 4368, "loss": 0.4588, "lr": 1.87780099359316e-05, "epoch": 3.9743589743589745, "percentage": 56.78, "elapsed_time": "4:39:28", "remaining_time": "3:32:45"} +{"current_steps": 2485, "total_steps": 4368, "loss": 0.435, "lr": 1.869825080845734e-05, "epoch": 3.9823717948717947, "percentage": 56.89, "elapsed_time": "4:40:01", "remaining_time": "3:32:11"} +{"current_steps": 2490, "total_steps": 4368, "loss": 0.4894, "lr": 1.8618512466479686e-05, "epoch": 3.9903846153846154, "percentage": 57.01, "elapsed_time": "4:40:32", "remaining_time": "3:31:35"} +{"current_steps": 2495, "total_steps": 4368, "loss": 0.4577, "lr": 1.8538796183209373e-05, "epoch": 3.998397435897436, "percentage": 57.12, "elapsed_time": "4:41:04", "remaining_time": "3:31:00"} +{"current_steps": 2500, "total_steps": 4368, "loss": 0.4366, "lr": 1.845910323150491e-05, "epoch": 4.006410256410256, "percentage": 57.23, "elapsed_time": "4:41:39", "remaining_time": "3:30:27"} +{"current_steps": 2505, "total_steps": 4368, "loss": 0.4037, "lr": 1.8379434883852255e-05, "epoch": 4.014423076923077, "percentage": 57.35, "elapsed_time": "4:42:11", "remaining_time": "3:29:52"} +{"current_steps": 2510, "total_steps": 4368, "loss": 0.3962, "lr": 1.8299792412344524e-05, "epoch": 4.022435897435898, "percentage": 57.46, "elapsed_time": "4:42:43", "remaining_time": "3:29:17"} +{"current_steps": 2515, "total_steps": 4368, "loss": 0.4001, "lr": 1.8220177088661635e-05, "epoch": 4.030448717948718, "percentage": 57.58, "elapsed_time": "4:43:16", "remaining_time": "3:28:42"} +{"current_steps": 2520, "total_steps": 4368, "loss": 0.4172, "lr": 1.814059018405004e-05, "epoch": 4.038461538461538, "percentage": 57.69, "elapsed_time": "4:43:47", "remaining_time": "3:28:07"} +{"current_steps": 2525, "total_steps": 4368, "loss": 0.437, "lr": 1.806103296930243e-05, "epoch": 4.046474358974359, "percentage": 57.81, "elapsed_time": "4:44:19", "remaining_time": "3:27:31"} +{"current_steps": 2530, "total_steps": 4368, "loss": 0.4118, "lr": 1.7981506714737392e-05, "epoch": 4.05448717948718, "percentage": 57.92, "elapsed_time": "4:44:52", "remaining_time": "3:26:57"} +{"current_steps": 2535, "total_steps": 4368, "loss": 0.4397, "lr": 1.7902012690179188e-05, "epoch": 4.0625, "percentage": 58.04, "elapsed_time": "4:45:27", "remaining_time": "3:26:24"} +{"current_steps": 2540, "total_steps": 4368, "loss": 0.4426, "lr": 1.7822552164937437e-05, "epoch": 4.07051282051282, "percentage": 58.15, "elapsed_time": "4:46:07", "remaining_time": "3:25:54"} +{"current_steps": 2545, "total_steps": 4368, "loss": 0.3975, "lr": 1.7743126407786873e-05, "epoch": 4.078525641025641, "percentage": 58.26, "elapsed_time": "4:46:43", "remaining_time": "3:25:22"} +{"current_steps": 2550, "total_steps": 4368, "loss": 0.4412, "lr": 1.766373668694707e-05, "epoch": 4.086538461538462, "percentage": 58.38, "elapsed_time": "4:47:20", "remaining_time": "3:24:51"} +{"current_steps": 2555, "total_steps": 4368, "loss": 0.4407, "lr": 1.7584384270062195e-05, "epoch": 4.094551282051282, "percentage": 58.49, "elapsed_time": "4:47:46", "remaining_time": "3:24:11"} +{"current_steps": 2560, "total_steps": 4368, "loss": 0.4317, "lr": 1.7505070424180772e-05, "epoch": 4.102564102564102, "percentage": 58.61, "elapsed_time": "4:48:26", "remaining_time": "3:23:42"} +{"current_steps": 2565, "total_steps": 4368, "loss": 0.4379, "lr": 1.7425796415735454e-05, "epoch": 4.110576923076923, "percentage": 58.72, "elapsed_time": "4:49:02", "remaining_time": "3:23:10"} +{"current_steps": 2570, "total_steps": 4368, "loss": 0.4002, "lr": 1.7346563510522783e-05, "epoch": 4.118589743589744, "percentage": 58.84, "elapsed_time": "4:49:39", "remaining_time": "3:22:38"} +{"current_steps": 2575, "total_steps": 4368, "loss": 0.4048, "lr": 1.7267372973682998e-05, "epoch": 4.126602564102564, "percentage": 58.95, "elapsed_time": "4:50:20", "remaining_time": "3:22:10"} +{"current_steps": 2580, "total_steps": 4368, "loss": 0.4158, "lr": 1.7188226069679834e-05, "epoch": 4.134615384615385, "percentage": 59.07, "elapsed_time": "4:50:51", "remaining_time": "3:21:34"} +{"current_steps": 2585, "total_steps": 4368, "loss": 0.4418, "lr": 1.7109124062280307e-05, "epoch": 4.142628205128205, "percentage": 59.18, "elapsed_time": "4:51:20", "remaining_time": "3:20:57"} +{"current_steps": 2590, "total_steps": 4368, "loss": 0.4229, "lr": 1.7030068214534567e-05, "epoch": 4.1506410256410255, "percentage": 59.29, "elapsed_time": "4:51:51", "remaining_time": "3:20:21"} +{"current_steps": 2595, "total_steps": 4368, "loss": 0.436, "lr": 1.695105978875572e-05, "epoch": 4.158653846153846, "percentage": 59.41, "elapsed_time": "4:52:19", "remaining_time": "3:19:43"} +{"current_steps": 2600, "total_steps": 4368, "loss": 0.4057, "lr": 1.687210004649965e-05, "epoch": 4.166666666666667, "percentage": 59.52, "elapsed_time": "4:52:56", "remaining_time": "3:19:11"} +{"current_steps": 2605, "total_steps": 4368, "loss": 0.4066, "lr": 1.679319024854491e-05, "epoch": 4.174679487179487, "percentage": 59.64, "elapsed_time": "4:53:24", "remaining_time": "3:18:34"} +{"current_steps": 2610, "total_steps": 4368, "loss": 0.4466, "lr": 1.6714331654872564e-05, "epoch": 4.1826923076923075, "percentage": 59.75, "elapsed_time": "4:53:54", "remaining_time": "3:17:58"} +{"current_steps": 2615, "total_steps": 4368, "loss": 0.4036, "lr": 1.663552552464609e-05, "epoch": 4.190705128205128, "percentage": 59.87, "elapsed_time": "4:54:23", "remaining_time": "3:17:21"} +{"current_steps": 2620, "total_steps": 4368, "loss": 0.4164, "lr": 1.6556773116191257e-05, "epoch": 4.198717948717949, "percentage": 59.98, "elapsed_time": "4:55:06", "remaining_time": "3:16:53"} +{"current_steps": 2625, "total_steps": 4368, "loss": 0.4334, "lr": 1.647807568697603e-05, "epoch": 4.206730769230769, "percentage": 60.1, "elapsed_time": "4:55:31", "remaining_time": "3:16:14"} +{"current_steps": 2630, "total_steps": 4368, "loss": 0.4207, "lr": 1.6399434493590524e-05, "epoch": 4.214743589743589, "percentage": 60.21, "elapsed_time": "4:56:05", "remaining_time": "3:15:39"} +{"current_steps": 2635, "total_steps": 4368, "loss": 0.4304, "lr": 1.6320850791726884e-05, "epoch": 4.222756410256411, "percentage": 60.33, "elapsed_time": "4:56:37", "remaining_time": "3:15:05"} +{"current_steps": 2640, "total_steps": 4368, "loss": 0.4399, "lr": 1.6242325836159304e-05, "epoch": 4.230769230769231, "percentage": 60.44, "elapsed_time": "4:57:15", "remaining_time": "3:14:34"} +{"current_steps": 2645, "total_steps": 4368, "loss": 0.4276, "lr": 1.6163860880723923e-05, "epoch": 4.238782051282051, "percentage": 60.55, "elapsed_time": "4:57:50", "remaining_time": "3:14:01"} +{"current_steps": 2650, "total_steps": 4368, "loss": 0.452, "lr": 1.6085457178298866e-05, "epoch": 4.246794871794872, "percentage": 60.67, "elapsed_time": "4:58:21", "remaining_time": "3:13:25"} +{"current_steps": 2655, "total_steps": 4368, "loss": 0.429, "lr": 1.6007115980784182e-05, "epoch": 4.2548076923076925, "percentage": 60.78, "elapsed_time": "4:58:59", "remaining_time": "3:12:54"} +{"current_steps": 2660, "total_steps": 4368, "loss": 0.4215, "lr": 1.592883853908188e-05, "epoch": 4.262820512820513, "percentage": 60.9, "elapsed_time": "4:59:31", "remaining_time": "3:12:19"} +{"current_steps": 2665, "total_steps": 4368, "loss": 0.4247, "lr": 1.585062610307599e-05, "epoch": 4.270833333333333, "percentage": 61.01, "elapsed_time": "5:00:08", "remaining_time": "3:11:47"} +{"current_steps": 2670, "total_steps": 4368, "loss": 0.3991, "lr": 1.5772479921612543e-05, "epoch": 4.278846153846154, "percentage": 61.13, "elapsed_time": "5:00:49", "remaining_time": "3:11:18"} +{"current_steps": 2675, "total_steps": 4368, "loss": 0.4286, "lr": 1.5694401242479677e-05, "epoch": 4.2868589743589745, "percentage": 61.24, "elapsed_time": "5:01:18", "remaining_time": "3:10:41"} +{"current_steps": 2680, "total_steps": 4368, "loss": 0.409, "lr": 1.5616391312387683e-05, "epoch": 4.294871794871795, "percentage": 61.36, "elapsed_time": "5:01:53", "remaining_time": "3:10:08"} +{"current_steps": 2685, "total_steps": 4368, "loss": 0.4099, "lr": 1.5538451376949106e-05, "epoch": 4.302884615384615, "percentage": 61.47, "elapsed_time": "5:02:33", "remaining_time": "3:09:38"} +{"current_steps": 2690, "total_steps": 4368, "loss": 0.422, "lr": 1.5460582680658888e-05, "epoch": 4.310897435897436, "percentage": 61.58, "elapsed_time": "5:03:10", "remaining_time": "3:09:06"} +{"current_steps": 2695, "total_steps": 4368, "loss": 0.4066, "lr": 1.5382786466874446e-05, "epoch": 4.318910256410256, "percentage": 61.7, "elapsed_time": "5:03:44", "remaining_time": "3:08:33"} +{"current_steps": 2700, "total_steps": 4368, "loss": 0.3739, "lr": 1.5305063977795856e-05, "epoch": 4.326923076923077, "percentage": 61.81, "elapsed_time": "5:04:19", "remaining_time": "3:08:00"} +{"current_steps": 2705, "total_steps": 4368, "loss": 0.4354, "lr": 1.5227416454445995e-05, "epoch": 4.334935897435898, "percentage": 61.93, "elapsed_time": "5:04:54", "remaining_time": "3:07:27"} +{"current_steps": 2710, "total_steps": 4368, "loss": 0.4251, "lr": 1.5149845136650748e-05, "epoch": 4.342948717948718, "percentage": 62.04, "elapsed_time": "5:05:26", "remaining_time": "3:06:52"} +{"current_steps": 2715, "total_steps": 4368, "loss": 0.4211, "lr": 1.5072351263019177e-05, "epoch": 4.350961538461538, "percentage": 62.16, "elapsed_time": "5:05:56", "remaining_time": "3:06:15"} +{"current_steps": 2720, "total_steps": 4368, "loss": 0.4328, "lr": 1.4994936070923784e-05, "epoch": 4.358974358974359, "percentage": 62.27, "elapsed_time": "5:06:26", "remaining_time": "3:05:39"} +{"current_steps": 2725, "total_steps": 4368, "loss": 0.4001, "lr": 1.4917600796480745e-05, "epoch": 4.36698717948718, "percentage": 62.39, "elapsed_time": "5:06:53", "remaining_time": "3:05:02"} +{"current_steps": 2730, "total_steps": 4368, "loss": 0.3949, "lr": 1.4840346674530122e-05, "epoch": 4.375, "percentage": 62.5, "elapsed_time": "5:07:35", "remaining_time": "3:04:33"} +{"current_steps": 2735, "total_steps": 4368, "loss": 0.415, "lr": 1.4763174938616232e-05, "epoch": 4.38301282051282, "percentage": 62.61, "elapsed_time": "5:08:13", "remaining_time": "3:04:01"} +{"current_steps": 2740, "total_steps": 4368, "loss": 0.4093, "lr": 1.4686086820967865e-05, "epoch": 4.391025641025641, "percentage": 62.73, "elapsed_time": "5:08:44", "remaining_time": "3:03:26"} +{"current_steps": 2745, "total_steps": 4368, "loss": 0.4401, "lr": 1.460908355247868e-05, "epoch": 4.399038461538462, "percentage": 62.84, "elapsed_time": "5:09:12", "remaining_time": "3:02:48"} +{"current_steps": 2750, "total_steps": 4368, "loss": 0.43, "lr": 1.4532166362687507e-05, "epoch": 4.407051282051282, "percentage": 62.96, "elapsed_time": "5:09:52", "remaining_time": "3:02:19"} +{"current_steps": 2755, "total_steps": 4368, "loss": 0.4013, "lr": 1.445533647975871e-05, "epoch": 4.415064102564102, "percentage": 63.07, "elapsed_time": "5:10:33", "remaining_time": "3:01:49"} +{"current_steps": 2760, "total_steps": 4368, "loss": 0.4162, "lr": 1.437859513046263e-05, "epoch": 4.423076923076923, "percentage": 63.19, "elapsed_time": "5:11:04", "remaining_time": "3:01:13"} +{"current_steps": 2765, "total_steps": 4368, "loss": 0.4158, "lr": 1.4301943540155914e-05, "epoch": 4.431089743589744, "percentage": 63.3, "elapsed_time": "5:11:38", "remaining_time": "3:00:40"} +{"current_steps": 2770, "total_steps": 4368, "loss": 0.4465, "lr": 1.4225382932762033e-05, "epoch": 4.439102564102564, "percentage": 63.42, "elapsed_time": "5:12:02", "remaining_time": "3:00:00"} +{"current_steps": 2775, "total_steps": 4368, "loss": 0.412, "lr": 1.4148914530751681e-05, "epoch": 4.447115384615385, "percentage": 63.53, "elapsed_time": "5:12:27", "remaining_time": "2:59:22"} +{"current_steps": 2780, "total_steps": 4368, "loss": 0.4245, "lr": 1.4072539555123292e-05, "epoch": 4.455128205128205, "percentage": 63.64, "elapsed_time": "5:12:58", "remaining_time": "2:58:46"} +{"current_steps": 2785, "total_steps": 4368, "loss": 0.4151, "lr": 1.3996259225383514e-05, "epoch": 4.4631410256410255, "percentage": 63.76, "elapsed_time": "5:13:30", "remaining_time": "2:58:11"} +{"current_steps": 2790, "total_steps": 4368, "loss": 0.4122, "lr": 1.3920074759527737e-05, "epoch": 4.471153846153846, "percentage": 63.87, "elapsed_time": "5:14:00", "remaining_time": "2:57:36"} +{"current_steps": 2795, "total_steps": 4368, "loss": 0.4642, "lr": 1.3843987374020689e-05, "epoch": 4.479166666666667, "percentage": 63.99, "elapsed_time": "5:14:38", "remaining_time": "2:57:04"} +{"current_steps": 2800, "total_steps": 4368, "loss": 0.4063, "lr": 1.376799828377696e-05, "epoch": 4.487179487179487, "percentage": 64.1, "elapsed_time": "5:15:10", "remaining_time": "2:56:29"} +{"current_steps": 2805, "total_steps": 4368, "loss": 0.439, "lr": 1.3692108702141642e-05, "epoch": 4.4951923076923075, "percentage": 64.22, "elapsed_time": "5:15:41", "remaining_time": "2:55:54"} +{"current_steps": 2810, "total_steps": 4368, "loss": 0.4091, "lr": 1.361631984087091e-05, "epoch": 4.503205128205128, "percentage": 64.33, "elapsed_time": "5:16:15", "remaining_time": "2:55:21"} +{"current_steps": 2815, "total_steps": 4368, "loss": 0.4402, "lr": 1.354063291011273e-05, "epoch": 4.511217948717949, "percentage": 64.45, "elapsed_time": "5:16:46", "remaining_time": "2:54:45"} +{"current_steps": 2820, "total_steps": 4368, "loss": 0.4303, "lr": 1.3465049118387486e-05, "epoch": 4.519230769230769, "percentage": 64.56, "elapsed_time": "5:17:15", "remaining_time": "2:54:09"} +{"current_steps": 2825, "total_steps": 4368, "loss": 0.4162, "lr": 1.3389569672568707e-05, "epoch": 4.527243589743589, "percentage": 64.67, "elapsed_time": "5:17:45", "remaining_time": "2:53:33"} +{"current_steps": 2830, "total_steps": 4368, "loss": 0.4221, "lr": 1.331419577786381e-05, "epoch": 4.535256410256411, "percentage": 64.79, "elapsed_time": "5:18:14", "remaining_time": "2:52:57"} +{"current_steps": 2835, "total_steps": 4368, "loss": 0.4486, "lr": 1.3238928637794816e-05, "epoch": 4.543269230769231, "percentage": 64.9, "elapsed_time": "5:18:42", "remaining_time": "2:52:20"} +{"current_steps": 2840, "total_steps": 4368, "loss": 0.41, "lr": 1.3163769454179183e-05, "epoch": 4.551282051282051, "percentage": 65.02, "elapsed_time": "5:19:12", "remaining_time": "2:51:44"} +{"current_steps": 2845, "total_steps": 4368, "loss": 0.413, "lr": 1.3088719427110552e-05, "epoch": 4.559294871794872, "percentage": 65.13, "elapsed_time": "5:19:41", "remaining_time": "2:51:08"} +{"current_steps": 2850, "total_steps": 4368, "loss": 0.423, "lr": 1.3013779754939666e-05, "epoch": 4.5673076923076925, "percentage": 65.25, "elapsed_time": "5:20:18", "remaining_time": "2:50:36"} +{"current_steps": 2855, "total_steps": 4368, "loss": 0.3971, "lr": 1.2938951634255164e-05, "epoch": 4.575320512820513, "percentage": 65.36, "elapsed_time": "5:20:49", "remaining_time": "2:50:01"} +{"current_steps": 2860, "total_steps": 4368, "loss": 0.4465, "lr": 1.2864236259864495e-05, "epoch": 4.583333333333333, "percentage": 65.48, "elapsed_time": "5:21:14", "remaining_time": "2:49:22"} +{"current_steps": 2865, "total_steps": 4368, "loss": 0.4301, "lr": 1.2789634824774887e-05, "epoch": 4.591346153846154, "percentage": 65.59, "elapsed_time": "5:21:45", "remaining_time": "2:48:47"} +{"current_steps": 2870, "total_steps": 4368, "loss": 0.4478, "lr": 1.2715148520174206e-05, "epoch": 4.5993589743589745, "percentage": 65.71, "elapsed_time": "5:22:14", "remaining_time": "2:48:11"} +{"current_steps": 2875, "total_steps": 4368, "loss": 0.433, "lr": 1.2640778535412036e-05, "epoch": 4.607371794871795, "percentage": 65.82, "elapsed_time": "5:22:51", "remaining_time": "2:47:39"} +{"current_steps": 2880, "total_steps": 4368, "loss": 0.4042, "lr": 1.2566526057980608e-05, "epoch": 4.615384615384615, "percentage": 65.93, "elapsed_time": "5:23:21", "remaining_time": "2:47:04"} +{"current_steps": 2885, "total_steps": 4368, "loss": 0.4385, "lr": 1.2492392273495879e-05, "epoch": 4.623397435897436, "percentage": 66.05, "elapsed_time": "5:23:57", "remaining_time": "2:46:31"} +{"current_steps": 2890, "total_steps": 4368, "loss": 0.4306, "lr": 1.2418378365678612e-05, "epoch": 4.631410256410256, "percentage": 66.16, "elapsed_time": "5:24:27", "remaining_time": "2:45:56"} +{"current_steps": 2895, "total_steps": 4368, "loss": 0.4431, "lr": 1.234448551633542e-05, "epoch": 4.639423076923077, "percentage": 66.28, "elapsed_time": "5:24:56", "remaining_time": "2:45:20"} +{"current_steps": 2900, "total_steps": 4368, "loss": 0.4445, "lr": 1.2270714905339969e-05, "epoch": 4.647435897435898, "percentage": 66.39, "elapsed_time": "5:25:29", "remaining_time": "2:44:46"} +{"current_steps": 2905, "total_steps": 4368, "loss": 0.429, "lr": 1.2197067710614075e-05, "epoch": 4.655448717948718, "percentage": 66.51, "elapsed_time": "5:26:05", "remaining_time": "2:44:13"} +{"current_steps": 2910, "total_steps": 4368, "loss": 0.4321, "lr": 1.2123545108108943e-05, "epoch": 4.663461538461538, "percentage": 66.62, "elapsed_time": "5:26:35", "remaining_time": "2:43:37"} +{"current_steps": 2915, "total_steps": 4368, "loss": 0.4267, "lr": 1.2050148271786348e-05, "epoch": 4.671474358974359, "percentage": 66.74, "elapsed_time": "5:27:15", "remaining_time": "2:43:07"} +{"current_steps": 2920, "total_steps": 4368, "loss": 0.417, "lr": 1.1976878373599928e-05, "epoch": 4.67948717948718, "percentage": 66.85, "elapsed_time": "5:27:42", "remaining_time": "2:42:30"} +{"current_steps": 2925, "total_steps": 4368, "loss": 0.4342, "lr": 1.1903736583476441e-05, "epoch": 4.6875, "percentage": 66.96, "elapsed_time": "5:28:17", "remaining_time": "2:41:57"} +{"current_steps": 2930, "total_steps": 4368, "loss": 0.4267, "lr": 1.1830724069297106e-05, "epoch": 4.69551282051282, "percentage": 67.08, "elapsed_time": "5:28:47", "remaining_time": "2:41:21"} +{"current_steps": 2935, "total_steps": 4368, "loss": 0.4266, "lr": 1.1757841996878957e-05, "epoch": 4.703525641025641, "percentage": 67.19, "elapsed_time": "5:29:24", "remaining_time": "2:40:50"} +{"current_steps": 2940, "total_steps": 4368, "loss": 0.44, "lr": 1.1685091529956187e-05, "epoch": 4.711538461538462, "percentage": 67.31, "elapsed_time": "5:30:00", "remaining_time": "2:40:17"} +{"current_steps": 2945, "total_steps": 4368, "loss": 0.4352, "lr": 1.161247383016163e-05, "epoch": 4.719551282051282, "percentage": 67.42, "elapsed_time": "5:30:30", "remaining_time": "2:39:41"} +{"current_steps": 2950, "total_steps": 4368, "loss": 0.4156, "lr": 1.1539990057008166e-05, "epoch": 4.727564102564102, "percentage": 67.54, "elapsed_time": "5:30:59", "remaining_time": "2:39:06"} +{"current_steps": 2955, "total_steps": 4368, "loss": 0.4371, "lr": 1.1467641367870198e-05, "epoch": 4.735576923076923, "percentage": 67.65, "elapsed_time": "5:31:30", "remaining_time": "2:38:30"} +{"current_steps": 2960, "total_steps": 4368, "loss": 0.4445, "lr": 1.1395428917965239e-05, "epoch": 4.743589743589744, "percentage": 67.77, "elapsed_time": "5:32:05", "remaining_time": "2:37:57"} +{"current_steps": 2965, "total_steps": 4368, "loss": 0.4272, "lr": 1.1323353860335385e-05, "epoch": 4.751602564102564, "percentage": 67.88, "elapsed_time": "5:32:36", "remaining_time": "2:37:23"} +{"current_steps": 2970, "total_steps": 4368, "loss": 0.4242, "lr": 1.1251417345828962e-05, "epoch": 4.759615384615385, "percentage": 67.99, "elapsed_time": "5:33:09", "remaining_time": "2:36:49"} +{"current_steps": 2975, "total_steps": 4368, "loss": 0.3973, "lr": 1.1179620523082107e-05, "epoch": 4.767628205128205, "percentage": 68.11, "elapsed_time": "5:33:45", "remaining_time": "2:36:16"} +{"current_steps": 2980, "total_steps": 4368, "loss": 0.4182, "lr": 1.110796453850047e-05, "epoch": 4.7756410256410255, "percentage": 68.22, "elapsed_time": "5:34:24", "remaining_time": "2:35:45"} +{"current_steps": 2985, "total_steps": 4368, "loss": 0.4096, "lr": 1.1036450536240877e-05, "epoch": 4.783653846153846, "percentage": 68.34, "elapsed_time": "5:34:54", "remaining_time": "2:35:10"} +{"current_steps": 2990, "total_steps": 4368, "loss": 0.4019, "lr": 1.0965079658193068e-05, "epoch": 4.791666666666667, "percentage": 68.45, "elapsed_time": "5:35:25", "remaining_time": "2:34:35"} +{"current_steps": 2995, "total_steps": 4368, "loss": 0.4373, "lr": 1.0893853043961475e-05, "epoch": 4.799679487179487, "percentage": 68.57, "elapsed_time": "5:35:58", "remaining_time": "2:34:01"} +{"current_steps": 3000, "total_steps": 4368, "loss": 0.4413, "lr": 1.0822771830847011e-05, "epoch": 4.8076923076923075, "percentage": 68.68, "elapsed_time": "5:36:35", "remaining_time": "2:33:29"} +{"current_steps": 3005, "total_steps": 4368, "loss": 0.4358, "lr": 1.0751837153828926e-05, "epoch": 4.815705128205128, "percentage": 68.8, "elapsed_time": "5:37:46", "remaining_time": "2:33:12"} +{"current_steps": 3010, "total_steps": 4368, "loss": 0.4125, "lr": 1.0681050145546666e-05, "epoch": 4.823717948717949, "percentage": 68.91, "elapsed_time": "5:38:21", "remaining_time": "2:32:39"} +{"current_steps": 3015, "total_steps": 4368, "loss": 0.4151, "lr": 1.0610411936281801e-05, "epoch": 4.831730769230769, "percentage": 69.02, "elapsed_time": "5:38:58", "remaining_time": "2:32:07"} +{"current_steps": 3020, "total_steps": 4368, "loss": 0.433, "lr": 1.0539923653939978e-05, "epoch": 4.839743589743589, "percentage": 69.14, "elapsed_time": "5:39:33", "remaining_time": "2:31:33"} +{"current_steps": 3025, "total_steps": 4368, "loss": 0.4063, "lr": 1.0469586424032903e-05, "epoch": 4.847756410256411, "percentage": 69.25, "elapsed_time": "5:40:14", "remaining_time": "2:31:03"} +{"current_steps": 3030, "total_steps": 4368, "loss": 0.4201, "lr": 1.0399401369660369e-05, "epoch": 4.855769230769231, "percentage": 69.37, "elapsed_time": "5:40:48", "remaining_time": "2:30:29"} +{"current_steps": 3035, "total_steps": 4368, "loss": 0.4508, "lr": 1.0329369611492334e-05, "epoch": 4.863782051282051, "percentage": 69.48, "elapsed_time": "5:41:13", "remaining_time": "2:29:52"} +{"current_steps": 3040, "total_steps": 4368, "loss": 0.4179, "lr": 1.0259492267751022e-05, "epoch": 4.871794871794872, "percentage": 69.6, "elapsed_time": "5:41:45", "remaining_time": "2:29:17"} +{"current_steps": 3045, "total_steps": 4368, "loss": 0.4212, "lr": 1.0189770454193052e-05, "epoch": 4.8798076923076925, "percentage": 69.71, "elapsed_time": "5:42:18", "remaining_time": "2:28:43"} +{"current_steps": 3050, "total_steps": 4368, "loss": 0.4287, "lr": 1.0120205284091673e-05, "epoch": 4.887820512820513, "percentage": 69.83, "elapsed_time": "5:42:48", "remaining_time": "2:28:08"} +{"current_steps": 3055, "total_steps": 4368, "loss": 0.3922, "lr": 1.0050797868218907e-05, "epoch": 4.895833333333333, "percentage": 69.94, "elapsed_time": "5:43:21", "remaining_time": "2:27:34"} +{"current_steps": 3060, "total_steps": 4368, "loss": 0.3927, "lr": 9.981549314827876e-06, "epoch": 4.903846153846154, "percentage": 70.05, "elapsed_time": "5:43:50", "remaining_time": "2:26:58"} +{"current_steps": 3065, "total_steps": 4368, "loss": 0.4313, "lr": 9.912460729635097e-06, "epoch": 4.9118589743589745, "percentage": 70.17, "elapsed_time": "5:44:27", "remaining_time": "2:26:26"} +{"current_steps": 3070, "total_steps": 4368, "loss": 0.4036, "lr": 9.843533215802796e-06, "epoch": 4.919871794871795, "percentage": 70.28, "elapsed_time": "5:45:02", "remaining_time": "2:25:52"} +{"current_steps": 3075, "total_steps": 4368, "loss": 0.4215, "lr": 9.774767873921357e-06, "epoch": 4.927884615384615, "percentage": 70.4, "elapsed_time": "5:45:39", "remaining_time": "2:25:20"} +{"current_steps": 3080, "total_steps": 4368, "loss": 0.4535, "lr": 9.706165801991651e-06, "epoch": 4.935897435897436, "percentage": 70.51, "elapsed_time": "5:46:15", "remaining_time": "2:24:47"} +{"current_steps": 3085, "total_steps": 4368, "loss": 0.4379, "lr": 9.637728095407593e-06, "epoch": 4.943910256410256, "percentage": 70.63, "elapsed_time": "5:46:44", "remaining_time": "2:24:12"} +{"current_steps": 3090, "total_steps": 4368, "loss": 0.4411, "lr": 9.56945584693861e-06, "epoch": 4.951923076923077, "percentage": 70.74, "elapsed_time": "5:47:08", "remaining_time": "2:23:34"} +{"current_steps": 3095, "total_steps": 4368, "loss": 0.3867, "lr": 9.501350146712193e-06, "epoch": 4.959935897435898, "percentage": 70.86, "elapsed_time": "5:47:40", "remaining_time": "2:23:00"} +{"current_steps": 3100, "total_steps": 4368, "loss": 0.4242, "lr": 9.433412082196527e-06, "epoch": 4.967948717948718, "percentage": 70.97, "elapsed_time": "5:48:17", "remaining_time": "2:22:27"} +{"current_steps": 3105, "total_steps": 4368, "loss": 0.4166, "lr": 9.365642738183044e-06, "epoch": 4.975961538461538, "percentage": 71.09, "elapsed_time": "5:48:55", "remaining_time": "2:21:55"} +{"current_steps": 3110, "total_steps": 4368, "loss": 0.4291, "lr": 9.298043196769217e-06, "epoch": 4.983974358974359, "percentage": 71.2, "elapsed_time": "5:49:26", "remaining_time": "2:21:21"} +{"current_steps": 3115, "total_steps": 4368, "loss": 0.4216, "lr": 9.230614537341167e-06, "epoch": 4.99198717948718, "percentage": 71.31, "elapsed_time": "5:50:02", "remaining_time": "2:20:48"} +{"current_steps": 3120, "total_steps": 4368, "loss": 0.4128, "lr": 9.163357836556498e-06, "epoch": 5.0, "percentage": 71.43, "elapsed_time": "5:50:31", "remaining_time": "2:20:12"} +{"current_steps": 3125, "total_steps": 4368, "loss": 0.3801, "lr": 9.096274168327122e-06, "epoch": 5.00801282051282, "percentage": 71.54, "elapsed_time": "5:51:06", "remaining_time": "2:19:39"} +{"current_steps": 3130, "total_steps": 4368, "loss": 0.3865, "lr": 9.029364603802017e-06, "epoch": 5.016025641025641, "percentage": 71.66, "elapsed_time": "5:51:36", "remaining_time": "2:19:04"} +{"current_steps": 3135, "total_steps": 4368, "loss": 0.3906, "lr": 8.962630211350248e-06, "epoch": 5.024038461538462, "percentage": 71.77, "elapsed_time": "5:52:06", "remaining_time": "2:18:29"} +{"current_steps": 3140, "total_steps": 4368, "loss": 0.4028, "lr": 8.89607205654378e-06, "epoch": 5.032051282051282, "percentage": 71.89, "elapsed_time": "5:52:38", "remaining_time": "2:17:54"} +{"current_steps": 3145, "total_steps": 4368, "loss": 0.4019, "lr": 8.829691202140591e-06, "epoch": 5.040064102564102, "percentage": 72.0, "elapsed_time": "5:53:18", "remaining_time": "2:17:23"} +{"current_steps": 3150, "total_steps": 4368, "loss": 0.3897, "lr": 8.763488708067604e-06, "epoch": 5.048076923076923, "percentage": 72.12, "elapsed_time": "5:53:50", "remaining_time": "2:16:49"} +{"current_steps": 3155, "total_steps": 4368, "loss": 0.4263, "lr": 8.69746563140379e-06, "epoch": 5.056089743589744, "percentage": 72.23, "elapsed_time": "5:54:26", "remaining_time": "2:16:16"} +{"current_steps": 3160, "total_steps": 4368, "loss": 0.4153, "lr": 8.631623026363331e-06, "epoch": 5.064102564102564, "percentage": 72.34, "elapsed_time": "5:54:58", "remaining_time": "2:15:42"} +{"current_steps": 3165, "total_steps": 4368, "loss": 0.4085, "lr": 8.56596194427873e-06, "epoch": 5.072115384615385, "percentage": 72.46, "elapsed_time": "5:55:33", "remaining_time": "2:15:08"} +{"current_steps": 3170, "total_steps": 4368, "loss": 0.4467, "lr": 8.500483433584054e-06, "epoch": 5.080128205128205, "percentage": 72.57, "elapsed_time": "5:56:09", "remaining_time": "2:14:35"} +{"current_steps": 3175, "total_steps": 4368, "loss": 0.4102, "lr": 8.435188539798187e-06, "epoch": 5.0881410256410255, "percentage": 72.69, "elapsed_time": "5:56:41", "remaining_time": "2:14:01"} +{"current_steps": 3180, "total_steps": 4368, "loss": 0.4244, "lr": 8.370078305508136e-06, "epoch": 5.096153846153846, "percentage": 72.8, "elapsed_time": "5:57:15", "remaining_time": "2:13:28"} +{"current_steps": 3185, "total_steps": 4368, "loss": 0.3796, "lr": 8.305153770352384e-06, "epoch": 5.104166666666667, "percentage": 72.92, "elapsed_time": "5:57:47", "remaining_time": "2:12:53"} +{"current_steps": 3190, "total_steps": 4368, "loss": 0.4005, "lr": 8.240415971004285e-06, "epoch": 5.112179487179487, "percentage": 73.03, "elapsed_time": "5:58:18", "remaining_time": "2:12:18"} +{"current_steps": 3195, "total_steps": 4368, "loss": 0.3831, "lr": 8.175865941155525e-06, "epoch": 5.1201923076923075, "percentage": 73.15, "elapsed_time": "5:58:50", "remaining_time": "2:11:44"} +{"current_steps": 3200, "total_steps": 4368, "loss": 0.4089, "lr": 8.111504711499598e-06, "epoch": 5.128205128205128, "percentage": 73.26, "elapsed_time": "5:59:25", "remaining_time": "2:11:11"} +{"current_steps": 3205, "total_steps": 4368, "loss": 0.3741, "lr": 8.04733330971536e-06, "epoch": 5.136217948717949, "percentage": 73.37, "elapsed_time": "5:59:53", "remaining_time": "2:10:35"} +{"current_steps": 3210, "total_steps": 4368, "loss": 0.4007, "lr": 7.983352760450618e-06, "epoch": 5.144230769230769, "percentage": 73.49, "elapsed_time": "6:00:28", "remaining_time": "2:10:02"} +{"current_steps": 3215, "total_steps": 4368, "loss": 0.41, "lr": 7.919564085305768e-06, "epoch": 5.152243589743589, "percentage": 73.6, "elapsed_time": "6:01:05", "remaining_time": "2:09:29"} +{"current_steps": 3220, "total_steps": 4368, "loss": 0.3867, "lr": 7.855968302817487e-06, "epoch": 5.160256410256411, "percentage": 73.72, "elapsed_time": "6:01:36", "remaining_time": "2:08:55"} +{"current_steps": 3225, "total_steps": 4368, "loss": 0.4052, "lr": 7.792566428442456e-06, "epoch": 5.168269230769231, "percentage": 73.83, "elapsed_time": "6:02:09", "remaining_time": "2:08:21"} +{"current_steps": 3230, "total_steps": 4368, "loss": 0.4287, "lr": 7.729359474541168e-06, "epoch": 5.176282051282051, "percentage": 73.95, "elapsed_time": "6:02:40", "remaining_time": "2:07:46"} +{"current_steps": 3235, "total_steps": 4368, "loss": 0.4148, "lr": 7.666348450361737e-06, "epoch": 5.184294871794872, "percentage": 74.06, "elapsed_time": "6:03:17", "remaining_time": "2:07:14"} +{"current_steps": 3240, "total_steps": 4368, "loss": 0.3988, "lr": 7.60353436202381e-06, "epoch": 5.1923076923076925, "percentage": 74.18, "elapsed_time": "6:03:52", "remaining_time": "2:06:41"} +{"current_steps": 3245, "total_steps": 4368, "loss": 0.3828, "lr": 7.540918212502479e-06, "epoch": 5.200320512820513, "percentage": 74.29, "elapsed_time": "6:04:22", "remaining_time": "2:06:06"} +{"current_steps": 3250, "total_steps": 4368, "loss": 0.4274, "lr": 7.478501001612281e-06, "epoch": 5.208333333333333, "percentage": 74.4, "elapsed_time": "6:04:58", "remaining_time": "2:05:32"} +{"current_steps": 3255, "total_steps": 4368, "loss": 0.4046, "lr": 7.416283725991229e-06, "epoch": 5.216346153846154, "percentage": 74.52, "elapsed_time": "6:05:31", "remaining_time": "2:04:59"} +{"current_steps": 3260, "total_steps": 4368, "loss": 0.376, "lr": 7.354267379084896e-06, "epoch": 5.2243589743589745, "percentage": 74.63, "elapsed_time": "6:06:06", "remaining_time": "2:04:25"} +{"current_steps": 3265, "total_steps": 4368, "loss": 0.4233, "lr": 7.292452951130548e-06, "epoch": 5.232371794871795, "percentage": 74.75, "elapsed_time": "6:06:37", "remaining_time": "2:03:51"} +{"current_steps": 3270, "total_steps": 4368, "loss": 0.3969, "lr": 7.230841429141347e-06, "epoch": 5.240384615384615, "percentage": 74.86, "elapsed_time": "6:07:13", "remaining_time": "2:03:18"} +{"current_steps": 3275, "total_steps": 4368, "loss": 0.4104, "lr": 7.169433796890595e-06, "epoch": 5.248397435897436, "percentage": 74.98, "elapsed_time": "6:07:48", "remaining_time": "2:02:45"} +{"current_steps": 3280, "total_steps": 4368, "loss": 0.4074, "lr": 7.108231034895976e-06, "epoch": 5.256410256410256, "percentage": 75.09, "elapsed_time": "6:08:26", "remaining_time": "2:02:12"} +{"current_steps": 3285, "total_steps": 4368, "loss": 0.3838, "lr": 7.047234120403972e-06, "epoch": 5.264423076923077, "percentage": 75.21, "elapsed_time": "6:09:02", "remaining_time": "2:01:40"} +{"current_steps": 3290, "total_steps": 4368, "loss": 0.4132, "lr": 6.986444027374211e-06, "epoch": 5.272435897435898, "percentage": 75.32, "elapsed_time": "6:09:32", "remaining_time": "2:01:04"} +{"current_steps": 3295, "total_steps": 4368, "loss": 0.3892, "lr": 6.925861726463919e-06, "epoch": 5.280448717948718, "percentage": 75.43, "elapsed_time": "6:10:05", "remaining_time": "2:00:31"} +{"current_steps": 3300, "total_steps": 4368, "loss": 0.3923, "lr": 6.865488185012464e-06, "epoch": 5.288461538461538, "percentage": 75.55, "elapsed_time": "6:10:39", "remaining_time": "1:59:57"} +{"current_steps": 3305, "total_steps": 4368, "loss": 0.4348, "lr": 6.805324367025825e-06, "epoch": 5.296474358974359, "percentage": 75.66, "elapsed_time": "6:11:10", "remaining_time": "1:59:22"} +{"current_steps": 3310, "total_steps": 4368, "loss": 0.4154, "lr": 6.745371233161309e-06, "epoch": 5.30448717948718, "percentage": 75.78, "elapsed_time": "6:11:51", "remaining_time": "1:58:51"} +{"current_steps": 3315, "total_steps": 4368, "loss": 0.4033, "lr": 6.685629740712103e-06, "epoch": 5.3125, "percentage": 75.89, "elapsed_time": "6:12:26", "remaining_time": "1:58:18"} +{"current_steps": 3320, "total_steps": 4368, "loss": 0.4305, "lr": 6.6261008435920605e-06, "epoch": 5.32051282051282, "percentage": 76.01, "elapsed_time": "6:13:00", "remaining_time": "1:57:44"} +{"current_steps": 3325, "total_steps": 4368, "loss": 0.3946, "lr": 6.566785492320471e-06, "epoch": 5.328525641025641, "percentage": 76.12, "elapsed_time": "6:13:35", "remaining_time": "1:57:11"} +{"current_steps": 3330, "total_steps": 4368, "loss": 0.4381, "lr": 6.507684634006815e-06, "epoch": 5.336538461538462, "percentage": 76.24, "elapsed_time": "6:14:11", "remaining_time": "1:56:38"} +{"current_steps": 3335, "total_steps": 4368, "loss": 0.3715, "lr": 6.448799212335734e-06, "epoch": 5.344551282051282, "percentage": 76.35, "elapsed_time": "6:14:46", "remaining_time": "1:56:04"} +{"current_steps": 3340, "total_steps": 4368, "loss": 0.4279, "lr": 6.390130167551869e-06, "epoch": 5.352564102564102, "percentage": 76.47, "elapsed_time": "6:15:13", "remaining_time": "1:55:29"} +{"current_steps": 3345, "total_steps": 4368, "loss": 0.4022, "lr": 6.331678436444939e-06, "epoch": 5.360576923076923, "percentage": 76.58, "elapsed_time": "6:15:44", "remaining_time": "1:54:54"} +{"current_steps": 3350, "total_steps": 4368, "loss": 0.3923, "lr": 6.273444952334713e-06, "epoch": 5.368589743589744, "percentage": 76.69, "elapsed_time": "6:16:16", "remaining_time": "1:54:20"} +{"current_steps": 3355, "total_steps": 4368, "loss": 0.4087, "lr": 6.2154306450561175e-06, "epoch": 5.376602564102564, "percentage": 76.81, "elapsed_time": "6:16:47", "remaining_time": "1:53:45"} +{"current_steps": 3360, "total_steps": 4368, "loss": 0.4346, "lr": 6.157636440944445e-06, "epoch": 5.384615384615385, "percentage": 76.92, "elapsed_time": "6:17:13", "remaining_time": "1:53:10"} +{"current_steps": 3365, "total_steps": 4368, "loss": 0.416, "lr": 6.100063262820474e-06, "epoch": 5.392628205128205, "percentage": 77.04, "elapsed_time": "6:17:44", "remaining_time": "1:52:35"} +{"current_steps": 3370, "total_steps": 4368, "loss": 0.4525, "lr": 6.0427120299758236e-06, "epoch": 5.4006410256410255, "percentage": 77.15, "elapsed_time": "6:18:12", "remaining_time": "1:52:00"} +{"current_steps": 3375, "total_steps": 4368, "loss": 0.41, "lr": 5.985583658158212e-06, "epoch": 5.408653846153846, "percentage": 77.27, "elapsed_time": "6:18:47", "remaining_time": "1:51:26"} +{"current_steps": 3380, "total_steps": 4368, "loss": 0.3973, "lr": 5.928679059556852e-06, "epoch": 5.416666666666667, "percentage": 77.38, "elapsed_time": "6:19:17", "remaining_time": "1:50:52"} +{"current_steps": 3385, "total_steps": 4368, "loss": 0.3936, "lr": 5.871999142787908e-06, "epoch": 5.424679487179487, "percentage": 77.5, "elapsed_time": "6:19:54", "remaining_time": "1:50:19"} +{"current_steps": 3390, "total_steps": 4368, "loss": 0.4003, "lr": 5.815544812879936e-06, "epoch": 5.4326923076923075, "percentage": 77.61, "elapsed_time": "6:20:27", "remaining_time": "1:49:45"} +{"current_steps": 3395, "total_steps": 4368, "loss": 0.4045, "lr": 5.759316971259503e-06, "epoch": 5.440705128205128, "percentage": 77.72, "elapsed_time": "6:20:58", "remaining_time": "1:49:11"} +{"current_steps": 3400, "total_steps": 4368, "loss": 0.4018, "lr": 5.703316515736734e-06, "epoch": 5.448717948717949, "percentage": 77.84, "elapsed_time": "6:21:33", "remaining_time": "1:48:38"} +{"current_steps": 3405, "total_steps": 4368, "loss": 0.3746, "lr": 5.647544340491007e-06, "epoch": 5.456730769230769, "percentage": 77.95, "elapsed_time": "6:22:10", "remaining_time": "1:48:05"} +{"current_steps": 3410, "total_steps": 4368, "loss": 0.4007, "lr": 5.592001336056659e-06, "epoch": 5.464743589743589, "percentage": 78.07, "elapsed_time": "6:22:41", "remaining_time": "1:47:30"} +{"current_steps": 3415, "total_steps": 4368, "loss": 0.4017, "lr": 5.536688389308782e-06, "epoch": 5.472756410256411, "percentage": 78.18, "elapsed_time": "6:23:13", "remaining_time": "1:46:56"} +{"current_steps": 3420, "total_steps": 4368, "loss": 0.4164, "lr": 5.4816063834490496e-06, "epoch": 5.480769230769231, "percentage": 78.3, "elapsed_time": "6:23:53", "remaining_time": "1:46:24"} +{"current_steps": 3425, "total_steps": 4368, "loss": 0.4074, "lr": 5.426756197991625e-06, "epoch": 5.488782051282051, "percentage": 78.41, "elapsed_time": "6:24:27", "remaining_time": "1:45:51"} +{"current_steps": 3430, "total_steps": 4368, "loss": 0.3879, "lr": 5.372138708749104e-06, "epoch": 5.496794871794872, "percentage": 78.53, "elapsed_time": "6:25:00", "remaining_time": "1:45:17"} +{"current_steps": 3435, "total_steps": 4368, "loss": 0.3736, "lr": 5.3177547878185436e-06, "epoch": 5.5048076923076925, "percentage": 78.64, "elapsed_time": "6:25:30", "remaining_time": "1:44:42"} +{"current_steps": 3440, "total_steps": 4368, "loss": 0.4153, "lr": 5.263605303567532e-06, "epoch": 5.512820512820513, "percentage": 78.75, "elapsed_time": "6:26:06", "remaining_time": "1:44:09"} +{"current_steps": 3445, "total_steps": 4368, "loss": 0.3831, "lr": 5.20969112062032e-06, "epoch": 5.520833333333333, "percentage": 78.87, "elapsed_time": "6:26:42", "remaining_time": "1:43:36"} +{"current_steps": 3450, "total_steps": 4368, "loss": 0.3868, "lr": 5.156013099844017e-06, "epoch": 5.528846153846154, "percentage": 78.98, "elapsed_time": "6:27:14", "remaining_time": "1:43:02"} +{"current_steps": 3455, "total_steps": 4368, "loss": 0.4109, "lr": 5.1025720983348544e-06, "epoch": 5.5368589743589745, "percentage": 79.1, "elapsed_time": "6:27:50", "remaining_time": "1:42:29"} +{"current_steps": 3460, "total_steps": 4368, "loss": 0.3905, "lr": 5.049368969404484e-06, "epoch": 5.544871794871795, "percentage": 79.21, "elapsed_time": "6:28:23", "remaining_time": "1:41:55"} +{"current_steps": 3465, "total_steps": 4368, "loss": 0.4139, "lr": 4.99640456256636e-06, "epoch": 5.552884615384615, "percentage": 79.33, "elapsed_time": "6:28:56", "remaining_time": "1:41:21"} +{"current_steps": 3470, "total_steps": 4368, "loss": 0.41, "lr": 4.9436797235221814e-06, "epoch": 5.560897435897436, "percentage": 79.44, "elapsed_time": "6:29:32", "remaining_time": "1:40:48"} +{"current_steps": 3475, "total_steps": 4368, "loss": 0.4136, "lr": 4.891195294148376e-06, "epoch": 5.568910256410256, "percentage": 79.56, "elapsed_time": "6:30:07", "remaining_time": "1:40:15"} +{"current_steps": 3480, "total_steps": 4368, "loss": 0.3984, "lr": 4.838952112482671e-06, "epoch": 5.576923076923077, "percentage": 79.67, "elapsed_time": "6:30:37", "remaining_time": "1:39:40"} +{"current_steps": 3485, "total_steps": 4368, "loss": 0.3803, "lr": 4.786951012710699e-06, "epoch": 5.584935897435898, "percentage": 79.78, "elapsed_time": "6:31:12", "remaining_time": "1:39:07"} +{"current_steps": 3490, "total_steps": 4368, "loss": 0.4046, "lr": 4.735192825152686e-06, "epoch": 5.592948717948718, "percentage": 79.9, "elapsed_time": "6:31:46", "remaining_time": "1:38:33"} +{"current_steps": 3495, "total_steps": 4368, "loss": 0.3944, "lr": 4.683678376250189e-06, "epoch": 5.600961538461538, "percentage": 80.01, "elapsed_time": "6:32:25", "remaining_time": "1:38:01"} +{"current_steps": 3500, "total_steps": 4368, "loss": 0.4091, "lr": 4.6324084885529086e-06, "epoch": 5.608974358974359, "percentage": 80.13, "elapsed_time": "6:32:51", "remaining_time": "1:37:25"} +{"current_steps": 3505, "total_steps": 4368, "loss": 0.4397, "lr": 4.581383980705538e-06, "epoch": 5.61698717948718, "percentage": 80.24, "elapsed_time": "6:33:21", "remaining_time": "1:36:51"} +{"current_steps": 3510, "total_steps": 4368, "loss": 0.4096, "lr": 4.530605667434727e-06, "epoch": 5.625, "percentage": 80.36, "elapsed_time": "6:33:59", "remaining_time": "1:36:18"} +{"current_steps": 3515, "total_steps": 4368, "loss": 0.4133, "lr": 4.480074359536013e-06, "epoch": 5.63301282051282, "percentage": 80.47, "elapsed_time": "6:34:28", "remaining_time": "1:35:43"} +{"current_steps": 3520, "total_steps": 4368, "loss": 0.391, "lr": 4.429790863860934e-06, "epoch": 5.641025641025641, "percentage": 80.59, "elapsed_time": "6:35:01", "remaining_time": "1:35:09"} +{"current_steps": 3525, "total_steps": 4368, "loss": 0.4122, "lr": 4.3797559833041146e-06, "epoch": 5.649038461538462, "percentage": 80.7, "elapsed_time": "6:35:32", "remaining_time": "1:34:35"} +{"current_steps": 3530, "total_steps": 4368, "loss": 0.4348, "lr": 4.329970516790447e-06, "epoch": 5.657051282051282, "percentage": 80.82, "elapsed_time": "6:36:01", "remaining_time": "1:34:00"} +{"current_steps": 3535, "total_steps": 4368, "loss": 0.4085, "lr": 4.280435259262363e-06, "epoch": 5.665064102564102, "percentage": 80.93, "elapsed_time": "6:36:35", "remaining_time": "1:33:27"} +{"current_steps": 3540, "total_steps": 4368, "loss": 0.4072, "lr": 4.231151001667077e-06, "epoch": 5.673076923076923, "percentage": 81.04, "elapsed_time": "6:37:17", "remaining_time": "1:32:55"} +{"current_steps": 3545, "total_steps": 4368, "loss": 0.4066, "lr": 4.182118530944044e-06, "epoch": 5.681089743589744, "percentage": 81.16, "elapsed_time": "6:37:47", "remaining_time": "1:32:20"} +{"current_steps": 3550, "total_steps": 4368, "loss": 0.4002, "lr": 4.133338630012307e-06, "epoch": 5.689102564102564, "percentage": 81.27, "elapsed_time": "6:38:17", "remaining_time": "1:31:46"} +{"current_steps": 3555, "total_steps": 4368, "loss": 0.3958, "lr": 4.0848120777580554e-06, "epoch": 5.697115384615385, "percentage": 81.39, "elapsed_time": "6:38:44", "remaining_time": "1:31:11"} +{"current_steps": 3560, "total_steps": 4368, "loss": 0.4021, "lr": 4.036539649022182e-06, "epoch": 5.705128205128205, "percentage": 81.5, "elapsed_time": "6:39:19", "remaining_time": "1:30:37"} +{"current_steps": 3565, "total_steps": 4368, "loss": 0.3859, "lr": 3.988522114587865e-06, "epoch": 5.7131410256410255, "percentage": 81.62, "elapsed_time": "6:39:52", "remaining_time": "1:30:04"} +{"current_steps": 3570, "total_steps": 4368, "loss": 0.4157, "lr": 3.940760241168331e-06, "epoch": 5.721153846153846, "percentage": 81.73, "elapsed_time": "6:40:28", "remaining_time": "1:29:30"} +{"current_steps": 3575, "total_steps": 4368, "loss": 0.4162, "lr": 3.893254791394541e-06, "epoch": 5.729166666666667, "percentage": 81.85, "elapsed_time": "6:41:05", "remaining_time": "1:28:58"} +{"current_steps": 3580, "total_steps": 4368, "loss": 0.3974, "lr": 3.846006523803074e-06, "epoch": 5.737179487179487, "percentage": 81.96, "elapsed_time": "6:41:41", "remaining_time": "1:28:24"} +{"current_steps": 3585, "total_steps": 4368, "loss": 0.3933, "lr": 3.799016192823981e-06, "epoch": 5.7451923076923075, "percentage": 82.07, "elapsed_time": "6:42:13", "remaining_time": "1:27:51"} +{"current_steps": 3590, "total_steps": 4368, "loss": 0.4092, "lr": 3.7522845487687276e-06, "epoch": 5.753205128205128, "percentage": 82.19, "elapsed_time": "6:42:48", "remaining_time": "1:27:17"} +{"current_steps": 3595, "total_steps": 4368, "loss": 0.3683, "lr": 3.7058123378182664e-06, "epoch": 5.761217948717949, "percentage": 82.3, "elapsed_time": "6:43:23", "remaining_time": "1:26:44"} +{"current_steps": 3600, "total_steps": 4368, "loss": 0.4169, "lr": 3.6596003020110636e-06, "epoch": 5.769230769230769, "percentage": 82.42, "elapsed_time": "6:43:52", "remaining_time": "1:26:09"} +{"current_steps": 3605, "total_steps": 4368, "loss": 0.4302, "lr": 3.613649179231287e-06, "epoch": 5.777243589743589, "percentage": 82.53, "elapsed_time": "6:44:24", "remaining_time": "1:25:35"} +{"current_steps": 3610, "total_steps": 4368, "loss": 0.3894, "lr": 3.5679597031970017e-06, "epoch": 5.785256410256411, "percentage": 82.65, "elapsed_time": "6:44:58", "remaining_time": "1:25:02"} +{"current_steps": 3615, "total_steps": 4368, "loss": 0.4108, "lr": 3.5225326034484764e-06, "epoch": 5.793269230769231, "percentage": 82.76, "elapsed_time": "6:45:30", "remaining_time": "1:24:27"} +{"current_steps": 3620, "total_steps": 4368, "loss": 0.4515, "lr": 3.4773686053365197e-06, "epoch": 5.801282051282051, "percentage": 82.88, "elapsed_time": "6:45:57", "remaining_time": "1:23:52"} +{"current_steps": 3625, "total_steps": 4368, "loss": 0.3831, "lr": 3.4324684300109003e-06, "epoch": 5.809294871794872, "percentage": 82.99, "elapsed_time": "6:46:40", "remaining_time": "1:23:21"} +{"current_steps": 3630, "total_steps": 4368, "loss": 0.4073, "lr": 3.387832794408832e-06, "epoch": 5.8173076923076925, "percentage": 83.1, "elapsed_time": "6:47:19", "remaining_time": "1:22:48"} +{"current_steps": 3635, "total_steps": 4368, "loss": 0.416, "lr": 3.3434624112435342e-06, "epoch": 5.825320512820513, "percentage": 83.22, "elapsed_time": "6:47:47", "remaining_time": "1:22:13"} +{"current_steps": 3640, "total_steps": 4368, "loss": 0.4223, "lr": 3.2993579889928397e-06, "epoch": 5.833333333333333, "percentage": 83.33, "elapsed_time": "6:48:23", "remaining_time": "1:21:40"} +{"current_steps": 3645, "total_steps": 4368, "loss": 0.3931, "lr": 3.25552023188789e-06, "epoch": 5.841346153846154, "percentage": 83.45, "elapsed_time": "6:48:56", "remaining_time": "1:21:06"} +{"current_steps": 3650, "total_steps": 4368, "loss": 0.3803, "lr": 3.211949839901889e-06, "epoch": 5.8493589743589745, "percentage": 83.56, "elapsed_time": "6:49:28", "remaining_time": "1:20:32"} +{"current_steps": 3655, "total_steps": 4368, "loss": 0.395, "lr": 3.168647508738927e-06, "epoch": 5.857371794871795, "percentage": 83.68, "elapsed_time": "6:50:07", "remaining_time": "1:20:00"} +{"current_steps": 3660, "total_steps": 4368, "loss": 0.4319, "lr": 3.125613929822866e-06, "epoch": 5.865384615384615, "percentage": 83.79, "elapsed_time": "6:50:37", "remaining_time": "1:19:26"} +{"current_steps": 3665, "total_steps": 4368, "loss": 0.3962, "lr": 3.0828497902863106e-06, "epoch": 5.873397435897436, "percentage": 83.91, "elapsed_time": "6:51:13", "remaining_time": "1:18:52"} +{"current_steps": 3670, "total_steps": 4368, "loss": 0.4114, "lr": 3.0403557729596267e-06, "epoch": 5.881410256410256, "percentage": 84.02, "elapsed_time": "6:51:42", "remaining_time": "1:18:18"} +{"current_steps": 3675, "total_steps": 4368, "loss": 0.4098, "lr": 2.998132556360038e-06, "epoch": 5.889423076923077, "percentage": 84.13, "elapsed_time": "6:52:07", "remaining_time": "1:17:42"} +{"current_steps": 3680, "total_steps": 4368, "loss": 0.4222, "lr": 2.9561808146808068e-06, "epoch": 5.897435897435898, "percentage": 84.25, "elapsed_time": "6:52:42", "remaining_time": "1:17:09"} +{"current_steps": 3685, "total_steps": 4368, "loss": 0.3919, "lr": 2.9145012177804476e-06, "epoch": 5.905448717948718, "percentage": 84.36, "elapsed_time": "6:53:15", "remaining_time": "1:16:35"} +{"current_steps": 3690, "total_steps": 4368, "loss": 0.3973, "lr": 2.8730944311720454e-06, "epoch": 5.913461538461538, "percentage": 84.48, "elapsed_time": "6:53:43", "remaining_time": "1:16:01"} +{"current_steps": 3695, "total_steps": 4368, "loss": 0.4237, "lr": 2.8319611160126226e-06, "epoch": 5.921474358974359, "percentage": 84.59, "elapsed_time": "6:54:09", "remaining_time": "1:15:26"} +{"current_steps": 3700, "total_steps": 4368, "loss": 0.3863, "lr": 2.791101929092592e-06, "epoch": 5.92948717948718, "percentage": 84.71, "elapsed_time": "6:54:44", "remaining_time": "1:14:52"} +{"current_steps": 3705, "total_steps": 4368, "loss": 0.4225, "lr": 2.750517522825251e-06, "epoch": 5.9375, "percentage": 84.82, "elapsed_time": "6:55:13", "remaining_time": "1:14:18"} +{"current_steps": 3710, "total_steps": 4368, "loss": 0.4267, "lr": 2.710208545236397e-06, "epoch": 5.94551282051282, "percentage": 84.94, "elapsed_time": "6:55:41", "remaining_time": "1:13:43"} +{"current_steps": 3715, "total_steps": 4368, "loss": 0.4145, "lr": 2.670175639953929e-06, "epoch": 5.953525641025641, "percentage": 85.05, "elapsed_time": "6:56:15", "remaining_time": "1:13:09"} +{"current_steps": 3720, "total_steps": 4368, "loss": 0.4144, "lr": 2.6304194461976207e-06, "epoch": 5.961538461538462, "percentage": 85.16, "elapsed_time": "6:56:50", "remaining_time": "1:12:36"} +{"current_steps": 3725, "total_steps": 4368, "loss": 0.3919, "lr": 2.5909405987688896e-06, "epoch": 5.969551282051282, "percentage": 85.28, "elapsed_time": "6:57:27", "remaining_time": "1:12:03"} +{"current_steps": 3730, "total_steps": 4368, "loss": 0.4356, "lr": 2.5517397280406565e-06, "epoch": 5.977564102564102, "percentage": 85.39, "elapsed_time": "6:57:59", "remaining_time": "1:11:29"} +{"current_steps": 3735, "total_steps": 4368, "loss": 0.4249, "lr": 2.512817459947312e-06, "epoch": 5.985576923076923, "percentage": 85.51, "elapsed_time": "6:58:30", "remaining_time": "1:10:55"} +{"current_steps": 3740, "total_steps": 4368, "loss": 0.4011, "lr": 2.4741744159746618e-06, "epoch": 5.993589743589744, "percentage": 85.62, "elapsed_time": "6:59:04", "remaining_time": "1:10:22"} +{"current_steps": 3745, "total_steps": 4368, "loss": 0.4217, "lr": 2.435811213150079e-06, "epoch": 6.001602564102564, "percentage": 85.74, "elapsed_time": "6:59:33", "remaining_time": "1:09:47"} +{"current_steps": 3750, "total_steps": 4368, "loss": 0.3776, "lr": 2.3977284640325805e-06, "epoch": 6.009615384615385, "percentage": 85.85, "elapsed_time": "7:00:03", "remaining_time": "1:09:13"} +{"current_steps": 3755, "total_steps": 4368, "loss": 0.4209, "lr": 2.359926776703092e-06, "epoch": 6.017628205128205, "percentage": 85.97, "elapsed_time": "7:00:29", "remaining_time": "1:08:38"} +{"current_steps": 3760, "total_steps": 4368, "loss": 0.3876, "lr": 2.3224067547547357e-06, "epoch": 6.0256410256410255, "percentage": 86.08, "elapsed_time": "7:00:57", "remaining_time": "1:08:04"} +{"current_steps": 3765, "total_steps": 4368, "loss": 0.412, "lr": 2.2851689972831536e-06, "epoch": 6.033653846153846, "percentage": 86.2, "elapsed_time": "7:01:38", "remaining_time": "1:07:31"} +{"current_steps": 3770, "total_steps": 4368, "loss": 0.3807, "lr": 2.248214098877002e-06, "epoch": 6.041666666666667, "percentage": 86.31, "elapsed_time": "7:02:07", "remaining_time": "1:06:57"} +{"current_steps": 3775, "total_steps": 4368, "loss": 0.3989, "lr": 2.2115426496083958e-06, "epoch": 6.049679487179487, "percentage": 86.42, "elapsed_time": "7:02:43", "remaining_time": "1:06:24"} +{"current_steps": 3780, "total_steps": 4368, "loss": 0.4132, "lr": 2.175155235023536e-06, "epoch": 6.0576923076923075, "percentage": 86.54, "elapsed_time": "7:03:16", "remaining_time": "1:05:50"} +{"current_steps": 3785, "total_steps": 4368, "loss": 0.3877, "lr": 2.1390524361333355e-06, "epoch": 6.065705128205129, "percentage": 86.65, "elapsed_time": "7:03:47", "remaining_time": "1:05:16"} +{"current_steps": 3790, "total_steps": 4368, "loss": 0.4136, "lr": 2.1032348294041305e-06, "epoch": 6.073717948717949, "percentage": 86.77, "elapsed_time": "7:04:26", "remaining_time": "1:04:43"} +{"current_steps": 3795, "total_steps": 4368, "loss": 0.3736, "lr": 2.067702986748521e-06, "epoch": 6.081730769230769, "percentage": 86.88, "elapsed_time": "7:04:58", "remaining_time": "1:04:09"} +{"current_steps": 3800, "total_steps": 4368, "loss": 0.3749, "lr": 2.0324574755161764e-06, "epoch": 6.089743589743589, "percentage": 87.0, "elapsed_time": "7:05:33", "remaining_time": "1:03:36"} +{"current_steps": 3805, "total_steps": 4368, "loss": 0.3897, "lr": 1.9974988584848385e-06, "epoch": 6.097756410256411, "percentage": 87.11, "elapsed_time": "7:06:05", "remaining_time": "1:03:02"} +{"current_steps": 3810, "total_steps": 4368, "loss": 0.4113, "lr": 1.96282769385129e-06, "epoch": 6.105769230769231, "percentage": 87.23, "elapsed_time": "7:06:40", "remaining_time": "1:02:29"} +{"current_steps": 3815, "total_steps": 4368, "loss": 0.4191, "lr": 1.9284445352224625e-06, "epoch": 6.113782051282051, "percentage": 87.34, "elapsed_time": "7:07:13", "remaining_time": "1:01:55"} +{"current_steps": 3820, "total_steps": 4368, "loss": 0.3713, "lr": 1.894349931606596e-06, "epoch": 6.121794871794871, "percentage": 87.45, "elapsed_time": "7:07:52", "remaining_time": "1:01:22"} +{"current_steps": 3825, "total_steps": 4368, "loss": 0.3957, "lr": 1.8605444274044493e-06, "epoch": 6.1298076923076925, "percentage": 87.57, "elapsed_time": "7:08:30", "remaining_time": "1:00:49"} +{"current_steps": 3830, "total_steps": 4368, "loss": 0.3877, "lr": 1.827028562400659e-06, "epoch": 6.137820512820513, "percentage": 87.68, "elapsed_time": "7:09:08", "remaining_time": "1:00:16"} +{"current_steps": 3835, "total_steps": 4368, "loss": 0.3694, "lr": 1.793802871755066e-06, "epoch": 6.145833333333333, "percentage": 87.8, "elapsed_time": "7:09:39", "remaining_time": "0:59:42"} +{"current_steps": 3840, "total_steps": 4368, "loss": 0.4413, "lr": 1.760867885994202e-06, "epoch": 6.153846153846154, "percentage": 87.91, "elapsed_time": "7:10:08", "remaining_time": "0:59:08"} +{"current_steps": 3845, "total_steps": 4368, "loss": 0.4076, "lr": 1.7282241310028047e-06, "epoch": 6.1618589743589745, "percentage": 88.03, "elapsed_time": "7:10:35", "remaining_time": "0:58:34"} +{"current_steps": 3850, "total_steps": 4368, "loss": 0.3926, "lr": 1.6958721280154232e-06, "epoch": 6.169871794871795, "percentage": 88.14, "elapsed_time": "7:11:12", "remaining_time": "0:58:01"} +{"current_steps": 3855, "total_steps": 4368, "loss": 0.3997, "lr": 1.6638123936081085e-06, "epoch": 6.177884615384615, "percentage": 88.26, "elapsed_time": "7:11:42", "remaining_time": "0:57:26"} +{"current_steps": 3860, "total_steps": 4368, "loss": 0.3961, "lr": 1.6320454396901463e-06, "epoch": 6.185897435897436, "percentage": 88.37, "elapsed_time": "7:12:11", "remaining_time": "0:56:52"} +{"current_steps": 3865, "total_steps": 4368, "loss": 0.4289, "lr": 1.6005717734958914e-06, "epoch": 6.193910256410256, "percentage": 88.48, "elapsed_time": "7:12:45", "remaining_time": "0:56:19"} +{"current_steps": 3870, "total_steps": 4368, "loss": 0.4043, "lr": 1.569391897576671e-06, "epoch": 6.201923076923077, "percentage": 88.6, "elapsed_time": "7:13:26", "remaining_time": "0:55:46"} +{"current_steps": 3875, "total_steps": 4368, "loss": 0.3779, "lr": 1.5385063097927533e-06, "epoch": 6.209935897435898, "percentage": 88.71, "elapsed_time": "7:13:52", "remaining_time": "0:55:11"} +{"current_steps": 3880, "total_steps": 4368, "loss": 0.3874, "lr": 1.5079155033054104e-06, "epoch": 6.217948717948718, "percentage": 88.83, "elapsed_time": "7:14:28", "remaining_time": "0:54:38"} +{"current_steps": 3885, "total_steps": 4368, "loss": 0.4278, "lr": 1.4776199665690239e-06, "epoch": 6.225961538461538, "percentage": 88.94, "elapsed_time": "7:15:07", "remaining_time": "0:54:05"} +{"current_steps": 3890, "total_steps": 4368, "loss": 0.4093, "lr": 1.4476201833233084e-06, "epoch": 6.233974358974359, "percentage": 89.06, "elapsed_time": "7:15:43", "remaining_time": "0:53:32"} +{"current_steps": 3895, "total_steps": 4368, "loss": 0.4119, "lr": 1.4179166325855676e-06, "epoch": 6.24198717948718, "percentage": 89.17, "elapsed_time": "7:16:17", "remaining_time": "0:52:58"} +{"current_steps": 3900, "total_steps": 4368, "loss": 0.4037, "lr": 1.3885097886430599e-06, "epoch": 6.25, "percentage": 89.29, "elapsed_time": "7:16:45", "remaining_time": "0:52:24"} +{"current_steps": 3905, "total_steps": 4368, "loss": 0.4288, "lr": 1.35940012104542e-06, "epoch": 6.25801282051282, "percentage": 89.4, "elapsed_time": "7:17:14", "remaining_time": "0:51:50"} +{"current_steps": 3910, "total_steps": 4368, "loss": 0.3933, "lr": 1.3305880945971583e-06, "epoch": 6.266025641025641, "percentage": 89.51, "elapsed_time": "7:17:51", "remaining_time": "0:51:17"} +{"current_steps": 3915, "total_steps": 4368, "loss": 0.3783, "lr": 1.3020741693502403e-06, "epoch": 6.274038461538462, "percentage": 89.63, "elapsed_time": "7:18:20", "remaining_time": "0:50:43"} +{"current_steps": 3920, "total_steps": 4368, "loss": 0.3999, "lr": 1.27385880059675e-06, "epoch": 6.282051282051282, "percentage": 89.74, "elapsed_time": "7:18:56", "remaining_time": "0:50:09"} +{"current_steps": 3925, "total_steps": 4368, "loss": 0.3803, "lr": 1.245942438861607e-06, "epoch": 6.290064102564102, "percentage": 89.86, "elapsed_time": "7:19:28", "remaining_time": "0:49:36"} +{"current_steps": 3930, "total_steps": 4368, "loss": 0.3727, "lr": 1.2183255298953788e-06, "epoch": 6.298076923076923, "percentage": 89.97, "elapsed_time": "7:20:03", "remaining_time": "0:49:02"} +{"current_steps": 3935, "total_steps": 4368, "loss": 0.4265, "lr": 1.1910085146671645e-06, "epoch": 6.306089743589744, "percentage": 90.09, "elapsed_time": "7:20:43", "remaining_time": "0:48:29"} +{"current_steps": 3940, "total_steps": 4368, "loss": 0.4004, "lr": 1.1639918293575492e-06, "epoch": 6.314102564102564, "percentage": 90.2, "elapsed_time": "7:21:20", "remaining_time": "0:47:56"} +{"current_steps": 3945, "total_steps": 4368, "loss": 0.3888, "lr": 1.1372759053516536e-06, "epoch": 6.322115384615385, "percentage": 90.32, "elapsed_time": "7:21:55", "remaining_time": "0:47:23"} +{"current_steps": 3950, "total_steps": 4368, "loss": 0.3717, "lr": 1.1108611692322157e-06, "epoch": 6.330128205128205, "percentage": 90.43, "elapsed_time": "7:22:29", "remaining_time": "0:46:49"} +{"current_steps": 3955, "total_steps": 4368, "loss": 0.3794, "lr": 1.0847480427728142e-06, "epoch": 6.3381410256410255, "percentage": 90.54, "elapsed_time": "7:23:03", "remaining_time": "0:46:15"} +{"current_steps": 3960, "total_steps": 4368, "loss": 0.4025, "lr": 1.0589369429311125e-06, "epoch": 6.346153846153846, "percentage": 90.66, "elapsed_time": "7:23:33", "remaining_time": "0:45:41"} +{"current_steps": 3965, "total_steps": 4368, "loss": 0.4148, "lr": 1.0334282818422037e-06, "epoch": 6.354166666666667, "percentage": 90.77, "elapsed_time": "7:24:08", "remaining_time": "0:45:08"} +{"current_steps": 3970, "total_steps": 4368, "loss": 0.3937, "lr": 1.008222466812041e-06, "epoch": 6.362179487179487, "percentage": 90.89, "elapsed_time": "7:24:39", "remaining_time": "0:44:34"} +{"current_steps": 3975, "total_steps": 4368, "loss": 0.3647, "lr": 9.83319900310915e-07, "epoch": 6.3701923076923075, "percentage": 91.0, "elapsed_time": "7:25:15", "remaining_time": "0:44:01"} +{"current_steps": 3980, "total_steps": 4368, "loss": 0.3931, "lr": 9.587209799670495e-07, "epoch": 6.378205128205128, "percentage": 91.12, "elapsed_time": "7:25:51", "remaining_time": "0:43:27"} +{"current_steps": 3985, "total_steps": 4368, "loss": 0.3828, "lr": 9.344260985602327e-07, "epoch": 6.386217948717949, "percentage": 91.23, "elapsed_time": "7:26:23", "remaining_time": "0:42:54"} +{"current_steps": 3990, "total_steps": 4368, "loss": 0.3975, "lr": 9.104356440155526e-07, "epoch": 6.394230769230769, "percentage": 91.35, "elapsed_time": "7:27:00", "remaining_time": "0:42:20"} +{"current_steps": 3995, "total_steps": 4368, "loss": 0.4013, "lr": 8.867499993972162e-07, "epoch": 6.402243589743589, "percentage": 91.46, "elapsed_time": "7:27:33", "remaining_time": "0:41:47"} +{"current_steps": 4000, "total_steps": 4368, "loss": 0.3851, "lr": 8.633695429024058e-07, "epoch": 6.410256410256411, "percentage": 91.58, "elapsed_time": "7:28:00", "remaining_time": "0:41:12"} +{"current_steps": 4005, "total_steps": 4368, "loss": 0.3947, "lr": 8.402946478552732e-07, "epoch": 6.418269230769231, "percentage": 91.69, "elapsed_time": "7:28:33", "remaining_time": "0:40:39"} +{"current_steps": 4010, "total_steps": 4368, "loss": 0.3696, "lr": 8.175256827009392e-07, "epoch": 6.426282051282051, "percentage": 91.8, "elapsed_time": "7:29:08", "remaining_time": "0:40:05"} +{"current_steps": 4015, "total_steps": 4368, "loss": 0.3904, "lr": 7.95063010999646e-07, "epoch": 6.434294871794872, "percentage": 91.92, "elapsed_time": "7:29:39", "remaining_time": "0:39:32"} +{"current_steps": 4020, "total_steps": 4368, "loss": 0.3813, "lr": 7.729069914209409e-07, "epoch": 6.4423076923076925, "percentage": 92.03, "elapsed_time": "7:30:20", "remaining_time": "0:38:59"} +{"current_steps": 4025, "total_steps": 4368, "loss": 0.3946, "lr": 7.510579777379345e-07, "epoch": 6.450320512820513, "percentage": 92.15, "elapsed_time": "7:30:53", "remaining_time": "0:38:25"} +{"current_steps": 4030, "total_steps": 4368, "loss": 0.4176, "lr": 7.295163188216792e-07, "epoch": 6.458333333333333, "percentage": 92.26, "elapsed_time": "7:31:28", "remaining_time": "0:37:51"} +{"current_steps": 4035, "total_steps": 4368, "loss": 0.4402, "lr": 7.08282358635568e-07, "epoch": 6.466346153846154, "percentage": 92.38, "elapsed_time": "7:31:57", "remaining_time": "0:37:17"} +{"current_steps": 4040, "total_steps": 4368, "loss": 0.3906, "lr": 6.87356436229869e-07, "epoch": 6.4743589743589745, "percentage": 92.49, "elapsed_time": "7:32:26", "remaining_time": "0:36:44"} +{"current_steps": 4045, "total_steps": 4368, "loss": 0.3799, "lr": 6.667388857362977e-07, "epoch": 6.482371794871795, "percentage": 92.61, "elapsed_time": "7:32:56", "remaining_time": "0:36:10"} +{"current_steps": 4050, "total_steps": 4368, "loss": 0.3893, "lr": 6.464300363626797e-07, "epoch": 6.490384615384615, "percentage": 92.72, "elapsed_time": "7:33:30", "remaining_time": "0:35:36"} +{"current_steps": 4055, "total_steps": 4368, "loss": 0.3768, "lr": 6.264302123877053e-07, "epoch": 6.498397435897436, "percentage": 92.83, "elapsed_time": "7:34:11", "remaining_time": "0:35:03"} +{"current_steps": 4060, "total_steps": 4368, "loss": 0.4053, "lr": 6.067397331557412e-07, "epoch": 6.506410256410256, "percentage": 92.95, "elapsed_time": "7:34:42", "remaining_time": "0:34:29"} +{"current_steps": 4065, "total_steps": 4368, "loss": 0.3862, "lr": 5.873589130717405e-07, "epoch": 6.514423076923077, "percentage": 93.06, "elapsed_time": "7:35:15", "remaining_time": "0:33:56"} +{"current_steps": 4070, "total_steps": 4368, "loss": 0.4036, "lr": 5.682880615962116e-07, "epoch": 6.522435897435898, "percentage": 93.18, "elapsed_time": "7:35:48", "remaining_time": "0:33:22"} +{"current_steps": 4075, "total_steps": 4368, "loss": 0.3654, "lr": 5.495274832402841e-07, "epoch": 6.530448717948718, "percentage": 93.29, "elapsed_time": "7:36:24", "remaining_time": "0:32:48"} +{"current_steps": 4080, "total_steps": 4368, "loss": 0.4087, "lr": 5.310774775608529e-07, "epoch": 6.538461538461538, "percentage": 93.41, "elapsed_time": "7:36:53", "remaining_time": "0:32:15"} +{"current_steps": 4085, "total_steps": 4368, "loss": 0.4054, "lr": 5.129383391557751e-07, "epoch": 6.546474358974359, "percentage": 93.52, "elapsed_time": "7:37:22", "remaining_time": "0:31:41"} +{"current_steps": 4090, "total_steps": 4368, "loss": 0.4025, "lr": 4.951103576591876e-07, "epoch": 6.55448717948718, "percentage": 93.64, "elapsed_time": "7:37:54", "remaining_time": "0:31:07"} +{"current_steps": 4095, "total_steps": 4368, "loss": 0.3995, "lr": 4.7759381773687e-07, "epoch": 6.5625, "percentage": 93.75, "elapsed_time": "7:38:30", "remaining_time": "0:30:34"} +{"current_steps": 4100, "total_steps": 4368, "loss": 0.3935, "lr": 4.6038899908170234e-07, "epoch": 6.57051282051282, "percentage": 93.86, "elapsed_time": "7:39:04", "remaining_time": "0:30:00"} +{"current_steps": 4105, "total_steps": 4368, "loss": 0.387, "lr": 4.4349617640920164e-07, "epoch": 6.578525641025641, "percentage": 93.98, "elapsed_time": "7:39:37", "remaining_time": "0:29:26"} +{"current_steps": 4110, "total_steps": 4368, "loss": 0.4247, "lr": 4.2691561945312764e-07, "epoch": 6.586538461538462, "percentage": 94.09, "elapsed_time": "7:40:11", "remaining_time": "0:28:53"} +{"current_steps": 4115, "total_steps": 4368, "loss": 0.4019, "lr": 4.106475929611886e-07, "epoch": 6.594551282051282, "percentage": 94.21, "elapsed_time": "7:40:40", "remaining_time": "0:28:19"} +{"current_steps": 4120, "total_steps": 4368, "loss": 0.3853, "lr": 3.9469235669080007e-07, "epoch": 6.602564102564102, "percentage": 94.32, "elapsed_time": "7:41:11", "remaining_time": "0:27:45"} +{"current_steps": 4125, "total_steps": 4368, "loss": 0.3941, "lr": 3.7905016540495053e-07, "epoch": 6.610576923076923, "percentage": 94.44, "elapsed_time": "7:41:46", "remaining_time": "0:27:12"} +{"current_steps": 4130, "total_steps": 4368, "loss": 0.3879, "lr": 3.63721268868118e-07, "epoch": 6.618589743589744, "percentage": 94.55, "elapsed_time": "7:42:21", "remaining_time": "0:26:38"} +{"current_steps": 4135, "total_steps": 4368, "loss": 0.3963, "lr": 3.487059118422997e-07, "epoch": 6.626602564102564, "percentage": 94.67, "elapsed_time": "7:42:58", "remaining_time": "0:26:05"} +{"current_steps": 4140, "total_steps": 4368, "loss": 0.4244, "lr": 3.3400433408308895e-07, "epoch": 6.634615384615385, "percentage": 94.78, "elapsed_time": "7:43:22", "remaining_time": "0:25:31"} +{"current_steps": 4145, "total_steps": 4368, "loss": 0.4013, "lr": 3.196167703358577e-07, "epoch": 6.642628205128205, "percentage": 94.89, "elapsed_time": "7:43:55", "remaining_time": "0:24:57"} +{"current_steps": 4150, "total_steps": 4368, "loss": 0.3891, "lr": 3.0554345033199985e-07, "epoch": 6.6506410256410255, "percentage": 95.01, "elapsed_time": "7:44:27", "remaining_time": "0:24:23"} +{"current_steps": 4155, "total_steps": 4368, "loss": 0.399, "lr": 2.917845987852652e-07, "epoch": 6.658653846153846, "percentage": 95.12, "elapsed_time": "7:45:03", "remaining_time": "0:23:50"} +{"current_steps": 4160, "total_steps": 4368, "loss": 0.3978, "lr": 2.783404353881758e-07, "epoch": 6.666666666666667, "percentage": 95.24, "elapsed_time": "7:45:38", "remaining_time": "0:23:16"} +{"current_steps": 4165, "total_steps": 4368, "loss": 0.3957, "lr": 2.652111748085151e-07, "epoch": 6.674679487179487, "percentage": 95.35, "elapsed_time": "7:46:16", "remaining_time": "0:22:43"} +{"current_steps": 4170, "total_steps": 4368, "loss": 0.3914, "lr": 2.523970266859044e-07, "epoch": 6.6826923076923075, "percentage": 95.47, "elapsed_time": "7:46:50", "remaining_time": "0:22:10"} +{"current_steps": 4175, "total_steps": 4368, "loss": 0.4228, "lr": 2.398981956284363e-07, "epoch": 6.690705128205128, "percentage": 95.58, "elapsed_time": "7:47:17", "remaining_time": "0:21:36"} +{"current_steps": 4180, "total_steps": 4368, "loss": 0.3807, "lr": 2.2771488120944207e-07, "epoch": 6.698717948717949, "percentage": 95.7, "elapsed_time": "7:47:46", "remaining_time": "0:21:02"} +{"current_steps": 4185, "total_steps": 4368, "loss": 0.4204, "lr": 2.1584727796427174e-07, "epoch": 6.706730769230769, "percentage": 95.81, "elapsed_time": "7:48:20", "remaining_time": "0:20:28"} +{"current_steps": 4190, "total_steps": 4368, "loss": 0.399, "lr": 2.0429557538720556e-07, "epoch": 6.714743589743589, "percentage": 95.92, "elapsed_time": "7:48:56", "remaining_time": "0:19:55"} +{"current_steps": 4195, "total_steps": 4368, "loss": 0.4038, "lr": 1.930599579284298e-07, "epoch": 6.722756410256411, "percentage": 96.04, "elapsed_time": "7:49:34", "remaining_time": "0:19:21"} +{"current_steps": 4200, "total_steps": 4368, "loss": 0.3882, "lr": 1.8214060499107679e-07, "epoch": 6.730769230769231, "percentage": 96.15, "elapsed_time": "7:50:08", "remaining_time": "0:18:48"} +{"current_steps": 4205, "total_steps": 4368, "loss": 0.386, "lr": 1.7153769092837614e-07, "epoch": 6.738782051282051, "percentage": 96.27, "elapsed_time": "7:50:41", "remaining_time": "0:18:14"} +{"current_steps": 4210, "total_steps": 4368, "loss": 0.3876, "lr": 1.6125138504086146e-07, "epoch": 6.746794871794872, "percentage": 96.38, "elapsed_time": "7:51:16", "remaining_time": "0:17:41"} +{"current_steps": 4215, "total_steps": 4368, "loss": 0.3885, "lr": 1.5128185157367247e-07, "epoch": 6.7548076923076925, "percentage": 96.5, "elapsed_time": "7:51:49", "remaining_time": "0:17:07"} +{"current_steps": 4220, "total_steps": 4368, "loss": 0.3976, "lr": 1.4162924971393044e-07, "epoch": 6.762820512820513, "percentage": 96.61, "elapsed_time": "7:52:22", "remaining_time": "0:16:33"} +{"current_steps": 4225, "total_steps": 4368, "loss": 0.3858, "lr": 1.322937335881891e-07, "epoch": 6.770833333333333, "percentage": 96.73, "elapsed_time": "7:52:51", "remaining_time": "0:16:00"} +{"current_steps": 4230, "total_steps": 4368, "loss": 0.392, "lr": 1.2327545225999215e-07, "epoch": 6.778846153846154, "percentage": 96.84, "elapsed_time": "7:53:18", "remaining_time": "0:15:26"} +{"current_steps": 4235, "total_steps": 4368, "loss": 0.4144, "lr": 1.145745497274664e-07, "epoch": 6.7868589743589745, "percentage": 96.96, "elapsed_time": "7:53:51", "remaining_time": "0:14:52"} +{"current_steps": 4240, "total_steps": 4368, "loss": 0.3853, "lr": 1.061911649210523e-07, "epoch": 6.794871794871795, "percentage": 97.07, "elapsed_time": "7:54:27", "remaining_time": "0:14:19"} +{"current_steps": 4245, "total_steps": 4368, "loss": 0.4517, "lr": 9.812543170126365e-08, "epoch": 6.802884615384615, "percentage": 97.18, "elapsed_time": "7:55:00", "remaining_time": "0:13:45"} +{"current_steps": 4250, "total_steps": 4368, "loss": 0.408, "lr": 9.03774788565559e-08, "epoch": 6.810897435897436, "percentage": 97.3, "elapsed_time": "7:55:35", "remaining_time": "0:13:12"} +{"current_steps": 4255, "total_steps": 4368, "loss": 0.3969, "lr": 8.294743010127448e-08, "epoch": 6.818910256410256, "percentage": 97.41, "elapsed_time": "7:56:11", "remaining_time": "0:12:38"} +{"current_steps": 4260, "total_steps": 4368, "loss": 0.4147, "lr": 7.583540407367418e-08, "epoch": 6.826923076923077, "percentage": 97.53, "elapsed_time": "7:56:40", "remaining_time": "0:12:05"} +{"current_steps": 4265, "total_steps": 4368, "loss": 0.4047, "lr": 6.904151433402728e-08, "epoch": 6.834935897435898, "percentage": 97.64, "elapsed_time": "7:57:13", "remaining_time": "0:11:31"} +{"current_steps": 4270, "total_steps": 4368, "loss": 0.3987, "lr": 6.256586936281172e-08, "epoch": 6.842948717948718, "percentage": 97.76, "elapsed_time": "7:57:46", "remaining_time": "0:10:57"} +{"current_steps": 4275, "total_steps": 4368, "loss": 0.4033, "lr": 5.6408572558972475e-08, "epoch": 6.850961538461538, "percentage": 97.87, "elapsed_time": "7:58:15", "remaining_time": "0:10:24"} +{"current_steps": 4280, "total_steps": 4368, "loss": 0.4235, "lr": 5.0569722238280605e-08, "epoch": 6.858974358974359, "percentage": 97.99, "elapsed_time": "7:58:47", "remaining_time": "0:09:50"} +{"current_steps": 4285, "total_steps": 4368, "loss": 0.3707, "lr": 4.504941163175236e-08, "epoch": 6.86698717948718, "percentage": 98.1, "elapsed_time": "7:59:22", "remaining_time": "0:09:17"} +{"current_steps": 4290, "total_steps": 4368, "loss": 0.397, "lr": 3.984772888417032e-08, "epoch": 6.875, "percentage": 98.21, "elapsed_time": "7:59:52", "remaining_time": "0:08:43"} +{"current_steps": 4295, "total_steps": 4368, "loss": 0.3893, "lr": 3.4964757052671216e-08, "epoch": 6.88301282051282, "percentage": 98.33, "elapsed_time": "8:00:24", "remaining_time": "0:08:09"} +{"current_steps": 4300, "total_steps": 4368, "loss": 0.4174, "lr": 3.0400574105415856e-08, "epoch": 6.891025641025641, "percentage": 98.44, "elapsed_time": "8:01:00", "remaining_time": "0:07:36"} +{"current_steps": 4305, "total_steps": 4368, "loss": 0.3741, "lr": 2.615525292035459e-08, "epoch": 6.899038461538462, "percentage": 98.56, "elapsed_time": "8:01:43", "remaining_time": "0:07:02"} +{"current_steps": 4310, "total_steps": 4368, "loss": 0.4015, "lr": 2.222886128405266e-08, "epoch": 6.907051282051282, "percentage": 98.67, "elapsed_time": "8:02:14", "remaining_time": "0:06:29"} +{"current_steps": 4315, "total_steps": 4368, "loss": 0.4098, "lr": 1.8621461890617752e-08, "epoch": 6.915064102564102, "percentage": 98.79, "elapsed_time": "8:02:46", "remaining_time": "0:05:55"} +{"current_steps": 4320, "total_steps": 4368, "loss": 0.3887, "lr": 1.5333112340687463e-08, "epoch": 6.923076923076923, "percentage": 98.9, "elapsed_time": "8:03:17", "remaining_time": "0:05:22"} +{"current_steps": 4325, "total_steps": 4368, "loss": 0.3959, "lr": 1.2363865140518905e-08, "epoch": 6.931089743589744, "percentage": 99.02, "elapsed_time": "8:03:44", "remaining_time": "0:04:48"} +{"current_steps": 4330, "total_steps": 4368, "loss": 0.3725, "lr": 9.713767701151621e-09, "epoch": 6.939102564102564, "percentage": 99.13, "elapsed_time": "8:04:20", "remaining_time": "0:04:15"} +{"current_steps": 4335, "total_steps": 4368, "loss": 0.3995, "lr": 7.382862337641516e-09, "epoch": 6.947115384615385, "percentage": 99.24, "elapsed_time": "8:04:51", "remaining_time": "0:03:41"} +{"current_steps": 4340, "total_steps": 4368, "loss": 0.3989, "lr": 5.371186268390283e-09, "epoch": 6.955128205128205, "percentage": 99.36, "elapsed_time": "8:05:26", "remaining_time": "0:03:07"} +{"current_steps": 4345, "total_steps": 4368, "loss": 0.4011, "lr": 3.678771614550325e-09, "epoch": 6.9631410256410255, "percentage": 99.47, "elapsed_time": "8:05:57", "remaining_time": "0:02:34"} +{"current_steps": 4350, "total_steps": 4368, "loss": 0.3919, "lr": 2.3056453995162763e-09, "epoch": 6.971153846153846, "percentage": 99.59, "elapsed_time": "8:06:28", "remaining_time": "0:02:00"} +{"current_steps": 4355, "total_steps": 4368, "loss": 0.4062, "lr": 1.2518295484875708e-09, "epoch": 6.979166666666667, "percentage": 99.7, "elapsed_time": "8:06:56", "remaining_time": "0:01:27"} +{"current_steps": 4360, "total_steps": 4368, "loss": 0.4064, "lr": 5.173408881198328e-10, "epoch": 6.987179487179487, "percentage": 99.82, "elapsed_time": "8:07:37", "remaining_time": "0:00:53"} +{"current_steps": 4365, "total_steps": 4368, "loss": 0.4175, "lr": 1.0219114625398263e-10, "epoch": 6.9951923076923075, "percentage": 99.93, "elapsed_time": "8:08:06", "remaining_time": "0:00:20"} +{"current_steps": 4368, "total_steps": 4368, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "8:09:06", "remaining_time": "0:00:00"} diff --git a/trainer_state.json b/trainer_state.json new file mode 100644 index 0000000..a9c71c9 --- /dev/null +++ b/trainer_state.json @@ -0,0 +1,9650 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 7.0, + "eval_steps": 500, + "global_step": 4368, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.008012820512820512, + "grad_norm": 8.08084235605047, + "learning_rate": 3.661327231121282e-07, + "loss": 0.9655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5593497157096863, + "step": 5, + "valid_targets_mean": 4383.4, + "valid_targets_min": 1122 + }, + { + "epoch": 0.016025641025641024, + "grad_norm": 9.044541289479458, + "learning_rate": 8.237986270022884e-07, + "loss": 0.9567, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5324016213417053, + "step": 10, + "valid_targets_mean": 3594.5, + "valid_targets_min": 2016 + }, + { + "epoch": 0.02403846153846154, + "grad_norm": 7.166559297656916, + "learning_rate": 1.2814645308924487e-06, + "loss": 0.9797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5059735774993896, + "step": 15, + "valid_targets_mean": 3787.8, + "valid_targets_min": 2258 + }, + { + "epoch": 0.03205128205128205, + "grad_norm": 5.452455576397862, + "learning_rate": 1.7391304347826088e-06, + "loss": 0.9819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.49549150466918945, + "step": 20, + "valid_targets_mean": 4717.0, + "valid_targets_min": 2632 + }, + { + "epoch": 0.04006410256410257, + "grad_norm": 4.851495340044941, + "learning_rate": 2.196796338672769e-06, + "loss": 0.943, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5397368669509888, + "step": 25, + "valid_targets_mean": 4225.4, + "valid_targets_min": 1708 + }, + { + "epoch": 0.04807692307692308, + "grad_norm": 3.7381624792446315, + "learning_rate": 2.654462242562929e-06, + "loss": 0.894, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.39446473121643066, + "step": 30, + "valid_targets_mean": 2677.9, + "valid_targets_min": 751 + }, + { + "epoch": 0.05608974358974359, + "grad_norm": 3.1826856383683095, + "learning_rate": 3.1121281464530894e-06, + "loss": 0.898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3918735086917877, + "step": 35, + "valid_targets_mean": 2681.5, + "valid_targets_min": 1297 + }, + { + "epoch": 0.0641025641025641, + "grad_norm": 2.0241717424816854, + "learning_rate": 3.56979405034325e-06, + "loss": 0.8348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42430946230888367, + "step": 40, + "valid_targets_mean": 4015.0, + "valid_targets_min": 2340 + }, + { + "epoch": 0.07211538461538461, + "grad_norm": 1.8342940717726537, + "learning_rate": 4.0274599542334094e-06, + "loss": 0.8442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42231297492980957, + "step": 45, + "valid_targets_mean": 2895.2, + "valid_targets_min": 1292 + }, + { + "epoch": 0.08012820512820513, + "grad_norm": 1.1921894888508766, + "learning_rate": 4.48512585812357e-06, + "loss": 0.8109, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4566168487071991, + "step": 50, + "valid_targets_mean": 4140.8, + "valid_targets_min": 1570 + }, + { + "epoch": 0.08814102564102565, + "grad_norm": 1.0790276870334963, + "learning_rate": 4.94279176201373e-06, + "loss": 0.8115, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42182424664497375, + "step": 55, + "valid_targets_mean": 3181.6, + "valid_targets_min": 1348 + }, + { + "epoch": 0.09615384615384616, + "grad_norm": 1.1056454806438016, + "learning_rate": 5.400457665903891e-06, + "loss": 0.7536, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42460107803344727, + "step": 60, + "valid_targets_mean": 3233.1, + "valid_targets_min": 1300 + }, + { + "epoch": 0.10416666666666667, + "grad_norm": 0.8566148102566176, + "learning_rate": 5.858123569794051e-06, + "loss": 0.7537, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33287298679351807, + "step": 65, + "valid_targets_mean": 3847.6, + "valid_targets_min": 1511 + }, + { + "epoch": 0.11217948717948718, + "grad_norm": 0.8924213627940077, + "learning_rate": 6.31578947368421e-06, + "loss": 0.7417, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.35968464612960815, + "step": 70, + "valid_targets_mean": 3229.1, + "valid_targets_min": 1225 + }, + { + "epoch": 0.1201923076923077, + "grad_norm": 0.9644620118142551, + "learning_rate": 6.773455377574372e-06, + "loss": 0.7748, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.40846461057662964, + "step": 75, + "valid_targets_mean": 2716.6, + "valid_targets_min": 888 + }, + { + "epoch": 0.1282051282051282, + "grad_norm": 0.9392567512162411, + "learning_rate": 7.231121281464531e-06, + "loss": 0.7666, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34942084550857544, + "step": 80, + "valid_targets_mean": 2539.5, + "valid_targets_min": 1587 + }, + { + "epoch": 0.1362179487179487, + "grad_norm": 0.7739844563733203, + "learning_rate": 7.688787185354691e-06, + "loss": 0.6723, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.35039928555488586, + "step": 85, + "valid_targets_mean": 3542.8, + "valid_targets_min": 1045 + }, + { + "epoch": 0.14423076923076922, + "grad_norm": 0.7724037800220916, + "learning_rate": 8.146453089244852e-06, + "loss": 0.7063, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37619489431381226, + "step": 90, + "valid_targets_mean": 3338.1, + "valid_targets_min": 1600 + }, + { + "epoch": 0.15224358974358973, + "grad_norm": 0.7354092804071887, + "learning_rate": 8.604118993135013e-06, + "loss": 0.7005, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.261581152677536, + "step": 95, + "valid_targets_mean": 2842.9, + "valid_targets_min": 896 + }, + { + "epoch": 0.16025641025641027, + "grad_norm": 0.7535948536747676, + "learning_rate": 9.061784897025172e-06, + "loss": 0.6994, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.45470109581947327, + "step": 100, + "valid_targets_mean": 4142.6, + "valid_targets_min": 2325 + }, + { + "epoch": 0.16826923076923078, + "grad_norm": 0.8212782082458024, + "learning_rate": 9.519450800915333e-06, + "loss": 0.7064, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3767229914665222, + "step": 105, + "valid_targets_mean": 3076.0, + "valid_targets_min": 1277 + }, + { + "epoch": 0.1762820512820513, + "grad_norm": 0.7612910527252497, + "learning_rate": 9.977116704805492e-06, + "loss": 0.6694, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2832098603248596, + "step": 110, + "valid_targets_mean": 3347.0, + "valid_targets_min": 1359 + }, + { + "epoch": 0.1842948717948718, + "grad_norm": 0.7188977006199924, + "learning_rate": 1.0434782608695653e-05, + "loss": 0.6781, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.44378337264060974, + "step": 115, + "valid_targets_mean": 4160.6, + "valid_targets_min": 3086 + }, + { + "epoch": 0.19230769230769232, + "grad_norm": 0.7831637105452718, + "learning_rate": 1.0892448512585814e-05, + "loss": 0.6593, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.38758358359336853, + "step": 120, + "valid_targets_mean": 3695.4, + "valid_targets_min": 2276 + }, + { + "epoch": 0.20032051282051283, + "grad_norm": 0.7095839726904716, + "learning_rate": 1.1350114416475973e-05, + "loss": 0.6411, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26519066095352173, + "step": 125, + "valid_targets_mean": 3381.9, + "valid_targets_min": 1495 + }, + { + "epoch": 0.20833333333333334, + "grad_norm": 0.8498216693406554, + "learning_rate": 1.1807780320366134e-05, + "loss": 0.6848, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3395300507545471, + "step": 130, + "valid_targets_mean": 2843.8, + "valid_targets_min": 1468 + }, + { + "epoch": 0.21634615384615385, + "grad_norm": 0.7785532316523031, + "learning_rate": 1.2265446224256295e-05, + "loss": 0.6421, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2892741858959198, + "step": 135, + "valid_targets_mean": 2567.4, + "valid_targets_min": 1605 + }, + { + "epoch": 0.22435897435897437, + "grad_norm": 0.9882123688352539, + "learning_rate": 1.2723112128146454e-05, + "loss": 0.6348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27983033657073975, + "step": 140, + "valid_targets_mean": 3580.8, + "valid_targets_min": 910 + }, + { + "epoch": 0.23237179487179488, + "grad_norm": 0.6893670858056347, + "learning_rate": 1.3180778032036615e-05, + "loss": 0.683, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23598313331604004, + "step": 145, + "valid_targets_mean": 2884.8, + "valid_targets_min": 1634 + }, + { + "epoch": 0.2403846153846154, + "grad_norm": 0.7926440711998264, + "learning_rate": 1.3638443935926776e-05, + "loss": 0.656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.43385663628578186, + "step": 150, + "valid_targets_mean": 2975.4, + "valid_targets_min": 1151 + }, + { + "epoch": 0.2483974358974359, + "grad_norm": 0.6979197898323751, + "learning_rate": 1.4096109839816933e-05, + "loss": 0.6734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4018835127353668, + "step": 155, + "valid_targets_mean": 4062.4, + "valid_targets_min": 2175 + }, + { + "epoch": 0.2564102564102564, + "grad_norm": 0.769838461478084, + "learning_rate": 1.4553775743707096e-05, + "loss": 0.6344, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30462324619293213, + "step": 160, + "valid_targets_mean": 2892.1, + "valid_targets_min": 983 + }, + { + "epoch": 0.2644230769230769, + "grad_norm": 1.1457483820172079, + "learning_rate": 1.5011441647597256e-05, + "loss": 0.6547, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.43081146478652954, + "step": 165, + "valid_targets_mean": 4230.0, + "valid_targets_min": 2025 + }, + { + "epoch": 0.2724358974358974, + "grad_norm": 0.7385179675194287, + "learning_rate": 1.5469107551487414e-05, + "loss": 0.6144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31631702184677124, + "step": 170, + "valid_targets_mean": 3773.2, + "valid_targets_min": 1766 + }, + { + "epoch": 0.28044871794871795, + "grad_norm": 0.696382838481809, + "learning_rate": 1.5926773455377575e-05, + "loss": 0.6454, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2497679740190506, + "step": 175, + "valid_targets_mean": 3122.0, + "valid_targets_min": 1341 + }, + { + "epoch": 0.28846153846153844, + "grad_norm": 0.6936387376847872, + "learning_rate": 1.6384439359267736e-05, + "loss": 0.6317, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4676125645637512, + "step": 180, + "valid_targets_mean": 4096.8, + "valid_targets_min": 940 + }, + { + "epoch": 0.296474358974359, + "grad_norm": 0.8654942859728572, + "learning_rate": 1.6842105263157896e-05, + "loss": 0.6377, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31465697288513184, + "step": 185, + "valid_targets_mean": 2974.5, + "valid_targets_min": 1109 + }, + { + "epoch": 0.30448717948717946, + "grad_norm": 0.6447647159220515, + "learning_rate": 1.7299771167048057e-05, + "loss": 0.6429, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3067771792411804, + "step": 190, + "valid_targets_mean": 4463.5, + "valid_targets_min": 2472 + }, + { + "epoch": 0.3125, + "grad_norm": 0.6513912597624796, + "learning_rate": 1.7757437070938218e-05, + "loss": 0.6447, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37445366382598877, + "step": 195, + "valid_targets_mean": 5337.8, + "valid_targets_min": 1597 + }, + { + "epoch": 0.32051282051282054, + "grad_norm": 0.7414288624146944, + "learning_rate": 1.8215102974828376e-05, + "loss": 0.6205, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3041468560695648, + "step": 200, + "valid_targets_mean": 3204.9, + "valid_targets_min": 1126 + }, + { + "epoch": 0.328525641025641, + "grad_norm": 0.8300806457519919, + "learning_rate": 1.8672768878718537e-05, + "loss": 0.611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.279873788356781, + "step": 205, + "valid_targets_mean": 2603.8, + "valid_targets_min": 1211 + }, + { + "epoch": 0.33653846153846156, + "grad_norm": 0.9862056058555818, + "learning_rate": 1.9130434782608697e-05, + "loss": 0.6301, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31322455406188965, + "step": 210, + "valid_targets_mean": 2928.4, + "valid_targets_min": 1082 + }, + { + "epoch": 0.34455128205128205, + "grad_norm": 0.8048081212369187, + "learning_rate": 1.9588100686498858e-05, + "loss": 0.5953, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.231235072016716, + "step": 215, + "valid_targets_mean": 2329.6, + "valid_targets_min": 1117 + }, + { + "epoch": 0.3525641025641026, + "grad_norm": 0.7324326237742053, + "learning_rate": 2.004576659038902e-05, + "loss": 0.656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31233295798301697, + "step": 220, + "valid_targets_mean": 3424.4, + "valid_targets_min": 1384 + }, + { + "epoch": 0.3605769230769231, + "grad_norm": 0.6765129410304772, + "learning_rate": 2.050343249427918e-05, + "loss": 0.6255, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30385103821754456, + "step": 225, + "valid_targets_mean": 3820.4, + "valid_targets_min": 813 + }, + { + "epoch": 0.3685897435897436, + "grad_norm": 0.6756252553004568, + "learning_rate": 2.0961098398169337e-05, + "loss": 0.614, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3490321934223175, + "step": 230, + "valid_targets_mean": 4262.9, + "valid_targets_min": 1626 + }, + { + "epoch": 0.3766025641025641, + "grad_norm": 0.8014553263282023, + "learning_rate": 2.14187643020595e-05, + "loss": 0.6089, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.46571505069732666, + "step": 235, + "valid_targets_mean": 4265.9, + "valid_targets_min": 2041 + }, + { + "epoch": 0.38461538461538464, + "grad_norm": 0.7692673587252775, + "learning_rate": 2.187643020594966e-05, + "loss": 0.6354, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.40585455298423767, + "step": 240, + "valid_targets_mean": 3458.2, + "valid_targets_min": 1601 + }, + { + "epoch": 0.3926282051282051, + "grad_norm": 0.7140886846825153, + "learning_rate": 2.2334096109839817e-05, + "loss": 0.6026, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30154934525489807, + "step": 245, + "valid_targets_mean": 3431.8, + "valid_targets_min": 1747 + }, + { + "epoch": 0.40064102564102566, + "grad_norm": 0.7185288490656583, + "learning_rate": 2.279176201372998e-05, + "loss": 0.6119, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37900102138519287, + "step": 250, + "valid_targets_mean": 4164.8, + "valid_targets_min": 1854 + }, + { + "epoch": 0.40865384615384615, + "grad_norm": 0.8715149750891712, + "learning_rate": 2.3249427917620138e-05, + "loss": 0.5854, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.39362844824790955, + "step": 255, + "valid_targets_mean": 3589.1, + "valid_targets_min": 1828 + }, + { + "epoch": 0.4166666666666667, + "grad_norm": 0.6863865980866662, + "learning_rate": 2.37070938215103e-05, + "loss": 0.5917, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2835404872894287, + "step": 260, + "valid_targets_mean": 3883.9, + "valid_targets_min": 1446 + }, + { + "epoch": 0.42467948717948717, + "grad_norm": 0.8440919105603253, + "learning_rate": 2.4164759725400463e-05, + "loss": 0.6211, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2947370707988739, + "step": 265, + "valid_targets_mean": 2854.9, + "valid_targets_min": 1171 + }, + { + "epoch": 0.4326923076923077, + "grad_norm": 0.7297722089036464, + "learning_rate": 2.462242562929062e-05, + "loss": 0.5775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28876638412475586, + "step": 270, + "valid_targets_mean": 3363.9, + "valid_targets_min": 1683 + }, + { + "epoch": 0.4407051282051282, + "grad_norm": 0.7563976413635497, + "learning_rate": 2.508009153318078e-05, + "loss": 0.5995, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3830539584159851, + "step": 275, + "valid_targets_mean": 3486.9, + "valid_targets_min": 983 + }, + { + "epoch": 0.44871794871794873, + "grad_norm": 0.7705900301879751, + "learning_rate": 2.5537757437070943e-05, + "loss": 0.5977, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23917141556739807, + "step": 280, + "valid_targets_mean": 2930.1, + "valid_targets_min": 702 + }, + { + "epoch": 0.4567307692307692, + "grad_norm": 0.9416935955279536, + "learning_rate": 2.59954233409611e-05, + "loss": 0.6142, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29815295338630676, + "step": 285, + "valid_targets_mean": 3551.5, + "valid_targets_min": 1989 + }, + { + "epoch": 0.46474358974358976, + "grad_norm": 0.7308814490332448, + "learning_rate": 2.645308924485126e-05, + "loss": 0.5716, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2769700288772583, + "step": 290, + "valid_targets_mean": 2907.6, + "valid_targets_min": 1249 + }, + { + "epoch": 0.47275641025641024, + "grad_norm": 0.7119530860074925, + "learning_rate": 2.6910755148741422e-05, + "loss": 0.5892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30059486627578735, + "step": 295, + "valid_targets_mean": 3402.6, + "valid_targets_min": 1165 + }, + { + "epoch": 0.4807692307692308, + "grad_norm": 0.7585289775760408, + "learning_rate": 2.7368421052631583e-05, + "loss": 0.6112, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4211341142654419, + "step": 300, + "valid_targets_mean": 3679.5, + "valid_targets_min": 1610 + }, + { + "epoch": 0.48878205128205127, + "grad_norm": 0.6358035082994052, + "learning_rate": 2.782608695652174e-05, + "loss": 0.5587, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24304768443107605, + "step": 305, + "valid_targets_mean": 3991.4, + "valid_targets_min": 1917 + }, + { + "epoch": 0.4967948717948718, + "grad_norm": 0.7121086650202204, + "learning_rate": 2.8283752860411904e-05, + "loss": 0.6255, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21161486208438873, + "step": 310, + "valid_targets_mean": 3236.9, + "valid_targets_min": 1305 + }, + { + "epoch": 0.5048076923076923, + "grad_norm": 0.7991833537985452, + "learning_rate": 2.8741418764302062e-05, + "loss": 0.6187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34455418586730957, + "step": 315, + "valid_targets_mean": 3130.6, + "valid_targets_min": 726 + }, + { + "epoch": 0.5128205128205128, + "grad_norm": 0.6784715400376627, + "learning_rate": 2.9199084668192223e-05, + "loss": 0.5639, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2229878306388855, + "step": 320, + "valid_targets_mean": 3083.2, + "valid_targets_min": 1632 + }, + { + "epoch": 0.5208333333333334, + "grad_norm": 0.6865751999804144, + "learning_rate": 2.9656750572082384e-05, + "loss": 0.5994, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2551387548446655, + "step": 325, + "valid_targets_mean": 3287.9, + "valid_targets_min": 1044 + }, + { + "epoch": 0.5288461538461539, + "grad_norm": 0.6388097941114113, + "learning_rate": 3.0114416475972544e-05, + "loss": 0.5855, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34533900022506714, + "step": 330, + "valid_targets_mean": 4452.8, + "valid_targets_min": 2903 + }, + { + "epoch": 0.5368589743589743, + "grad_norm": 0.7313634812205421, + "learning_rate": 3.05720823798627e-05, + "loss": 0.5956, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3254767060279846, + "step": 335, + "valid_targets_mean": 3634.8, + "valid_targets_min": 1310 + }, + { + "epoch": 0.5448717948717948, + "grad_norm": 0.801255367788867, + "learning_rate": 3.102974828375286e-05, + "loss": 0.5798, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19807271659374237, + "step": 340, + "valid_targets_mean": 2020.0, + "valid_targets_min": 1191 + }, + { + "epoch": 0.5528846153846154, + "grad_norm": 0.7221088406680793, + "learning_rate": 3.1487414187643024e-05, + "loss": 0.5892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30220168828964233, + "step": 345, + "valid_targets_mean": 3140.1, + "valid_targets_min": 1105 + }, + { + "epoch": 0.5608974358974359, + "grad_norm": 0.7433702602085418, + "learning_rate": 3.1945080091533184e-05, + "loss": 0.5648, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32824429869651794, + "step": 350, + "valid_targets_mean": 3396.4, + "valid_targets_min": 1318 + }, + { + "epoch": 0.5689102564102564, + "grad_norm": 0.7889124411819611, + "learning_rate": 3.240274599542334e-05, + "loss": 0.567, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2554265260696411, + "step": 355, + "valid_targets_mean": 2997.5, + "valid_targets_min": 1277 + }, + { + "epoch": 0.5769230769230769, + "grad_norm": 0.7136883610030275, + "learning_rate": 3.2860411899313506e-05, + "loss": 0.6071, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30628496408462524, + "step": 360, + "valid_targets_mean": 3329.5, + "valid_targets_min": 1300 + }, + { + "epoch": 0.5849358974358975, + "grad_norm": 0.7608774745754158, + "learning_rate": 3.331807780320366e-05, + "loss": 0.6091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27490195631980896, + "step": 365, + "valid_targets_mean": 3102.2, + "valid_targets_min": 1458 + }, + { + "epoch": 0.592948717948718, + "grad_norm": 0.731910433713548, + "learning_rate": 3.377574370709382e-05, + "loss": 0.5993, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2733514606952667, + "step": 370, + "valid_targets_mean": 3291.9, + "valid_targets_min": 1109 + }, + { + "epoch": 0.6009615384615384, + "grad_norm": 0.6730063976162348, + "learning_rate": 3.423340961098399e-05, + "loss": 0.582, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2592528462409973, + "step": 375, + "valid_targets_mean": 3596.5, + "valid_targets_min": 1080 + }, + { + "epoch": 0.6089743589743589, + "grad_norm": 0.854474523055332, + "learning_rate": 3.469107551487414e-05, + "loss": 0.5572, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3616791367530823, + "step": 380, + "valid_targets_mean": 4046.5, + "valid_targets_min": 2860 + }, + { + "epoch": 0.6169871794871795, + "grad_norm": 0.6831231776793484, + "learning_rate": 3.5148741418764304e-05, + "loss": 0.5954, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3057277202606201, + "step": 385, + "valid_targets_mean": 3469.0, + "valid_targets_min": 1780 + }, + { + "epoch": 0.625, + "grad_norm": 0.8269093106749031, + "learning_rate": 3.5606407322654464e-05, + "loss": 0.5986, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33814537525177, + "step": 390, + "valid_targets_mean": 2999.0, + "valid_targets_min": 1572 + }, + { + "epoch": 0.6330128205128205, + "grad_norm": 0.9073771966241339, + "learning_rate": 3.6064073226544625e-05, + "loss": 0.5691, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27034977078437805, + "step": 395, + "valid_targets_mean": 2510.2, + "valid_targets_min": 1566 + }, + { + "epoch": 0.6410256410256411, + "grad_norm": 0.7758020143156383, + "learning_rate": 3.6521739130434786e-05, + "loss": 0.5678, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24109980463981628, + "step": 400, + "valid_targets_mean": 2395.4, + "valid_targets_min": 1419 + }, + { + "epoch": 0.6490384615384616, + "grad_norm": 0.8340729368403325, + "learning_rate": 3.697940503432495e-05, + "loss": 0.5637, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2517845034599304, + "step": 405, + "valid_targets_mean": 2416.4, + "valid_targets_min": 832 + }, + { + "epoch": 0.657051282051282, + "grad_norm": 0.7066664087292727, + "learning_rate": 3.743707093821511e-05, + "loss": 0.5706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24300941824913025, + "step": 410, + "valid_targets_mean": 3324.0, + "valid_targets_min": 1449 + }, + { + "epoch": 0.6650641025641025, + "grad_norm": 0.7778424231871143, + "learning_rate": 3.789473684210526e-05, + "loss": 0.6069, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2633519470691681, + "step": 415, + "valid_targets_mean": 2464.2, + "valid_targets_min": 1917 + }, + { + "epoch": 0.6730769230769231, + "grad_norm": 0.8409922086813345, + "learning_rate": 3.835240274599543e-05, + "loss": 0.5823, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2964377999305725, + "step": 420, + "valid_targets_mean": 3425.6, + "valid_targets_min": 1431 + }, + { + "epoch": 0.6810897435897436, + "grad_norm": 0.6894834332506039, + "learning_rate": 3.8810068649885584e-05, + "loss": 0.5661, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24966856837272644, + "step": 425, + "valid_targets_mean": 2792.6, + "valid_targets_min": 1731 + }, + { + "epoch": 0.6891025641025641, + "grad_norm": 0.7400767696813589, + "learning_rate": 3.9267734553775745e-05, + "loss": 0.5573, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3029435873031616, + "step": 430, + "valid_targets_mean": 3512.1, + "valid_targets_min": 1132 + }, + { + "epoch": 0.6971153846153846, + "grad_norm": 0.7098042958395291, + "learning_rate": 3.9725400457665905e-05, + "loss": 0.5766, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3583853542804718, + "step": 435, + "valid_targets_mean": 3790.5, + "valid_targets_min": 1222 + }, + { + "epoch": 0.7051282051282052, + "grad_norm": 0.8267285592318016, + "learning_rate": 3.999997445219712e-05, + "loss": 0.5625, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3317101299762726, + "step": 440, + "valid_targets_mean": 3392.5, + "valid_targets_min": 1529 + }, + { + "epoch": 0.7131410256410257, + "grad_norm": 0.6951674348117636, + "learning_rate": 3.999968704016428e-05, + "loss": 0.5622, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2969678044319153, + "step": 445, + "valid_targets_mean": 3420.4, + "valid_targets_min": 1673 + }, + { + "epoch": 0.7211538461538461, + "grad_norm": 0.7954953485363763, + "learning_rate": 3.9999080285949514e-05, + "loss": 0.5867, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.238167405128479, + "step": 450, + "valid_targets_mean": 2271.5, + "valid_targets_min": 1538 + }, + { + "epoch": 0.7291666666666666, + "grad_norm": 0.711381876343244, + "learning_rate": 3.999815419924108e-05, + "loss": 0.545, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.272437185049057, + "step": 455, + "valid_targets_mean": 3888.2, + "valid_targets_min": 1862 + }, + { + "epoch": 0.7371794871794872, + "grad_norm": 0.7417617077526117, + "learning_rate": 3.999690879482614e-05, + "loss": 0.6004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3043525218963623, + "step": 460, + "valid_targets_mean": 2486.0, + "valid_targets_min": 1393 + }, + { + "epoch": 0.7451923076923077, + "grad_norm": 0.7402990680812328, + "learning_rate": 3.9995344092590506e-05, + "loss": 0.5831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3080860674381256, + "step": 465, + "valid_targets_mean": 3419.1, + "valid_targets_min": 1412 + }, + { + "epoch": 0.7532051282051282, + "grad_norm": 0.6684658510073965, + "learning_rate": 3.999346011751835e-05, + "loss": 0.5489, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25505971908569336, + "step": 470, + "valid_targets_mean": 3117.9, + "valid_targets_min": 1208 + }, + { + "epoch": 0.7612179487179487, + "grad_norm": 0.6108916090579026, + "learning_rate": 3.999125689969176e-05, + "loss": 0.5706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24049857258796692, + "step": 475, + "valid_targets_mean": 3742.9, + "valid_targets_min": 975 + }, + { + "epoch": 0.7692307692307693, + "grad_norm": 0.8697710143115589, + "learning_rate": 3.9988734474290324e-05, + "loss": 0.5563, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30834394693374634, + "step": 480, + "valid_targets_mean": 3330.5, + "valid_targets_min": 1352 + }, + { + "epoch": 0.7772435897435898, + "grad_norm": 0.8178039384606676, + "learning_rate": 3.9985892881590513e-05, + "loss": 0.5372, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23117424547672272, + "step": 485, + "valid_targets_mean": 2532.6, + "valid_targets_min": 1511 + }, + { + "epoch": 0.7852564102564102, + "grad_norm": 1.0293996633669094, + "learning_rate": 3.9982732166965054e-05, + "loss": 0.6105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26724469661712646, + "step": 490, + "valid_targets_mean": 3244.2, + "valid_targets_min": 1421 + }, + { + "epoch": 0.7932692307692307, + "grad_norm": 0.6602804682902749, + "learning_rate": 3.997925238088221e-05, + "loss": 0.6127, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3438800871372223, + "step": 495, + "valid_targets_mean": 4257.4, + "valid_targets_min": 2336 + }, + { + "epoch": 0.8012820512820513, + "grad_norm": 0.7238258648359763, + "learning_rate": 3.9975453578904975e-05, + "loss": 0.5466, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31390106678009033, + "step": 500, + "valid_targets_mean": 3550.0, + "valid_targets_min": 1400 + }, + { + "epoch": 0.8092948717948718, + "grad_norm": 0.7574772765096325, + "learning_rate": 3.997133582169018e-05, + "loss": 0.5801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.301613986492157, + "step": 505, + "valid_targets_mean": 2869.1, + "valid_targets_min": 1258 + }, + { + "epoch": 0.8173076923076923, + "grad_norm": 0.7314431081040959, + "learning_rate": 3.996689917498754e-05, + "loss": 0.5661, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3173075020313263, + "step": 510, + "valid_targets_mean": 2976.9, + "valid_targets_min": 1819 + }, + { + "epoch": 0.8253205128205128, + "grad_norm": 0.704961796384467, + "learning_rate": 3.9962143709638585e-05, + "loss": 0.5698, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2546284794807434, + "step": 515, + "valid_targets_mean": 3917.4, + "valid_targets_min": 1207 + }, + { + "epoch": 0.8333333333333334, + "grad_norm": 0.7236498923856338, + "learning_rate": 3.995706950157554e-05, + "loss": 0.5813, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2781308591365814, + "step": 520, + "valid_targets_mean": 2678.0, + "valid_targets_min": 1177 + }, + { + "epoch": 0.8413461538461539, + "grad_norm": 0.6879750332716809, + "learning_rate": 3.995167663182008e-05, + "loss": 0.561, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2687014937400818, + "step": 525, + "valid_targets_mean": 3366.2, + "valid_targets_min": 1598 + }, + { + "epoch": 0.8493589743589743, + "grad_norm": 0.7876640148534456, + "learning_rate": 3.994596518648214e-05, + "loss": 0.5973, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27589696645736694, + "step": 530, + "valid_targets_mean": 2870.8, + "valid_targets_min": 1915 + }, + { + "epoch": 0.8573717948717948, + "grad_norm": 0.7441322134066838, + "learning_rate": 3.993993525675838e-05, + "loss": 0.5672, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29585331678390503, + "step": 535, + "valid_targets_mean": 3032.0, + "valid_targets_min": 1765 + }, + { + "epoch": 0.8653846153846154, + "grad_norm": 0.6796067438679934, + "learning_rate": 3.993358693893086e-05, + "loss": 0.5876, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2950208783149719, + "step": 540, + "valid_targets_mean": 3695.9, + "valid_targets_min": 1773 + }, + { + "epoch": 0.8733974358974359, + "grad_norm": 0.6876505154980246, + "learning_rate": 3.9926920334365457e-05, + "loss": 0.5817, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32896968722343445, + "step": 545, + "valid_targets_mean": 3169.4, + "valid_targets_min": 1222 + }, + { + "epoch": 0.8814102564102564, + "grad_norm": 0.6730762747339177, + "learning_rate": 3.991993554951023e-05, + "loss": 0.5339, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24927011132240295, + "step": 550, + "valid_targets_mean": 3364.6, + "valid_targets_min": 1786 + }, + { + "epoch": 0.8894230769230769, + "grad_norm": 0.7118986936736847, + "learning_rate": 3.991263269589376e-05, + "loss": 0.5681, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3718481659889221, + "step": 555, + "valid_targets_mean": 3274.1, + "valid_targets_min": 1530 + }, + { + "epoch": 0.8974358974358975, + "grad_norm": 0.6808706754982825, + "learning_rate": 3.990501189012332e-05, + "loss": 0.5348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2935500144958496, + "step": 560, + "valid_targets_mean": 3658.0, + "valid_targets_min": 1915 + }, + { + "epoch": 0.905448717948718, + "grad_norm": 0.697026766946665, + "learning_rate": 3.989707325388305e-05, + "loss": 0.5521, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2970616817474365, + "step": 565, + "valid_targets_mean": 3312.1, + "valid_targets_min": 963 + }, + { + "epoch": 0.9134615384615384, + "grad_norm": 0.7306580931343188, + "learning_rate": 3.9888816913932016e-05, + "loss": 0.5455, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2733103632926941, + "step": 570, + "valid_targets_mean": 2979.2, + "valid_targets_min": 1961 + }, + { + "epoch": 0.9214743589743589, + "grad_norm": 0.6123927313586777, + "learning_rate": 3.988024300210215e-05, + "loss": 0.5792, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2773978114128113, + "step": 575, + "valid_targets_mean": 3638.0, + "valid_targets_min": 1751 + }, + { + "epoch": 0.9294871794871795, + "grad_norm": 0.6869256098073525, + "learning_rate": 3.987135165529618e-05, + "loss": 0.5451, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32810235023498535, + "step": 580, + "valid_targets_mean": 4129.8, + "valid_targets_min": 1050 + }, + { + "epoch": 0.9375, + "grad_norm": 0.6619390784130208, + "learning_rate": 3.9862143015485446e-05, + "loss": 0.5535, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26915961503982544, + "step": 585, + "valid_targets_mean": 3638.8, + "valid_targets_min": 1792 + }, + { + "epoch": 0.9455128205128205, + "grad_norm": 0.6948700895501717, + "learning_rate": 3.985261722970759e-05, + "loss": 0.5085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20365601778030396, + "step": 590, + "valid_targets_mean": 3601.9, + "valid_targets_min": 975 + }, + { + "epoch": 0.9535256410256411, + "grad_norm": 0.7212645188353665, + "learning_rate": 3.984277445006426e-05, + "loss": 0.5729, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29846131801605225, + "step": 595, + "valid_targets_mean": 3109.9, + "valid_targets_min": 1552 + }, + { + "epoch": 0.9615384615384616, + "grad_norm": 0.6553829898385685, + "learning_rate": 3.9832614833718654e-05, + "loss": 0.5466, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2243560403585434, + "step": 600, + "valid_targets_mean": 3010.2, + "valid_targets_min": 890 + }, + { + "epoch": 0.969551282051282, + "grad_norm": 0.6401541506747835, + "learning_rate": 3.9822138542893005e-05, + "loss": 0.5446, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24172303080558777, + "step": 605, + "valid_targets_mean": 3785.6, + "valid_targets_min": 1326 + }, + { + "epoch": 0.9775641025641025, + "grad_norm": 0.7576921180597798, + "learning_rate": 3.9811345744866014e-05, + "loss": 0.552, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22569139301776886, + "step": 610, + "valid_targets_mean": 2737.8, + "valid_targets_min": 1701 + }, + { + "epoch": 0.9855769230769231, + "grad_norm": 0.780420438895869, + "learning_rate": 3.980023661197016e-05, + "loss": 0.6033, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.37722402811050415, + "step": 615, + "valid_targets_mean": 3719.5, + "valid_targets_min": 929 + }, + { + "epoch": 0.9935897435897436, + "grad_norm": 0.6990111689939135, + "learning_rate": 3.978881132158896e-05, + "loss": 0.5286, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31383395195007324, + "step": 620, + "valid_targets_mean": 2973.2, + "valid_targets_min": 1123 + }, + { + "epoch": 1.001602564102564, + "grad_norm": 0.6693535983646552, + "learning_rate": 3.9777070056154124e-05, + "loss": 0.5658, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.260606586933136, + "step": 625, + "valid_targets_mean": 3676.0, + "valid_targets_min": 2480 + }, + { + "epoch": 1.0096153846153846, + "grad_norm": 0.7098353962836165, + "learning_rate": 3.976501300314264e-05, + "loss": 0.5828, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2818909287452698, + "step": 630, + "valid_targets_mean": 3143.8, + "valid_targets_min": 1277 + }, + { + "epoch": 1.017628205128205, + "grad_norm": 0.663090582972122, + "learning_rate": 3.9752640355073825e-05, + "loss": 0.541, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29184168577194214, + "step": 635, + "valid_targets_mean": 4174.9, + "valid_targets_min": 2280 + }, + { + "epoch": 1.0256410256410255, + "grad_norm": 0.6544669358630107, + "learning_rate": 3.9739952309506175e-05, + "loss": 0.5157, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23627276718616486, + "step": 640, + "valid_targets_mean": 3359.8, + "valid_targets_min": 2015 + }, + { + "epoch": 1.0336538461538463, + "grad_norm": 0.7161004320937852, + "learning_rate": 3.972694906903427e-05, + "loss": 0.489, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21363291144371033, + "step": 645, + "valid_targets_mean": 3059.6, + "valid_targets_min": 2177 + }, + { + "epoch": 1.0416666666666667, + "grad_norm": 0.6883847742008947, + "learning_rate": 3.971363084128552e-05, + "loss": 0.5247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33157476782798767, + "step": 650, + "valid_targets_mean": 3736.0, + "valid_targets_min": 1348 + }, + { + "epoch": 1.0496794871794872, + "grad_norm": 0.6627380227713346, + "learning_rate": 3.969999783891685e-05, + "loss": 0.5068, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22591212391853333, + "step": 655, + "valid_targets_mean": 3279.5, + "valid_targets_min": 2484 + }, + { + "epoch": 1.0576923076923077, + "grad_norm": 0.7863795958962202, + "learning_rate": 3.96860502796113e-05, + "loss": 0.5414, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32518288493156433, + "step": 660, + "valid_targets_mean": 2789.1, + "valid_targets_min": 1314 + }, + { + "epoch": 1.0657051282051282, + "grad_norm": 0.8607621847459621, + "learning_rate": 3.967178838607456e-05, + "loss": 0.5552, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27951347827911377, + "step": 665, + "valid_targets_mean": 2587.4, + "valid_targets_min": 1132 + }, + { + "epoch": 1.0737179487179487, + "grad_norm": 0.7090441828077002, + "learning_rate": 3.965721238603139e-05, + "loss": 0.5292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2804470658302307, + "step": 670, + "valid_targets_mean": 3493.5, + "valid_targets_min": 1369 + }, + { + "epoch": 1.0817307692307692, + "grad_norm": 0.7303161801257573, + "learning_rate": 3.964232251222203e-05, + "loss": 0.5445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2374742031097412, + "step": 675, + "valid_targets_mean": 2382.2, + "valid_targets_min": 1830 + }, + { + "epoch": 1.0897435897435896, + "grad_norm": 0.6536572646610735, + "learning_rate": 3.962711900239844e-05, + "loss": 0.5318, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27097436785697937, + "step": 680, + "valid_targets_mean": 3724.1, + "valid_targets_min": 1150 + }, + { + "epoch": 1.0977564102564104, + "grad_norm": 0.7641607167806285, + "learning_rate": 3.961160209932051e-05, + "loss": 0.5339, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26693660020828247, + "step": 685, + "valid_targets_mean": 2725.9, + "valid_targets_min": 890 + }, + { + "epoch": 1.1057692307692308, + "grad_norm": 0.7787423353552926, + "learning_rate": 3.95957720507522e-05, + "loss": 0.5215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27508217096328735, + "step": 690, + "valid_targets_mean": 2505.0, + "valid_targets_min": 1296 + }, + { + "epoch": 1.1137820512820513, + "grad_norm": 0.7589181556991755, + "learning_rate": 3.957962910945759e-05, + "loss": 0.526, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24967530369758606, + "step": 695, + "valid_targets_mean": 2899.0, + "valid_targets_min": 1441 + }, + { + "epoch": 1.1217948717948718, + "grad_norm": 0.6978637028300161, + "learning_rate": 3.9563173533196805e-05, + "loss": 0.5335, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2285114973783493, + "step": 700, + "valid_targets_mean": 2929.0, + "valid_targets_min": 1364 + }, + { + "epoch": 1.1298076923076923, + "grad_norm": 0.6463768734605275, + "learning_rate": 3.954640558472195e-05, + "loss": 0.5364, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2634209990501404, + "step": 705, + "valid_targets_mean": 3257.6, + "valid_targets_min": 2146 + }, + { + "epoch": 1.1378205128205128, + "grad_norm": 0.6742776679241153, + "learning_rate": 3.952932553177287e-05, + "loss": 0.5172, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3074655532836914, + "step": 710, + "valid_targets_mean": 4021.2, + "valid_targets_min": 2139 + }, + { + "epoch": 1.1458333333333333, + "grad_norm": 0.6406102168762207, + "learning_rate": 3.95119336470729e-05, + "loss": 0.5399, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24337883293628693, + "step": 715, + "valid_targets_mean": 3200.2, + "valid_targets_min": 1139 + }, + { + "epoch": 1.1538461538461537, + "grad_norm": 0.6470392948348622, + "learning_rate": 3.949423020832451e-05, + "loss": 0.5593, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2592815160751343, + "step": 720, + "valid_targets_mean": 3938.0, + "valid_targets_min": 1996 + }, + { + "epoch": 1.1618589743589745, + "grad_norm": 0.6438135697706928, + "learning_rate": 3.947621549820485e-05, + "loss": 0.5088, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21767720580101013, + "step": 725, + "valid_targets_mean": 3209.6, + "valid_targets_min": 1806 + }, + { + "epoch": 1.169871794871795, + "grad_norm": 0.7101640126566555, + "learning_rate": 3.945788980436129e-05, + "loss": 0.5143, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2374275177717209, + "step": 730, + "valid_targets_mean": 3167.0, + "valid_targets_min": 1306 + }, + { + "epoch": 1.1778846153846154, + "grad_norm": 0.7235040211741174, + "learning_rate": 3.943925341940673e-05, + "loss": 0.5198, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22331471741199493, + "step": 735, + "valid_targets_mean": 2386.0, + "valid_targets_min": 954 + }, + { + "epoch": 1.185897435897436, + "grad_norm": 0.7415491424178813, + "learning_rate": 3.942030664091503e-05, + "loss": 0.5195, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3045019209384918, + "step": 740, + "valid_targets_mean": 3849.6, + "valid_targets_min": 1105 + }, + { + "epoch": 1.1939102564102564, + "grad_norm": 0.688111048663959, + "learning_rate": 3.9401049771416214e-05, + "loss": 0.5314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27955904603004456, + "step": 745, + "valid_targets_mean": 3316.6, + "valid_targets_min": 1595 + }, + { + "epoch": 1.2019230769230769, + "grad_norm": 0.6801217648301756, + "learning_rate": 3.938148311839162e-05, + "loss": 0.5278, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2693995237350464, + "step": 750, + "valid_targets_mean": 3905.8, + "valid_targets_min": 1303 + }, + { + "epoch": 1.2099358974358974, + "grad_norm": 0.6292118806078293, + "learning_rate": 3.9361606994269014e-05, + "loss": 0.517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2977278232574463, + "step": 755, + "valid_targets_mean": 4322.6, + "valid_targets_min": 2718 + }, + { + "epoch": 1.217948717948718, + "grad_norm": 0.6223360071859224, + "learning_rate": 3.934142171641763e-05, + "loss": 0.5336, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3141154944896698, + "step": 760, + "valid_targets_mean": 4818.8, + "valid_targets_min": 2504 + }, + { + "epoch": 1.2259615384615385, + "grad_norm": 0.7008654525673063, + "learning_rate": 3.9320927607143003e-05, + "loss": 0.544, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27953052520751953, + "step": 765, + "valid_targets_mean": 3073.0, + "valid_targets_min": 1637 + }, + { + "epoch": 1.233974358974359, + "grad_norm": 0.7228308757060894, + "learning_rate": 3.9300124993681976e-05, + "loss": 0.5579, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3026982545852661, + "step": 770, + "valid_targets_mean": 3506.4, + "valid_targets_min": 2496 + }, + { + "epoch": 1.2419871794871795, + "grad_norm": 0.7243923899991324, + "learning_rate": 3.9279014208197317e-05, + "loss": 0.5541, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19519637525081635, + "step": 775, + "valid_targets_mean": 2395.1, + "valid_targets_min": 920 + }, + { + "epoch": 1.25, + "grad_norm": 0.6158625841705502, + "learning_rate": 3.925759558777252e-05, + "loss": 0.5336, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22651325166225433, + "step": 780, + "valid_targets_mean": 3587.2, + "valid_targets_min": 1765 + }, + { + "epoch": 1.2580128205128205, + "grad_norm": 0.7059552328656222, + "learning_rate": 3.923586947440639e-05, + "loss": 0.5268, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2652747631072998, + "step": 785, + "valid_targets_mean": 3377.8, + "valid_targets_min": 1294 + }, + { + "epoch": 1.266025641025641, + "grad_norm": 0.6195789364544204, + "learning_rate": 3.921383621500758e-05, + "loss": 0.5041, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26103103160858154, + "step": 790, + "valid_targets_mean": 3506.1, + "valid_targets_min": 1368 + }, + { + "epoch": 1.2740384615384617, + "grad_norm": 0.64102800430927, + "learning_rate": 3.919149616138906e-05, + "loss": 0.5446, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27611809968948364, + "step": 795, + "valid_targets_mean": 3347.2, + "valid_targets_min": 1410 + }, + { + "epoch": 1.282051282051282, + "grad_norm": 0.634026707134316, + "learning_rate": 3.916884967026246e-05, + "loss": 0.5421, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3282822370529175, + "step": 800, + "valid_targets_mean": 3828.8, + "valid_targets_min": 965 + }, + { + "epoch": 1.2900641025641026, + "grad_norm": 0.6124529838275565, + "learning_rate": 3.914589710323245e-05, + "loss": 0.4931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2654038071632385, + "step": 805, + "valid_targets_mean": 3368.5, + "valid_targets_min": 2025 + }, + { + "epoch": 1.2980769230769231, + "grad_norm": 0.5798169056754897, + "learning_rate": 3.912263882679091e-05, + "loss": 0.5118, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27689328789711, + "step": 810, + "valid_targets_mean": 3953.1, + "valid_targets_min": 1147 + }, + { + "epoch": 1.3060897435897436, + "grad_norm": 0.7360585563860171, + "learning_rate": 3.9099075212311076e-05, + "loss": 0.5049, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2761501967906952, + "step": 815, + "valid_targets_mean": 3296.1, + "valid_targets_min": 1718 + }, + { + "epoch": 1.314102564102564, + "grad_norm": 0.7369777371156171, + "learning_rate": 3.9075206636041646e-05, + "loss": 0.5442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23401018977165222, + "step": 820, + "valid_targets_mean": 2535.9, + "valid_targets_min": 1576 + }, + { + "epoch": 1.3221153846153846, + "grad_norm": 0.6139072781311388, + "learning_rate": 3.905103347910075e-05, + "loss": 0.5011, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22416281700134277, + "step": 825, + "valid_targets_mean": 3596.5, + "valid_targets_min": 2253 + }, + { + "epoch": 1.330128205128205, + "grad_norm": 0.695393374549869, + "learning_rate": 3.902655612746985e-05, + "loss": 0.5498, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3552865982055664, + "step": 830, + "valid_targets_mean": 3999.4, + "valid_targets_min": 1396 + }, + { + "epoch": 1.3381410256410255, + "grad_norm": 0.6343396009128547, + "learning_rate": 3.900177497198761e-05, + "loss": 0.5102, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22865472733974457, + "step": 835, + "valid_targets_mean": 3107.8, + "valid_targets_min": 1061 + }, + { + "epoch": 1.3461538461538463, + "grad_norm": 0.662665718958983, + "learning_rate": 3.8976690408343635e-05, + "loss": 0.5582, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26849493384361267, + "step": 840, + "valid_targets_mean": 3564.8, + "valid_targets_min": 1627 + }, + { + "epoch": 1.3541666666666667, + "grad_norm": 0.6264358191163301, + "learning_rate": 3.8951302837072165e-05, + "loss": 0.5099, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22350981831550598, + "step": 845, + "valid_targets_mean": 3560.2, + "valid_targets_min": 1225 + }, + { + "epoch": 1.3621794871794872, + "grad_norm": 0.6400671426580623, + "learning_rate": 3.892561266354566e-05, + "loss": 0.5142, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2687187194824219, + "step": 850, + "valid_targets_mean": 3577.8, + "valid_targets_min": 1820 + }, + { + "epoch": 1.3701923076923077, + "grad_norm": 0.7260804634489083, + "learning_rate": 3.889962029796833e-05, + "loss": 0.5306, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28317689895629883, + "step": 855, + "valid_targets_mean": 2975.5, + "valid_targets_min": 1586 + }, + { + "epoch": 1.3782051282051282, + "grad_norm": 0.7251345187758271, + "learning_rate": 3.887332615536962e-05, + "loss": 0.5239, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3422122001647949, + "step": 860, + "valid_targets_mean": 3427.1, + "valid_targets_min": 1786 + }, + { + "epoch": 1.3862179487179487, + "grad_norm": 0.6399218646571189, + "learning_rate": 3.8846730655597535e-05, + "loss": 0.5186, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34644943475723267, + "step": 865, + "valid_targets_mean": 4054.6, + "valid_targets_min": 1610 + }, + { + "epoch": 1.3942307692307692, + "grad_norm": 0.6922616704310385, + "learning_rate": 3.881983422331198e-05, + "loss": 0.5468, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23638592660427094, + "step": 870, + "valid_targets_mean": 3441.1, + "valid_targets_min": 1526 + }, + { + "epoch": 1.4022435897435899, + "grad_norm": 0.869942916137242, + "learning_rate": 3.879263728797792e-05, + "loss": 0.5015, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25148484110832214, + "step": 875, + "valid_targets_mean": 2407.4, + "valid_targets_min": 1359 + }, + { + "epoch": 1.4102564102564101, + "grad_norm": 0.6198340929906818, + "learning_rate": 3.876514028385861e-05, + "loss": 0.4915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21183347702026367, + "step": 880, + "valid_targets_mean": 3489.0, + "valid_targets_min": 1342 + }, + { + "epoch": 1.4182692307692308, + "grad_norm": 0.771740909392385, + "learning_rate": 3.873734365000857e-05, + "loss": 0.5515, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.307413786649704, + "step": 885, + "valid_targets_mean": 3279.5, + "valid_targets_min": 1683 + }, + { + "epoch": 1.4262820512820513, + "grad_norm": 0.657804162466593, + "learning_rate": 3.870924783026663e-05, + "loss": 0.515, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27265098690986633, + "step": 890, + "valid_targets_mean": 3807.9, + "valid_targets_min": 1678 + }, + { + "epoch": 1.4342948717948718, + "grad_norm": 0.6096577149655553, + "learning_rate": 3.8680853273248826e-05, + "loss": 0.5177, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2563854157924652, + "step": 895, + "valid_targets_mean": 3631.6, + "valid_targets_min": 2078 + }, + { + "epoch": 1.4423076923076923, + "grad_norm": 0.6335175697292809, + "learning_rate": 3.865216043234126e-05, + "loss": 0.4936, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28854525089263916, + "step": 900, + "valid_targets_mean": 3971.0, + "valid_targets_min": 1851 + }, + { + "epoch": 1.4503205128205128, + "grad_norm": 0.689067326457323, + "learning_rate": 3.862316976569281e-05, + "loss": 0.5145, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3020906448364258, + "step": 905, + "valid_targets_mean": 3140.1, + "valid_targets_min": 1201 + }, + { + "epoch": 1.4583333333333333, + "grad_norm": 0.5888442051030699, + "learning_rate": 3.859388173620785e-05, + "loss": 0.4967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23648330569267273, + "step": 910, + "valid_targets_mean": 3785.5, + "valid_targets_min": 2554 + }, + { + "epoch": 1.4663461538461537, + "grad_norm": 0.6819048703373433, + "learning_rate": 3.8564296811538874e-05, + "loss": 0.5083, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21770158410072327, + "step": 915, + "valid_targets_mean": 3246.5, + "valid_targets_min": 1489 + }, + { + "epoch": 1.4743589743589745, + "grad_norm": 0.603547484916333, + "learning_rate": 3.853441546407898e-05, + "loss": 0.5424, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24904467165470123, + "step": 920, + "valid_targets_mean": 3848.2, + "valid_targets_min": 1383 + }, + { + "epoch": 1.482371794871795, + "grad_norm": 0.6070001286859679, + "learning_rate": 3.850423817095438e-05, + "loss": 0.5221, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2704816460609436, + "step": 925, + "valid_targets_mean": 3577.2, + "valid_targets_min": 1606 + }, + { + "epoch": 1.4903846153846154, + "grad_norm": 0.7339144876408403, + "learning_rate": 3.847376541401674e-05, + "loss": 0.5446, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25435471534729004, + "step": 930, + "valid_targets_mean": 2866.5, + "valid_targets_min": 1254 + }, + { + "epoch": 1.498397435897436, + "grad_norm": 0.6748847092208896, + "learning_rate": 3.844299767983551e-05, + "loss": 0.5334, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2652042508125305, + "step": 935, + "valid_targets_mean": 2942.2, + "valid_targets_min": 1673 + }, + { + "epoch": 1.5064102564102564, + "grad_norm": 0.5810912239888301, + "learning_rate": 3.841193545969015e-05, + "loss": 0.5161, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27438345551490784, + "step": 940, + "valid_targets_mean": 3674.2, + "valid_targets_min": 1399 + }, + { + "epoch": 1.5144230769230769, + "grad_norm": 0.7935476830979005, + "learning_rate": 3.8380579249562265e-05, + "loss": 0.5105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27584928274154663, + "step": 945, + "valid_targets_mean": 2631.9, + "valid_targets_min": 1151 + }, + { + "epoch": 1.5224358974358974, + "grad_norm": 0.6468795075216801, + "learning_rate": 3.8348929550127734e-05, + "loss": 0.4847, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23122943937778473, + "step": 950, + "valid_targets_mean": 2821.5, + "valid_targets_min": 1173 + }, + { + "epoch": 1.530448717948718, + "grad_norm": 0.6417076870537934, + "learning_rate": 3.831698686674866e-05, + "loss": 0.5687, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34728938341140747, + "step": 955, + "valid_targets_mean": 3951.6, + "valid_targets_min": 1145 + }, + { + "epoch": 1.5384615384615383, + "grad_norm": 0.6429228268225295, + "learning_rate": 3.828475170946534e-05, + "loss": 0.5203, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.277060329914093, + "step": 960, + "valid_targets_mean": 3545.1, + "valid_targets_min": 1347 + }, + { + "epoch": 1.546474358974359, + "grad_norm": 0.6845959324077293, + "learning_rate": 3.8252224592988087e-05, + "loss": 0.5172, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2928009033203125, + "step": 965, + "valid_targets_mean": 3707.2, + "valid_targets_min": 1248 + }, + { + "epoch": 1.5544871794871795, + "grad_norm": 0.644417122790312, + "learning_rate": 3.821940603668906e-05, + "loss": 0.529, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25426438450813293, + "step": 970, + "valid_targets_mean": 3476.6, + "valid_targets_min": 2036 + }, + { + "epoch": 1.5625, + "grad_norm": 0.6364515272988177, + "learning_rate": 3.8186296564593924e-05, + "loss": 0.5386, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2075498253107071, + "step": 975, + "valid_targets_mean": 3171.4, + "valid_targets_min": 777 + }, + { + "epoch": 1.5705128205128205, + "grad_norm": 0.5688757500865562, + "learning_rate": 3.815289670537351e-05, + "loss": 0.4983, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29371678829193115, + "step": 980, + "valid_targets_mean": 4655.0, + "valid_targets_min": 1551 + }, + { + "epoch": 1.578525641025641, + "grad_norm": 0.6901000947224574, + "learning_rate": 3.811920699233535e-05, + "loss": 0.5514, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27325496077537537, + "step": 985, + "valid_targets_mean": 3302.6, + "valid_targets_min": 1358 + }, + { + "epoch": 1.5865384615384617, + "grad_norm": 0.7154589912655275, + "learning_rate": 3.8085227963415186e-05, + "loss": 0.5246, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.34031611680984497, + "step": 990, + "valid_targets_mean": 4200.9, + "valid_targets_min": 1625 + }, + { + "epoch": 1.594551282051282, + "grad_norm": 0.5849892762277402, + "learning_rate": 3.805096016116838e-05, + "loss": 0.5319, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24774928390979767, + "step": 995, + "valid_targets_mean": 3870.5, + "valid_targets_min": 1829 + }, + { + "epoch": 1.6025641025641026, + "grad_norm": 0.6436387301055653, + "learning_rate": 3.801640413276121e-05, + "loss": 0.5422, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20658418536186218, + "step": 1000, + "valid_targets_mean": 2460.6, + "valid_targets_min": 1116 + }, + { + "epoch": 1.6105769230769231, + "grad_norm": 0.5861496751447847, + "learning_rate": 3.7981560429962204e-05, + "loss": 0.487, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2669129967689514, + "step": 1005, + "valid_targets_mean": 4720.4, + "valid_targets_min": 1498 + }, + { + "epoch": 1.6185897435897436, + "grad_norm": 0.6321853703205694, + "learning_rate": 3.7946429609133274e-05, + "loss": 0.5211, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20208066701889038, + "step": 1010, + "valid_targets_mean": 2828.6, + "valid_targets_min": 1458 + }, + { + "epoch": 1.626602564102564, + "grad_norm": 0.5938934570569941, + "learning_rate": 3.791101223122084e-05, + "loss": 0.5188, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20370927453041077, + "step": 1015, + "valid_targets_mean": 2865.6, + "valid_targets_min": 1169 + }, + { + "epoch": 1.6346153846153846, + "grad_norm": 0.716207864226055, + "learning_rate": 3.787530886174688e-05, + "loss": 0.5088, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21031898260116577, + "step": 1020, + "valid_targets_mean": 3724.8, + "valid_targets_min": 1258 + }, + { + "epoch": 1.6426282051282053, + "grad_norm": 0.6046762418058222, + "learning_rate": 3.783932007079992e-05, + "loss": 0.5486, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.277350515127182, + "step": 1025, + "valid_targets_mean": 4345.4, + "valid_targets_min": 2128 + }, + { + "epoch": 1.6506410256410255, + "grad_norm": 0.6715149255707481, + "learning_rate": 3.7803046433025905e-05, + "loss": 0.4929, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2609095871448517, + "step": 1030, + "valid_targets_mean": 2605.5, + "valid_targets_min": 1028 + }, + { + "epoch": 1.6586538461538463, + "grad_norm": 0.634880125309771, + "learning_rate": 3.7766488527619024e-05, + "loss": 0.52, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29789501428604126, + "step": 1035, + "valid_targets_mean": 3737.0, + "valid_targets_min": 1540 + }, + { + "epoch": 1.6666666666666665, + "grad_norm": 0.5910363110251786, + "learning_rate": 3.772964693831247e-05, + "loss": 0.4955, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2488008439540863, + "step": 1040, + "valid_targets_mean": 4164.1, + "valid_targets_min": 2806 + }, + { + "epoch": 1.6746794871794872, + "grad_norm": 0.6271135303025038, + "learning_rate": 3.7692522253369136e-05, + "loss": 0.5276, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23178881406784058, + "step": 1045, + "valid_targets_mean": 3305.2, + "valid_targets_min": 1684 + }, + { + "epoch": 1.6826923076923077, + "grad_norm": 0.6863528385490293, + "learning_rate": 3.7655115065572194e-05, + "loss": 0.5139, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2752208709716797, + "step": 1050, + "valid_targets_mean": 3002.0, + "valid_targets_min": 708 + }, + { + "epoch": 1.6907051282051282, + "grad_norm": 0.569826556935927, + "learning_rate": 3.7617425972215626e-05, + "loss": 0.4682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2438904047012329, + "step": 1055, + "valid_targets_mean": 3605.6, + "valid_targets_min": 1895 + }, + { + "epoch": 1.6987179487179487, + "grad_norm": 0.6240684658771203, + "learning_rate": 3.757945557509472e-05, + "loss": 0.5348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28922080993652344, + "step": 1060, + "valid_targets_mean": 3543.9, + "valid_targets_min": 2065 + }, + { + "epoch": 1.7067307692307692, + "grad_norm": 0.6094842835966949, + "learning_rate": 3.7541204480496444e-05, + "loss": 0.5192, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26414355635643005, + "step": 1065, + "valid_targets_mean": 4152.5, + "valid_targets_min": 1563 + }, + { + "epoch": 1.7147435897435899, + "grad_norm": 0.6359779653852059, + "learning_rate": 3.7502673299189745e-05, + "loss": 0.5074, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29333317279815674, + "step": 1070, + "valid_targets_mean": 4449.2, + "valid_targets_min": 2160 + }, + { + "epoch": 1.7227564102564101, + "grad_norm": 0.6390054563573255, + "learning_rate": 3.746386264641583e-05, + "loss": 0.5368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2948388159275055, + "step": 1075, + "valid_targets_mean": 3832.0, + "valid_targets_min": 924 + }, + { + "epoch": 1.7307692307692308, + "grad_norm": 0.7028585699486477, + "learning_rate": 3.7424773141878324e-05, + "loss": 0.5334, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3039441704750061, + "step": 1080, + "valid_targets_mean": 3563.4, + "valid_targets_min": 1631 + }, + { + "epoch": 1.7387820512820513, + "grad_norm": 0.6539803018002867, + "learning_rate": 3.738540540973338e-05, + "loss": 0.5129, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3273501992225647, + "step": 1085, + "valid_targets_mean": 3519.8, + "valid_targets_min": 1334 + }, + { + "epoch": 1.7467948717948718, + "grad_norm": 0.662427070423614, + "learning_rate": 3.7345760078579695e-05, + "loss": 0.5256, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3128707706928253, + "step": 1090, + "valid_targets_mean": 3408.9, + "valid_targets_min": 1636 + }, + { + "epoch": 1.7548076923076923, + "grad_norm": 0.5879525182017116, + "learning_rate": 3.730583778144852e-05, + "loss": 0.5195, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23824341595172882, + "step": 1095, + "valid_targets_mean": 3399.2, + "valid_targets_min": 1656 + }, + { + "epoch": 1.7628205128205128, + "grad_norm": 0.5912265123471225, + "learning_rate": 3.7265639155793494e-05, + "loss": 0.522, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2527797520160675, + "step": 1100, + "valid_targets_mean": 3880.9, + "valid_targets_min": 1801 + }, + { + "epoch": 1.7708333333333335, + "grad_norm": 0.641077623227018, + "learning_rate": 3.7225164843480503e-05, + "loss": 0.528, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26728448271751404, + "step": 1105, + "valid_targets_mean": 3456.8, + "valid_targets_min": 1424 + }, + { + "epoch": 1.7788461538461537, + "grad_norm": 0.6376128785180524, + "learning_rate": 3.7184415490777426e-05, + "loss": 0.5171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25470465421676636, + "step": 1110, + "valid_targets_mean": 3636.0, + "valid_targets_min": 1641 + }, + { + "epoch": 1.7868589743589745, + "grad_norm": 0.601744936260028, + "learning_rate": 3.714339174834379e-05, + "loss": 0.5287, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25488728284835815, + "step": 1115, + "valid_targets_mean": 3772.8, + "valid_targets_min": 1646 + }, + { + "epoch": 1.7948717948717947, + "grad_norm": 0.6203437577351082, + "learning_rate": 3.710209427122044e-05, + "loss": 0.493, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2158557027578354, + "step": 1120, + "valid_targets_mean": 2691.9, + "valid_targets_min": 1446 + }, + { + "epoch": 1.8028846153846154, + "grad_norm": 0.6535039740122612, + "learning_rate": 3.7060523718819e-05, + "loss": 0.5003, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30899614095687866, + "step": 1125, + "valid_targets_mean": 3524.8, + "valid_targets_min": 1706 + }, + { + "epoch": 1.810897435897436, + "grad_norm": 0.6994516188114505, + "learning_rate": 3.701868075491139e-05, + "loss": 0.5375, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2746647894382477, + "step": 1130, + "valid_targets_mean": 3063.6, + "valid_targets_min": 1632 + }, + { + "epoch": 1.8189102564102564, + "grad_norm": 0.7394909436899711, + "learning_rate": 3.697656604761926e-05, + "loss": 0.5086, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28098416328430176, + "step": 1135, + "valid_targets_mean": 2982.4, + "valid_targets_min": 1559 + }, + { + "epoch": 1.8269230769230769, + "grad_norm": 0.604356109493262, + "learning_rate": 3.693418026940325e-05, + "loss": 0.5056, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20717529952526093, + "step": 1140, + "valid_targets_mean": 3181.8, + "valid_targets_min": 1438 + }, + { + "epoch": 1.8349358974358974, + "grad_norm": 0.6491307283184545, + "learning_rate": 3.689152409705229e-05, + "loss": 0.4896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24789538979530334, + "step": 1145, + "valid_targets_mean": 3535.5, + "valid_targets_min": 1301 + }, + { + "epoch": 1.842948717948718, + "grad_norm": 0.6916515768964382, + "learning_rate": 3.6848598211672794e-05, + "loss": 0.4998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2222422957420349, + "step": 1150, + "valid_targets_mean": 2472.4, + "valid_targets_min": 1713 + }, + { + "epoch": 1.8509615384615383, + "grad_norm": 0.7209202823039291, + "learning_rate": 3.6805403298677797e-05, + "loss": 0.4874, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2711182236671448, + "step": 1155, + "valid_targets_mean": 2882.9, + "valid_targets_min": 1797 + }, + { + "epoch": 1.858974358974359, + "grad_norm": 0.6803774343041409, + "learning_rate": 3.6761940047775966e-05, + "loss": 0.5301, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27900511026382446, + "step": 1160, + "valid_targets_mean": 3306.1, + "valid_targets_min": 1147 + }, + { + "epoch": 1.8669871794871795, + "grad_norm": 0.7119991437456121, + "learning_rate": 3.671820915296063e-05, + "loss": 0.5226, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23256748914718628, + "step": 1165, + "valid_targets_mean": 2749.6, + "valid_targets_min": 1697 + }, + { + "epoch": 1.875, + "grad_norm": 0.6412309344299336, + "learning_rate": 3.667421131249869e-05, + "loss": 0.5085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2757815420627594, + "step": 1170, + "valid_targets_mean": 3552.4, + "valid_targets_min": 2082 + }, + { + "epoch": 1.8830128205128205, + "grad_norm": 0.631900359779816, + "learning_rate": 3.662994722891946e-05, + "loss": 0.5034, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2032797932624817, + "step": 1175, + "valid_targets_mean": 2642.8, + "valid_targets_min": 1413 + }, + { + "epoch": 1.891025641025641, + "grad_norm": 0.6501888947810458, + "learning_rate": 3.658541760900344e-05, + "loss": 0.5253, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2954370379447937, + "step": 1180, + "valid_targets_mean": 3798.0, + "valid_targets_min": 1592 + }, + { + "epoch": 1.8990384615384617, + "grad_norm": 0.5951925718012213, + "learning_rate": 3.654062316377106e-05, + "loss": 0.4923, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2648494243621826, + "step": 1185, + "valid_targets_mean": 3257.1, + "valid_targets_min": 2271 + }, + { + "epoch": 1.907051282051282, + "grad_norm": 0.6344922879371704, + "learning_rate": 3.649556460847131e-05, + "loss": 0.4968, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2527773380279541, + "step": 1190, + "valid_targets_mean": 3105.8, + "valid_targets_min": 1905 + }, + { + "epoch": 1.9150641025641026, + "grad_norm": 0.666863022512596, + "learning_rate": 3.6450242662570314e-05, + "loss": 0.5464, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22034277021884918, + "step": 1195, + "valid_targets_mean": 2297.0, + "valid_targets_min": 1675 + }, + { + "epoch": 1.9230769230769231, + "grad_norm": 0.634393921153026, + "learning_rate": 3.6404658049739854e-05, + "loss": 0.5017, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33139002323150635, + "step": 1200, + "valid_targets_mean": 3685.6, + "valid_targets_min": 1547 + }, + { + "epoch": 1.9310897435897436, + "grad_norm": 0.603420777981733, + "learning_rate": 3.63588114978458e-05, + "loss": 0.5117, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.268787682056427, + "step": 1205, + "valid_targets_mean": 3892.5, + "valid_targets_min": 2533 + }, + { + "epoch": 1.939102564102564, + "grad_norm": 0.5758119768467407, + "learning_rate": 3.6312703738936504e-05, + "loss": 0.4967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24584612250328064, + "step": 1210, + "valid_targets_mean": 3714.8, + "valid_targets_min": 1949 + }, + { + "epoch": 1.9471153846153846, + "grad_norm": 0.6473776443066526, + "learning_rate": 3.626633550923111e-05, + "loss": 0.4842, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19100603461265564, + "step": 1215, + "valid_targets_mean": 2641.8, + "valid_targets_min": 1372 + }, + { + "epoch": 1.9551282051282053, + "grad_norm": 0.6295920932789495, + "learning_rate": 3.621970754910778e-05, + "loss": 0.4889, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1859293133020401, + "step": 1220, + "valid_targets_mean": 2846.8, + "valid_targets_min": 1881 + }, + { + "epoch": 1.9631410256410255, + "grad_norm": 0.5638732663009508, + "learning_rate": 3.6172820603091885e-05, + "loss": 0.4942, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27063971757888794, + "step": 1225, + "valid_targets_mean": 4205.5, + "valid_targets_min": 3027 + }, + { + "epoch": 1.9711538461538463, + "grad_norm": 0.7013738798900168, + "learning_rate": 3.612567541984413e-05, + "loss": 0.4815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19793829321861267, + "step": 1230, + "valid_targets_mean": 2972.1, + "valid_targets_min": 1169 + }, + { + "epoch": 1.9791666666666665, + "grad_norm": 0.6992819909391444, + "learning_rate": 3.6078272752148574e-05, + "loss": 0.5211, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2753292918205261, + "step": 1235, + "valid_targets_mean": 3305.8, + "valid_targets_min": 1057 + }, + { + "epoch": 1.9871794871794872, + "grad_norm": 0.6654447125313435, + "learning_rate": 3.6030613356900635e-05, + "loss": 0.5182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24492907524108887, + "step": 1240, + "valid_targets_mean": 3007.1, + "valid_targets_min": 1583 + }, + { + "epoch": 1.9951923076923077, + "grad_norm": 0.6293757629100242, + "learning_rate": 3.598269799509498e-05, + "loss": 0.4999, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17128071188926697, + "step": 1245, + "valid_targets_mean": 2834.0, + "valid_targets_min": 1164 + }, + { + "epoch": 2.003205128205128, + "grad_norm": 0.5855951997570703, + "learning_rate": 3.5934527431813385e-05, + "loss": 0.5314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22649341821670532, + "step": 1250, + "valid_targets_mean": 3951.8, + "valid_targets_min": 1789 + }, + { + "epoch": 2.011217948717949, + "grad_norm": 0.745592364024377, + "learning_rate": 3.5886102436212536e-05, + "loss": 0.4891, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32366320490837097, + "step": 1255, + "valid_targets_mean": 3810.0, + "valid_targets_min": 1102 + }, + { + "epoch": 2.019230769230769, + "grad_norm": 0.6474533150807608, + "learning_rate": 3.583742378151171e-05, + "loss": 0.4868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3340919315814972, + "step": 1260, + "valid_targets_mean": 4271.2, + "valid_targets_min": 1313 + }, + { + "epoch": 2.02724358974359, + "grad_norm": 0.9231892933611271, + "learning_rate": 3.5788492244980464e-05, + "loss": 0.4944, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23930369317531586, + "step": 1265, + "valid_targets_mean": 3406.1, + "valid_targets_min": 751 + }, + { + "epoch": 2.03525641025641, + "grad_norm": 0.5967116221953936, + "learning_rate": 3.573930860792621e-05, + "loss": 0.4826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23936381936073303, + "step": 1270, + "valid_targets_mean": 3996.6, + "valid_targets_min": 1487 + }, + { + "epoch": 2.043269230769231, + "grad_norm": 0.6845436282944499, + "learning_rate": 3.568987365568173e-05, + "loss": 0.4955, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2506715655326843, + "step": 1275, + "valid_targets_mean": 2830.9, + "valid_targets_min": 1239 + }, + { + "epoch": 2.051282051282051, + "grad_norm": 0.6603489942252528, + "learning_rate": 3.564018817759266e-05, + "loss": 0.4912, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2369793951511383, + "step": 1280, + "valid_targets_mean": 2957.4, + "valid_targets_min": 1512 + }, + { + "epoch": 2.059294871794872, + "grad_norm": 0.6745587821383664, + "learning_rate": 3.559025296700484e-05, + "loss": 0.4789, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2429685890674591, + "step": 1285, + "valid_targets_mean": 2875.2, + "valid_targets_min": 1442 + }, + { + "epoch": 2.0673076923076925, + "grad_norm": 0.6428739323975855, + "learning_rate": 3.554006882125173e-05, + "loss": 0.4888, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31421977281570435, + "step": 1290, + "valid_targets_mean": 3832.6, + "valid_targets_min": 1904 + }, + { + "epoch": 2.0753205128205128, + "grad_norm": 0.6812964971727579, + "learning_rate": 3.5489636541641586e-05, + "loss": 0.4921, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23777496814727783, + "step": 1295, + "valid_targets_mean": 2616.4, + "valid_targets_min": 755 + }, + { + "epoch": 2.0833333333333335, + "grad_norm": 0.6098267836001501, + "learning_rate": 3.543895693344472e-05, + "loss": 0.478, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21908503770828247, + "step": 1300, + "valid_targets_mean": 3908.2, + "valid_targets_min": 1184 + }, + { + "epoch": 2.0913461538461537, + "grad_norm": 0.634276858890153, + "learning_rate": 3.538803080588063e-05, + "loss": 0.4908, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23947657644748688, + "step": 1305, + "valid_targets_mean": 3484.8, + "valid_targets_min": 1347 + }, + { + "epoch": 2.0993589743589745, + "grad_norm": 0.6249208103974271, + "learning_rate": 3.5336858972105076e-05, + "loss": 0.4789, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23222315311431885, + "step": 1310, + "valid_targets_mean": 3566.4, + "valid_targets_min": 1418 + }, + { + "epoch": 2.1073717948717947, + "grad_norm": 0.6801231960769616, + "learning_rate": 3.528544224919708e-05, + "loss": 0.4858, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22080151736736298, + "step": 1315, + "valid_targets_mean": 3204.1, + "valid_targets_min": 813 + }, + { + "epoch": 2.1153846153846154, + "grad_norm": 0.6604190934606706, + "learning_rate": 3.5233781458145934e-05, + "loss": 0.4942, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27828842401504517, + "step": 1320, + "valid_targets_mean": 3578.8, + "valid_targets_min": 1475 + }, + { + "epoch": 2.123397435897436, + "grad_norm": 0.680105569391109, + "learning_rate": 3.5181877423838034e-05, + "loss": 0.4767, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2696499228477478, + "step": 1325, + "valid_targets_mean": 3212.0, + "valid_targets_min": 2139 + }, + { + "epoch": 2.1314102564102564, + "grad_norm": 0.6665552362175731, + "learning_rate": 3.512973097504371e-05, + "loss": 0.4903, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2727278769016266, + "step": 1330, + "valid_targets_mean": 3753.0, + "valid_targets_min": 1625 + }, + { + "epoch": 2.139423076923077, + "grad_norm": 0.6804702494796768, + "learning_rate": 3.507734294440403e-05, + "loss": 0.4905, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.205522820353508, + "step": 1335, + "valid_targets_mean": 3127.8, + "valid_targets_min": 2073 + }, + { + "epoch": 2.1474358974358974, + "grad_norm": 0.6375255747120028, + "learning_rate": 3.50247141684175e-05, + "loss": 0.4654, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2905551493167877, + "step": 1340, + "valid_targets_mean": 4370.0, + "valid_targets_min": 1780 + }, + { + "epoch": 2.155448717948718, + "grad_norm": 0.6131745809070184, + "learning_rate": 3.497184548742667e-05, + "loss": 0.4741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2416357845067978, + "step": 1345, + "valid_targets_mean": 3570.8, + "valid_targets_min": 2443 + }, + { + "epoch": 2.1634615384615383, + "grad_norm": 0.5958674671234483, + "learning_rate": 3.491873774560473e-05, + "loss": 0.4819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2461138367652893, + "step": 1350, + "valid_targets_mean": 3949.4, + "valid_targets_min": 2656 + }, + { + "epoch": 2.171474358974359, + "grad_norm": 0.6036915539550539, + "learning_rate": 3.486539179094208e-05, + "loss": 0.4889, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21640899777412415, + "step": 1355, + "valid_targets_mean": 3787.8, + "valid_targets_min": 2028 + }, + { + "epoch": 2.1794871794871793, + "grad_norm": 0.7089712793497042, + "learning_rate": 3.481180847523272e-05, + "loss": 0.496, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23603364825248718, + "step": 1360, + "valid_targets_mean": 2813.6, + "valid_targets_min": 1744 + }, + { + "epoch": 2.1875, + "grad_norm": 0.6238521641092111, + "learning_rate": 3.4757988654060684e-05, + "loss": 0.4668, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2260483205318451, + "step": 1365, + "valid_targets_mean": 3394.8, + "valid_targets_min": 1416 + }, + { + "epoch": 2.1955128205128207, + "grad_norm": 0.7216276761727595, + "learning_rate": 3.470393318678637e-05, + "loss": 0.4848, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21823830902576447, + "step": 1370, + "valid_targets_mean": 2100.1, + "valid_targets_min": 1550 + }, + { + "epoch": 2.203525641025641, + "grad_norm": 0.6377450071965831, + "learning_rate": 3.4649642936532836e-05, + "loss": 0.4529, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23668509721755981, + "step": 1375, + "valid_targets_mean": 3202.6, + "valid_targets_min": 1712 + }, + { + "epoch": 2.2115384615384617, + "grad_norm": 0.6601831606118221, + "learning_rate": 3.4595118770171984e-05, + "loss": 0.4754, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2398017942905426, + "step": 1380, + "valid_targets_mean": 3331.2, + "valid_targets_min": 1343 + }, + { + "epoch": 2.219551282051282, + "grad_norm": 0.644002951558391, + "learning_rate": 3.454036155831077e-05, + "loss": 0.4613, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17118749022483826, + "step": 1385, + "valid_targets_mean": 2805.0, + "valid_targets_min": 1326 + }, + { + "epoch": 2.2275641025641026, + "grad_norm": 0.6343086929278687, + "learning_rate": 3.4485372175277236e-05, + "loss": 0.4866, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2244950830936432, + "step": 1390, + "valid_targets_mean": 3561.5, + "valid_targets_min": 1653 + }, + { + "epoch": 2.235576923076923, + "grad_norm": 0.6315324156998673, + "learning_rate": 3.44301514991066e-05, + "loss": 0.4799, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18784162402153015, + "step": 1395, + "valid_targets_mean": 2765.4, + "valid_targets_min": 664 + }, + { + "epoch": 2.2435897435897436, + "grad_norm": 0.5997921576736266, + "learning_rate": 3.4374700411527225e-05, + "loss": 0.4898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2700258791446686, + "step": 1400, + "valid_targets_mean": 3945.1, + "valid_targets_min": 2218 + }, + { + "epoch": 2.251602564102564, + "grad_norm": 0.6153494528342913, + "learning_rate": 3.431901979794653e-05, + "loss": 0.5002, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2444554567337036, + "step": 1405, + "valid_targets_mean": 3688.9, + "valid_targets_min": 1375 + }, + { + "epoch": 2.2596153846153846, + "grad_norm": 0.6534648396215342, + "learning_rate": 3.426311054743685e-05, + "loss": 0.4801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24272340536117554, + "step": 1410, + "valid_targets_mean": 3171.4, + "valid_targets_min": 1129 + }, + { + "epoch": 2.2676282051282053, + "grad_norm": 0.6583553494828702, + "learning_rate": 3.420697355272127e-05, + "loss": 0.4954, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2562671899795532, + "step": 1415, + "valid_targets_mean": 3280.4, + "valid_targets_min": 1777 + }, + { + "epoch": 2.2756410256410255, + "grad_norm": 0.7513984890287411, + "learning_rate": 3.415060971015933e-05, + "loss": 0.4855, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26656612753868103, + "step": 1420, + "valid_targets_mean": 3355.4, + "valid_targets_min": 1545 + }, + { + "epoch": 2.2836538461538463, + "grad_norm": 0.6471585522918397, + "learning_rate": 3.4094019919732736e-05, + "loss": 0.4701, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22290436923503876, + "step": 1425, + "valid_targets_mean": 3149.8, + "valid_targets_min": 1360 + }, + { + "epoch": 2.2916666666666665, + "grad_norm": 0.6297019444456764, + "learning_rate": 3.403720508503098e-05, + "loss": 0.4814, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16554424166679382, + "step": 1430, + "valid_targets_mean": 2819.4, + "valid_targets_min": 1481 + }, + { + "epoch": 2.2996794871794872, + "grad_norm": 0.6573066417027729, + "learning_rate": 3.398016611323693e-05, + "loss": 0.4613, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22763067483901978, + "step": 1435, + "valid_targets_mean": 3083.5, + "valid_targets_min": 1683 + }, + { + "epoch": 2.3076923076923075, + "grad_norm": 0.6315994762606028, + "learning_rate": 3.392290391511232e-05, + "loss": 0.4826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25256437063217163, + "step": 1440, + "valid_targets_mean": 3372.2, + "valid_targets_min": 1832 + }, + { + "epoch": 2.315705128205128, + "grad_norm": 0.6981833866754609, + "learning_rate": 3.386541940498322e-05, + "loss": 0.4692, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.306129515171051, + "step": 1445, + "valid_targets_mean": 3764.6, + "valid_targets_min": 2139 + }, + { + "epoch": 2.323717948717949, + "grad_norm": 0.6724697636556377, + "learning_rate": 3.380771350072543e-05, + "loss": 0.4883, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17205148935317993, + "step": 1450, + "valid_targets_mean": 2479.1, + "valid_targets_min": 1307 + }, + { + "epoch": 2.331730769230769, + "grad_norm": 0.6431726267233159, + "learning_rate": 3.374978712374986e-05, + "loss": 0.4807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2723473310470581, + "step": 1455, + "valid_targets_mean": 4011.1, + "valid_targets_min": 1358 + }, + { + "epoch": 2.33974358974359, + "grad_norm": 0.7142290531752585, + "learning_rate": 3.369164119898774e-05, + "loss": 0.4669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28070539236068726, + "step": 1460, + "valid_targets_mean": 3313.4, + "valid_targets_min": 1580 + }, + { + "epoch": 2.34775641025641, + "grad_norm": 0.5994021259339507, + "learning_rate": 3.363327665487593e-05, + "loss": 0.5001, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25300800800323486, + "step": 1465, + "valid_targets_mean": 3993.4, + "valid_targets_min": 2477 + }, + { + "epoch": 2.355769230769231, + "grad_norm": 0.6950110115699835, + "learning_rate": 3.357469442334206e-05, + "loss": 0.5013, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2721792757511139, + "step": 1470, + "valid_targets_mean": 3043.1, + "valid_targets_min": 910 + }, + { + "epoch": 2.363782051282051, + "grad_norm": 0.6097608478834989, + "learning_rate": 3.351589543978965e-05, + "loss": 0.4724, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21048158407211304, + "step": 1475, + "valid_targets_mean": 3166.6, + "valid_targets_min": 1816 + }, + { + "epoch": 2.371794871794872, + "grad_norm": 0.6436941603055468, + "learning_rate": 3.345688064308317e-05, + "loss": 0.4442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23861803114414215, + "step": 1480, + "valid_targets_mean": 3306.6, + "valid_targets_min": 1261 + }, + { + "epoch": 2.3798076923076925, + "grad_norm": 0.6260306873785256, + "learning_rate": 3.339765097553307e-05, + "loss": 0.4715, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24460147321224213, + "step": 1485, + "valid_targets_mean": 3560.1, + "valid_targets_min": 1606 + }, + { + "epoch": 2.3878205128205128, + "grad_norm": 0.6589590154735876, + "learning_rate": 3.33382073828807e-05, + "loss": 0.4965, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25442594289779663, + "step": 1490, + "valid_targets_mean": 3744.2, + "valid_targets_min": 1543 + }, + { + "epoch": 2.3958333333333335, + "grad_norm": 0.5988572352920845, + "learning_rate": 3.327855081428326e-05, + "loss": 0.4658, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16690203547477722, + "step": 1495, + "valid_targets_mean": 3643.4, + "valid_targets_min": 906 + }, + { + "epoch": 2.4038461538461537, + "grad_norm": 0.6291529080047887, + "learning_rate": 3.3218682222298584e-05, + "loss": 0.4668, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27912992238998413, + "step": 1500, + "valid_targets_mean": 3125.2, + "valid_targets_min": 1181 + }, + { + "epoch": 2.4118589743589745, + "grad_norm": 0.5515056254042542, + "learning_rate": 3.315860256286996e-05, + "loss": 0.4531, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21268051862716675, + "step": 1505, + "valid_targets_mean": 4042.5, + "valid_targets_min": 2131 + }, + { + "epoch": 2.4198717948717947, + "grad_norm": 0.8394975915539936, + "learning_rate": 3.3098312795310894e-05, + "loss": 0.4799, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22699277102947235, + "step": 1510, + "valid_targets_mean": 2835.9, + "valid_targets_min": 1528 + }, + { + "epoch": 2.4278846153846154, + "grad_norm": 0.5886952468262564, + "learning_rate": 3.303781388228974e-05, + "loss": 0.4576, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21781568229198456, + "step": 1515, + "valid_targets_mean": 3517.4, + "valid_targets_min": 1799 + }, + { + "epoch": 2.435897435897436, + "grad_norm": 0.6712039996381571, + "learning_rate": 3.297710678981435e-05, + "loss": 0.4892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2826261520385742, + "step": 1520, + "valid_targets_mean": 3315.8, + "valid_targets_min": 2004 + }, + { + "epoch": 2.4439102564102564, + "grad_norm": 0.6082197125869999, + "learning_rate": 3.291619248721667e-05, + "loss": 0.503, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.251178503036499, + "step": 1525, + "valid_targets_mean": 3918.4, + "valid_targets_min": 1865 + }, + { + "epoch": 2.451923076923077, + "grad_norm": 0.6826070301946944, + "learning_rate": 3.285507194713724e-05, + "loss": 0.4559, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20794638991355896, + "step": 1530, + "valid_targets_mean": 2633.5, + "valid_targets_min": 672 + }, + { + "epoch": 2.4599358974358974, + "grad_norm": 0.6314752384427056, + "learning_rate": 3.279374614550966e-05, + "loss": 0.4832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1810711771249771, + "step": 1535, + "valid_targets_mean": 2679.8, + "valid_targets_min": 1530 + }, + { + "epoch": 2.467948717948718, + "grad_norm": 0.7875887471981011, + "learning_rate": 3.2732216061545e-05, + "loss": 0.4852, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22018864750862122, + "step": 1540, + "valid_targets_mean": 2626.0, + "valid_targets_min": 1068 + }, + { + "epoch": 2.4759615384615383, + "grad_norm": 0.6283699206065749, + "learning_rate": 3.2670482677716214e-05, + "loss": 0.5175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2622578740119934, + "step": 1545, + "valid_targets_mean": 3574.9, + "valid_targets_min": 1708 + }, + { + "epoch": 2.483974358974359, + "grad_norm": 0.6177707766180375, + "learning_rate": 3.2608546979742394e-05, + "loss": 0.4546, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18695585429668427, + "step": 1550, + "valid_targets_mean": 3022.8, + "valid_targets_min": 1620 + }, + { + "epoch": 2.4919871794871793, + "grad_norm": 0.6394582243550946, + "learning_rate": 3.254640995657307e-05, + "loss": 0.498, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2070213258266449, + "step": 1555, + "valid_targets_mean": 2989.0, + "valid_targets_min": 1219 + }, + { + "epoch": 2.5, + "grad_norm": 0.6839525276461119, + "learning_rate": 3.248407260037239e-05, + "loss": 0.4668, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24948441982269287, + "step": 1560, + "valid_targets_mean": 3143.9, + "valid_targets_min": 1352 + }, + { + "epoch": 2.5080128205128203, + "grad_norm": 0.6279579852506579, + "learning_rate": 3.24215359065033e-05, + "loss": 0.4543, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18786096572875977, + "step": 1565, + "valid_targets_mean": 3372.6, + "valid_targets_min": 373 + }, + { + "epoch": 2.516025641025641, + "grad_norm": 0.625428034900398, + "learning_rate": 3.235880087351164e-05, + "loss": 0.4411, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2752924859523773, + "step": 1570, + "valid_targets_mean": 3777.4, + "valid_targets_min": 2610 + }, + { + "epoch": 2.5240384615384617, + "grad_norm": 0.6047588892174584, + "learning_rate": 3.2295868503110184e-05, + "loss": 0.5194, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1983676701784134, + "step": 1575, + "valid_targets_mean": 3342.6, + "valid_targets_min": 1310 + }, + { + "epoch": 2.532051282051282, + "grad_norm": 0.7001008115821391, + "learning_rate": 3.22327398001627e-05, + "loss": 0.4734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20834968984127045, + "step": 1580, + "valid_targets_mean": 2329.0, + "valid_targets_min": 300 + }, + { + "epoch": 2.5400641025641026, + "grad_norm": 0.6890090652193042, + "learning_rate": 3.216941577266783e-05, + "loss": 0.4595, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27099794149398804, + "step": 1585, + "valid_targets_mean": 3480.2, + "valid_targets_min": 2370 + }, + { + "epoch": 2.5480769230769234, + "grad_norm": 0.6271601404225648, + "learning_rate": 3.210589743174308e-05, + "loss": 0.4874, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24052384495735168, + "step": 1590, + "valid_targets_mean": 3597.1, + "valid_targets_min": 1556 + }, + { + "epoch": 2.5560897435897436, + "grad_norm": 0.692554209321863, + "learning_rate": 3.204218579160857e-05, + "loss": 0.4805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23645177483558655, + "step": 1595, + "valid_targets_mean": 3246.9, + "valid_targets_min": 1711 + }, + { + "epoch": 2.564102564102564, + "grad_norm": 0.7153518914954438, + "learning_rate": 3.197828186957094e-05, + "loss": 0.4735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2892233729362488, + "step": 1600, + "valid_targets_mean": 2725.1, + "valid_targets_min": 1333 + }, + { + "epoch": 2.5721153846153846, + "grad_norm": 0.6272664436981538, + "learning_rate": 3.191418668600705e-05, + "loss": 0.4776, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2429126650094986, + "step": 1605, + "valid_targets_mean": 2919.0, + "valid_targets_min": 1512 + }, + { + "epoch": 2.5801282051282053, + "grad_norm": 0.6843942254859565, + "learning_rate": 3.184990126434771e-05, + "loss": 0.523, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2766910195350647, + "step": 1610, + "valid_targets_mean": 3214.1, + "valid_targets_min": 1525 + }, + { + "epoch": 2.5881410256410255, + "grad_norm": 0.6438227602163282, + "learning_rate": 3.178542663106131e-05, + "loss": 0.4761, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17935946583747864, + "step": 1615, + "valid_targets_mean": 2904.4, + "valid_targets_min": 1028 + }, + { + "epoch": 2.5961538461538463, + "grad_norm": 0.5846835259172851, + "learning_rate": 3.172076381563748e-05, + "loss": 0.521, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2687412202358246, + "step": 1620, + "valid_targets_mean": 4076.4, + "valid_targets_min": 1246 + }, + { + "epoch": 2.6041666666666665, + "grad_norm": 0.6259210229847506, + "learning_rate": 3.165591385057058e-05, + "loss": 0.4565, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2709156274795532, + "step": 1625, + "valid_targets_mean": 3459.8, + "valid_targets_min": 1586 + }, + { + "epoch": 2.6121794871794872, + "grad_norm": 0.6899990112048433, + "learning_rate": 3.1590877771343316e-05, + "loss": 0.4784, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22439658641815186, + "step": 1630, + "valid_targets_mean": 2552.8, + "valid_targets_min": 1451 + }, + { + "epoch": 2.6201923076923075, + "grad_norm": 0.7352353373228997, + "learning_rate": 3.152565661641008e-05, + "loss": 0.4735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1965309977531433, + "step": 1635, + "valid_targets_mean": 2373.1, + "valid_targets_min": 1311 + }, + { + "epoch": 2.628205128205128, + "grad_norm": 0.6899705443819116, + "learning_rate": 3.1460251427180474e-05, + "loss": 0.4606, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22880998253822327, + "step": 1640, + "valid_targets_mean": 2794.4, + "valid_targets_min": 1001 + }, + { + "epoch": 2.636217948717949, + "grad_norm": 0.5726113495324329, + "learning_rate": 3.139466324800263e-05, + "loss": 0.4749, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19087274372577667, + "step": 1645, + "valid_targets_mean": 3492.6, + "valid_targets_min": 1692 + }, + { + "epoch": 2.644230769230769, + "grad_norm": 0.6194098681429608, + "learning_rate": 3.132889312614655e-05, + "loss": 0.4964, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20216143131256104, + "step": 1650, + "valid_targets_mean": 2909.5, + "valid_targets_min": 1328 + }, + { + "epoch": 2.65224358974359, + "grad_norm": 0.6423755681040674, + "learning_rate": 3.126294211178737e-05, + "loss": 0.4728, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2490118443965912, + "step": 1655, + "valid_targets_mean": 3506.1, + "valid_targets_min": 1424 + }, + { + "epoch": 2.66025641025641, + "grad_norm": 0.7298413615296814, + "learning_rate": 3.1196811257988634e-05, + "loss": 0.4895, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2908113896846771, + "step": 1660, + "valid_targets_mean": 2868.0, + "valid_targets_min": 1262 + }, + { + "epoch": 2.668269230769231, + "grad_norm": 0.6384350596688362, + "learning_rate": 3.1130501620685394e-05, + "loss": 0.4868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26503297686576843, + "step": 1665, + "valid_targets_mean": 3551.0, + "valid_targets_min": 874 + }, + { + "epoch": 2.676282051282051, + "grad_norm": 0.6440922048029468, + "learning_rate": 3.106401425866745e-05, + "loss": 0.4694, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18802538514137268, + "step": 1670, + "valid_targets_mean": 3070.6, + "valid_targets_min": 1567 + }, + { + "epoch": 2.684294871794872, + "grad_norm": 0.5941320328012714, + "learning_rate": 3.099735023356236e-05, + "loss": 0.4814, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2637801170349121, + "step": 1675, + "valid_targets_mean": 3670.4, + "valid_targets_min": 1615 + }, + { + "epoch": 2.6923076923076925, + "grad_norm": 0.6079591201458582, + "learning_rate": 3.0930510609818564e-05, + "loss": 0.4999, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21088537573814392, + "step": 1680, + "valid_targets_mean": 3003.0, + "valid_targets_min": 1147 + }, + { + "epoch": 2.7003205128205128, + "grad_norm": 0.645154817306774, + "learning_rate": 3.086349645468831e-05, + "loss": 0.48, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24316158890724182, + "step": 1685, + "valid_targets_mean": 3658.9, + "valid_targets_min": 1718 + }, + { + "epoch": 2.7083333333333335, + "grad_norm": 0.776856593857071, + "learning_rate": 3.079630883821067e-05, + "loss": 0.4785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28521084785461426, + "step": 1690, + "valid_targets_mean": 3276.6, + "valid_targets_min": 1656 + }, + { + "epoch": 2.7163461538461537, + "grad_norm": 0.6626403496222938, + "learning_rate": 3.0728948833194436e-05, + "loss": 0.4662, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2650311291217804, + "step": 1695, + "valid_targets_mean": 3819.5, + "valid_targets_min": 1393 + }, + { + "epoch": 2.7243589743589745, + "grad_norm": 0.7755751457991342, + "learning_rate": 3.066141751520099e-05, + "loss": 0.4703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19013823568820953, + "step": 1700, + "valid_targets_mean": 3997.9, + "valid_targets_min": 2073 + }, + { + "epoch": 2.7323717948717947, + "grad_norm": 0.7056418504813926, + "learning_rate": 3.059371596252712e-05, + "loss": 0.4707, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22634246945381165, + "step": 1705, + "valid_targets_mean": 3447.9, + "valid_targets_min": 2449 + }, + { + "epoch": 2.7403846153846154, + "grad_norm": 0.5841570980318216, + "learning_rate": 3.0525845256187834e-05, + "loss": 0.4602, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2625887989997864, + "step": 1710, + "valid_targets_mean": 4130.5, + "valid_targets_min": 2982 + }, + { + "epoch": 2.748397435897436, + "grad_norm": 0.6917063870213431, + "learning_rate": 3.0457806479899044e-05, + "loss": 0.4771, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2650618851184845, + "step": 1715, + "valid_targets_mean": 3134.5, + "valid_targets_min": 1396 + }, + { + "epoch": 2.7564102564102564, + "grad_norm": 0.5869876548634071, + "learning_rate": 3.0389600720060318e-05, + "loss": 0.4698, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2087412178516388, + "step": 1720, + "valid_targets_mean": 3215.0, + "valid_targets_min": 1330 + }, + { + "epoch": 2.7644230769230766, + "grad_norm": 0.6084361850045575, + "learning_rate": 3.0321229065737522e-05, + "loss": 0.4479, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18068155646324158, + "step": 1725, + "valid_targets_mean": 3030.2, + "valid_targets_min": 1458 + }, + { + "epoch": 2.7724358974358974, + "grad_norm": 0.5796728448062316, + "learning_rate": 3.0252692608645384e-05, + "loss": 0.4466, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21863749623298645, + "step": 1730, + "valid_targets_mean": 3512.4, + "valid_targets_min": 1055 + }, + { + "epoch": 2.780448717948718, + "grad_norm": 0.6714841093317424, + "learning_rate": 3.0183992443130127e-05, + "loss": 0.4691, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25895217061042786, + "step": 1735, + "valid_targets_mean": 2736.4, + "valid_targets_min": 1289 + }, + { + "epoch": 2.7884615384615383, + "grad_norm": 0.6972007685774616, + "learning_rate": 3.011512966615195e-05, + "loss": 0.4869, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3409285545349121, + "step": 1740, + "valid_targets_mean": 3147.4, + "valid_targets_min": 1481 + }, + { + "epoch": 2.796474358974359, + "grad_norm": 0.6277081607857874, + "learning_rate": 3.0046105377267523e-05, + "loss": 0.4615, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24109208583831787, + "step": 1745, + "valid_targets_mean": 3923.9, + "valid_targets_min": 2040 + }, + { + "epoch": 2.8044871794871797, + "grad_norm": 0.6223055407586584, + "learning_rate": 2.9976920678612456e-05, + "loss": 0.4963, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27366888523101807, + "step": 1750, + "valid_targets_mean": 3521.5, + "valid_targets_min": 1425 + }, + { + "epoch": 2.8125, + "grad_norm": 0.6244554364099718, + "learning_rate": 2.9907576674883664e-05, + "loss": 0.4788, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28175079822540283, + "step": 1755, + "valid_targets_mean": 3831.1, + "valid_targets_min": 1897 + }, + { + "epoch": 2.8205128205128203, + "grad_norm": 0.682083588647277, + "learning_rate": 2.983807447332174e-05, + "loss": 0.4996, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1731472909450531, + "step": 1760, + "valid_targets_mean": 2523.9, + "valid_targets_min": 1631 + }, + { + "epoch": 2.828525641025641, + "grad_norm": 0.5639258121786636, + "learning_rate": 2.9768415183693293e-05, + "loss": 0.509, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2450920045375824, + "step": 1765, + "valid_targets_mean": 4051.0, + "valid_targets_min": 2041 + }, + { + "epoch": 2.8365384615384617, + "grad_norm": 0.6013335047341235, + "learning_rate": 2.9698599918273197e-05, + "loss": 0.4845, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2414018213748932, + "step": 1770, + "valid_targets_mean": 3637.2, + "valid_targets_min": 870 + }, + { + "epoch": 2.844551282051282, + "grad_norm": 0.6738346231611319, + "learning_rate": 2.962862979182686e-05, + "loss": 0.4682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23030608892440796, + "step": 1775, + "valid_targets_mean": 3231.8, + "valid_targets_min": 1300 + }, + { + "epoch": 2.8525641025641026, + "grad_norm": 0.5889962877403477, + "learning_rate": 2.95585059215924e-05, + "loss": 0.4888, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23207280039787292, + "step": 1780, + "valid_targets_mean": 3977.5, + "valid_targets_min": 1866 + }, + { + "epoch": 2.8605769230769234, + "grad_norm": 0.6580078057152334, + "learning_rate": 2.948822942726284e-05, + "loss": 0.4737, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21945837140083313, + "step": 1785, + "valid_targets_mean": 3338.9, + "valid_targets_min": 1346 + }, + { + "epoch": 2.8685897435897436, + "grad_norm": 0.6133284564858451, + "learning_rate": 2.941780143096817e-05, + "loss": 0.4671, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2510795295238495, + "step": 1790, + "valid_targets_mean": 3602.9, + "valid_targets_min": 1430 + }, + { + "epoch": 2.876602564102564, + "grad_norm": 0.6525261456377198, + "learning_rate": 2.9347223057257505e-05, + "loss": 0.4892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22132179141044617, + "step": 1795, + "valid_targets_mean": 2901.2, + "valid_targets_min": 1862 + }, + { + "epoch": 2.8846153846153846, + "grad_norm": 0.5802703075236487, + "learning_rate": 2.927649543308106e-05, + "loss": 0.4479, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21268293261528015, + "step": 1800, + "valid_targets_mean": 3473.4, + "valid_targets_min": 1830 + }, + { + "epoch": 2.8926282051282053, + "grad_norm": 0.603509875782638, + "learning_rate": 2.9205619687772212e-05, + "loss": 0.5043, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1973789632320404, + "step": 1805, + "valid_targets_mean": 2862.9, + "valid_targets_min": 1109 + }, + { + "epoch": 2.9006410256410255, + "grad_norm": 0.6447137106666401, + "learning_rate": 2.9134596953029413e-05, + "loss": 0.4703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27511319518089294, + "step": 1810, + "valid_targets_mean": 3472.2, + "valid_targets_min": 1400 + }, + { + "epoch": 2.9086538461538463, + "grad_norm": 0.6232929329973954, + "learning_rate": 2.9063428362898168e-05, + "loss": 0.4819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.251217782497406, + "step": 1815, + "valid_targets_mean": 3711.0, + "valid_targets_min": 2019 + }, + { + "epoch": 2.9166666666666665, + "grad_norm": 0.6099864342632423, + "learning_rate": 2.8992115053752905e-05, + "loss": 0.4456, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1781488060951233, + "step": 1820, + "valid_targets_mean": 2532.1, + "valid_targets_min": 1393 + }, + { + "epoch": 2.9246794871794872, + "grad_norm": 0.6049656079029271, + "learning_rate": 2.8920658164278816e-05, + "loss": 0.4694, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20879614353179932, + "step": 1825, + "valid_targets_mean": 3086.2, + "valid_targets_min": 1461 + }, + { + "epoch": 2.9326923076923075, + "grad_norm": 0.6133689582188944, + "learning_rate": 2.884905883545373e-05, + "loss": 0.4882, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21167677640914917, + "step": 1830, + "valid_targets_mean": 3229.1, + "valid_targets_min": 1506 + }, + { + "epoch": 2.940705128205128, + "grad_norm": 0.5559493895214943, + "learning_rate": 2.877731821052981e-05, + "loss": 0.4582, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22521714866161346, + "step": 1835, + "valid_targets_mean": 4057.2, + "valid_targets_min": 1994 + }, + { + "epoch": 2.948717948717949, + "grad_norm": 0.5761005092354734, + "learning_rate": 2.8705437435015375e-05, + "loss": 0.4697, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24761134386062622, + "step": 1840, + "valid_targets_mean": 4159.9, + "valid_targets_min": 1836 + }, + { + "epoch": 2.956730769230769, + "grad_norm": 0.6308088784307772, + "learning_rate": 2.8633417656656566e-05, + "loss": 0.4801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23453302681446075, + "step": 1845, + "valid_targets_mean": 3038.5, + "valid_targets_min": 965 + }, + { + "epoch": 2.96474358974359, + "grad_norm": 0.6641627856292389, + "learning_rate": 2.8561260025419036e-05, + "loss": 0.5233, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28335341811180115, + "step": 1850, + "valid_targets_mean": 3738.2, + "valid_targets_min": 1768 + }, + { + "epoch": 2.97275641025641, + "grad_norm": 0.6448886806409363, + "learning_rate": 2.8488965693469583e-05, + "loss": 0.49, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2826387882232666, + "step": 1855, + "valid_targets_mean": 3662.5, + "valid_targets_min": 1283 + }, + { + "epoch": 2.980769230769231, + "grad_norm": 0.5795804288022112, + "learning_rate": 2.8416535815157763e-05, + "loss": 0.4719, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22267763316631317, + "step": 1860, + "valid_targets_mean": 3452.4, + "valid_targets_min": 1734 + }, + { + "epoch": 2.988782051282051, + "grad_norm": 0.6166006251109611, + "learning_rate": 2.8343971546997434e-05, + "loss": 0.4551, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22959822416305542, + "step": 1865, + "valid_targets_mean": 3563.2, + "valid_targets_min": 1801 + }, + { + "epoch": 2.996794871794872, + "grad_norm": 0.6002863297041605, + "learning_rate": 2.827127404764831e-05, + "loss": 0.473, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24405668675899506, + "step": 1870, + "valid_targets_mean": 3947.8, + "valid_targets_min": 1462 + }, + { + "epoch": 3.0048076923076925, + "grad_norm": 0.6263396810031966, + "learning_rate": 2.8198444477897467e-05, + "loss": 0.4595, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2550222873687744, + "step": 1875, + "valid_targets_mean": 2879.2, + "valid_targets_min": 808 + }, + { + "epoch": 3.0128205128205128, + "grad_norm": 0.6727617745783966, + "learning_rate": 2.8125484000640787e-05, + "loss": 0.4477, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2327970564365387, + "step": 1880, + "valid_targets_mean": 3547.8, + "valid_targets_min": 2337 + }, + { + "epoch": 3.0208333333333335, + "grad_norm": 0.6318288775500677, + "learning_rate": 2.8052393780864394e-05, + "loss": 0.4509, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2146126627922058, + "step": 1885, + "valid_targets_mean": 3456.6, + "valid_targets_min": 1552 + }, + { + "epoch": 3.0288461538461537, + "grad_norm": 0.6377400173545166, + "learning_rate": 2.797917498562607e-05, + "loss": 0.4363, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2266453355550766, + "step": 1890, + "valid_targets_mean": 3601.6, + "valid_targets_min": 1265 + }, + { + "epoch": 3.0368589743589745, + "grad_norm": 0.7000619286919751, + "learning_rate": 2.7905828784036596e-05, + "loss": 0.4465, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2918475568294525, + "step": 1895, + "valid_targets_mean": 3394.5, + "valid_targets_min": 1451 + }, + { + "epoch": 3.0448717948717947, + "grad_norm": 0.6452704093297165, + "learning_rate": 2.78323563472411e-05, + "loss": 0.4506, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22940441966056824, + "step": 1900, + "valid_targets_mean": 3473.6, + "valid_targets_min": 2195 + }, + { + "epoch": 3.0528846153846154, + "grad_norm": 0.6454782257184184, + "learning_rate": 2.7758758848400354e-05, + "loss": 0.4415, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18762445449829102, + "step": 1905, + "valid_targets_mean": 3490.9, + "valid_targets_min": 2252 + }, + { + "epoch": 3.0608974358974357, + "grad_norm": 0.6849970569805061, + "learning_rate": 2.7685037462672043e-05, + "loss": 0.45, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26085925102233887, + "step": 1910, + "valid_targets_mean": 3437.1, + "valid_targets_min": 1613 + }, + { + "epoch": 3.0689102564102564, + "grad_norm": 0.6186227843082133, + "learning_rate": 2.7611193367191993e-05, + "loss": 0.4363, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2334415465593338, + "step": 1915, + "valid_targets_mean": 3855.5, + "valid_targets_min": 594 + }, + { + "epoch": 3.076923076923077, + "grad_norm": 0.7765991054039258, + "learning_rate": 2.7537227741055378e-05, + "loss": 0.4282, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2089974582195282, + "step": 1920, + "valid_targets_mean": 2916.4, + "valid_targets_min": 1222 + }, + { + "epoch": 3.0849358974358974, + "grad_norm": 0.6703851068982292, + "learning_rate": 2.746314176529791e-05, + "loss": 0.4557, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2197643220424652, + "step": 1925, + "valid_targets_mean": 3478.4, + "valid_targets_min": 1424 + }, + { + "epoch": 3.092948717948718, + "grad_norm": 0.637206353524805, + "learning_rate": 2.7388936622876957e-05, + "loss": 0.4514, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21921581029891968, + "step": 1930, + "valid_targets_mean": 3605.5, + "valid_targets_min": 1709 + }, + { + "epoch": 3.1009615384615383, + "grad_norm": 0.6790802911388166, + "learning_rate": 2.7314613498652663e-05, + "loss": 0.4517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2148890495300293, + "step": 1935, + "valid_targets_mean": 2826.6, + "valid_targets_min": 1646 + }, + { + "epoch": 3.108974358974359, + "grad_norm": 0.6782524691145373, + "learning_rate": 2.7240173579369025e-05, + "loss": 0.4373, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21406295895576477, + "step": 1940, + "valid_targets_mean": 3344.8, + "valid_targets_min": 1953 + }, + { + "epoch": 3.1169871794871793, + "grad_norm": 0.6566919536684225, + "learning_rate": 2.7165618053634962e-05, + "loss": 0.4448, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20008143782615662, + "step": 1945, + "valid_targets_mean": 3073.1, + "valid_targets_min": 1547 + }, + { + "epoch": 3.125, + "grad_norm": 0.6432468480015572, + "learning_rate": 2.7090948111905304e-05, + "loss": 0.4531, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21931028366088867, + "step": 1950, + "valid_targets_mean": 3487.8, + "valid_targets_min": 1742 + }, + { + "epoch": 3.1330128205128207, + "grad_norm": 0.6584676772921859, + "learning_rate": 2.701616494646183e-05, + "loss": 0.4368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20845025777816772, + "step": 1955, + "valid_targets_mean": 3234.8, + "valid_targets_min": 1849 + }, + { + "epoch": 3.141025641025641, + "grad_norm": 0.6407395025792776, + "learning_rate": 2.6941269751394174e-05, + "loss": 0.4155, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19711872935295105, + "step": 1960, + "valid_targets_mean": 3211.1, + "valid_targets_min": 1634 + }, + { + "epoch": 3.1490384615384617, + "grad_norm": 0.6179534449055757, + "learning_rate": 2.686626372258081e-05, + "loss": 0.429, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20131778717041016, + "step": 1965, + "valid_targets_mean": 3845.0, + "valid_targets_min": 1612 + }, + { + "epoch": 3.157051282051282, + "grad_norm": 0.7235826105010092, + "learning_rate": 2.6791148057669913e-05, + "loss": 0.4498, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3174961805343628, + "step": 1970, + "valid_targets_mean": 3182.8, + "valid_targets_min": 1111 + }, + { + "epoch": 3.1650641025641026, + "grad_norm": 0.6367357387639193, + "learning_rate": 2.671592395606027e-05, + "loss": 0.4441, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1891191005706787, + "step": 1975, + "valid_targets_mean": 3340.5, + "valid_targets_min": 1742 + }, + { + "epoch": 3.173076923076923, + "grad_norm": 0.681000510394323, + "learning_rate": 2.6640592618882114e-05, + "loss": 0.4213, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21171995997428894, + "step": 1980, + "valid_targets_mean": 3007.5, + "valid_targets_min": 1666 + }, + { + "epoch": 3.1810897435897436, + "grad_norm": 0.6404689417206875, + "learning_rate": 2.656515524897795e-05, + "loss": 0.4915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3128596246242523, + "step": 1985, + "valid_targets_mean": 4163.2, + "valid_targets_min": 1610 + }, + { + "epoch": 3.189102564102564, + "grad_norm": 0.660881627713433, + "learning_rate": 2.6489613050883343e-05, + "loss": 0.4379, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17708660662174225, + "step": 1990, + "valid_targets_mean": 2993.0, + "valid_targets_min": 1700 + }, + { + "epoch": 3.1971153846153846, + "grad_norm": 0.6210413896130281, + "learning_rate": 2.6413967230807677e-05, + "loss": 0.4356, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22479748725891113, + "step": 1995, + "valid_targets_mean": 3692.4, + "valid_targets_min": 1251 + }, + { + "epoch": 3.2051282051282053, + "grad_norm": 0.6605070784535979, + "learning_rate": 2.6338218996614924e-05, + "loss": 0.4592, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18355093896389008, + "step": 2000, + "valid_targets_mean": 3071.5, + "valid_targets_min": 1773 + }, + { + "epoch": 3.2131410256410255, + "grad_norm": 0.6525509389453794, + "learning_rate": 2.6262369557804325e-05, + "loss": 0.4566, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2559788227081299, + "step": 2005, + "valid_targets_mean": 4331.2, + "valid_targets_min": 1559 + }, + { + "epoch": 3.2211538461538463, + "grad_norm": 0.6897579390909623, + "learning_rate": 2.6186420125491094e-05, + "loss": 0.4709, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25256770849227905, + "step": 2010, + "valid_targets_mean": 3229.6, + "valid_targets_min": 1343 + }, + { + "epoch": 3.2291666666666665, + "grad_norm": 0.6821282947800485, + "learning_rate": 2.6110371912387083e-05, + "loss": 0.4659, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24340540170669556, + "step": 2015, + "valid_targets_mean": 3707.6, + "valid_targets_min": 1459 + }, + { + "epoch": 3.2371794871794872, + "grad_norm": 0.6421380884436009, + "learning_rate": 2.6034226132781407e-05, + "loss": 0.4449, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2082211673259735, + "step": 2020, + "valid_targets_mean": 3628.4, + "valid_targets_min": 1899 + }, + { + "epoch": 3.2451923076923075, + "grad_norm": 0.6364701124310044, + "learning_rate": 2.5957984002521066e-05, + "loss": 0.4912, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19811758399009705, + "step": 2025, + "valid_targets_mean": 3573.5, + "valid_targets_min": 1665 + }, + { + "epoch": 3.253205128205128, + "grad_norm": 0.6624229005543212, + "learning_rate": 2.588164673899151e-05, + "loss": 0.4263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2016112506389618, + "step": 2030, + "valid_targets_mean": 3093.8, + "valid_targets_min": 929 + }, + { + "epoch": 3.261217948717949, + "grad_norm": 0.6365823083571853, + "learning_rate": 2.580521556109724e-05, + "loss": 0.4229, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16492486000061035, + "step": 2035, + "valid_targets_mean": 3244.5, + "valid_targets_min": 1489 + }, + { + "epoch": 3.269230769230769, + "grad_norm": 0.6131047892600353, + "learning_rate": 2.57286916892423e-05, + "loss": 0.4726, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19335848093032837, + "step": 2040, + "valid_targets_mean": 2849.9, + "valid_targets_min": 1437 + }, + { + "epoch": 3.27724358974359, + "grad_norm": 0.7166514707307369, + "learning_rate": 2.5652076345310822e-05, + "loss": 0.425, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21478678286075592, + "step": 2045, + "valid_targets_mean": 2827.6, + "valid_targets_min": 1484 + }, + { + "epoch": 3.28525641025641, + "grad_norm": 0.6974166372198569, + "learning_rate": 2.5575370752647507e-05, + "loss": 0.4533, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2548741400241852, + "step": 2050, + "valid_targets_mean": 3825.1, + "valid_targets_min": 1537 + }, + { + "epoch": 3.293269230769231, + "grad_norm": 0.6492909611498847, + "learning_rate": 2.5498576136038077e-05, + "loss": 0.4375, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2354935258626938, + "step": 2055, + "valid_targets_mean": 3419.4, + "valid_targets_min": 1112 + }, + { + "epoch": 3.301282051282051, + "grad_norm": 0.5977545216727056, + "learning_rate": 2.542169372168976e-05, + "loss": 0.4532, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25061458349227905, + "step": 2060, + "valid_targets_mean": 3754.5, + "valid_targets_min": 459 + }, + { + "epoch": 3.309294871794872, + "grad_norm": 0.7405567425568698, + "learning_rate": 2.5344724737211646e-05, + "loss": 0.4632, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24346590042114258, + "step": 2065, + "valid_targets_mean": 3129.0, + "valid_targets_min": 1399 + }, + { + "epoch": 3.3173076923076925, + "grad_norm": 0.6547669107740941, + "learning_rate": 2.5267670411595152e-05, + "loss": 0.4304, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29133039712905884, + "step": 2070, + "valid_targets_mean": 4111.4, + "valid_targets_min": 1545 + }, + { + "epoch": 3.3253205128205128, + "grad_norm": 0.6739007469231487, + "learning_rate": 2.5190531975194345e-05, + "loss": 0.4502, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24122922122478485, + "step": 2075, + "valid_targets_mean": 3440.5, + "valid_targets_min": 1429 + }, + { + "epoch": 3.3333333333333335, + "grad_norm": 0.7533581022167966, + "learning_rate": 2.5113310659706322e-05, + "loss": 0.4721, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20918479561805725, + "step": 2080, + "valid_targets_mean": 1995.5, + "valid_targets_min": 1007 + }, + { + "epoch": 3.3413461538461537, + "grad_norm": 0.5978785503552605, + "learning_rate": 2.5036007698151553e-05, + "loss": 0.4301, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19538165628910065, + "step": 2085, + "valid_targets_mean": 3868.0, + "valid_targets_min": 1863 + }, + { + "epoch": 3.3493589743589745, + "grad_norm": 0.5706930639868784, + "learning_rate": 2.4958624324854185e-05, + "loss": 0.4438, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23559853434562683, + "step": 2090, + "valid_targets_mean": 4552.6, + "valid_targets_min": 1885 + }, + { + "epoch": 3.3573717948717947, + "grad_norm": 0.6486239927449475, + "learning_rate": 2.4881161775422303e-05, + "loss": 0.4625, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20744717121124268, + "step": 2095, + "valid_targets_mean": 2924.4, + "valid_targets_min": 1469 + }, + { + "epoch": 3.3653846153846154, + "grad_norm": 0.6758747789571541, + "learning_rate": 2.480362128672824e-05, + "loss": 0.5032, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2253543883562088, + "step": 2100, + "valid_targets_mean": 2974.6, + "valid_targets_min": 1198 + }, + { + "epoch": 3.373397435897436, + "grad_norm": 0.6539840394430934, + "learning_rate": 2.4726004096888817e-05, + "loss": 0.4448, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2474648356437683, + "step": 2105, + "valid_targets_mean": 3957.6, + "valid_targets_min": 1138 + }, + { + "epoch": 3.3814102564102564, + "grad_norm": 0.678786426827947, + "learning_rate": 2.4648311445245558e-05, + "loss": 0.4615, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2906879782676697, + "step": 2110, + "valid_targets_mean": 3849.1, + "valid_targets_min": 1883 + }, + { + "epoch": 3.389423076923077, + "grad_norm": 0.6971307938338155, + "learning_rate": 2.457054457234493e-05, + "loss": 0.4455, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18562695384025574, + "step": 2115, + "valid_targets_mean": 2733.9, + "valid_targets_min": 1348 + }, + { + "epoch": 3.3974358974358974, + "grad_norm": 0.5917409575549772, + "learning_rate": 2.4492704719918497e-05, + "loss": 0.4104, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20136576890945435, + "step": 2120, + "valid_targets_mean": 3688.6, + "valid_targets_min": 2152 + }, + { + "epoch": 3.405448717948718, + "grad_norm": 0.664348479777749, + "learning_rate": 2.4414793130863134e-05, + "loss": 0.4627, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17913828790187836, + "step": 2125, + "valid_targets_mean": 2386.6, + "valid_targets_min": 1295 + }, + { + "epoch": 3.4134615384615383, + "grad_norm": 0.5397233425981973, + "learning_rate": 2.433681104922114e-05, + "loss": 0.4218, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22439253330230713, + "step": 2130, + "valid_targets_mean": 4057.6, + "valid_targets_min": 2313 + }, + { + "epoch": 3.421474358974359, + "grad_norm": 0.6816089789857905, + "learning_rate": 2.4258759720160412e-05, + "loss": 0.4261, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2554217278957367, + "step": 2135, + "valid_targets_mean": 3368.1, + "valid_targets_min": 2143 + }, + { + "epoch": 3.4294871794871793, + "grad_norm": 0.7196067177185654, + "learning_rate": 2.4180640389954534e-05, + "loss": 0.4442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19523802399635315, + "step": 2140, + "valid_targets_mean": 2441.0, + "valid_targets_min": 1254 + }, + { + "epoch": 3.4375, + "grad_norm": 0.6116035709986902, + "learning_rate": 2.4102454305962892e-05, + "loss": 0.475, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.32390114665031433, + "step": 2145, + "valid_targets_mean": 4818.0, + "valid_targets_min": 2458 + }, + { + "epoch": 3.4455128205128207, + "grad_norm": 0.7198986904603069, + "learning_rate": 2.402420271661076e-05, + "loss": 0.4877, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23523953557014465, + "step": 2150, + "valid_targets_mean": 3893.0, + "valid_targets_min": 2248 + }, + { + "epoch": 3.453525641025641, + "grad_norm": 0.672020281280272, + "learning_rate": 2.3945886871369338e-05, + "loss": 0.4468, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23118053376674652, + "step": 2155, + "valid_targets_mean": 2928.2, + "valid_targets_min": 1130 + }, + { + "epoch": 3.4615384615384617, + "grad_norm": 0.7569600229336744, + "learning_rate": 2.3867508020735865e-05, + "loss": 0.4411, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26120418310165405, + "step": 2160, + "valid_targets_mean": 2751.4, + "valid_targets_min": 1396 + }, + { + "epoch": 3.469551282051282, + "grad_norm": 0.6554687409142441, + "learning_rate": 2.3789067416213568e-05, + "loss": 0.45, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23005196452140808, + "step": 2165, + "valid_targets_mean": 3382.0, + "valid_targets_min": 1397 + }, + { + "epoch": 3.4775641025641026, + "grad_norm": 0.747028905682346, + "learning_rate": 2.3710566310291733e-05, + "loss": 0.4333, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18495328724384308, + "step": 2170, + "valid_targets_mean": 2332.0, + "valid_targets_min": 1156 + }, + { + "epoch": 3.485576923076923, + "grad_norm": 0.7100166126510623, + "learning_rate": 2.36320059564257e-05, + "loss": 0.4327, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25588464736938477, + "step": 2175, + "valid_targets_mean": 3583.8, + "valid_targets_min": 1498 + }, + { + "epoch": 3.4935897435897436, + "grad_norm": 0.6545686028989137, + "learning_rate": 2.3553387609016833e-05, + "loss": 0.4424, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19589433073997498, + "step": 2180, + "valid_targets_mean": 2560.0, + "valid_targets_min": 1213 + }, + { + "epoch": 3.501602564102564, + "grad_norm": 0.65773072360802, + "learning_rate": 2.347471252339252e-05, + "loss": 0.4338, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20293955504894257, + "step": 2185, + "valid_targets_mean": 3022.8, + "valid_targets_min": 1082 + }, + { + "epoch": 3.5096153846153846, + "grad_norm": 0.6177409745438089, + "learning_rate": 2.339598195578608e-05, + "loss": 0.4342, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25875529646873474, + "step": 2190, + "valid_targets_mean": 4052.6, + "valid_targets_min": 1801 + }, + { + "epoch": 3.5176282051282053, + "grad_norm": 0.6329092510074122, + "learning_rate": 2.3317197163316757e-05, + "loss": 0.461, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24638238549232483, + "step": 2195, + "valid_targets_mean": 3553.9, + "valid_targets_min": 1596 + }, + { + "epoch": 3.5256410256410255, + "grad_norm": 0.6529483555107919, + "learning_rate": 2.3238359403969608e-05, + "loss": 0.4543, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18306535482406616, + "step": 2200, + "valid_targets_mean": 3056.1, + "valid_targets_min": 1417 + }, + { + "epoch": 3.5336538461538463, + "grad_norm": 0.6267577658348751, + "learning_rate": 2.315946993657543e-05, + "loss": 0.442, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21580994129180908, + "step": 2205, + "valid_targets_mean": 3201.5, + "valid_targets_min": 1336 + }, + { + "epoch": 3.5416666666666665, + "grad_norm": 0.6005105626455489, + "learning_rate": 2.3080530020790673e-05, + "loss": 0.4852, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2399841994047165, + "step": 2210, + "valid_targets_mean": 4254.6, + "valid_targets_min": 1462 + }, + { + "epoch": 3.5496794871794872, + "grad_norm": 0.6552387500620179, + "learning_rate": 2.300154091707731e-05, + "loss": 0.4376, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24262598156929016, + "step": 2215, + "valid_targets_mean": 3319.6, + "valid_targets_min": 2144 + }, + { + "epoch": 3.5576923076923075, + "grad_norm": 0.6060809840934867, + "learning_rate": 2.2922503886682706e-05, + "loss": 0.4586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20404544472694397, + "step": 2220, + "valid_targets_mean": 3787.6, + "valid_targets_min": 1371 + }, + { + "epoch": 3.565705128205128, + "grad_norm": 0.6299731710514318, + "learning_rate": 2.28434201916195e-05, + "loss": 0.4556, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2249678075313568, + "step": 2225, + "valid_targets_mean": 3228.4, + "valid_targets_min": 2044 + }, + { + "epoch": 3.573717948717949, + "grad_norm": 0.6029742132361011, + "learning_rate": 2.2764291094645446e-05, + "loss": 0.4144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2010929435491562, + "step": 2230, + "valid_targets_mean": 3879.8, + "valid_targets_min": 1956 + }, + { + "epoch": 3.581730769230769, + "grad_norm": 0.6873767677218213, + "learning_rate": 2.2685117859243223e-05, + "loss": 0.4595, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27158740162849426, + "step": 2235, + "valid_targets_mean": 3189.4, + "valid_targets_min": 1513 + }, + { + "epoch": 3.58974358974359, + "grad_norm": 1.1710449580528, + "learning_rate": 2.2605901749600312e-05, + "loss": 0.4554, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.221965491771698, + "step": 2240, + "valid_targets_mean": 3651.8, + "valid_targets_min": 2278 + }, + { + "epoch": 3.59775641025641, + "grad_norm": 0.6619076853535136, + "learning_rate": 2.2526644030588764e-05, + "loss": 0.473, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2691589891910553, + "step": 2245, + "valid_targets_mean": 3569.5, + "valid_targets_min": 1693 + }, + { + "epoch": 3.605769230769231, + "grad_norm": 0.6132468991635431, + "learning_rate": 2.2447345967745036e-05, + "loss": 0.4373, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2110455334186554, + "step": 2250, + "valid_targets_mean": 3703.0, + "valid_targets_min": 1927 + }, + { + "epoch": 3.613782051282051, + "grad_norm": 0.666763065129391, + "learning_rate": 2.2368008827249756e-05, + "loss": 0.4463, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1922965943813324, + "step": 2255, + "valid_targets_mean": 2739.9, + "valid_targets_min": 906 + }, + { + "epoch": 3.621794871794872, + "grad_norm": 0.6373842745925598, + "learning_rate": 2.228863387590752e-05, + "loss": 0.4369, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20126289129257202, + "step": 2260, + "valid_targets_mean": 3573.1, + "valid_targets_min": 1615 + }, + { + "epoch": 3.6298076923076925, + "grad_norm": 0.6474342528681002, + "learning_rate": 2.2209222381126687e-05, + "loss": 0.4472, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2593004107475281, + "step": 2265, + "valid_targets_mean": 3669.8, + "valid_targets_min": 1782 + }, + { + "epoch": 3.6378205128205128, + "grad_norm": 0.5913169448210792, + "learning_rate": 2.212977561089908e-05, + "loss": 0.4591, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16557088494300842, + "step": 2270, + "valid_targets_mean": 3567.5, + "valid_targets_min": 1666 + }, + { + "epoch": 3.6458333333333335, + "grad_norm": 0.6656921193211783, + "learning_rate": 2.20502948337798e-05, + "loss": 0.4355, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24777624011039734, + "step": 2275, + "valid_targets_mean": 3180.9, + "valid_targets_min": 1248 + }, + { + "epoch": 3.6538461538461537, + "grad_norm": 0.6717945927532133, + "learning_rate": 2.1970781318866953e-05, + "loss": 0.4439, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.242192342877388, + "step": 2280, + "valid_targets_mean": 2964.2, + "valid_targets_min": 1102 + }, + { + "epoch": 3.6618589743589745, + "grad_norm": 0.6797241275716768, + "learning_rate": 2.1891236335781363e-05, + "loss": 0.4445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23175741732120514, + "step": 2285, + "valid_targets_mean": 3787.2, + "valid_targets_min": 1027 + }, + { + "epoch": 3.6698717948717947, + "grad_norm": 0.7134571544398269, + "learning_rate": 2.1811661154646332e-05, + "loss": 0.4735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2182222455739975, + "step": 2290, + "valid_targets_mean": 3151.6, + "valid_targets_min": 2148 + }, + { + "epoch": 3.6778846153846154, + "grad_norm": 0.6496708866677491, + "learning_rate": 2.173205704606735e-05, + "loss": 0.4986, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27931785583496094, + "step": 2295, + "valid_targets_mean": 3918.9, + "valid_targets_min": 1684 + }, + { + "epoch": 3.685897435897436, + "grad_norm": 0.6323427887195299, + "learning_rate": 2.1652425281111785e-05, + "loss": 0.4557, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20083309710025787, + "step": 2300, + "valid_targets_mean": 3181.9, + "valid_targets_min": 1414 + }, + { + "epoch": 3.6939102564102564, + "grad_norm": 0.6216268656137096, + "learning_rate": 2.1572767131288607e-05, + "loss": 0.4787, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2681775689125061, + "step": 2305, + "valid_targets_mean": 4084.2, + "valid_targets_min": 1622 + }, + { + "epoch": 3.7019230769230766, + "grad_norm": 0.6732498170511206, + "learning_rate": 2.1493083868528095e-05, + "loss": 0.4328, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21949903666973114, + "step": 2310, + "valid_targets_mean": 3151.4, + "valid_targets_min": 1513 + }, + { + "epoch": 3.7099358974358974, + "grad_norm": 0.6287673686444623, + "learning_rate": 2.141337676516151e-05, + "loss": 0.4616, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2609504461288452, + "step": 2315, + "valid_targets_mean": 3938.0, + "valid_targets_min": 1673 + }, + { + "epoch": 3.717948717948718, + "grad_norm": 0.688967980961017, + "learning_rate": 2.1333647093900772e-05, + "loss": 0.4168, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2067280262708664, + "step": 2320, + "valid_targets_mean": 3478.5, + "valid_targets_min": 1245 + }, + { + "epoch": 3.7259615384615383, + "grad_norm": 0.6123641377842816, + "learning_rate": 2.1253896127818175e-05, + "loss": 0.431, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23997995257377625, + "step": 2325, + "valid_targets_mean": 3909.1, + "valid_targets_min": 1035 + }, + { + "epoch": 3.733974358974359, + "grad_norm": 0.615172533051529, + "learning_rate": 2.1174125140326013e-05, + "loss": 0.4593, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20165491104125977, + "step": 2330, + "valid_targets_mean": 3072.5, + "valid_targets_min": 1683 + }, + { + "epoch": 3.7419871794871797, + "grad_norm": 0.7780983242931223, + "learning_rate": 2.1094335405156277e-05, + "loss": 0.4249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22177544236183167, + "step": 2335, + "valid_targets_mean": 2725.5, + "valid_targets_min": 1370 + }, + { + "epoch": 3.75, + "grad_norm": 0.6096345797489205, + "learning_rate": 2.1014528196340316e-05, + "loss": 0.4456, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21743535995483398, + "step": 2340, + "valid_targets_mean": 3668.5, + "valid_targets_min": 1845 + }, + { + "epoch": 3.7580128205128203, + "grad_norm": 0.6645513239786645, + "learning_rate": 2.093470478818847e-05, + "loss": 0.4293, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25073322653770447, + "step": 2345, + "valid_targets_mean": 3488.8, + "valid_targets_min": 1500 + }, + { + "epoch": 3.766025641025641, + "grad_norm": 0.6070270739552414, + "learning_rate": 2.0854866455269756e-05, + "loss": 0.4573, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21000362932682037, + "step": 2350, + "valid_targets_mean": 3843.8, + "valid_targets_min": 1573 + }, + { + "epoch": 3.7740384615384617, + "grad_norm": 0.6727675248882965, + "learning_rate": 2.0775014472391496e-05, + "loss": 0.4343, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20167317986488342, + "step": 2355, + "valid_targets_mean": 2754.1, + "valid_targets_min": 1554 + }, + { + "epoch": 3.782051282051282, + "grad_norm": 0.6122852507941956, + "learning_rate": 2.0695150114578958e-05, + "loss": 0.4277, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16956289112567902, + "step": 2360, + "valid_targets_mean": 3181.2, + "valid_targets_min": 962 + }, + { + "epoch": 3.7900641025641026, + "grad_norm": 0.6388133658246149, + "learning_rate": 2.061527465705502e-05, + "loss": 0.4593, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20931920409202576, + "step": 2365, + "valid_targets_mean": 3005.4, + "valid_targets_min": 1171 + }, + { + "epoch": 3.7980769230769234, + "grad_norm": 0.5682912163577616, + "learning_rate": 2.0535389375219773e-05, + "loss": 0.4198, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1814444363117218, + "step": 2370, + "valid_targets_mean": 3885.2, + "valid_targets_min": 1395 + }, + { + "epoch": 3.8060897435897436, + "grad_norm": 0.6574722980505796, + "learning_rate": 2.045549554463019e-05, + "loss": 0.4433, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24636971950531006, + "step": 2375, + "valid_targets_mean": 4014.1, + "valid_targets_min": 1998 + }, + { + "epoch": 3.814102564102564, + "grad_norm": 0.6273095709433691, + "learning_rate": 2.0375594440979744e-05, + "loss": 0.4185, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22933214902877808, + "step": 2380, + "valid_targets_mean": 3326.8, + "valid_targets_min": 1139 + }, + { + "epoch": 3.8221153846153846, + "grad_norm": 0.8005277631684277, + "learning_rate": 2.0295687340078037e-05, + "loss": 0.4298, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20167329907417297, + "step": 2385, + "valid_targets_mean": 2439.2, + "valid_targets_min": 1559 + }, + { + "epoch": 3.8301282051282053, + "grad_norm": 0.6893751775357772, + "learning_rate": 2.0215775517830437e-05, + "loss": 0.4531, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.201649010181427, + "step": 2390, + "valid_targets_mean": 2689.5, + "valid_targets_min": 1646 + }, + { + "epoch": 3.8381410256410255, + "grad_norm": 0.6557278620462597, + "learning_rate": 2.013586025021769e-05, + "loss": 0.4517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2141844630241394, + "step": 2395, + "valid_targets_mean": 3022.9, + "valid_targets_min": 1898 + }, + { + "epoch": 3.8461538461538463, + "grad_norm": 0.7001408281016362, + "learning_rate": 2.0055942813275564e-05, + "loss": 0.4619, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2770041823387146, + "step": 2400, + "valid_targets_mean": 3905.4, + "valid_targets_min": 2320 + }, + { + "epoch": 3.8541666666666665, + "grad_norm": 0.6381369984137532, + "learning_rate": 1.9976024483074456e-05, + "loss": 0.4574, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23646041750907898, + "step": 2405, + "valid_targets_mean": 4021.9, + "valid_targets_min": 1446 + }, + { + "epoch": 3.8621794871794872, + "grad_norm": 0.7099349983584301, + "learning_rate": 1.9896106535699025e-05, + "loss": 0.4382, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2050849199295044, + "step": 2410, + "valid_targets_mean": 2474.8, + "valid_targets_min": 1082 + }, + { + "epoch": 3.8701923076923075, + "grad_norm": 0.7232764021711042, + "learning_rate": 1.9816190247227834e-05, + "loss": 0.451, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25825926661491394, + "step": 2415, + "valid_targets_mean": 3910.2, + "valid_targets_min": 2320 + }, + { + "epoch": 3.878205128205128, + "grad_norm": 0.7047224706964702, + "learning_rate": 1.9736276893712954e-05, + "loss": 0.4264, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22713282704353333, + "step": 2420, + "valid_targets_mean": 3444.4, + "valid_targets_min": 1360 + }, + { + "epoch": 3.886217948717949, + "grad_norm": 0.6716350824217531, + "learning_rate": 1.9656367751159565e-05, + "loss": 0.4344, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22322410345077515, + "step": 2425, + "valid_targets_mean": 3021.6, + "valid_targets_min": 1347 + }, + { + "epoch": 3.894230769230769, + "grad_norm": 0.563267042897286, + "learning_rate": 1.957646409550565e-05, + "loss": 0.4489, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17717109620571136, + "step": 2430, + "valid_targets_mean": 3728.2, + "valid_targets_min": 1451 + }, + { + "epoch": 3.90224358974359, + "grad_norm": 0.6291340525425678, + "learning_rate": 1.9496567202601545e-05, + "loss": 0.4735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3077090382575989, + "step": 2435, + "valid_targets_mean": 3830.9, + "valid_targets_min": 1372 + }, + { + "epoch": 3.91025641025641, + "grad_norm": 0.7050752155641052, + "learning_rate": 1.9416678348189627e-05, + "loss": 0.4378, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2817001938819885, + "step": 2440, + "valid_targets_mean": 3444.0, + "valid_targets_min": 1917 + }, + { + "epoch": 3.918269230769231, + "grad_norm": 0.6555555133872296, + "learning_rate": 1.9336798807883907e-05, + "loss": 0.4167, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1960204392671585, + "step": 2445, + "valid_targets_mean": 3324.8, + "valid_targets_min": 1572 + }, + { + "epoch": 3.926282051282051, + "grad_norm": 0.6321924413414791, + "learning_rate": 1.9256929857149686e-05, + "loss": 0.4526, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2062862366437912, + "step": 2450, + "valid_targets_mean": 3591.9, + "valid_targets_min": 550 + }, + { + "epoch": 3.934294871794872, + "grad_norm": 0.616346001652963, + "learning_rate": 1.9177072771283167e-05, + "loss": 0.4541, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24890682101249695, + "step": 2455, + "valid_targets_mean": 3909.6, + "valid_targets_min": 1039 + }, + { + "epoch": 3.9423076923076925, + "grad_norm": 0.6799112655166439, + "learning_rate": 1.9097228825391087e-05, + "loss": 0.4309, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21679437160491943, + "step": 2460, + "valid_targets_mean": 2652.1, + "valid_targets_min": 1238 + }, + { + "epoch": 3.9503205128205128, + "grad_norm": 0.6583026813097685, + "learning_rate": 1.9017399294370413e-05, + "loss": 0.4793, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22348573803901672, + "step": 2465, + "valid_targets_mean": 2977.6, + "valid_targets_min": 1343 + }, + { + "epoch": 3.9583333333333335, + "grad_norm": 0.6670441809816142, + "learning_rate": 1.893758545288791e-05, + "loss": 0.4545, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21193484961986542, + "step": 2470, + "valid_targets_mean": 3141.6, + "valid_targets_min": 1694 + }, + { + "epoch": 3.9663461538461537, + "grad_norm": 0.61851995126912, + "learning_rate": 1.8857788575359847e-05, + "loss": 0.4811, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19258682429790497, + "step": 2475, + "valid_targets_mean": 3049.1, + "valid_targets_min": 1476 + }, + { + "epoch": 3.9743589743589745, + "grad_norm": 0.6794924164380063, + "learning_rate": 1.87780099359316e-05, + "loss": 0.4588, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27653396129608154, + "step": 2480, + "valid_targets_mean": 3255.9, + "valid_targets_min": 1908 + }, + { + "epoch": 3.9823717948717947, + "grad_norm": 0.7169543641596569, + "learning_rate": 1.869825080845734e-05, + "loss": 0.435, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19775667786598206, + "step": 2485, + "valid_targets_mean": 4001.0, + "valid_targets_min": 2498 + }, + { + "epoch": 3.9903846153846154, + "grad_norm": 0.6671341375603711, + "learning_rate": 1.8618512466479686e-05, + "loss": 0.4894, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21765673160552979, + "step": 2490, + "valid_targets_mean": 3365.1, + "valid_targets_min": 1805 + }, + { + "epoch": 3.998397435897436, + "grad_norm": 0.6765061948636006, + "learning_rate": 1.8538796183209373e-05, + "loss": 0.4577, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2636738419532776, + "step": 2495, + "valid_targets_mean": 4012.4, + "valid_targets_min": 1428 + }, + { + "epoch": 4.006410256410256, + "grad_norm": 0.6655155710231736, + "learning_rate": 1.845910323150491e-05, + "loss": 0.4366, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24367494881153107, + "step": 2500, + "valid_targets_mean": 3231.6, + "valid_targets_min": 1219 + }, + { + "epoch": 4.014423076923077, + "grad_norm": 0.7059345393440495, + "learning_rate": 1.8379434883852255e-05, + "loss": 0.4037, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26208043098449707, + "step": 2505, + "valid_targets_mean": 3079.8, + "valid_targets_min": 986 + }, + { + "epoch": 4.022435897435898, + "grad_norm": 0.6062537531967671, + "learning_rate": 1.8299792412344524e-05, + "loss": 0.3962, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21878264844417572, + "step": 2510, + "valid_targets_mean": 4089.1, + "valid_targets_min": 1814 + }, + { + "epoch": 4.030448717948718, + "grad_norm": 0.6246330390817048, + "learning_rate": 1.8220177088661635e-05, + "loss": 0.4001, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21555614471435547, + "step": 2515, + "valid_targets_mean": 3795.8, + "valid_targets_min": 1825 + }, + { + "epoch": 4.038461538461538, + "grad_norm": 0.639561324678776, + "learning_rate": 1.814059018405004e-05, + "loss": 0.4172, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21151083707809448, + "step": 2520, + "valid_targets_mean": 4458.9, + "valid_targets_min": 1846 + }, + { + "epoch": 4.046474358974359, + "grad_norm": 0.7403920395651158, + "learning_rate": 1.806103296930243e-05, + "loss": 0.437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33656492829322815, + "step": 2525, + "valid_targets_mean": 3656.4, + "valid_targets_min": 1344 + }, + { + "epoch": 4.05448717948718, + "grad_norm": 0.6825577996306195, + "learning_rate": 1.7981506714737392e-05, + "loss": 0.4118, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20105662941932678, + "step": 2530, + "valid_targets_mean": 3150.6, + "valid_targets_min": 1136 + }, + { + "epoch": 4.0625, + "grad_norm": 0.749708965824278, + "learning_rate": 1.7902012690179188e-05, + "loss": 0.4397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20164191722869873, + "step": 2535, + "valid_targets_mean": 2514.1, + "valid_targets_min": 1387 + }, + { + "epoch": 4.07051282051282, + "grad_norm": 0.7097724382282349, + "learning_rate": 1.7822552164937437e-05, + "loss": 0.4426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.199723482131958, + "step": 2540, + "valid_targets_mean": 3137.9, + "valid_targets_min": 1292 + }, + { + "epoch": 4.078525641025641, + "grad_norm": 0.6773219534138569, + "learning_rate": 1.7743126407786873e-05, + "loss": 0.3975, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22100375592708588, + "step": 2545, + "valid_targets_mean": 3075.5, + "valid_targets_min": 1410 + }, + { + "epoch": 4.086538461538462, + "grad_norm": 0.6702225878356237, + "learning_rate": 1.766373668694707e-05, + "loss": 0.4412, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24414443969726562, + "step": 2550, + "valid_targets_mean": 3953.2, + "valid_targets_min": 2485 + }, + { + "epoch": 4.094551282051282, + "grad_norm": 0.6972193610913691, + "learning_rate": 1.7584384270062195e-05, + "loss": 0.4407, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1885961890220642, + "step": 2555, + "valid_targets_mean": 2996.1, + "valid_targets_min": 1835 + }, + { + "epoch": 4.102564102564102, + "grad_norm": 0.6975292697185156, + "learning_rate": 1.7505070424180772e-05, + "loss": 0.4317, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18920600414276123, + "step": 2560, + "valid_targets_mean": 2883.2, + "valid_targets_min": 1551 + }, + { + "epoch": 4.110576923076923, + "grad_norm": 0.6577599409862241, + "learning_rate": 1.7425796415735454e-05, + "loss": 0.4379, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20840083062648773, + "step": 2565, + "valid_targets_mean": 4064.6, + "valid_targets_min": 1982 + }, + { + "epoch": 4.118589743589744, + "grad_norm": 0.6695196774253919, + "learning_rate": 1.7346563510522783e-05, + "loss": 0.4002, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18286213278770447, + "step": 2570, + "valid_targets_mean": 2274.0, + "valid_targets_min": 1095 + }, + { + "epoch": 4.126602564102564, + "grad_norm": 0.6537862606131949, + "learning_rate": 1.7267372973682998e-05, + "loss": 0.4048, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19385072588920593, + "step": 2575, + "valid_targets_mean": 3719.2, + "valid_targets_min": 1474 + }, + { + "epoch": 4.134615384615385, + "grad_norm": 0.5868498848485982, + "learning_rate": 1.7188226069679834e-05, + "loss": 0.4158, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21755146980285645, + "step": 2580, + "valid_targets_mean": 4564.9, + "valid_targets_min": 1938 + }, + { + "epoch": 4.142628205128205, + "grad_norm": 0.6767048420725426, + "learning_rate": 1.7109124062280307e-05, + "loss": 0.4418, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22366130352020264, + "step": 2585, + "valid_targets_mean": 3031.8, + "valid_targets_min": 1243 + }, + { + "epoch": 4.1506410256410255, + "grad_norm": 0.7120830994901551, + "learning_rate": 1.7030068214534567e-05, + "loss": 0.4229, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23689591884613037, + "step": 2590, + "valid_targets_mean": 3208.6, + "valid_targets_min": 1150 + }, + { + "epoch": 4.158653846153846, + "grad_norm": 0.703777570197176, + "learning_rate": 1.695105978875572e-05, + "loss": 0.436, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24541151523590088, + "step": 2595, + "valid_targets_mean": 3488.1, + "valid_targets_min": 1595 + }, + { + "epoch": 4.166666666666667, + "grad_norm": 0.5966577197558491, + "learning_rate": 1.687210004649965e-05, + "loss": 0.4057, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15391811728477478, + "step": 2600, + "valid_targets_mean": 3095.4, + "valid_targets_min": 1222 + }, + { + "epoch": 4.174679487179487, + "grad_norm": 0.6152942652644364, + "learning_rate": 1.679319024854491e-05, + "loss": 0.4066, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19409197568893433, + "step": 2605, + "valid_targets_mean": 3663.8, + "valid_targets_min": 2129 + }, + { + "epoch": 4.1826923076923075, + "grad_norm": 0.6704321714725356, + "learning_rate": 1.6714331654872564e-05, + "loss": 0.4466, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18893314898014069, + "step": 2610, + "valid_targets_mean": 3231.8, + "valid_targets_min": 1416 + }, + { + "epoch": 4.190705128205128, + "grad_norm": 0.6790183832725902, + "learning_rate": 1.663552552464609e-05, + "loss": 0.4036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18782873451709747, + "step": 2615, + "valid_targets_mean": 2877.1, + "valid_targets_min": 1577 + }, + { + "epoch": 4.198717948717949, + "grad_norm": 0.7432291147508001, + "learning_rate": 1.6556773116191257e-05, + "loss": 0.4164, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1897243857383728, + "step": 2620, + "valid_targets_mean": 3462.6, + "valid_targets_min": 1822 + }, + { + "epoch": 4.206730769230769, + "grad_norm": 0.7419012457950234, + "learning_rate": 1.647807568697603e-05, + "loss": 0.4334, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2460288107395172, + "step": 2625, + "valid_targets_mean": 3369.2, + "valid_targets_min": 1473 + }, + { + "epoch": 4.214743589743589, + "grad_norm": 0.6997547380176556, + "learning_rate": 1.6399434493590524e-05, + "loss": 0.4207, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26003655791282654, + "step": 2630, + "valid_targets_mean": 2828.1, + "valid_targets_min": 1672 + }, + { + "epoch": 4.222756410256411, + "grad_norm": 0.6408145422593703, + "learning_rate": 1.6320850791726884e-05, + "loss": 0.4304, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20207910239696503, + "step": 2635, + "valid_targets_mean": 4211.0, + "valid_targets_min": 2612 + }, + { + "epoch": 4.230769230769231, + "grad_norm": 0.6519080842758559, + "learning_rate": 1.6242325836159304e-05, + "loss": 0.4399, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19569715857505798, + "step": 2640, + "valid_targets_mean": 3958.1, + "valid_targets_min": 2131 + }, + { + "epoch": 4.238782051282051, + "grad_norm": 0.7440505840964299, + "learning_rate": 1.6163860880723923e-05, + "loss": 0.4276, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20437708497047424, + "step": 2645, + "valid_targets_mean": 3260.0, + "valid_targets_min": 1898 + }, + { + "epoch": 4.246794871794872, + "grad_norm": 0.6477870134403431, + "learning_rate": 1.6085457178298866e-05, + "loss": 0.452, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18539613485336304, + "step": 2650, + "valid_targets_mean": 3409.0, + "valid_targets_min": 1298 + }, + { + "epoch": 4.2548076923076925, + "grad_norm": 0.7027555667225593, + "learning_rate": 1.6007115980784182e-05, + "loss": 0.429, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25074416399002075, + "step": 2655, + "valid_targets_mean": 3670.0, + "valid_targets_min": 1460 + }, + { + "epoch": 4.262820512820513, + "grad_norm": 0.6616349670095881, + "learning_rate": 1.592883853908188e-05, + "loss": 0.4215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24440480768680573, + "step": 2660, + "valid_targets_mean": 3443.6, + "valid_targets_min": 1700 + }, + { + "epoch": 4.270833333333333, + "grad_norm": 0.6700230014541894, + "learning_rate": 1.585062610307599e-05, + "loss": 0.4247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23861554265022278, + "step": 2665, + "valid_targets_mean": 4433.6, + "valid_targets_min": 1694 + }, + { + "epoch": 4.278846153846154, + "grad_norm": 0.6309841423986773, + "learning_rate": 1.5772479921612543e-05, + "loss": 0.3991, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20668131113052368, + "step": 2670, + "valid_targets_mean": 3856.2, + "valid_targets_min": 1328 + }, + { + "epoch": 4.2868589743589745, + "grad_norm": 0.7039155998764873, + "learning_rate": 1.5694401242479677e-05, + "loss": 0.4286, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24892082810401917, + "step": 2675, + "valid_targets_mean": 3544.2, + "valid_targets_min": 1305 + }, + { + "epoch": 4.294871794871795, + "grad_norm": 0.6587387758222163, + "learning_rate": 1.5616391312387683e-05, + "loss": 0.409, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.270779550075531, + "step": 2680, + "valid_targets_mean": 3489.4, + "valid_targets_min": 499 + }, + { + "epoch": 4.302884615384615, + "grad_norm": 0.6497113518798093, + "learning_rate": 1.5538451376949106e-05, + "loss": 0.4099, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14637237787246704, + "step": 2685, + "valid_targets_mean": 3210.9, + "valid_targets_min": 1280 + }, + { + "epoch": 4.310897435897436, + "grad_norm": 0.6759898185472816, + "learning_rate": 1.5460582680658888e-05, + "loss": 0.422, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.264614462852478, + "step": 2690, + "valid_targets_mean": 3662.5, + "valid_targets_min": 958 + }, + { + "epoch": 4.318910256410256, + "grad_norm": 0.817264426638095, + "learning_rate": 1.5382786466874446e-05, + "loss": 0.4066, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21336670219898224, + "step": 2695, + "valid_targets_mean": 3901.5, + "valid_targets_min": 2171 + }, + { + "epoch": 4.326923076923077, + "grad_norm": 0.6562994101403469, + "learning_rate": 1.5305063977795856e-05, + "loss": 0.3739, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15959221124649048, + "step": 2700, + "valid_targets_mean": 3191.8, + "valid_targets_min": 414 + }, + { + "epoch": 4.334935897435898, + "grad_norm": 0.7131764046047239, + "learning_rate": 1.5227416454445995e-05, + "loss": 0.4354, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22022467851638794, + "step": 2705, + "valid_targets_mean": 3388.1, + "valid_targets_min": 1956 + }, + { + "epoch": 4.342948717948718, + "grad_norm": 0.6862361843603952, + "learning_rate": 1.5149845136650748e-05, + "loss": 0.4251, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23837068676948547, + "step": 2710, + "valid_targets_mean": 3612.4, + "valid_targets_min": 2013 + }, + { + "epoch": 4.350961538461538, + "grad_norm": 0.6766264991957116, + "learning_rate": 1.5072351263019177e-05, + "loss": 0.4211, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18963858485221863, + "step": 2715, + "valid_targets_mean": 3206.9, + "valid_targets_min": 1246 + }, + { + "epoch": 4.358974358974359, + "grad_norm": 0.6423632252482322, + "learning_rate": 1.4994936070923784e-05, + "loss": 0.4328, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21200747787952423, + "step": 2720, + "valid_targets_mean": 3951.4, + "valid_targets_min": 1737 + }, + { + "epoch": 4.36698717948718, + "grad_norm": 0.72947251277914, + "learning_rate": 1.4917600796480745e-05, + "loss": 0.4001, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23151524364948273, + "step": 2725, + "valid_targets_mean": 3062.0, + "valid_targets_min": 1930 + }, + { + "epoch": 4.375, + "grad_norm": 0.6808810668536788, + "learning_rate": 1.4840346674530122e-05, + "loss": 0.3949, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18378394842147827, + "step": 2730, + "valid_targets_mean": 3416.1, + "valid_targets_min": 1432 + }, + { + "epoch": 4.38301282051282, + "grad_norm": 0.6956159730218684, + "learning_rate": 1.4763174938616232e-05, + "loss": 0.415, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23507966101169586, + "step": 2735, + "valid_targets_mean": 3285.8, + "valid_targets_min": 1288 + }, + { + "epoch": 4.391025641025641, + "grad_norm": 0.7593064935847421, + "learning_rate": 1.4686086820967865e-05, + "loss": 0.4093, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17585715651512146, + "step": 2740, + "valid_targets_mean": 2403.0, + "valid_targets_min": 1542 + }, + { + "epoch": 4.399038461538462, + "grad_norm": 0.6879660078700143, + "learning_rate": 1.460908355247868e-05, + "loss": 0.4401, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20843517780303955, + "step": 2745, + "valid_targets_mean": 3105.5, + "valid_targets_min": 863 + }, + { + "epoch": 4.407051282051282, + "grad_norm": 0.6873460704516899, + "learning_rate": 1.4532166362687507e-05, + "loss": 0.43, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20860859751701355, + "step": 2750, + "valid_targets_mean": 3617.4, + "valid_targets_min": 1439 + }, + { + "epoch": 4.415064102564102, + "grad_norm": 0.6317930419680684, + "learning_rate": 1.445533647975871e-05, + "loss": 0.4013, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17258626222610474, + "step": 2755, + "valid_targets_mean": 3327.8, + "valid_targets_min": 1072 + }, + { + "epoch": 4.423076923076923, + "grad_norm": 0.6669263848270824, + "learning_rate": 1.437859513046263e-05, + "loss": 0.4162, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22215047478675842, + "step": 2760, + "valid_targets_mean": 3657.6, + "valid_targets_min": 2037 + }, + { + "epoch": 4.431089743589744, + "grad_norm": 0.7486743173364545, + "learning_rate": 1.4301943540155914e-05, + "loss": 0.4158, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2349042296409607, + "step": 2765, + "valid_targets_mean": 2998.9, + "valid_targets_min": 1132 + }, + { + "epoch": 4.439102564102564, + "grad_norm": 0.6889074075352274, + "learning_rate": 1.4225382932762033e-05, + "loss": 0.4465, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.269346684217453, + "step": 2770, + "valid_targets_mean": 4051.9, + "valid_targets_min": 1045 + }, + { + "epoch": 4.447115384615385, + "grad_norm": 0.6137490359997793, + "learning_rate": 1.4148914530751681e-05, + "loss": 0.412, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15939438343048096, + "step": 2775, + "valid_targets_mean": 3874.0, + "valid_targets_min": 1534 + }, + { + "epoch": 4.455128205128205, + "grad_norm": 0.6504916131404131, + "learning_rate": 1.4072539555123292e-05, + "loss": 0.4245, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22253087162971497, + "step": 2780, + "valid_targets_mean": 3175.6, + "valid_targets_min": 954 + }, + { + "epoch": 4.4631410256410255, + "grad_norm": 0.6896233473844445, + "learning_rate": 1.3996259225383514e-05, + "loss": 0.4151, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19968242943286896, + "step": 2785, + "valid_targets_mean": 3552.5, + "valid_targets_min": 1764 + }, + { + "epoch": 4.471153846153846, + "grad_norm": 0.6992123547782415, + "learning_rate": 1.3920074759527737e-05, + "loss": 0.4122, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17509326338768005, + "step": 2790, + "valid_targets_mean": 2779.2, + "valid_targets_min": 1373 + }, + { + "epoch": 4.479166666666667, + "grad_norm": 0.6750521919582315, + "learning_rate": 1.3843987374020689e-05, + "loss": 0.4642, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.262592077255249, + "step": 2795, + "valid_targets_mean": 3841.5, + "valid_targets_min": 2576 + }, + { + "epoch": 4.487179487179487, + "grad_norm": 0.6857465728937038, + "learning_rate": 1.376799828377696e-05, + "loss": 0.4063, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16925640404224396, + "step": 2800, + "valid_targets_mean": 3173.4, + "valid_targets_min": 1814 + }, + { + "epoch": 4.4951923076923075, + "grad_norm": 0.6765970937646676, + "learning_rate": 1.3692108702141642e-05, + "loss": 0.439, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23575451970100403, + "step": 2805, + "valid_targets_mean": 4194.6, + "valid_targets_min": 1251 + }, + { + "epoch": 4.503205128205128, + "grad_norm": 0.6553531608498788, + "learning_rate": 1.361631984087091e-05, + "loss": 0.4091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1784219890832901, + "step": 2810, + "valid_targets_mean": 3642.9, + "valid_targets_min": 1717 + }, + { + "epoch": 4.511217948717949, + "grad_norm": 0.6676890139605697, + "learning_rate": 1.354063291011273e-05, + "loss": 0.4402, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2754717469215393, + "step": 2815, + "valid_targets_mean": 3976.6, + "valid_targets_min": 1437 + }, + { + "epoch": 4.519230769230769, + "grad_norm": 0.7254927172252034, + "learning_rate": 1.3465049118387486e-05, + "loss": 0.4303, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22217559814453125, + "step": 2820, + "valid_targets_mean": 2873.2, + "valid_targets_min": 1542 + }, + { + "epoch": 4.527243589743589, + "grad_norm": 0.6692788323218626, + "learning_rate": 1.3389569672568707e-05, + "loss": 0.4162, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20454254746437073, + "step": 2825, + "valid_targets_mean": 3318.4, + "valid_targets_min": 1186 + }, + { + "epoch": 4.535256410256411, + "grad_norm": 0.7232464763746403, + "learning_rate": 1.331419577786381e-05, + "loss": 0.4221, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20352435111999512, + "step": 2830, + "valid_targets_mean": 2910.6, + "valid_targets_min": 1001 + }, + { + "epoch": 4.543269230769231, + "grad_norm": 0.742131444756949, + "learning_rate": 1.3238928637794816e-05, + "loss": 0.4486, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19160433113574982, + "step": 2835, + "valid_targets_mean": 2323.6, + "valid_targets_min": 1262 + }, + { + "epoch": 4.551282051282051, + "grad_norm": 0.7121044176017662, + "learning_rate": 1.3163769454179183e-05, + "loss": 0.41, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21123504638671875, + "step": 2840, + "valid_targets_mean": 3215.4, + "valid_targets_min": 1144 + }, + { + "epoch": 4.559294871794872, + "grad_norm": 0.6925068843691634, + "learning_rate": 1.3088719427110552e-05, + "loss": 0.413, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24364648759365082, + "step": 2845, + "valid_targets_mean": 3563.1, + "valid_targets_min": 1962 + }, + { + "epoch": 4.5673076923076925, + "grad_norm": 0.816363568822839, + "learning_rate": 1.3013779754939666e-05, + "loss": 0.423, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22845833003520966, + "step": 2850, + "valid_targets_mean": 2455.4, + "valid_targets_min": 550 + }, + { + "epoch": 4.575320512820513, + "grad_norm": 0.6877719670747243, + "learning_rate": 1.2938951634255164e-05, + "loss": 0.3971, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15317656099796295, + "step": 2855, + "valid_targets_mean": 3912.5, + "valid_targets_min": 2289 + }, + { + "epoch": 4.583333333333333, + "grad_norm": 0.6810657065494565, + "learning_rate": 1.2864236259864495e-05, + "loss": 0.4465, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18710535764694214, + "step": 2860, + "valid_targets_mean": 3079.2, + "valid_targets_min": 1062 + }, + { + "epoch": 4.591346153846154, + "grad_norm": 0.7388686288448452, + "learning_rate": 1.2789634824774887e-05, + "loss": 0.4301, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17414233088493347, + "step": 2865, + "valid_targets_mean": 2205.1, + "valid_targets_min": 888 + }, + { + "epoch": 4.5993589743589745, + "grad_norm": 0.6908940420707356, + "learning_rate": 1.2715148520174206e-05, + "loss": 0.4478, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18807990849018097, + "step": 2870, + "valid_targets_mean": 3430.1, + "valid_targets_min": 1486 + }, + { + "epoch": 4.607371794871795, + "grad_norm": 0.7244653586868361, + "learning_rate": 1.2640778535412036e-05, + "loss": 0.433, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25073760747909546, + "step": 2875, + "valid_targets_mean": 2897.6, + "valid_targets_min": 1399 + }, + { + "epoch": 4.615384615384615, + "grad_norm": 0.6742654847596043, + "learning_rate": 1.2566526057980608e-05, + "loss": 0.4042, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16925884783267975, + "step": 2880, + "valid_targets_mean": 3137.2, + "valid_targets_min": 1471 + }, + { + "epoch": 4.623397435897436, + "grad_norm": 0.662283326948339, + "learning_rate": 1.2492392273495879e-05, + "loss": 0.4385, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2201838195323944, + "step": 2885, + "valid_targets_mean": 3038.2, + "valid_targets_min": 1431 + }, + { + "epoch": 4.631410256410256, + "grad_norm": 0.6330280165928103, + "learning_rate": 1.2418378365678612e-05, + "loss": 0.4306, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18156296014785767, + "step": 2890, + "valid_targets_mean": 3015.2, + "valid_targets_min": 1585 + }, + { + "epoch": 4.639423076923077, + "grad_norm": 0.763424932007766, + "learning_rate": 1.234448551633542e-05, + "loss": 0.4431, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23698724806308746, + "step": 2895, + "valid_targets_mean": 2702.5, + "valid_targets_min": 1279 + }, + { + "epoch": 4.647435897435898, + "grad_norm": 0.7242998627972805, + "learning_rate": 1.2270714905339969e-05, + "loss": 0.4445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3452325761318207, + "step": 2900, + "valid_targets_mean": 3654.0, + "valid_targets_min": 2077 + }, + { + "epoch": 4.655448717948718, + "grad_norm": 0.7707549750614332, + "learning_rate": 1.2197067710614075e-05, + "loss": 0.429, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1654444932937622, + "step": 2905, + "valid_targets_mean": 3281.2, + "valid_targets_min": 1639 + }, + { + "epoch": 4.663461538461538, + "grad_norm": 0.7253164300928779, + "learning_rate": 1.2123545108108943e-05, + "loss": 0.4321, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2323237657546997, + "step": 2910, + "valid_targets_mean": 3340.8, + "valid_targets_min": 1600 + }, + { + "epoch": 4.671474358974359, + "grad_norm": 0.612699318047636, + "learning_rate": 1.2050148271786348e-05, + "loss": 0.4267, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19045571982860565, + "step": 2915, + "valid_targets_mean": 3440.0, + "valid_targets_min": 1826 + }, + { + "epoch": 4.67948717948718, + "grad_norm": 0.735080951398492, + "learning_rate": 1.1976878373599928e-05, + "loss": 0.417, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19636553525924683, + "step": 2920, + "valid_targets_mean": 2708.1, + "valid_targets_min": 979 + }, + { + "epoch": 4.6875, + "grad_norm": 0.657443931557103, + "learning_rate": 1.1903736583476441e-05, + "loss": 0.4342, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21986836194992065, + "step": 2925, + "valid_targets_mean": 3358.6, + "valid_targets_min": 1883 + }, + { + "epoch": 4.69551282051282, + "grad_norm": 0.682993970543838, + "learning_rate": 1.1830724069297106e-05, + "loss": 0.4267, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2588217854499817, + "step": 2930, + "valid_targets_mean": 3359.2, + "valid_targets_min": 1359 + }, + { + "epoch": 4.703525641025641, + "grad_norm": 0.6474784546938326, + "learning_rate": 1.1757841996878957e-05, + "loss": 0.4266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2362293303012848, + "step": 2935, + "valid_targets_mean": 4250.5, + "valid_targets_min": 2993 + }, + { + "epoch": 4.711538461538462, + "grad_norm": 0.7501659574509763, + "learning_rate": 1.1685091529956187e-05, + "loss": 0.44, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2783074975013733, + "step": 2940, + "valid_targets_mean": 3313.5, + "valid_targets_min": 1262 + }, + { + "epoch": 4.719551282051282, + "grad_norm": 0.6978281308325148, + "learning_rate": 1.161247383016163e-05, + "loss": 0.4352, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19544479250907898, + "step": 2945, + "valid_targets_mean": 2959.8, + "valid_targets_min": 1005 + }, + { + "epoch": 4.727564102564102, + "grad_norm": 0.6908017823409119, + "learning_rate": 1.1539990057008166e-05, + "loss": 0.4156, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19036749005317688, + "step": 2950, + "valid_targets_mean": 3262.1, + "valid_targets_min": 1725 + }, + { + "epoch": 4.735576923076923, + "grad_norm": 0.7014147475823345, + "learning_rate": 1.1467641367870198e-05, + "loss": 0.4371, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24416306614875793, + "step": 2955, + "valid_targets_mean": 3036.5, + "valid_targets_min": 1567 + }, + { + "epoch": 4.743589743589744, + "grad_norm": 0.6439502519640261, + "learning_rate": 1.1395428917965239e-05, + "loss": 0.4445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20787352323532104, + "step": 2960, + "valid_targets_mean": 3304.0, + "valid_targets_min": 1978 + }, + { + "epoch": 4.751602564102564, + "grad_norm": 0.6757167663877847, + "learning_rate": 1.1323353860335385e-05, + "loss": 0.4272, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20633909106254578, + "step": 2965, + "valid_targets_mean": 3118.0, + "valid_targets_min": 1276 + }, + { + "epoch": 4.759615384615385, + "grad_norm": 0.8015474096966113, + "learning_rate": 1.1251417345828962e-05, + "loss": 0.4242, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26273638010025024, + "step": 2970, + "valid_targets_mean": 3264.5, + "valid_targets_min": 1424 + }, + { + "epoch": 4.767628205128205, + "grad_norm": 0.653174346750347, + "learning_rate": 1.1179620523082107e-05, + "loss": 0.3973, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16476941108703613, + "step": 2975, + "valid_targets_mean": 3145.1, + "valid_targets_min": 1541 + }, + { + "epoch": 4.7756410256410255, + "grad_norm": 0.9842938465268183, + "learning_rate": 1.110796453850047e-05, + "loss": 0.4182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20752543210983276, + "step": 2980, + "valid_targets_mean": 3191.5, + "valid_targets_min": 1345 + }, + { + "epoch": 4.783653846153846, + "grad_norm": 0.6083086270292675, + "learning_rate": 1.1036450536240877e-05, + "loss": 0.4096, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21362093091011047, + "step": 2985, + "valid_targets_mean": 4176.8, + "valid_targets_min": 3208 + }, + { + "epoch": 4.791666666666667, + "grad_norm": 0.6968763238899879, + "learning_rate": 1.0965079658193068e-05, + "loss": 0.4019, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19594931602478027, + "step": 2990, + "valid_targets_mean": 3102.4, + "valid_targets_min": 1687 + }, + { + "epoch": 4.799679487179487, + "grad_norm": 0.7839558880300881, + "learning_rate": 1.0893853043961475e-05, + "loss": 0.4373, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23440927267074585, + "step": 2995, + "valid_targets_mean": 2671.4, + "valid_targets_min": 1873 + }, + { + "epoch": 4.8076923076923075, + "grad_norm": 0.6586122407734638, + "learning_rate": 1.0822771830847011e-05, + "loss": 0.4413, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18959015607833862, + "step": 3000, + "valid_targets_mean": 3147.2, + "valid_targets_min": 2148 + }, + { + "epoch": 4.815705128205128, + "grad_norm": 0.678352819999638, + "learning_rate": 1.0751837153828926e-05, + "loss": 0.4358, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21638141572475433, + "step": 3005, + "valid_targets_mean": 3559.1, + "valid_targets_min": 1652 + }, + { + "epoch": 4.823717948717949, + "grad_norm": 0.6939760841212267, + "learning_rate": 1.0681050145546666e-05, + "loss": 0.4125, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20729690790176392, + "step": 3010, + "valid_targets_mean": 3605.6, + "valid_targets_min": 1453 + }, + { + "epoch": 4.831730769230769, + "grad_norm": 0.6497288464232186, + "learning_rate": 1.0610411936281801e-05, + "loss": 0.4151, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17099347710609436, + "step": 3015, + "valid_targets_mean": 2961.2, + "valid_targets_min": 1492 + }, + { + "epoch": 4.839743589743589, + "grad_norm": 0.6746147685566729, + "learning_rate": 1.0539923653939978e-05, + "loss": 0.433, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22235651314258575, + "step": 3020, + "valid_targets_mean": 3617.4, + "valid_targets_min": 1139 + }, + { + "epoch": 4.847756410256411, + "grad_norm": 0.6061124819950633, + "learning_rate": 1.0469586424032903e-05, + "loss": 0.4063, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22783973813056946, + "step": 3025, + "valid_targets_mean": 3897.8, + "valid_targets_min": 1709 + }, + { + "epoch": 4.855769230769231, + "grad_norm": 0.6502422422037057, + "learning_rate": 1.0399401369660369e-05, + "loss": 0.4201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24566122889518738, + "step": 3030, + "valid_targets_mean": 3775.1, + "valid_targets_min": 1358 + }, + { + "epoch": 4.863782051282051, + "grad_norm": 0.7756849146002172, + "learning_rate": 1.0329369611492334e-05, + "loss": 0.4508, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3204527199268341, + "step": 3035, + "valid_targets_mean": 3318.0, + "valid_targets_min": 1533 + }, + { + "epoch": 4.871794871794872, + "grad_norm": 0.7128943391519159, + "learning_rate": 1.0259492267751022e-05, + "loss": 0.4179, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26550376415252686, + "step": 3040, + "valid_targets_mean": 3249.8, + "valid_targets_min": 1759 + }, + { + "epoch": 4.8798076923076925, + "grad_norm": 0.6257102605335594, + "learning_rate": 1.0189770454193052e-05, + "loss": 0.4212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17940127849578857, + "step": 3045, + "valid_targets_mean": 3007.9, + "valid_targets_min": 1605 + }, + { + "epoch": 4.887820512820513, + "grad_norm": 0.6658037649834273, + "learning_rate": 1.0120205284091673e-05, + "loss": 0.4287, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2461349219083786, + "step": 3050, + "valid_targets_mean": 3405.4, + "valid_targets_min": 1610 + }, + { + "epoch": 4.895833333333333, + "grad_norm": 0.6088992492443039, + "learning_rate": 1.0050797868218907e-05, + "loss": 0.3922, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21799695491790771, + "step": 3055, + "valid_targets_mean": 4130.9, + "valid_targets_min": 3140 + }, + { + "epoch": 4.903846153846154, + "grad_norm": 0.720631787482109, + "learning_rate": 9.981549314827876e-06, + "loss": 0.3927, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26906195282936096, + "step": 3060, + "valid_targets_mean": 4083.5, + "valid_targets_min": 1684 + }, + { + "epoch": 4.9118589743589745, + "grad_norm": 0.7438701142315824, + "learning_rate": 9.912460729635097e-06, + "loss": 0.4313, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28831809759140015, + "step": 3065, + "valid_targets_mean": 3378.2, + "valid_targets_min": 2101 + }, + { + "epoch": 4.919871794871795, + "grad_norm": 0.6929518865026288, + "learning_rate": 9.843533215802796e-06, + "loss": 0.4036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1646612584590912, + "step": 3070, + "valid_targets_mean": 2657.8, + "valid_targets_min": 1281 + }, + { + "epoch": 4.927884615384615, + "grad_norm": 0.7264560851780181, + "learning_rate": 9.774767873921357e-06, + "loss": 0.4215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20341382920742035, + "step": 3075, + "valid_targets_mean": 2974.9, + "valid_targets_min": 1139 + }, + { + "epoch": 4.935897435897436, + "grad_norm": 0.7233257329986168, + "learning_rate": 9.706165801991651e-06, + "loss": 0.4535, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22906440496444702, + "step": 3080, + "valid_targets_mean": 2782.8, + "valid_targets_min": 1780 + }, + { + "epoch": 4.943910256410256, + "grad_norm": 0.699799002799183, + "learning_rate": 9.637728095407593e-06, + "loss": 0.4379, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22438335418701172, + "step": 3085, + "valid_targets_mean": 3167.8, + "valid_targets_min": 1102 + }, + { + "epoch": 4.951923076923077, + "grad_norm": 0.6238344864682764, + "learning_rate": 9.56945584693861e-06, + "loss": 0.4411, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2419654130935669, + "step": 3090, + "valid_targets_mean": 4029.0, + "valid_targets_min": 1959 + }, + { + "epoch": 4.959935897435898, + "grad_norm": 0.7211972903417753, + "learning_rate": 9.501350146712193e-06, + "loss": 0.3867, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21571232378482819, + "step": 3095, + "valid_targets_mean": 3564.9, + "valid_targets_min": 1100 + }, + { + "epoch": 4.967948717948718, + "grad_norm": 0.7053492464736898, + "learning_rate": 9.433412082196527e-06, + "loss": 0.4242, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18533562123775482, + "step": 3100, + "valid_targets_mean": 3007.1, + "valid_targets_min": 1388 + }, + { + "epoch": 4.975961538461538, + "grad_norm": 0.6849707876117926, + "learning_rate": 9.365642738183044e-06, + "loss": 0.4166, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19545109570026398, + "step": 3105, + "valid_targets_mean": 3327.9, + "valid_targets_min": 2159 + }, + { + "epoch": 4.983974358974359, + "grad_norm": 0.6370873534930793, + "learning_rate": 9.298043196769217e-06, + "loss": 0.4291, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19308175146579742, + "step": 3110, + "valid_targets_mean": 3785.6, + "valid_targets_min": 2033 + }, + { + "epoch": 4.99198717948718, + "grad_norm": 0.6454443176511807, + "learning_rate": 9.230614537341167e-06, + "loss": 0.4216, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1699100136756897, + "step": 3115, + "valid_targets_mean": 3461.2, + "valid_targets_min": 2214 + }, + { + "epoch": 5.0, + "grad_norm": 0.6429695062279672, + "learning_rate": 9.163357836556498e-06, + "loss": 0.4128, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1951141357421875, + "step": 3120, + "valid_targets_mean": 3430.5, + "valid_targets_min": 1091 + }, + { + "epoch": 5.00801282051282, + "grad_norm": 0.66720250544281, + "learning_rate": 9.096274168327122e-06, + "loss": 0.3801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1902952492237091, + "step": 3125, + "valid_targets_mean": 2890.1, + "valid_targets_min": 1327 + }, + { + "epoch": 5.016025641025641, + "grad_norm": 0.669454231915622, + "learning_rate": 9.029364603802017e-06, + "loss": 0.3865, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18216481804847717, + "step": 3130, + "valid_targets_mean": 3393.0, + "valid_targets_min": 1481 + }, + { + "epoch": 5.024038461538462, + "grad_norm": 0.6661302890123278, + "learning_rate": 8.962630211350248e-06, + "loss": 0.3906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16417013108730316, + "step": 3135, + "valid_targets_mean": 3326.0, + "valid_targets_min": 1581 + }, + { + "epoch": 5.032051282051282, + "grad_norm": 0.7925350846744128, + "learning_rate": 8.89607205654378e-06, + "loss": 0.4028, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1825258433818817, + "step": 3140, + "valid_targets_mean": 2510.0, + "valid_targets_min": 1164 + }, + { + "epoch": 5.040064102564102, + "grad_norm": 0.7013136630014062, + "learning_rate": 8.829691202140591e-06, + "loss": 0.4019, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19449245929718018, + "step": 3145, + "valid_targets_mean": 3181.2, + "valid_targets_min": 1486 + }, + { + "epoch": 5.048076923076923, + "grad_norm": 0.7829605133852731, + "learning_rate": 8.763488708067604e-06, + "loss": 0.3897, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19113725423812866, + "step": 3150, + "valid_targets_mean": 2806.6, + "valid_targets_min": 1211 + }, + { + "epoch": 5.056089743589744, + "grad_norm": 0.7321570896705135, + "learning_rate": 8.69746563140379e-06, + "loss": 0.4263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17530694603919983, + "step": 3155, + "valid_targets_mean": 2612.5, + "valid_targets_min": 1538 + }, + { + "epoch": 5.064102564102564, + "grad_norm": 0.714247520403382, + "learning_rate": 8.631623026363331e-06, + "loss": 0.4153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20394036173820496, + "step": 3160, + "valid_targets_mean": 2703.5, + "valid_targets_min": 1376 + }, + { + "epoch": 5.072115384615385, + "grad_norm": 0.7484531732138497, + "learning_rate": 8.56596194427873e-06, + "loss": 0.4085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21379892528057098, + "step": 3165, + "valid_targets_mean": 2677.2, + "valid_targets_min": 1336 + }, + { + "epoch": 5.080128205128205, + "grad_norm": 0.7060537296523256, + "learning_rate": 8.500483433584054e-06, + "loss": 0.4467, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2669312357902527, + "step": 3170, + "valid_targets_mean": 3217.6, + "valid_targets_min": 1213 + }, + { + "epoch": 5.0881410256410255, + "grad_norm": 0.7260585860653015, + "learning_rate": 8.435188539798187e-06, + "loss": 0.4102, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26159393787384033, + "step": 3175, + "valid_targets_mean": 3665.5, + "valid_targets_min": 1497 + }, + { + "epoch": 5.096153846153846, + "grad_norm": 0.72499578714397, + "learning_rate": 8.370078305508136e-06, + "loss": 0.4244, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21100249886512756, + "step": 3180, + "valid_targets_mean": 3384.0, + "valid_targets_min": 2226 + }, + { + "epoch": 5.104166666666667, + "grad_norm": 0.6887260663196959, + "learning_rate": 8.305153770352384e-06, + "loss": 0.3796, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22591131925582886, + "step": 3185, + "valid_targets_mean": 4126.4, + "valid_targets_min": 2272 + }, + { + "epoch": 5.112179487179487, + "grad_norm": 0.6989200526949411, + "learning_rate": 8.240415971004285e-06, + "loss": 0.4005, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16576944291591644, + "step": 3190, + "valid_targets_mean": 3220.1, + "valid_targets_min": 1444 + }, + { + "epoch": 5.1201923076923075, + "grad_norm": 0.6864972332131725, + "learning_rate": 8.175865941155525e-06, + "loss": 0.3831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16296260058879852, + "step": 3195, + "valid_targets_mean": 3324.0, + "valid_targets_min": 910 + }, + { + "epoch": 5.128205128205128, + "grad_norm": 0.7543519847756472, + "learning_rate": 8.111504711499598e-06, + "loss": 0.4089, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22263804078102112, + "step": 3200, + "valid_targets_mean": 2618.9, + "valid_targets_min": 1818 + }, + { + "epoch": 5.136217948717949, + "grad_norm": 0.7212508777376027, + "learning_rate": 8.04733330971536e-06, + "loss": 0.3741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1671215444803238, + "step": 3205, + "valid_targets_mean": 3066.6, + "valid_targets_min": 1704 + }, + { + "epoch": 5.144230769230769, + "grad_norm": 0.796742920785991, + "learning_rate": 7.983352760450618e-06, + "loss": 0.4007, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17478328943252563, + "step": 3210, + "valid_targets_mean": 2288.9, + "valid_targets_min": 1145 + }, + { + "epoch": 5.152243589743589, + "grad_norm": 0.7580241319638324, + "learning_rate": 7.919564085305768e-06, + "loss": 0.41, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20330554246902466, + "step": 3215, + "valid_targets_mean": 2943.9, + "valid_targets_min": 1139 + }, + { + "epoch": 5.160256410256411, + "grad_norm": 0.6680107869642323, + "learning_rate": 7.855968302817487e-06, + "loss": 0.3867, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15371182560920715, + "step": 3220, + "valid_targets_mean": 3359.5, + "valid_targets_min": 1095 + }, + { + "epoch": 5.168269230769231, + "grad_norm": 0.6968368566588543, + "learning_rate": 7.792566428442456e-06, + "loss": 0.4052, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2633320987224579, + "step": 3225, + "valid_targets_mean": 4066.5, + "valid_targets_min": 1513 + }, + { + "epoch": 5.176282051282051, + "grad_norm": 0.6875394470534935, + "learning_rate": 7.729359474541168e-06, + "loss": 0.4287, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2854374647140503, + "step": 3230, + "valid_targets_mean": 4442.9, + "valid_targets_min": 2500 + }, + { + "epoch": 5.184294871794872, + "grad_norm": 0.6925330861753394, + "learning_rate": 7.666348450361737e-06, + "loss": 0.4148, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17162266373634338, + "step": 3235, + "valid_targets_mean": 2721.9, + "valid_targets_min": 1968 + }, + { + "epoch": 5.1923076923076925, + "grad_norm": 0.7570677907029336, + "learning_rate": 7.60353436202381e-06, + "loss": 0.3988, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1700199842453003, + "step": 3240, + "valid_targets_mean": 2320.1, + "valid_targets_min": 1441 + }, + { + "epoch": 5.200320512820513, + "grad_norm": 0.7158614972007095, + "learning_rate": 7.540918212502479e-06, + "loss": 0.3828, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1725841462612152, + "step": 3245, + "valid_targets_mean": 3038.8, + "valid_targets_min": 1377 + }, + { + "epoch": 5.208333333333333, + "grad_norm": 0.7181978160840652, + "learning_rate": 7.478501001612281e-06, + "loss": 0.4274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21954363584518433, + "step": 3250, + "valid_targets_mean": 3661.5, + "valid_targets_min": 2180 + }, + { + "epoch": 5.216346153846154, + "grad_norm": 0.6783801521788718, + "learning_rate": 7.416283725991229e-06, + "loss": 0.4046, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17134936153888702, + "step": 3255, + "valid_targets_mean": 3496.2, + "valid_targets_min": 1997 + }, + { + "epoch": 5.2243589743589745, + "grad_norm": 0.6875655682468811, + "learning_rate": 7.354267379084896e-06, + "loss": 0.376, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1433093249797821, + "step": 3260, + "valid_targets_mean": 2500.0, + "valid_targets_min": 1442 + }, + { + "epoch": 5.232371794871795, + "grad_norm": 0.6590823909859274, + "learning_rate": 7.292452951130548e-06, + "loss": 0.4233, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18068474531173706, + "step": 3265, + "valid_targets_mean": 3639.6, + "valid_targets_min": 1665 + }, + { + "epoch": 5.240384615384615, + "grad_norm": 0.7609055986341658, + "learning_rate": 7.230841429141347e-06, + "loss": 0.3969, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2031887322664261, + "step": 3270, + "valid_targets_mean": 2639.9, + "valid_targets_min": 1116 + }, + { + "epoch": 5.248397435897436, + "grad_norm": 0.7297704086568011, + "learning_rate": 7.169433796890595e-06, + "loss": 0.4104, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2634013593196869, + "step": 3275, + "valid_targets_mean": 3474.2, + "valid_targets_min": 1055 + }, + { + "epoch": 5.256410256410256, + "grad_norm": 0.7292817007924668, + "learning_rate": 7.108231034895976e-06, + "loss": 0.4074, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1796400547027588, + "step": 3280, + "valid_targets_mean": 2613.4, + "valid_targets_min": 1002 + }, + { + "epoch": 5.264423076923077, + "grad_norm": 0.7705299567696228, + "learning_rate": 7.047234120403972e-06, + "loss": 0.3838, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18816229701042175, + "step": 3285, + "valid_targets_mean": 2420.5, + "valid_targets_min": 1169 + }, + { + "epoch": 5.272435897435898, + "grad_norm": 0.7370356355398553, + "learning_rate": 6.986444027374211e-06, + "loss": 0.4132, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18651533126831055, + "step": 3290, + "valid_targets_mean": 2874.6, + "valid_targets_min": 1239 + }, + { + "epoch": 5.280448717948718, + "grad_norm": 0.7135916882470296, + "learning_rate": 6.925861726463919e-06, + "loss": 0.3892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17134974896907806, + "step": 3295, + "valid_targets_mean": 3150.8, + "valid_targets_min": 1575 + }, + { + "epoch": 5.288461538461538, + "grad_norm": 0.6797397189653621, + "learning_rate": 6.865488185012464e-06, + "loss": 0.3923, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25696736574172974, + "step": 3300, + "valid_targets_mean": 3962.4, + "valid_targets_min": 2025 + }, + { + "epoch": 5.296474358974359, + "grad_norm": 0.6808708641726515, + "learning_rate": 6.805324367025825e-06, + "loss": 0.4348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20182277262210846, + "step": 3305, + "valid_targets_mean": 3453.0, + "valid_targets_min": 1826 + }, + { + "epoch": 5.30448717948718, + "grad_norm": 0.6955667805354783, + "learning_rate": 6.745371233161309e-06, + "loss": 0.4154, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17542634904384613, + "step": 3310, + "valid_targets_mean": 3122.4, + "valid_targets_min": 1289 + }, + { + "epoch": 5.3125, + "grad_norm": 0.7669626631188224, + "learning_rate": 6.685629740712103e-06, + "loss": 0.4033, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1757771074771881, + "step": 3315, + "valid_targets_mean": 2404.1, + "valid_targets_min": 1245 + }, + { + "epoch": 5.32051282051282, + "grad_norm": 0.6364145933456632, + "learning_rate": 6.6261008435920605e-06, + "loss": 0.4305, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19072768092155457, + "step": 3320, + "valid_targets_mean": 4167.0, + "valid_targets_min": 1782 + }, + { + "epoch": 5.328525641025641, + "grad_norm": 0.7385381709298334, + "learning_rate": 6.566785492320471e-06, + "loss": 0.3946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19371137022972107, + "step": 3325, + "valid_targets_mean": 3276.1, + "valid_targets_min": 1877 + }, + { + "epoch": 5.336538461538462, + "grad_norm": 0.6406519414101312, + "learning_rate": 6.507684634006815e-06, + "loss": 0.4381, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18672963976860046, + "step": 3330, + "valid_targets_mean": 3630.0, + "valid_targets_min": 1553 + }, + { + "epoch": 5.344551282051282, + "grad_norm": 1.097982858267181, + "learning_rate": 6.448799212335734e-06, + "loss": 0.3715, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.162650927901268, + "step": 3335, + "valid_targets_mean": 3895.4, + "valid_targets_min": 1285 + }, + { + "epoch": 5.352564102564102, + "grad_norm": 0.6997463934239558, + "learning_rate": 6.390130167551869e-06, + "loss": 0.4279, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18873977661132812, + "step": 3340, + "valid_targets_mean": 3904.0, + "valid_targets_min": 1570 + }, + { + "epoch": 5.360576923076923, + "grad_norm": 1.1100846001553775, + "learning_rate": 6.331678436444939e-06, + "loss": 0.4022, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18822500109672546, + "step": 3345, + "valid_targets_mean": 2726.0, + "valid_targets_min": 952 + }, + { + "epoch": 5.368589743589744, + "grad_norm": 0.7153858561529954, + "learning_rate": 6.273444952334713e-06, + "loss": 0.3923, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21593938767910004, + "step": 3350, + "valid_targets_mean": 3549.6, + "valid_targets_min": 1163 + }, + { + "epoch": 5.376602564102564, + "grad_norm": 0.7190662273232792, + "learning_rate": 6.2154306450561175e-06, + "loss": 0.4087, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2505274713039398, + "step": 3355, + "valid_targets_mean": 3911.0, + "valid_targets_min": 1846 + }, + { + "epoch": 5.384615384615385, + "grad_norm": 0.7831513366233006, + "learning_rate": 6.157636440944445e-06, + "loss": 0.4346, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2270057499408722, + "step": 3360, + "valid_targets_mean": 3002.4, + "valid_targets_min": 1543 + }, + { + "epoch": 5.392628205128205, + "grad_norm": 0.7321437000204511, + "learning_rate": 6.100063262820474e-06, + "loss": 0.416, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2008737325668335, + "step": 3365, + "valid_targets_mean": 2756.5, + "valid_targets_min": 1558 + }, + { + "epoch": 5.4006410256410255, + "grad_norm": 0.7304501731128669, + "learning_rate": 6.0427120299758236e-06, + "loss": 0.4525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23144027590751648, + "step": 3370, + "valid_targets_mean": 3062.9, + "valid_targets_min": 1384 + }, + { + "epoch": 5.408653846153846, + "grad_norm": 0.7235528319730333, + "learning_rate": 5.985583658158212e-06, + "loss": 0.41, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18171024322509766, + "step": 3375, + "valid_targets_mean": 3140.8, + "valid_targets_min": 1222 + }, + { + "epoch": 5.416666666666667, + "grad_norm": 0.7141083464217536, + "learning_rate": 5.928679059556852e-06, + "loss": 0.3973, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1845255345106125, + "step": 3380, + "valid_targets_mean": 3194.8, + "valid_targets_min": 2115 + }, + { + "epoch": 5.424679487179487, + "grad_norm": 0.6483089967406473, + "learning_rate": 5.871999142787908e-06, + "loss": 0.3936, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18002232909202576, + "step": 3385, + "valid_targets_mean": 3681.8, + "valid_targets_min": 1596 + }, + { + "epoch": 5.4326923076923075, + "grad_norm": 0.8151582962335311, + "learning_rate": 5.815544812879936e-06, + "loss": 0.4003, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22642439603805542, + "step": 3390, + "valid_targets_mean": 2856.9, + "valid_targets_min": 1199 + }, + { + "epoch": 5.440705128205128, + "grad_norm": 0.7596988380076639, + "learning_rate": 5.759316971259503e-06, + "loss": 0.4045, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17470070719718933, + "step": 3395, + "valid_targets_mean": 2373.8, + "valid_targets_min": 1417 + }, + { + "epoch": 5.448717948717949, + "grad_norm": 0.7329093752754443, + "learning_rate": 5.703316515736734e-06, + "loss": 0.4018, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1921035200357437, + "step": 3400, + "valid_targets_mean": 3090.9, + "valid_targets_min": 1336 + }, + { + "epoch": 5.456730769230769, + "grad_norm": 0.7452909007835515, + "learning_rate": 5.647544340491007e-06, + "loss": 0.3746, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19344599545001984, + "step": 3405, + "valid_targets_mean": 3248.2, + "valid_targets_min": 1252 + }, + { + "epoch": 5.464743589743589, + "grad_norm": 0.7126975137481713, + "learning_rate": 5.592001336056659e-06, + "loss": 0.4007, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19665846228599548, + "step": 3410, + "valid_targets_mean": 3343.5, + "valid_targets_min": 1677 + }, + { + "epoch": 5.472756410256411, + "grad_norm": 0.6833668634232472, + "learning_rate": 5.536688389308782e-06, + "loss": 0.4017, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17852634191513062, + "step": 3415, + "valid_targets_mean": 3677.0, + "valid_targets_min": 1612 + }, + { + "epoch": 5.480769230769231, + "grad_norm": 0.748433644741738, + "learning_rate": 5.4816063834490496e-06, + "loss": 0.4164, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18009144067764282, + "step": 3420, + "valid_targets_mean": 3052.0, + "valid_targets_min": 1286 + }, + { + "epoch": 5.488782051282051, + "grad_norm": 0.812620413026482, + "learning_rate": 5.426756197991625e-06, + "loss": 0.4074, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18821418285369873, + "step": 3425, + "valid_targets_mean": 2160.9, + "valid_targets_min": 1126 + }, + { + "epoch": 5.496794871794872, + "grad_norm": 0.7068608539186978, + "learning_rate": 5.372138708749104e-06, + "loss": 0.3879, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18039929866790771, + "step": 3430, + "valid_targets_mean": 2598.0, + "valid_targets_min": 1057 + }, + { + "epoch": 5.5048076923076925, + "grad_norm": 0.7028548405633532, + "learning_rate": 5.3177547878185436e-06, + "loss": 0.3736, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20030388236045837, + "step": 3435, + "valid_targets_mean": 3529.9, + "valid_targets_min": 1279 + }, + { + "epoch": 5.512820512820513, + "grad_norm": 0.7252628700686375, + "learning_rate": 5.263605303567532e-06, + "loss": 0.4153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1996767818927765, + "step": 3440, + "valid_targets_mean": 3796.0, + "valid_targets_min": 1994 + }, + { + "epoch": 5.520833333333333, + "grad_norm": 0.6769767972947832, + "learning_rate": 5.20969112062032e-06, + "loss": 0.3831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1702975332736969, + "step": 3445, + "valid_targets_mean": 3751.6, + "valid_targets_min": 2180 + }, + { + "epoch": 5.528846153846154, + "grad_norm": 0.6367400226538237, + "learning_rate": 5.156013099844017e-06, + "loss": 0.3868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18816307187080383, + "step": 3450, + "valid_targets_mean": 3761.1, + "valid_targets_min": 2067 + }, + { + "epoch": 5.5368589743589745, + "grad_norm": 0.762007661528858, + "learning_rate": 5.1025720983348544e-06, + "loss": 0.4109, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18160797655582428, + "step": 3455, + "valid_targets_mean": 2774.4, + "valid_targets_min": 1854 + }, + { + "epoch": 5.544871794871795, + "grad_norm": 0.6491196559358456, + "learning_rate": 5.049368969404484e-06, + "loss": 0.3905, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20168176293373108, + "step": 3460, + "valid_targets_mean": 4139.0, + "valid_targets_min": 2287 + }, + { + "epoch": 5.552884615384615, + "grad_norm": 0.7426585440275758, + "learning_rate": 4.99640456256636e-06, + "loss": 0.4139, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24590930342674255, + "step": 3465, + "valid_targets_mean": 4485.9, + "valid_targets_min": 2496 + }, + { + "epoch": 5.560897435897436, + "grad_norm": 0.7219434318095365, + "learning_rate": 4.9436797235221814e-06, + "loss": 0.41, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1727789044380188, + "step": 3470, + "valid_targets_mean": 3258.4, + "valid_targets_min": 1371 + }, + { + "epoch": 5.568910256410256, + "grad_norm": 0.6818676323113749, + "learning_rate": 4.891195294148376e-06, + "loss": 0.4136, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25181639194488525, + "step": 3475, + "valid_targets_mean": 4208.4, + "valid_targets_min": 1723 + }, + { + "epoch": 5.576923076923077, + "grad_norm": 0.6992763627546373, + "learning_rate": 4.838952112482671e-06, + "loss": 0.3984, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19770342111587524, + "step": 3480, + "valid_targets_mean": 3358.4, + "valid_targets_min": 1219 + }, + { + "epoch": 5.584935897435898, + "grad_norm": 0.7552459263829787, + "learning_rate": 4.786951012710699e-06, + "loss": 0.3803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1997116506099701, + "step": 3485, + "valid_targets_mean": 2835.9, + "valid_targets_min": 1751 + }, + { + "epoch": 5.592948717948718, + "grad_norm": 0.7498720530405246, + "learning_rate": 4.735192825152686e-06, + "loss": 0.4046, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22662481665611267, + "step": 3490, + "valid_targets_mean": 2884.1, + "valid_targets_min": 1326 + }, + { + "epoch": 5.600961538461538, + "grad_norm": 0.6838970190093023, + "learning_rate": 4.683678376250189e-06, + "loss": 0.3944, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1867828518152237, + "step": 3495, + "valid_targets_mean": 3274.0, + "valid_targets_min": 1449 + }, + { + "epoch": 5.608974358974359, + "grad_norm": 0.7179281563554628, + "learning_rate": 4.6324084885529086e-06, + "loss": 0.4091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1755140721797943, + "step": 3500, + "valid_targets_mean": 2720.6, + "valid_targets_min": 1161 + }, + { + "epoch": 5.61698717948718, + "grad_norm": 0.7132290780954255, + "learning_rate": 4.581383980705538e-06, + "loss": 0.4397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2771325707435608, + "step": 3505, + "valid_targets_mean": 3815.5, + "valid_targets_min": 1349 + }, + { + "epoch": 5.625, + "grad_norm": 0.7968947927990316, + "learning_rate": 4.530605667434727e-06, + "loss": 0.4096, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22923137247562408, + "step": 3510, + "valid_targets_mean": 2863.4, + "valid_targets_min": 936 + }, + { + "epoch": 5.63301282051282, + "grad_norm": 0.6979932817564295, + "learning_rate": 4.480074359536013e-06, + "loss": 0.4133, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18550440669059753, + "step": 3515, + "valid_targets_mean": 2944.8, + "valid_targets_min": 930 + }, + { + "epoch": 5.641025641025641, + "grad_norm": 0.7219225740101636, + "learning_rate": 4.429790863860934e-06, + "loss": 0.391, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2050371766090393, + "step": 3520, + "valid_targets_mean": 3712.1, + "valid_targets_min": 1453 + }, + { + "epoch": 5.649038461538462, + "grad_norm": 0.6965210669379445, + "learning_rate": 4.3797559833041146e-06, + "loss": 0.4122, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1735437661409378, + "step": 3525, + "valid_targets_mean": 3219.4, + "valid_targets_min": 2132 + }, + { + "epoch": 5.657051282051282, + "grad_norm": 0.8428872784634172, + "learning_rate": 4.329970516790447e-06, + "loss": 0.4348, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18732038140296936, + "step": 3530, + "valid_targets_mean": 2791.6, + "valid_targets_min": 336 + }, + { + "epoch": 5.665064102564102, + "grad_norm": 0.7150607650814045, + "learning_rate": 4.280435259262363e-06, + "loss": 0.4085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20566867291927338, + "step": 3535, + "valid_targets_mean": 3396.0, + "valid_targets_min": 1699 + }, + { + "epoch": 5.673076923076923, + "grad_norm": 0.7116633801835935, + "learning_rate": 4.231151001667077e-06, + "loss": 0.4072, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2252032309770584, + "step": 3540, + "valid_targets_mean": 3112.0, + "valid_targets_min": 1476 + }, + { + "epoch": 5.681089743589744, + "grad_norm": 0.7511725036698694, + "learning_rate": 4.182118530944044e-06, + "loss": 0.4066, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2120506465435028, + "step": 3545, + "valid_targets_mean": 3328.6, + "valid_targets_min": 936 + }, + { + "epoch": 5.689102564102564, + "grad_norm": 0.7876564837635224, + "learning_rate": 4.133338630012307e-06, + "loss": 0.4002, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1560036987066269, + "step": 3550, + "valid_targets_mean": 2496.6, + "valid_targets_min": 1106 + }, + { + "epoch": 5.697115384615385, + "grad_norm": 0.6763762178805722, + "learning_rate": 4.0848120777580554e-06, + "loss": 0.3958, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22274337708950043, + "step": 3555, + "valid_targets_mean": 3814.1, + "valid_targets_min": 2584 + }, + { + "epoch": 5.705128205128205, + "grad_norm": 0.704942903116712, + "learning_rate": 4.036539649022182e-06, + "loss": 0.4021, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2151128500699997, + "step": 3560, + "valid_targets_mean": 3590.1, + "valid_targets_min": 2213 + }, + { + "epoch": 5.7131410256410255, + "grad_norm": 0.6795435394572237, + "learning_rate": 3.988522114587865e-06, + "loss": 0.3859, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20813924074172974, + "step": 3565, + "valid_targets_mean": 3533.2, + "valid_targets_min": 705 + }, + { + "epoch": 5.721153846153846, + "grad_norm": 0.7939700789734224, + "learning_rate": 3.940760241168331e-06, + "loss": 0.4157, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22058705985546112, + "step": 3570, + "valid_targets_mean": 2726.1, + "valid_targets_min": 1556 + }, + { + "epoch": 5.729166666666667, + "grad_norm": 0.9570747396437661, + "learning_rate": 3.893254791394541e-06, + "loss": 0.4162, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18854621052742004, + "step": 3575, + "valid_targets_mean": 3211.8, + "valid_targets_min": 1412 + }, + { + "epoch": 5.737179487179487, + "grad_norm": 0.7579852115219292, + "learning_rate": 3.846006523803074e-06, + "loss": 0.3974, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2074403613805771, + "step": 3580, + "valid_targets_mean": 3376.0, + "valid_targets_min": 1343 + }, + { + "epoch": 5.7451923076923075, + "grad_norm": 0.6845106183070118, + "learning_rate": 3.799016192823981e-06, + "loss": 0.3933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19570234417915344, + "step": 3585, + "valid_targets_mean": 3755.8, + "valid_targets_min": 2177 + }, + { + "epoch": 5.753205128205128, + "grad_norm": 0.8004361786362002, + "learning_rate": 3.7522845487687276e-06, + "loss": 0.4092, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14786019921302795, + "step": 3590, + "valid_targets_mean": 2022.4, + "valid_targets_min": 1277 + }, + { + "epoch": 5.761217948717949, + "grad_norm": 0.7046916418363054, + "learning_rate": 3.7058123378182664e-06, + "loss": 0.3683, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23868544399738312, + "step": 3595, + "valid_targets_mean": 4221.0, + "valid_targets_min": 1016 + }, + { + "epoch": 5.769230769230769, + "grad_norm": 0.6955381635070105, + "learning_rate": 3.6596003020110636e-06, + "loss": 0.4169, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2173866182565689, + "step": 3600, + "valid_targets_mean": 3665.2, + "valid_targets_min": 2237 + }, + { + "epoch": 5.777243589743589, + "grad_norm": 0.7436163771450518, + "learning_rate": 3.613649179231287e-06, + "loss": 0.4302, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21486309170722961, + "step": 3605, + "valid_targets_mean": 3161.4, + "valid_targets_min": 1151 + }, + { + "epoch": 5.785256410256411, + "grad_norm": 0.7829331148056964, + "learning_rate": 3.5679597031970017e-06, + "loss": 0.3894, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18953225016593933, + "step": 3610, + "valid_targets_mean": 3401.4, + "valid_targets_min": 1370 + }, + { + "epoch": 5.793269230769231, + "grad_norm": 0.6950516414350525, + "learning_rate": 3.5225326034484764e-06, + "loss": 0.4108, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13954158127307892, + "step": 3615, + "valid_targets_mean": 3077.5, + "valid_targets_min": 1339 + }, + { + "epoch": 5.801282051282051, + "grad_norm": 0.6235832757051263, + "learning_rate": 3.4773686053365197e-06, + "loss": 0.4515, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2791977524757385, + "step": 3620, + "valid_targets_mean": 4062.9, + "valid_targets_min": 2367 + }, + { + "epoch": 5.809294871794872, + "grad_norm": 0.6458964170822857, + "learning_rate": 3.4324684300109003e-06, + "loss": 0.3831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17912498116493225, + "step": 3625, + "valid_targets_mean": 3829.0, + "valid_targets_min": 1869 + }, + { + "epoch": 5.8173076923076925, + "grad_norm": 0.8189307575750902, + "learning_rate": 3.387832794408832e-06, + "loss": 0.4073, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17674008011817932, + "step": 3630, + "valid_targets_mean": 2375.8, + "valid_targets_min": 1262 + }, + { + "epoch": 5.825320512820513, + "grad_norm": 0.6833961747412601, + "learning_rate": 3.3434624112435342e-06, + "loss": 0.416, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16443946957588196, + "step": 3635, + "valid_targets_mean": 3370.0, + "valid_targets_min": 2367 + }, + { + "epoch": 5.833333333333333, + "grad_norm": 0.6669342669580356, + "learning_rate": 3.2993579889928397e-06, + "loss": 0.4223, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20291730761528015, + "step": 3640, + "valid_targets_mean": 4021.2, + "valid_targets_min": 1763 + }, + { + "epoch": 5.841346153846154, + "grad_norm": 0.6881699647444639, + "learning_rate": 3.25552023188789e-06, + "loss": 0.3931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.202553853392601, + "step": 3645, + "valid_targets_mean": 3183.1, + "valid_targets_min": 2049 + }, + { + "epoch": 5.8493589743589745, + "grad_norm": 0.7150030760695406, + "learning_rate": 3.211949839901889e-06, + "loss": 0.3803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17485833168029785, + "step": 3650, + "valid_targets_mean": 3227.4, + "valid_targets_min": 1470 + }, + { + "epoch": 5.857371794871795, + "grad_norm": 0.6634402802873007, + "learning_rate": 3.168647508738927e-06, + "loss": 0.395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19096645712852478, + "step": 3655, + "valid_targets_mean": 3900.9, + "valid_targets_min": 1916 + }, + { + "epoch": 5.865384615384615, + "grad_norm": 0.7434378217779124, + "learning_rate": 3.125613929822866e-06, + "loss": 0.4319, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2568639814853668, + "step": 3660, + "valid_targets_mean": 3375.5, + "valid_targets_min": 1327 + }, + { + "epoch": 5.873397435897436, + "grad_norm": 0.7867797075510559, + "learning_rate": 3.0828497902863106e-06, + "loss": 0.3962, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14944416284561157, + "step": 3665, + "valid_targets_mean": 2277.9, + "valid_targets_min": 624 + }, + { + "epoch": 5.881410256410256, + "grad_norm": 0.6864014515846244, + "learning_rate": 3.0403557729596267e-06, + "loss": 0.4114, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2673938572406769, + "step": 3670, + "valid_targets_mean": 4138.4, + "valid_targets_min": 1669 + }, + { + "epoch": 5.889423076923077, + "grad_norm": 0.8379222723904577, + "learning_rate": 2.998132556360038e-06, + "loss": 0.4098, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22460472583770752, + "step": 3675, + "valid_targets_mean": 2899.4, + "valid_targets_min": 1615 + }, + { + "epoch": 5.897435897435898, + "grad_norm": 0.7268337278878783, + "learning_rate": 2.9561808146808068e-06, + "loss": 0.4222, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23322021961212158, + "step": 3680, + "valid_targets_mean": 3700.5, + "valid_targets_min": 1337 + }, + { + "epoch": 5.905448717948718, + "grad_norm": 0.6501800965550159, + "learning_rate": 2.9145012177804476e-06, + "loss": 0.3919, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18246079981327057, + "step": 3685, + "valid_targets_mean": 4446.0, + "valid_targets_min": 1308 + }, + { + "epoch": 5.913461538461538, + "grad_norm": 0.7429320013230565, + "learning_rate": 2.8730944311720454e-06, + "loss": 0.3973, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2123229056596756, + "step": 3690, + "valid_targets_mean": 3597.4, + "valid_targets_min": 2036 + }, + { + "epoch": 5.921474358974359, + "grad_norm": 0.7648587017773072, + "learning_rate": 2.8319611160126226e-06, + "loss": 0.4237, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2272709310054779, + "step": 3695, + "valid_targets_mean": 3211.0, + "valid_targets_min": 895 + }, + { + "epoch": 5.92948717948718, + "grad_norm": 0.697802310529263, + "learning_rate": 2.791101929092592e-06, + "loss": 0.3863, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18288114666938782, + "step": 3700, + "valid_targets_mean": 2783.2, + "valid_targets_min": 1506 + }, + { + "epoch": 5.9375, + "grad_norm": 0.733930232603034, + "learning_rate": 2.750517522825251e-06, + "loss": 0.4225, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15559916198253632, + "step": 3705, + "valid_targets_mean": 2322.0, + "valid_targets_min": 1248 + }, + { + "epoch": 5.94551282051282, + "grad_norm": 0.682039055258773, + "learning_rate": 2.710208545236397e-06, + "loss": 0.4267, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16459454596042633, + "step": 3710, + "valid_targets_mean": 3524.4, + "valid_targets_min": 1393 + }, + { + "epoch": 5.953525641025641, + "grad_norm": 0.711753873522178, + "learning_rate": 2.670175639953929e-06, + "loss": 0.4145, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19619062542915344, + "step": 3715, + "valid_targets_mean": 3609.4, + "valid_targets_min": 2155 + }, + { + "epoch": 5.961538461538462, + "grad_norm": 0.6651298775299185, + "learning_rate": 2.6304194461976207e-06, + "loss": 0.4144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1850934624671936, + "step": 3720, + "valid_targets_mean": 3501.0, + "valid_targets_min": 2052 + }, + { + "epoch": 5.969551282051282, + "grad_norm": 0.676670763166187, + "learning_rate": 2.5909405987688896e-06, + "loss": 0.3919, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19190022349357605, + "step": 3725, + "valid_targets_mean": 3892.2, + "valid_targets_min": 1738 + }, + { + "epoch": 5.977564102564102, + "grad_norm": 0.770137323827435, + "learning_rate": 2.5517397280406565e-06, + "loss": 0.4356, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21874959766864777, + "step": 3730, + "valid_targets_mean": 3535.2, + "valid_targets_min": 1201 + }, + { + "epoch": 5.985576923076923, + "grad_norm": 0.6868383217634174, + "learning_rate": 2.512817459947312e-06, + "loss": 0.4249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19356504082679749, + "step": 3735, + "valid_targets_mean": 3605.0, + "valid_targets_min": 1703 + }, + { + "epoch": 5.993589743589744, + "grad_norm": 0.6647711475312037, + "learning_rate": 2.4741744159746618e-06, + "loss": 0.4011, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17853152751922607, + "step": 3740, + "valid_targets_mean": 3820.6, + "valid_targets_min": 1967 + }, + { + "epoch": 6.001602564102564, + "grad_norm": 0.6724973582003144, + "learning_rate": 2.435811213150079e-06, + "loss": 0.4217, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2188216745853424, + "step": 3745, + "valid_targets_mean": 4232.6, + "valid_targets_min": 1149 + }, + { + "epoch": 6.009615384615385, + "grad_norm": 0.6421125033326872, + "learning_rate": 2.3977284640325805e-06, + "loss": 0.3776, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1879805475473404, + "step": 3750, + "valid_targets_mean": 3535.6, + "valid_targets_min": 1255 + }, + { + "epoch": 6.017628205128205, + "grad_norm": 0.8342671557065863, + "learning_rate": 2.359926776703092e-06, + "loss": 0.4209, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17423753440380096, + "step": 3755, + "valid_targets_mean": 2113.4, + "valid_targets_min": 848 + }, + { + "epoch": 6.0256410256410255, + "grad_norm": 0.7563965603486668, + "learning_rate": 2.3224067547547357e-06, + "loss": 0.3876, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17095929384231567, + "step": 3760, + "valid_targets_mean": 3024.0, + "valid_targets_min": 1276 + }, + { + "epoch": 6.033653846153846, + "grad_norm": 0.6784015835078847, + "learning_rate": 2.2851689972831536e-06, + "loss": 0.412, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17948806285858154, + "step": 3765, + "valid_targets_mean": 3347.2, + "valid_targets_min": 1456 + }, + { + "epoch": 6.041666666666667, + "grad_norm": 0.6781808012849021, + "learning_rate": 2.248214098877002e-06, + "loss": 0.3807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22467254102230072, + "step": 3770, + "valid_targets_mean": 4056.4, + "valid_targets_min": 1885 + }, + { + "epoch": 6.049679487179487, + "grad_norm": 0.7366760832153734, + "learning_rate": 2.2115426496083958e-06, + "loss": 0.3989, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22732368111610413, + "step": 3775, + "valid_targets_mean": 3288.6, + "valid_targets_min": 1605 + }, + { + "epoch": 6.0576923076923075, + "grad_norm": 0.801043492145877, + "learning_rate": 2.175155235023536e-06, + "loss": 0.4132, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2163580358028412, + "step": 3780, + "valid_targets_mean": 3162.5, + "valid_targets_min": 1326 + }, + { + "epoch": 6.065705128205129, + "grad_norm": 0.703706419142761, + "learning_rate": 2.1390524361333355e-06, + "loss": 0.3877, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17650943994522095, + "step": 3785, + "valid_targets_mean": 3476.1, + "valid_targets_min": 1130 + }, + { + "epoch": 6.073717948717949, + "grad_norm": 0.7875867053477588, + "learning_rate": 2.1032348294041305e-06, + "loss": 0.4136, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1701028048992157, + "step": 3790, + "valid_targets_mean": 2707.0, + "valid_targets_min": 1362 + }, + { + "epoch": 6.081730769230769, + "grad_norm": 0.6733381000673836, + "learning_rate": 2.067702986748521e-06, + "loss": 0.3736, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1501770317554474, + "step": 3795, + "valid_targets_mean": 3571.0, + "valid_targets_min": 1989 + }, + { + "epoch": 6.089743589743589, + "grad_norm": 0.7099796374212306, + "learning_rate": 2.0324574755161764e-06, + "loss": 0.3749, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1544666886329651, + "step": 3800, + "valid_targets_mean": 2638.5, + "valid_targets_min": 1038 + }, + { + "epoch": 6.097756410256411, + "grad_norm": 0.6828029012457236, + "learning_rate": 1.9974988584848385e-06, + "loss": 0.3897, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21423310041427612, + "step": 3805, + "valid_targets_mean": 3630.1, + "valid_targets_min": 1303 + }, + { + "epoch": 6.105769230769231, + "grad_norm": 0.7504893323422612, + "learning_rate": 1.96282769385129e-06, + "loss": 0.4113, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1843530386686325, + "step": 3810, + "valid_targets_mean": 3000.9, + "valid_targets_min": 843 + }, + { + "epoch": 6.113782051282051, + "grad_norm": 0.7404384900203205, + "learning_rate": 1.9284445352224625e-06, + "loss": 0.4191, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1834920346736908, + "step": 3815, + "valid_targets_mean": 3260.9, + "valid_targets_min": 1771 + }, + { + "epoch": 6.121794871794871, + "grad_norm": 0.7481210705246947, + "learning_rate": 1.894349931606596e-06, + "loss": 0.3713, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17029988765716553, + "step": 3820, + "valid_targets_mean": 2760.6, + "valid_targets_min": 1499 + }, + { + "epoch": 6.1298076923076925, + "grad_norm": 0.6724932528999034, + "learning_rate": 1.8605444274044493e-06, + "loss": 0.3957, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1943618506193161, + "step": 3825, + "valid_targets_mean": 4008.2, + "valid_targets_min": 1426 + }, + { + "epoch": 6.137820512820513, + "grad_norm": 0.8144849901236596, + "learning_rate": 1.827028562400659e-06, + "loss": 0.3877, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17849445343017578, + "step": 3830, + "valid_targets_mean": 2744.8, + "valid_targets_min": 1629 + }, + { + "epoch": 6.145833333333333, + "grad_norm": 0.7275604869963993, + "learning_rate": 1.793802871755066e-06, + "loss": 0.3694, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14773519337177277, + "step": 3835, + "valid_targets_mean": 2291.6, + "valid_targets_min": 1343 + }, + { + "epoch": 6.153846153846154, + "grad_norm": 0.703371937759423, + "learning_rate": 1.760867885994202e-06, + "loss": 0.4413, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2328527867794037, + "step": 3840, + "valid_targets_mean": 3447.9, + "valid_targets_min": 1240 + }, + { + "epoch": 6.1618589743589745, + "grad_norm": 0.6884386699501874, + "learning_rate": 1.7282241310028047e-06, + "loss": 0.4076, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15755273401737213, + "step": 3845, + "valid_targets_mean": 3017.5, + "valid_targets_min": 1376 + }, + { + "epoch": 6.169871794871795, + "grad_norm": 0.7459410694033667, + "learning_rate": 1.6958721280154232e-06, + "loss": 0.3926, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20226158201694489, + "step": 3850, + "valid_targets_mean": 3149.0, + "valid_targets_min": 1446 + }, + { + "epoch": 6.177884615384615, + "grad_norm": 0.7475501528726831, + "learning_rate": 1.6638123936081085e-06, + "loss": 0.3997, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1955845206975937, + "step": 3855, + "valid_targets_mean": 2637.0, + "valid_targets_min": 1211 + }, + { + "epoch": 6.185897435897436, + "grad_norm": 0.7562060493300558, + "learning_rate": 1.6320454396901463e-06, + "loss": 0.3961, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2613011300563812, + "step": 3860, + "valid_targets_mean": 3303.9, + "valid_targets_min": 1643 + }, + { + "epoch": 6.193910256410256, + "grad_norm": 0.7684439719769421, + "learning_rate": 1.6005717734958914e-06, + "loss": 0.4289, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2163650095462799, + "step": 3865, + "valid_targets_mean": 3717.5, + "valid_targets_min": 857 + }, + { + "epoch": 6.201923076923077, + "grad_norm": 0.7071461956673178, + "learning_rate": 1.569391897576671e-06, + "loss": 0.4043, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2408888339996338, + "step": 3870, + "valid_targets_mean": 3808.2, + "valid_targets_min": 1934 + }, + { + "epoch": 6.209935897435898, + "grad_norm": 0.7189402668097614, + "learning_rate": 1.5385063097927533e-06, + "loss": 0.3779, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19687172770500183, + "step": 3875, + "valid_targets_mean": 3274.5, + "valid_targets_min": 1807 + }, + { + "epoch": 6.217948717948718, + "grad_norm": 0.7589661003974408, + "learning_rate": 1.5079155033054104e-06, + "loss": 0.3874, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17770133912563324, + "step": 3880, + "valid_targets_mean": 2759.2, + "valid_targets_min": 937 + }, + { + "epoch": 6.225961538461538, + "grad_norm": 0.7508953076752961, + "learning_rate": 1.4776199665690239e-06, + "loss": 0.4278, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20443370938301086, + "step": 3885, + "valid_targets_mean": 3022.6, + "valid_targets_min": 1543 + }, + { + "epoch": 6.233974358974359, + "grad_norm": 0.7463451530972014, + "learning_rate": 1.4476201833233084e-06, + "loss": 0.4093, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1829492449760437, + "step": 3890, + "valid_targets_mean": 3447.4, + "valid_targets_min": 1318 + }, + { + "epoch": 6.24198717948718, + "grad_norm": 0.6961007244867831, + "learning_rate": 1.4179166325855676e-06, + "loss": 0.4119, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18456020951271057, + "step": 3895, + "valid_targets_mean": 3653.0, + "valid_targets_min": 1334 + }, + { + "epoch": 6.25, + "grad_norm": 0.6776195289692457, + "learning_rate": 1.3885097886430599e-06, + "loss": 0.4037, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1324504315853119, + "step": 3900, + "valid_targets_mean": 2858.4, + "valid_targets_min": 1048 + }, + { + "epoch": 6.25801282051282, + "grad_norm": 0.8331224228804128, + "learning_rate": 1.35940012104542e-06, + "loss": 0.4288, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2164047211408615, + "step": 3905, + "valid_targets_mean": 2775.4, + "valid_targets_min": 1477 + }, + { + "epoch": 6.266025641025641, + "grad_norm": 0.6810115407410244, + "learning_rate": 1.3305880945971583e-06, + "loss": 0.3933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21648231148719788, + "step": 3910, + "valid_targets_mean": 3720.4, + "valid_targets_min": 1497 + }, + { + "epoch": 6.274038461538462, + "grad_norm": 0.7455051883127036, + "learning_rate": 1.3020741693502403e-06, + "loss": 0.3783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16055342555046082, + "step": 3915, + "valid_targets_mean": 3078.0, + "valid_targets_min": 1753 + }, + { + "epoch": 6.282051282051282, + "grad_norm": 0.6801506869686841, + "learning_rate": 1.27385880059675e-06, + "loss": 0.3999, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20525869727134705, + "step": 3920, + "valid_targets_mean": 4212.9, + "valid_targets_min": 3128 + }, + { + "epoch": 6.290064102564102, + "grad_norm": 0.6759846770961159, + "learning_rate": 1.245942438861607e-06, + "loss": 0.3803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18482154607772827, + "step": 3925, + "valid_targets_mean": 3070.4, + "valid_targets_min": 750 + }, + { + "epoch": 6.298076923076923, + "grad_norm": 0.6877356380491443, + "learning_rate": 1.2183255298953788e-06, + "loss": 0.3727, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14156484603881836, + "step": 3930, + "valid_targets_mean": 3444.5, + "valid_targets_min": 1002 + }, + { + "epoch": 6.306089743589744, + "grad_norm": 0.6790325407656655, + "learning_rate": 1.1910085146671645e-06, + "loss": 0.4265, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18031473457813263, + "step": 3935, + "valid_targets_mean": 3309.8, + "valid_targets_min": 1642 + }, + { + "epoch": 6.314102564102564, + "grad_norm": 0.6311879678368827, + "learning_rate": 1.1639918293575492e-06, + "loss": 0.4004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13983756303787231, + "step": 3940, + "valid_targets_mean": 3745.8, + "valid_targets_min": 1785 + }, + { + "epoch": 6.322115384615385, + "grad_norm": 0.6925389776120398, + "learning_rate": 1.1372759053516536e-06, + "loss": 0.3888, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18344758450984955, + "step": 3945, + "valid_targets_mean": 3588.8, + "valid_targets_min": 1497 + }, + { + "epoch": 6.330128205128205, + "grad_norm": 0.7652196351968208, + "learning_rate": 1.1108611692322157e-06, + "loss": 0.3717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17916885018348694, + "step": 3950, + "valid_targets_mean": 2705.9, + "valid_targets_min": 1142 + }, + { + "epoch": 6.3381410256410255, + "grad_norm": 0.7624725998178393, + "learning_rate": 1.0847480427728142e-06, + "loss": 0.3794, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15216022729873657, + "step": 3955, + "valid_targets_mean": 2712.1, + "valid_targets_min": 1354 + }, + { + "epoch": 6.346153846153846, + "grad_norm": 0.7126002830588238, + "learning_rate": 1.0589369429311125e-06, + "loss": 0.4025, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17520037293434143, + "step": 3960, + "valid_targets_mean": 3364.6, + "valid_targets_min": 1861 + }, + { + "epoch": 6.354166666666667, + "grad_norm": 0.8475402741421544, + "learning_rate": 1.0334282818422037e-06, + "loss": 0.4148, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18458299338817596, + "step": 3965, + "valid_targets_mean": 2706.2, + "valid_targets_min": 1297 + }, + { + "epoch": 6.362179487179487, + "grad_norm": 0.770326912986349, + "learning_rate": 1.008222466812041e-06, + "loss": 0.3937, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1886075735092163, + "step": 3970, + "valid_targets_mean": 3014.1, + "valid_targets_min": 2323 + }, + { + "epoch": 6.3701923076923075, + "grad_norm": 0.7559760166338257, + "learning_rate": 9.83319900310915e-07, + "loss": 0.3647, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.157384991645813, + "step": 3975, + "valid_targets_mean": 2676.9, + "valid_targets_min": 606 + }, + { + "epoch": 6.378205128205128, + "grad_norm": 0.7324023394507626, + "learning_rate": 9.587209799670495e-07, + "loss": 0.3931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16191917657852173, + "step": 3980, + "valid_targets_mean": 2999.0, + "valid_targets_min": 1001 + }, + { + "epoch": 6.386217948717949, + "grad_norm": 0.6471063628263564, + "learning_rate": 9.344260985602327e-07, + "loss": 0.3828, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18219345808029175, + "step": 3985, + "valid_targets_mean": 3464.9, + "valid_targets_min": 1111 + }, + { + "epoch": 6.394230769230769, + "grad_norm": 0.6753667970176015, + "learning_rate": 9.104356440155526e-07, + "loss": 0.3975, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2515168786048889, + "step": 3990, + "valid_targets_mean": 3609.6, + "valid_targets_min": 1639 + }, + { + "epoch": 6.402243589743589, + "grad_norm": 0.6643185181542197, + "learning_rate": 8.867499993972162e-07, + "loss": 0.4013, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16642878949642181, + "step": 3995, + "valid_targets_mean": 3342.1, + "valid_targets_min": 841 + }, + { + "epoch": 6.410256410256411, + "grad_norm": 0.7068887498043743, + "learning_rate": 8.633695429024058e-07, + "loss": 0.3851, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16288138926029205, + "step": 4000, + "valid_targets_mean": 3288.5, + "valid_targets_min": 2207 + }, + { + "epoch": 6.418269230769231, + "grad_norm": 0.6787090072741736, + "learning_rate": 8.402946478552732e-07, + "loss": 0.3947, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22672399878501892, + "step": 4005, + "valid_targets_mean": 4131.4, + "valid_targets_min": 1793 + }, + { + "epoch": 6.426282051282051, + "grad_norm": 0.7269205952380333, + "learning_rate": 8.175256827009392e-07, + "loss": 0.3696, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1724364310503006, + "step": 4010, + "valid_targets_mean": 3227.5, + "valid_targets_min": 1388 + }, + { + "epoch": 6.434294871794872, + "grad_norm": 0.7214579371010499, + "learning_rate": 7.95063010999646e-07, + "loss": 0.3904, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22047564387321472, + "step": 4015, + "valid_targets_mean": 3751.2, + "valid_targets_min": 1777 + }, + { + "epoch": 6.4423076923076925, + "grad_norm": 0.6964492357488418, + "learning_rate": 7.729069914209409e-07, + "loss": 0.3813, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22362810373306274, + "step": 4020, + "valid_targets_mean": 3862.2, + "valid_targets_min": 2420 + }, + { + "epoch": 6.450320512820513, + "grad_norm": 0.68628670955385, + "learning_rate": 7.510579777379345e-07, + "loss": 0.3946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1704484224319458, + "step": 4025, + "valid_targets_mean": 3376.9, + "valid_targets_min": 2440 + }, + { + "epoch": 6.458333333333333, + "grad_norm": 0.6857811263770821, + "learning_rate": 7.295163188216792e-07, + "loss": 0.4176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16381393373012543, + "step": 4030, + "valid_targets_mean": 3568.1, + "valid_targets_min": 1514 + }, + { + "epoch": 6.466346153846154, + "grad_norm": 0.7297573072558087, + "learning_rate": 7.08282358635568e-07, + "loss": 0.4402, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24479550123214722, + "step": 4035, + "valid_targets_mean": 3851.8, + "valid_targets_min": 1843 + }, + { + "epoch": 6.4743589743589745, + "grad_norm": 0.6755480606273508, + "learning_rate": 6.87356436229869e-07, + "loss": 0.3906, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1633291393518448, + "step": 4040, + "valid_targets_mean": 3708.0, + "valid_targets_min": 2294 + }, + { + "epoch": 6.482371794871795, + "grad_norm": 0.7233973877284747, + "learning_rate": 6.667388857362977e-07, + "loss": 0.3799, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1944011151790619, + "step": 4045, + "valid_targets_mean": 3464.2, + "valid_targets_min": 1360 + }, + { + "epoch": 6.490384615384615, + "grad_norm": 0.6807412653837345, + "learning_rate": 6.464300363626797e-07, + "loss": 0.3893, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20194770395755768, + "step": 4050, + "valid_targets_mean": 4048.5, + "valid_targets_min": 1966 + }, + { + "epoch": 6.498397435897436, + "grad_norm": 0.6231614994754106, + "learning_rate": 6.264302123877053e-07, + "loss": 0.3768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2024296224117279, + "step": 4055, + "valid_targets_mean": 4000.9, + "valid_targets_min": 2778 + }, + { + "epoch": 6.506410256410256, + "grad_norm": 0.6591033176896715, + "learning_rate": 6.067397331557412e-07, + "loss": 0.4053, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18981048464775085, + "step": 4060, + "valid_targets_mean": 4078.8, + "valid_targets_min": 1592 + }, + { + "epoch": 6.514423076923077, + "grad_norm": 0.8006962757881702, + "learning_rate": 5.873589130717405e-07, + "loss": 0.3862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1924881786108017, + "step": 4065, + "valid_targets_mean": 3103.1, + "valid_targets_min": 1366 + }, + { + "epoch": 6.522435897435898, + "grad_norm": 0.7514868984089133, + "learning_rate": 5.682880615962116e-07, + "loss": 0.4036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20322997868061066, + "step": 4070, + "valid_targets_mean": 3077.5, + "valid_targets_min": 1396 + }, + { + "epoch": 6.530448717948718, + "grad_norm": 0.6750224643279723, + "learning_rate": 5.495274832402841e-07, + "loss": 0.3654, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1637900024652481, + "step": 4075, + "valid_targets_mean": 3921.8, + "valid_targets_min": 1934 + }, + { + "epoch": 6.538461538461538, + "grad_norm": 0.6576171730869497, + "learning_rate": 5.310774775608529e-07, + "loss": 0.4087, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2523769736289978, + "step": 4080, + "valid_targets_mean": 4599.2, + "valid_targets_min": 1951 + }, + { + "epoch": 6.546474358974359, + "grad_norm": 0.8245129253655102, + "learning_rate": 5.129383391557751e-07, + "loss": 0.4054, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16760532557964325, + "step": 4085, + "valid_targets_mean": 2562.9, + "valid_targets_min": 1129 + }, + { + "epoch": 6.55448717948718, + "grad_norm": 0.700290634081726, + "learning_rate": 4.951103576591876e-07, + "loss": 0.4025, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18336446583271027, + "step": 4090, + "valid_targets_mean": 3294.5, + "valid_targets_min": 1729 + }, + { + "epoch": 6.5625, + "grad_norm": 0.7467548274381719, + "learning_rate": 4.7759381773687e-07, + "loss": 0.3995, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17399880290031433, + "step": 4095, + "valid_targets_mean": 2699.5, + "valid_targets_min": 1082 + }, + { + "epoch": 6.57051282051282, + "grad_norm": 0.6566259534958416, + "learning_rate": 4.6038899908170234e-07, + "loss": 0.3935, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12733715772628784, + "step": 4100, + "valid_targets_mean": 3105.4, + "valid_targets_min": 1622 + }, + { + "epoch": 6.578525641025641, + "grad_norm": 0.7751893319031072, + "learning_rate": 4.4349617640920164e-07, + "loss": 0.387, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23506727814674377, + "step": 4105, + "valid_targets_mean": 3360.1, + "valid_targets_min": 1158 + }, + { + "epoch": 6.586538461538462, + "grad_norm": 0.755720907024601, + "learning_rate": 4.2691561945312764e-07, + "loss": 0.4247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15511149168014526, + "step": 4110, + "valid_targets_mean": 2486.6, + "valid_targets_min": 1260 + }, + { + "epoch": 6.594551282051282, + "grad_norm": 0.7315585217907202, + "learning_rate": 4.106475929611886e-07, + "loss": 0.4019, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22886908054351807, + "step": 4115, + "valid_targets_mean": 3884.1, + "valid_targets_min": 1791 + }, + { + "epoch": 6.602564102564102, + "grad_norm": 0.6970936621798591, + "learning_rate": 3.9469235669080007e-07, + "loss": 0.3853, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1670103371143341, + "step": 4120, + "valid_targets_mean": 3127.8, + "valid_targets_min": 1762 + }, + { + "epoch": 6.610576923076923, + "grad_norm": 0.806524142823936, + "learning_rate": 3.7905016540495053e-07, + "loss": 0.3941, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1694411188364029, + "step": 4125, + "valid_targets_mean": 2750.0, + "valid_targets_min": 1419 + }, + { + "epoch": 6.618589743589744, + "grad_norm": 0.6836955718164335, + "learning_rate": 3.63721268868118e-07, + "loss": 0.3879, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16425782442092896, + "step": 4130, + "valid_targets_mean": 3494.2, + "valid_targets_min": 1125 + }, + { + "epoch": 6.626602564102564, + "grad_norm": 0.6743534060410575, + "learning_rate": 3.487059118422997e-07, + "loss": 0.3963, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21351942420005798, + "step": 4135, + "valid_targets_mean": 3498.5, + "valid_targets_min": 1239 + }, + { + "epoch": 6.634615384615385, + "grad_norm": 0.8085363296092258, + "learning_rate": 3.3400433408308895e-07, + "loss": 0.4244, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2399795949459076, + "step": 4140, + "valid_targets_mean": 2882.0, + "valid_targets_min": 1277 + }, + { + "epoch": 6.642628205128205, + "grad_norm": 0.7714581386901284, + "learning_rate": 3.196167703358577e-07, + "loss": 0.4013, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21508635580539703, + "step": 4145, + "valid_targets_mean": 2410.8, + "valid_targets_min": 1169 + }, + { + "epoch": 6.6506410256410255, + "grad_norm": 0.7718070962483563, + "learning_rate": 3.0554345033199985e-07, + "loss": 0.3891, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1982671320438385, + "step": 4150, + "valid_targets_mean": 2493.1, + "valid_targets_min": 1132 + }, + { + "epoch": 6.658653846153846, + "grad_norm": 0.7902947548147179, + "learning_rate": 2.917845987852652e-07, + "loss": 0.399, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1802562177181244, + "step": 4155, + "valid_targets_mean": 4092.2, + "valid_targets_min": 2150 + }, + { + "epoch": 6.666666666666667, + "grad_norm": 0.7892202600782438, + "learning_rate": 2.783404353881758e-07, + "loss": 0.3978, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.148856058716774, + "step": 4160, + "valid_targets_mean": 2618.0, + "valid_targets_min": 1489 + }, + { + "epoch": 6.674679487179487, + "grad_norm": 0.6515101224593849, + "learning_rate": 2.652111748085151e-07, + "loss": 0.3957, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2379886507987976, + "step": 4165, + "valid_targets_mean": 4368.2, + "valid_targets_min": 1703 + }, + { + "epoch": 6.6826923076923075, + "grad_norm": 0.7025288372677095, + "learning_rate": 2.523970266859044e-07, + "loss": 0.3914, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17233210802078247, + "step": 4170, + "valid_targets_mean": 2906.8, + "valid_targets_min": 1814 + }, + { + "epoch": 6.690705128205128, + "grad_norm": 0.7520472523237106, + "learning_rate": 2.398981956284363e-07, + "loss": 0.4228, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2593967914581299, + "step": 4175, + "valid_targets_mean": 3382.5, + "valid_targets_min": 2159 + }, + { + "epoch": 6.698717948717949, + "grad_norm": 0.709415220363152, + "learning_rate": 2.2771488120944207e-07, + "loss": 0.3807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22311896085739136, + "step": 4180, + "valid_targets_mean": 3894.2, + "valid_targets_min": 2024 + }, + { + "epoch": 6.706730769230769, + "grad_norm": 0.6574375138351698, + "learning_rate": 2.1584727796427174e-07, + "loss": 0.4204, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2038302719593048, + "step": 4185, + "valid_targets_mean": 4063.6, + "valid_targets_min": 1676 + }, + { + "epoch": 6.714743589743589, + "grad_norm": 0.6612862506540996, + "learning_rate": 2.0429557538720556e-07, + "loss": 0.399, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25420081615448, + "step": 4190, + "valid_targets_mean": 4168.0, + "valid_targets_min": 2179 + }, + { + "epoch": 6.722756410256411, + "grad_norm": 0.8546933379712878, + "learning_rate": 1.930599579284298e-07, + "loss": 0.4038, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2427080273628235, + "step": 4195, + "valid_targets_mean": 2907.1, + "valid_targets_min": 1381 + }, + { + "epoch": 6.730769230769231, + "grad_norm": 0.6585275995914507, + "learning_rate": 1.8214060499107679e-07, + "loss": 0.3882, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1734529286623001, + "step": 4200, + "valid_targets_mean": 4035.6, + "valid_targets_min": 2046 + }, + { + "epoch": 6.738782051282051, + "grad_norm": 0.7556929827396288, + "learning_rate": 1.7153769092837614e-07, + "loss": 0.386, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1649579405784607, + "step": 4205, + "valid_targets_mean": 2988.5, + "valid_targets_min": 1635 + }, + { + "epoch": 6.746794871794872, + "grad_norm": 0.7424334041336462, + "learning_rate": 1.6125138504086146e-07, + "loss": 0.3876, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1976759433746338, + "step": 4210, + "valid_targets_mean": 3453.6, + "valid_targets_min": 2056 + }, + { + "epoch": 6.7548076923076925, + "grad_norm": 0.8368074359783422, + "learning_rate": 1.5128185157367247e-07, + "loss": 0.3885, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2033853679895401, + "step": 4215, + "valid_targets_mean": 2341.9, + "valid_targets_min": 1205 + }, + { + "epoch": 6.762820512820513, + "grad_norm": 0.7483250380183626, + "learning_rate": 1.4162924971393044e-07, + "loss": 0.3976, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1938762068748474, + "step": 4220, + "valid_targets_mean": 2908.9, + "valid_targets_min": 1186 + }, + { + "epoch": 6.770833333333333, + "grad_norm": 0.8496445236000816, + "learning_rate": 1.322937335881891e-07, + "loss": 0.3858, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16094794869422913, + "step": 4225, + "valid_targets_mean": 2377.2, + "valid_targets_min": 1117 + }, + { + "epoch": 6.778846153846154, + "grad_norm": 0.7861985520056147, + "learning_rate": 1.2327545225999215e-07, + "loss": 0.392, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20814895629882812, + "step": 4230, + "valid_targets_mean": 3071.9, + "valid_targets_min": 1171 + }, + { + "epoch": 6.7868589743589745, + "grad_norm": 0.658518765440376, + "learning_rate": 1.145745497274664e-07, + "loss": 0.4144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22170805931091309, + "step": 4235, + "valid_targets_mean": 4468.6, + "valid_targets_min": 2283 + }, + { + "epoch": 6.794871794871795, + "grad_norm": 0.6996290654736144, + "learning_rate": 1.061911649210523e-07, + "loss": 0.3853, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24156290292739868, + "step": 4240, + "valid_targets_mean": 3769.2, + "valid_targets_min": 1106 + }, + { + "epoch": 6.802884615384615, + "grad_norm": 0.75020573344263, + "learning_rate": 9.812543170126365e-08, + "loss": 0.4517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18836921453475952, + "step": 4245, + "valid_targets_mean": 3356.6, + "valid_targets_min": 1324 + }, + { + "epoch": 6.810897435897436, + "grad_norm": 0.7057110537016175, + "learning_rate": 9.03774788565559e-08, + "loss": 0.408, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1661943793296814, + "step": 4250, + "valid_targets_mean": 3314.0, + "valid_targets_min": 1702 + }, + { + "epoch": 6.818910256410256, + "grad_norm": 0.7618906742510079, + "learning_rate": 8.294743010127448e-08, + "loss": 0.3969, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23056551814079285, + "step": 4255, + "valid_targets_mean": 3316.4, + "valid_targets_min": 1368 + }, + { + "epoch": 6.826923076923077, + "grad_norm": 0.7712745947369309, + "learning_rate": 7.583540407367418e-08, + "loss": 0.4147, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.262591689825058, + "step": 4260, + "valid_targets_mean": 3469.9, + "valid_targets_min": 1251 + }, + { + "epoch": 6.834935897435898, + "grad_norm": 0.7790046224298568, + "learning_rate": 6.904151433402728e-08, + "loss": 0.4047, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20210416615009308, + "step": 4265, + "valid_targets_mean": 3110.0, + "valid_targets_min": 1286 + }, + { + "epoch": 6.842948717948718, + "grad_norm": 0.7051872992061925, + "learning_rate": 6.256586936281172e-08, + "loss": 0.3987, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2403348833322525, + "step": 4270, + "valid_targets_mean": 3841.6, + "valid_targets_min": 1170 + }, + { + "epoch": 6.850961538461538, + "grad_norm": 0.7316428843704131, + "learning_rate": 5.6408572558972475e-08, + "loss": 0.4033, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2063056081533432, + "step": 4275, + "valid_targets_mean": 2832.6, + "valid_targets_min": 1156 + }, + { + "epoch": 6.858974358974359, + "grad_norm": 0.7364353488373431, + "learning_rate": 5.0569722238280605e-08, + "loss": 0.4235, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17579138278961182, + "step": 4280, + "valid_targets_mean": 3365.0, + "valid_targets_min": 2075 + }, + { + "epoch": 6.86698717948718, + "grad_norm": 0.7629127105534976, + "learning_rate": 4.504941163175236e-08, + "loss": 0.3707, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18533465266227722, + "step": 4285, + "valid_targets_mean": 3188.2, + "valid_targets_min": 1318 + }, + { + "epoch": 6.875, + "grad_norm": 0.7169236615885344, + "learning_rate": 3.984772888417032e-08, + "loss": 0.397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15749841928482056, + "step": 4290, + "valid_targets_mean": 3382.8, + "valid_targets_min": 1596 + }, + { + "epoch": 6.88301282051282, + "grad_norm": 0.7382909097644158, + "learning_rate": 3.4964757052671216e-08, + "loss": 0.3893, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18395844101905823, + "step": 4295, + "valid_targets_mean": 3502.5, + "valid_targets_min": 1417 + }, + { + "epoch": 6.891025641025641, + "grad_norm": 0.7414021394734007, + "learning_rate": 3.0400574105415856e-08, + "loss": 0.4174, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22825759649276733, + "step": 4300, + "valid_targets_mean": 3385.8, + "valid_targets_min": 1789 + }, + { + "epoch": 6.899038461538462, + "grad_norm": 0.7078117125768987, + "learning_rate": 2.615525292035459e-08, + "loss": 0.3741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20435480773448944, + "step": 4305, + "valid_targets_mean": 3789.8, + "valid_targets_min": 1731 + }, + { + "epoch": 6.907051282051282, + "grad_norm": 0.6824500541358194, + "learning_rate": 2.222886128405266e-08, + "loss": 0.4015, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15906572341918945, + "step": 4310, + "valid_targets_mean": 3111.2, + "valid_targets_min": 1941 + }, + { + "epoch": 6.915064102564102, + "grad_norm": 0.6606673657085689, + "learning_rate": 1.8621461890617752e-08, + "loss": 0.4098, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16551586985588074, + "step": 4315, + "valid_targets_mean": 3464.6, + "valid_targets_min": 1782 + }, + { + "epoch": 6.923076923076923, + "grad_norm": 0.6994563800894192, + "learning_rate": 1.5333112340687463e-08, + "loss": 0.3887, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1654829978942871, + "step": 4320, + "valid_targets_mean": 4049.9, + "valid_targets_min": 1998 + }, + { + "epoch": 6.931089743589744, + "grad_norm": 0.8091322982481507, + "learning_rate": 1.2363865140518905e-08, + "loss": 0.3959, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1469999998807907, + "step": 4325, + "valid_targets_mean": 2206.2, + "valid_targets_min": 1102 + }, + { + "epoch": 6.939102564102564, + "grad_norm": 0.6903938557496052, + "learning_rate": 9.713767701151621e-09, + "loss": 0.3725, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17113935947418213, + "step": 4330, + "valid_targets_mean": 3694.0, + "valid_targets_min": 2731 + }, + { + "epoch": 6.947115384615385, + "grad_norm": 0.74060855257773, + "learning_rate": 7.382862337641516e-09, + "loss": 0.3995, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16735701262950897, + "step": 4335, + "valid_targets_mean": 2850.9, + "valid_targets_min": 1159 + }, + { + "epoch": 6.955128205128205, + "grad_norm": 0.7435862559808525, + "learning_rate": 5.371186268390283e-09, + "loss": 0.3989, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.180478036403656, + "step": 4340, + "valid_targets_mean": 3129.8, + "valid_targets_min": 1130 + }, + { + "epoch": 6.9631410256410255, + "grad_norm": 0.6696426077893213, + "learning_rate": 3.678771614550325e-09, + "loss": 0.4011, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2068789303302765, + "step": 4345, + "valid_targets_mean": 3965.0, + "valid_targets_min": 1890 + }, + { + "epoch": 6.971153846153846, + "grad_norm": 0.7403246762511096, + "learning_rate": 2.3056453995162763e-09, + "loss": 0.3919, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22071704268455505, + "step": 4350, + "valid_targets_mean": 3568.9, + "valid_targets_min": 1634 + }, + { + "epoch": 6.979166666666667, + "grad_norm": 0.6814469462606922, + "learning_rate": 1.2518295484875708e-09, + "loss": 0.4062, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.194525808095932, + "step": 4355, + "valid_targets_mean": 3148.6, + "valid_targets_min": 1569 + }, + { + "epoch": 6.987179487179487, + "grad_norm": 0.7683617478158435, + "learning_rate": 5.173408881198328e-10, + "loss": 0.4064, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2039300501346588, + "step": 4360, + "valid_targets_mean": 2865.8, + "valid_targets_min": 1173 + }, + { + "epoch": 6.9951923076923075, + "grad_norm": 0.6301277572038875, + "learning_rate": 1.0219114625398263e-10, + "loss": 0.4175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12411817908287048, + "step": 4365, + "valid_targets_mean": 2867.1, + "valid_targets_min": 2086 + }, + { + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18683922290802002, + "step": 4368, + "total_flos": 1.0544849440651346e+18, + "train_loss": 0.4722494138848214, + "train_runtime": 29349.8434, + "train_samples_per_second": 2.381, + "train_steps_per_second": 0.149, + "valid_targets_mean": 3776.8, + "valid_targets_min": 2389 + } + ], + "logging_steps": 5, + "max_steps": 4368, + "num_input_tokens_seen": 0, + "num_train_epochs": 7, + "save_steps": 1500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 1.0544849440651346e+18, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..2898bb0 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d330997b455901ea10dee2e751aa436c1c61073d9148d3e0a98f4d21f95d1226 +size 8657 diff --git a/training_loss.png b/training_loss.png new file mode 100644 index 0000000..02b0abb Binary files /dev/null and b/training_loss.png differ diff --git a/vocab.json b/vocab.json new file mode 100644 index 0000000..6c49fc6 --- /dev/null +++ b/vocab.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910 +size 2776833