commit 14752ce038124182443861254fe7d5828421193b Author: ModelHub XC Date: Tue Sep 8 23:32:13 2026 +0800 初始化项目,由ModelHub XC社区提供模型 Model: laion/glm46-swesmith-maxeps-131k Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..99c35af --- /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 + +merges.txt filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text +training_args.bin filter=lfs diff=lfs merge=lfs -text +model-00004-of-00004.safetensors 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 +vocab.json filter=lfs diff=lfs merge=lfs -text +model-00001-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..9190df1 --- /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: glm46-swesmith-maxeps-131k + results: [] +--- + + + +# glm46-swesmith-maxeps-131k + +This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on the DCAgent2/glm46-neulab-mind2web-sandboxes-maxeps-131k 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 0.22.1 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..7650f3b --- /dev/null +++ b/all_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 4.020718520229712, + "achieved_tflops_per_gpu_theoretical": 257.48074677456077, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15556100010871887, + "mfu_percent": 0.40654383419916196, + "mfu_percent_theoretical": 26.0344536678019, + "total_flos": 1.820182243006546e+18, + "train_loss": 0.2264533686425352, + "train_runtime": 56587.5923, + "train_samples_per_second": 1.212, + "train_steps_per_second": 0.076, + "valid_targets_mean": 4983.8, + "valid_targets_min": 1127 +} \ 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..0474cba --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:113bfd734751b49a280a93b863a35e41de6856256e96935ac6091ed23a986833 +size 4902257696 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..2a63732 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aab45028cf72da67767747ed3e3b65b4713e4128e30bc6271bbd72af467f48aa +size 4915960368 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..c661513 --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e42328c6c82bca33ba7925e7de625b78bfb1539fc16959fe1ef03b46a1151b63 +size 4983068496 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..664e728 --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ca507d6e1274f2b37b32d7cbcc0a8836caad9d2f6e608f8d39384b27c88fe93 +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..2d65063 --- /dev/null +++ b/run_summary.json @@ -0,0 +1,12 @@ +{ + "agent_name": null, + "training_start": null, + "training_end": null, + "created_by": "DCAgent", + "base_model_name": "Qwen/Qwen3-8B", + "dataset_name": "DCAgent2/glm46-neulab-mind2web-sandboxes-maxeps-131k", + "training_type": "SFT", + "training_parameters": "https://huggingface.co/laion/glm46-swesmith-maxeps-131k/blob/main/config.json", + "wandb_link": "https://wandb.ai/DCAgent/OpenThoughts-Agent/runs/glm46-neulab-mind2web-sandboxes-maxeps-131k_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..7650f3b --- /dev/null +++ b/train_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 4.020718520229712, + "achieved_tflops_per_gpu_theoretical": 257.48074677456077, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15556100010871887, + "mfu_percent": 0.40654383419916196, + "mfu_percent_theoretical": 26.0344536678019, + "total_flos": 1.820182243006546e+18, + "train_loss": 0.2264533686425352, + "train_runtime": 56587.5923, + "train_samples_per_second": 1.212, + "train_steps_per_second": 0.076, + "valid_targets_mean": 4983.8, + "valid_targets_min": 1127 +} \ No newline at end of file diff --git a/trainer_log.jsonl b/trainer_log.jsonl new file mode 100644 index 0000000..48a52f8 --- /dev/null +++ b/trainer_log.jsonl @@ -0,0 +1,859 @@ +{"current_steps": 5, "total_steps": 4291, "loss": 0.75, "lr": 3.7209302325581396e-07, "epoch": 0.00816326530612245, "percentage": 0.12, "elapsed_time": "0:01:06", "remaining_time": "15:45:08"} +{"current_steps": 10, "total_steps": 4291, "loss": 0.7413, "lr": 8.372093023255814e-07, "epoch": 0.0163265306122449, "percentage": 0.23, "elapsed_time": "0:02:11", "remaining_time": "15:37:54"} +{"current_steps": 15, "total_steps": 4291, "loss": 0.7263, "lr": 1.302325581395349e-06, "epoch": 0.024489795918367346, "percentage": 0.35, "elapsed_time": "0:03:17", "remaining_time": "15:40:38"} +{"current_steps": 20, "total_steps": 4291, "loss": 0.7189, "lr": 1.7674418604651164e-06, "epoch": 0.0326530612244898, "percentage": 0.47, "elapsed_time": "0:04:23", "remaining_time": "15:36:50"} +{"current_steps": 25, "total_steps": 4291, "loss": 0.6782, "lr": 2.232558139534884e-06, "epoch": 0.04081632653061224, "percentage": 0.58, "elapsed_time": "0:05:30", "remaining_time": "15:39:31"} +{"current_steps": 30, "total_steps": 4291, "loss": 0.5857, "lr": 2.6976744186046517e-06, "epoch": 0.04897959183673469, "percentage": 0.7, "elapsed_time": "0:06:41", "remaining_time": "15:50:45"} +{"current_steps": 35, "total_steps": 4291, "loss": 0.5584, "lr": 3.1627906976744187e-06, "epoch": 0.05714285714285714, "percentage": 0.82, "elapsed_time": "0:07:45", "remaining_time": "15:43:13"} +{"current_steps": 40, "total_steps": 4291, "loss": 0.5306, "lr": 3.6279069767441866e-06, "epoch": 0.0653061224489796, "percentage": 0.93, "elapsed_time": "0:08:58", "remaining_time": "15:53:53"} +{"current_steps": 45, "total_steps": 4291, "loss": 0.5073, "lr": 4.0930232558139536e-06, "epoch": 0.07346938775510205, "percentage": 1.05, "elapsed_time": "0:10:04", "remaining_time": "15:50:18"} +{"current_steps": 50, "total_steps": 4291, "loss": 0.463, "lr": 4.558139534883721e-06, "epoch": 0.08163265306122448, "percentage": 1.17, "elapsed_time": "0:11:21", "remaining_time": "16:02:49"} +{"current_steps": 55, "total_steps": 4291, "loss": 0.4861, "lr": 5.023255813953489e-06, "epoch": 0.08979591836734693, "percentage": 1.28, "elapsed_time": "0:12:22", "remaining_time": "15:53:44"} +{"current_steps": 60, "total_steps": 4291, "loss": 0.4585, "lr": 5.488372093023256e-06, "epoch": 0.09795918367346938, "percentage": 1.4, "elapsed_time": "0:13:29", "remaining_time": "15:51:05"} +{"current_steps": 65, "total_steps": 4291, "loss": 0.4672, "lr": 5.953488372093023e-06, "epoch": 0.10612244897959183, "percentage": 1.51, "elapsed_time": "0:14:34", "remaining_time": "15:47:54"} +{"current_steps": 70, "total_steps": 4291, "loss": 0.44, "lr": 6.418604651162791e-06, "epoch": 0.11428571428571428, "percentage": 1.63, "elapsed_time": "0:15:44", "remaining_time": "15:49:41"} +{"current_steps": 75, "total_steps": 4291, "loss": 0.4608, "lr": 6.883720930232559e-06, "epoch": 0.12244897959183673, "percentage": 1.75, "elapsed_time": "0:16:46", "remaining_time": "15:42:53"} +{"current_steps": 80, "total_steps": 4291, "loss": 0.4456, "lr": 7.348837209302326e-06, "epoch": 0.1306122448979592, "percentage": 1.86, "elapsed_time": "0:17:47", "remaining_time": "15:36:27"} +{"current_steps": 85, "total_steps": 4291, "loss": 0.4337, "lr": 7.813953488372094e-06, "epoch": 0.13877551020408163, "percentage": 1.98, "elapsed_time": "0:18:47", "remaining_time": "15:30:12"} +{"current_steps": 90, "total_steps": 4291, "loss": 0.3833, "lr": 8.279069767441861e-06, "epoch": 0.1469387755102041, "percentage": 2.1, "elapsed_time": "0:19:54", "remaining_time": "15:29:32"} +{"current_steps": 95, "total_steps": 4291, "loss": 0.4174, "lr": 8.74418604651163e-06, "epoch": 0.15510204081632653, "percentage": 2.21, "elapsed_time": "0:20:57", "remaining_time": "15:25:43"} +{"current_steps": 100, "total_steps": 4291, "loss": 0.4086, "lr": 9.209302325581397e-06, "epoch": 0.16326530612244897, "percentage": 2.33, "elapsed_time": "0:22:07", "remaining_time": "15:26:55"} +{"current_steps": 105, "total_steps": 4291, "loss": 0.3998, "lr": 9.674418604651164e-06, "epoch": 0.17142857142857143, "percentage": 2.45, "elapsed_time": "0:23:03", "remaining_time": "15:19:17"} +{"current_steps": 110, "total_steps": 4291, "loss": 0.3843, "lr": 1.0139534883720932e-05, "epoch": 0.17959183673469387, "percentage": 2.56, "elapsed_time": "0:24:14", "remaining_time": "15:21:07"} +{"current_steps": 115, "total_steps": 4291, "loss": 0.3954, "lr": 1.0604651162790698e-05, "epoch": 0.18775510204081633, "percentage": 2.68, "elapsed_time": "0:25:23", "remaining_time": "15:22:06"} +{"current_steps": 120, "total_steps": 4291, "loss": 0.3875, "lr": 1.1069767441860466e-05, "epoch": 0.19591836734693877, "percentage": 2.8, "elapsed_time": "0:26:23", "remaining_time": "15:17:02"} +{"current_steps": 125, "total_steps": 4291, "loss": 0.3829, "lr": 1.1534883720930235e-05, "epoch": 0.20408163265306123, "percentage": 2.91, "elapsed_time": "0:27:31", "remaining_time": "15:17:24"} +{"current_steps": 130, "total_steps": 4291, "loss": 0.3573, "lr": 1.2e-05, "epoch": 0.21224489795918366, "percentage": 3.03, "elapsed_time": "0:28:37", "remaining_time": "15:16:01"} +{"current_steps": 135, "total_steps": 4291, "loss": 0.3791, "lr": 1.2465116279069769e-05, "epoch": 0.22040816326530613, "percentage": 3.15, "elapsed_time": "0:29:44", "remaining_time": "15:15:47"} +{"current_steps": 140, "total_steps": 4291, "loss": 0.3689, "lr": 1.2930232558139534e-05, "epoch": 0.22857142857142856, "percentage": 3.26, "elapsed_time": "0:30:55", "remaining_time": "15:16:56"} +{"current_steps": 145, "total_steps": 4291, "loss": 0.3841, "lr": 1.3395348837209303e-05, "epoch": 0.23673469387755103, "percentage": 3.38, "elapsed_time": "0:31:55", "remaining_time": "15:12:41"} +{"current_steps": 150, "total_steps": 4291, "loss": 0.3655, "lr": 1.3860465116279072e-05, "epoch": 0.24489795918367346, "percentage": 3.5, "elapsed_time": "0:32:59", "remaining_time": "15:10:57"} +{"current_steps": 155, "total_steps": 4291, "loss": 0.3682, "lr": 1.4325581395348837e-05, "epoch": 0.2530612244897959, "percentage": 3.61, "elapsed_time": "0:34:00", "remaining_time": "15:07:27"} +{"current_steps": 160, "total_steps": 4291, "loss": 0.3562, "lr": 1.4790697674418606e-05, "epoch": 0.2612244897959184, "percentage": 3.73, "elapsed_time": "0:35:01", "remaining_time": "15:04:20"} +{"current_steps": 165, "total_steps": 4291, "loss": 0.3692, "lr": 1.5255813953488374e-05, "epoch": 0.2693877551020408, "percentage": 3.85, "elapsed_time": "0:36:06", "remaining_time": "15:03:04"} +{"current_steps": 170, "total_steps": 4291, "loss": 0.3689, "lr": 1.572093023255814e-05, "epoch": 0.27755102040816326, "percentage": 3.96, "elapsed_time": "0:37:10", "remaining_time": "15:01:11"} +{"current_steps": 175, "total_steps": 4291, "loss": 0.3431, "lr": 1.618604651162791e-05, "epoch": 0.2857142857142857, "percentage": 4.08, "elapsed_time": "0:38:12", "remaining_time": "14:58:31"} +{"current_steps": 180, "total_steps": 4291, "loss": 0.3154, "lr": 1.6651162790697674e-05, "epoch": 0.2938775510204082, "percentage": 4.19, "elapsed_time": "0:39:26", "remaining_time": "15:00:54"} +{"current_steps": 185, "total_steps": 4291, "loss": 0.3333, "lr": 1.7116279069767444e-05, "epoch": 0.3020408163265306, "percentage": 4.31, "elapsed_time": "0:40:27", "remaining_time": "14:57:53"} +{"current_steps": 190, "total_steps": 4291, "loss": 0.3352, "lr": 1.758139534883721e-05, "epoch": 0.31020408163265306, "percentage": 4.43, "elapsed_time": "0:41:36", "remaining_time": "14:58:08"} +{"current_steps": 195, "total_steps": 4291, "loss": 0.345, "lr": 1.8046511627906978e-05, "epoch": 0.3183673469387755, "percentage": 4.54, "elapsed_time": "0:42:41", "remaining_time": "14:56:35"} +{"current_steps": 200, "total_steps": 4291, "loss": 0.3586, "lr": 1.8511627906976745e-05, "epoch": 0.32653061224489793, "percentage": 4.66, "elapsed_time": "0:43:43", "remaining_time": "14:54:31"} +{"current_steps": 205, "total_steps": 4291, "loss": 0.3368, "lr": 1.8976744186046516e-05, "epoch": 0.3346938775510204, "percentage": 4.78, "elapsed_time": "0:45:27", "remaining_time": "15:05:57"} +{"current_steps": 210, "total_steps": 4291, "loss": 0.3404, "lr": 1.944186046511628e-05, "epoch": 0.34285714285714286, "percentage": 4.89, "elapsed_time": "0:46:33", "remaining_time": "15:04:37"} +{"current_steps": 215, "total_steps": 4291, "loss": 0.3455, "lr": 1.990697674418605e-05, "epoch": 0.3510204081632653, "percentage": 5.01, "elapsed_time": "0:47:36", "remaining_time": "15:02:30"} +{"current_steps": 220, "total_steps": 4291, "loss": 0.3256, "lr": 2.0372093023255813e-05, "epoch": 0.35918367346938773, "percentage": 5.13, "elapsed_time": "0:48:37", "remaining_time": "14:59:54"} +{"current_steps": 225, "total_steps": 4291, "loss": 0.2986, "lr": 2.0837209302325584e-05, "epoch": 0.3673469387755102, "percentage": 5.24, "elapsed_time": "0:49:48", "remaining_time": "15:00:02"} +{"current_steps": 230, "total_steps": 4291, "loss": 0.3275, "lr": 2.130232558139535e-05, "epoch": 0.37551020408163266, "percentage": 5.36, "elapsed_time": "0:50:43", "remaining_time": "14:55:44"} +{"current_steps": 235, "total_steps": 4291, "loss": 0.3489, "lr": 2.176744186046512e-05, "epoch": 0.3836734693877551, "percentage": 5.48, "elapsed_time": "0:51:40", "remaining_time": "14:52:01"} +{"current_steps": 240, "total_steps": 4291, "loss": 0.333, "lr": 2.2232558139534888e-05, "epoch": 0.39183673469387753, "percentage": 5.59, "elapsed_time": "0:52:38", "remaining_time": "14:48:33"} +{"current_steps": 245, "total_steps": 4291, "loss": 0.3198, "lr": 2.269767441860465e-05, "epoch": 0.4, "percentage": 5.71, "elapsed_time": "0:53:47", "remaining_time": "14:48:20"} +{"current_steps": 250, "total_steps": 4291, "loss": 0.3368, "lr": 2.316279069767442e-05, "epoch": 0.40816326530612246, "percentage": 5.83, "elapsed_time": "0:54:56", "remaining_time": "14:48:02"} +{"current_steps": 255, "total_steps": 4291, "loss": 0.3345, "lr": 2.362790697674419e-05, "epoch": 0.4163265306122449, "percentage": 5.94, "elapsed_time": "0:56:04", "remaining_time": "14:47:32"} +{"current_steps": 260, "total_steps": 4291, "loss": 0.3444, "lr": 2.4093023255813956e-05, "epoch": 0.42448979591836733, "percentage": 6.06, "elapsed_time": "0:57:01", "remaining_time": "14:44:11"} +{"current_steps": 265, "total_steps": 4291, "loss": 0.3311, "lr": 2.4558139534883726e-05, "epoch": 0.4326530612244898, "percentage": 6.18, "elapsed_time": "0:58:12", "remaining_time": "14:44:15"} +{"current_steps": 270, "total_steps": 4291, "loss": 0.3395, "lr": 2.502325581395349e-05, "epoch": 0.44081632653061226, "percentage": 6.29, "elapsed_time": "0:59:15", "remaining_time": "14:42:27"} +{"current_steps": 275, "total_steps": 4291, "loss": 0.3273, "lr": 2.5488372093023257e-05, "epoch": 0.4489795918367347, "percentage": 6.41, "elapsed_time": "1:00:15", "remaining_time": "14:39:58"} +{"current_steps": 280, "total_steps": 4291, "loss": 0.3111, "lr": 2.5953488372093024e-05, "epoch": 0.45714285714285713, "percentage": 6.53, "elapsed_time": "1:01:17", "remaining_time": "14:37:59"} +{"current_steps": 285, "total_steps": 4291, "loss": 0.2959, "lr": 2.6418604651162794e-05, "epoch": 0.46530612244897956, "percentage": 6.64, "elapsed_time": "1:02:29", "remaining_time": "14:38:17"} +{"current_steps": 290, "total_steps": 4291, "loss": 0.3089, "lr": 2.688372093023256e-05, "epoch": 0.47346938775510206, "percentage": 6.76, "elapsed_time": "1:03:42", "remaining_time": "14:38:55"} +{"current_steps": 295, "total_steps": 4291, "loss": 0.3217, "lr": 2.7348837209302325e-05, "epoch": 0.4816326530612245, "percentage": 6.87, "elapsed_time": "1:04:51", "remaining_time": "14:38:28"} +{"current_steps": 300, "total_steps": 4291, "loss": 0.3085, "lr": 2.7813953488372095e-05, "epoch": 0.4897959183673469, "percentage": 6.99, "elapsed_time": "1:06:00", "remaining_time": "14:38:12"} +{"current_steps": 305, "total_steps": 4291, "loss": 0.3217, "lr": 2.8279069767441862e-05, "epoch": 0.49795918367346936, "percentage": 7.11, "elapsed_time": "1:06:59", "remaining_time": "14:35:24"} +{"current_steps": 310, "total_steps": 4291, "loss": 0.3227, "lr": 2.874418604651163e-05, "epoch": 0.5061224489795918, "percentage": 7.22, "elapsed_time": "1:07:53", "remaining_time": "14:31:56"} +{"current_steps": 315, "total_steps": 4291, "loss": 0.3193, "lr": 2.92093023255814e-05, "epoch": 0.5142857142857142, "percentage": 7.34, "elapsed_time": "1:08:47", "remaining_time": "14:28:12"} +{"current_steps": 320, "total_steps": 4291, "loss": 0.3192, "lr": 2.9674418604651167e-05, "epoch": 0.5224489795918368, "percentage": 7.46, "elapsed_time": "1:09:53", "remaining_time": "14:27:14"} +{"current_steps": 325, "total_steps": 4291, "loss": 0.3121, "lr": 3.013953488372093e-05, "epoch": 0.5306122448979592, "percentage": 7.57, "elapsed_time": "1:10:53", "remaining_time": "14:25:08"} +{"current_steps": 330, "total_steps": 4291, "loss": 0.3149, "lr": 3.06046511627907e-05, "epoch": 0.5387755102040817, "percentage": 7.69, "elapsed_time": "1:11:57", "remaining_time": "14:23:41"} +{"current_steps": 335, "total_steps": 4291, "loss": 0.3143, "lr": 3.1069767441860465e-05, "epoch": 0.5469387755102041, "percentage": 7.81, "elapsed_time": "1:12:59", "remaining_time": "14:21:51"} +{"current_steps": 340, "total_steps": 4291, "loss": 0.3014, "lr": 3.1534883720930235e-05, "epoch": 0.5551020408163265, "percentage": 7.92, "elapsed_time": "1:13:58", "remaining_time": "14:19:42"} +{"current_steps": 345, "total_steps": 4291, "loss": 0.313, "lr": 3.2000000000000005e-05, "epoch": 0.563265306122449, "percentage": 8.04, "elapsed_time": "1:15:07", "remaining_time": "14:19:14"} +{"current_steps": 350, "total_steps": 4291, "loss": 0.2967, "lr": 3.246511627906977e-05, "epoch": 0.5714285714285714, "percentage": 8.16, "elapsed_time": "1:16:18", "remaining_time": "14:19:09"} +{"current_steps": 355, "total_steps": 4291, "loss": 0.3015, "lr": 3.293023255813954e-05, "epoch": 0.5795918367346938, "percentage": 8.27, "elapsed_time": "1:17:22", "remaining_time": "14:17:50"} +{"current_steps": 360, "total_steps": 4291, "loss": 0.3067, "lr": 3.33953488372093e-05, "epoch": 0.5877551020408164, "percentage": 8.39, "elapsed_time": "1:18:33", "remaining_time": "14:17:52"} +{"current_steps": 365, "total_steps": 4291, "loss": 0.3107, "lr": 3.386046511627907e-05, "epoch": 0.5959183673469388, "percentage": 8.51, "elapsed_time": "1:19:36", "remaining_time": "14:16:14"} +{"current_steps": 370, "total_steps": 4291, "loss": 0.2927, "lr": 3.4325581395348844e-05, "epoch": 0.6040816326530613, "percentage": 8.62, "elapsed_time": "1:20:50", "remaining_time": "14:16:44"} +{"current_steps": 375, "total_steps": 4291, "loss": 0.306, "lr": 3.479069767441861e-05, "epoch": 0.6122448979591837, "percentage": 8.74, "elapsed_time": "1:21:59", "remaining_time": "14:16:12"} +{"current_steps": 380, "total_steps": 4291, "loss": 0.3136, "lr": 3.525581395348837e-05, "epoch": 0.6204081632653061, "percentage": 8.86, "elapsed_time": "1:22:58", "remaining_time": "14:14:01"} +{"current_steps": 385, "total_steps": 4291, "loss": 0.2908, "lr": 3.572093023255814e-05, "epoch": 0.6285714285714286, "percentage": 8.97, "elapsed_time": "1:24:10", "remaining_time": "14:14:00"} +{"current_steps": 390, "total_steps": 4291, "loss": 0.3002, "lr": 3.618604651162791e-05, "epoch": 0.636734693877551, "percentage": 9.09, "elapsed_time": "1:25:19", "remaining_time": "14:13:25"} +{"current_steps": 395, "total_steps": 4291, "loss": 0.3036, "lr": 3.6651162790697675e-05, "epoch": 0.6448979591836734, "percentage": 9.21, "elapsed_time": "1:26:23", "remaining_time": "14:12:04"} +{"current_steps": 400, "total_steps": 4291, "loss": 0.2896, "lr": 3.7116279069767446e-05, "epoch": 0.6530612244897959, "percentage": 9.32, "elapsed_time": "1:27:29", "remaining_time": "14:11:08"} +{"current_steps": 405, "total_steps": 4291, "loss": 0.3105, "lr": 3.758139534883721e-05, "epoch": 0.6612244897959184, "percentage": 9.44, "elapsed_time": "1:29:19", "remaining_time": "14:17:06"} +{"current_steps": 410, "total_steps": 4291, "loss": 0.3186, "lr": 3.804651162790698e-05, "epoch": 0.6693877551020408, "percentage": 9.55, "elapsed_time": "1:30:16", "remaining_time": "14:14:27"} +{"current_steps": 415, "total_steps": 4291, "loss": 0.29, "lr": 3.851162790697675e-05, "epoch": 0.6775510204081633, "percentage": 9.67, "elapsed_time": "1:31:20", "remaining_time": "14:13:05"} +{"current_steps": 420, "total_steps": 4291, "loss": 0.2986, "lr": 3.8976744186046514e-05, "epoch": 0.6857142857142857, "percentage": 9.79, "elapsed_time": "1:32:27", "remaining_time": "14:12:11"} +{"current_steps": 425, "total_steps": 4291, "loss": 0.2987, "lr": 3.9441860465116284e-05, "epoch": 0.6938775510204082, "percentage": 9.9, "elapsed_time": "1:33:31", "remaining_time": "14:10:41"} +{"current_steps": 430, "total_steps": 4291, "loss": 0.2894, "lr": 3.990697674418605e-05, "epoch": 0.7020408163265306, "percentage": 10.02, "elapsed_time": "1:34:28", "remaining_time": "14:08:22"} +{"current_steps": 435, "total_steps": 4291, "loss": 0.3196, "lr": 3.99998940698124e-05, "epoch": 0.710204081632653, "percentage": 10.14, "elapsed_time": "1:35:29", "remaining_time": "14:06:28"} +{"current_steps": 440, "total_steps": 4291, "loss": 0.2733, "lr": 3.999946373034843e-05, "epoch": 0.7183673469387755, "percentage": 10.25, "elapsed_time": "1:36:43", "remaining_time": "14:06:30"} +{"current_steps": 445, "total_steps": 4291, "loss": 0.2884, "lr": 3.999870236808868e-05, "epoch": 0.726530612244898, "percentage": 10.37, "elapsed_time": "1:37:59", "remaining_time": "14:06:53"} +{"current_steps": 450, "total_steps": 4291, "loss": 0.2935, "lr": 3.999760999563491e-05, "epoch": 0.7346938775510204, "percentage": 10.49, "elapsed_time": "1:39:10", "remaining_time": "14:06:26"} +{"current_steps": 455, "total_steps": 4291, "loss": 0.2933, "lr": 3.999618663106761e-05, "epoch": 0.7428571428571429, "percentage": 10.6, "elapsed_time": "1:40:03", "remaining_time": "14:03:30"} +{"current_steps": 460, "total_steps": 4291, "loss": 0.2951, "lr": 3.999443229794572e-05, "epoch": 0.7510204081632653, "percentage": 10.72, "elapsed_time": "1:41:03", "remaining_time": "14:01:38"} +{"current_steps": 465, "total_steps": 4291, "loss": 0.2931, "lr": 3.9992347025306236e-05, "epoch": 0.7591836734693878, "percentage": 10.84, "elapsed_time": "1:42:03", "remaining_time": "13:59:47"} +{"current_steps": 470, "total_steps": 4291, "loss": 0.2778, "lr": 3.9989930847663706e-05, "epoch": 0.7673469387755102, "percentage": 10.95, "elapsed_time": "1:43:15", "remaining_time": "13:59:30"} +{"current_steps": 475, "total_steps": 4291, "loss": 0.292, "lr": 3.998718380500971e-05, "epoch": 0.7755102040816326, "percentage": 11.07, "elapsed_time": "1:44:16", "remaining_time": "13:57:44"} +{"current_steps": 480, "total_steps": 4291, "loss": 0.2923, "lr": 3.998410594281215e-05, "epoch": 0.7836734693877551, "percentage": 11.19, "elapsed_time": "1:45:21", "remaining_time": "13:56:31"} +{"current_steps": 485, "total_steps": 4291, "loss": 0.2956, "lr": 3.998069731201452e-05, "epoch": 0.7918367346938775, "percentage": 11.3, "elapsed_time": "1:46:30", "remaining_time": "13:55:47"} +{"current_steps": 490, "total_steps": 4291, "loss": 0.2764, "lr": 3.997695796903506e-05, "epoch": 0.8, "percentage": 11.42, "elapsed_time": "1:47:46", "remaining_time": "13:55:59"} +{"current_steps": 495, "total_steps": 4291, "loss": 0.2857, "lr": 3.997288797576581e-05, "epoch": 0.8081632653061225, "percentage": 11.54, "elapsed_time": "1:48:58", "remaining_time": "13:55:41"} +{"current_steps": 500, "total_steps": 4291, "loss": 0.2943, "lr": 3.996848739957162e-05, "epoch": 0.8163265306122449, "percentage": 11.65, "elapsed_time": "1:49:51", "remaining_time": "13:52:56"} +{"current_steps": 505, "total_steps": 4291, "loss": 0.3024, "lr": 3.996375631328898e-05, "epoch": 0.8244897959183674, "percentage": 11.77, "elapsed_time": "1:50:59", "remaining_time": "13:52:08"} +{"current_steps": 510, "total_steps": 4291, "loss": 0.3079, "lr": 3.995869479522486e-05, "epoch": 0.8326530612244898, "percentage": 11.89, "elapsed_time": "1:51:56", "remaining_time": "13:49:56"} +{"current_steps": 515, "total_steps": 4291, "loss": 0.2807, "lr": 3.99533029291554e-05, "epoch": 0.8408163265306122, "percentage": 12.0, "elapsed_time": "1:52:57", "remaining_time": "13:48:15"} +{"current_steps": 520, "total_steps": 4291, "loss": 0.2915, "lr": 3.994758080432453e-05, "epoch": 0.8489795918367347, "percentage": 12.12, "elapsed_time": "1:54:11", "remaining_time": "13:48:08"} +{"current_steps": 525, "total_steps": 4291, "loss": 0.287, "lr": 3.994152851544248e-05, "epoch": 0.8571428571428571, "percentage": 12.23, "elapsed_time": "1:55:13", "remaining_time": "13:46:33"} +{"current_steps": 530, "total_steps": 4291, "loss": 0.2786, "lr": 3.9935146162684206e-05, "epoch": 0.8653061224489796, "percentage": 12.35, "elapsed_time": "1:56:21", "remaining_time": "13:45:40"} +{"current_steps": 535, "total_steps": 4291, "loss": 0.2746, "lr": 3.992843385168775e-05, "epoch": 0.8734693877551021, "percentage": 12.47, "elapsed_time": "1:57:25", "remaining_time": "13:44:25"} +{"current_steps": 540, "total_steps": 4291, "loss": 0.2898, "lr": 3.99213916935525e-05, "epoch": 0.8816326530612245, "percentage": 12.58, "elapsed_time": "1:58:32", "remaining_time": "13:43:25"} +{"current_steps": 545, "total_steps": 4291, "loss": 0.2792, "lr": 3.991401980483731e-05, "epoch": 0.889795918367347, "percentage": 12.7, "elapsed_time": "1:59:42", "remaining_time": "13:42:49"} +{"current_steps": 550, "total_steps": 4291, "loss": 0.2863, "lr": 3.990631830755863e-05, "epoch": 0.8979591836734694, "percentage": 12.82, "elapsed_time": "2:00:42", "remaining_time": "13:41:04"} +{"current_steps": 555, "total_steps": 4291, "loss": 0.292, "lr": 3.9898287329188426e-05, "epoch": 0.9061224489795918, "percentage": 12.93, "elapsed_time": "2:01:52", "remaining_time": "13:40:24"} +{"current_steps": 560, "total_steps": 4291, "loss": 0.2751, "lr": 3.988992700265211e-05, "epoch": 0.9142857142857143, "percentage": 13.05, "elapsed_time": "2:02:49", "remaining_time": "13:38:17"} +{"current_steps": 565, "total_steps": 4291, "loss": 0.2664, "lr": 3.988123746632632e-05, "epoch": 0.9224489795918367, "percentage": 13.17, "elapsed_time": "2:03:57", "remaining_time": "13:37:30"} +{"current_steps": 570, "total_steps": 4291, "loss": 0.2805, "lr": 3.9872218864036656e-05, "epoch": 0.9306122448979591, "percentage": 13.28, "elapsed_time": "2:04:56", "remaining_time": "13:35:35"} +{"current_steps": 575, "total_steps": 4291, "loss": 0.2866, "lr": 3.986287134505525e-05, "epoch": 0.9387755102040817, "percentage": 13.4, "elapsed_time": "2:05:59", "remaining_time": "13:34:15"} +{"current_steps": 580, "total_steps": 4291, "loss": 0.2864, "lr": 3.985319506409836e-05, "epoch": 0.9469387755102041, "percentage": 13.52, "elapsed_time": "2:06:54", "remaining_time": "13:32:02"} +{"current_steps": 585, "total_steps": 4291, "loss": 0.2805, "lr": 3.9843190181323744e-05, "epoch": 0.9551020408163265, "percentage": 13.63, "elapsed_time": "2:08:00", "remaining_time": "13:30:57"} +{"current_steps": 590, "total_steps": 4291, "loss": 0.2883, "lr": 3.9832856862328074e-05, "epoch": 0.963265306122449, "percentage": 13.75, "elapsed_time": "2:09:10", "remaining_time": "13:30:20"} +{"current_steps": 595, "total_steps": 4291, "loss": 0.2788, "lr": 3.982219527814413e-05, "epoch": 0.9714285714285714, "percentage": 13.87, "elapsed_time": "2:10:21", "remaining_time": "13:29:43"} +{"current_steps": 600, "total_steps": 4291, "loss": 0.2731, "lr": 3.981120560523802e-05, "epoch": 0.9795918367346939, "percentage": 13.98, "elapsed_time": "2:11:31", "remaining_time": "13:29:05"} +{"current_steps": 605, "total_steps": 4291, "loss": 0.2832, "lr": 3.979988802550622e-05, "epoch": 0.9877551020408163, "percentage": 14.1, "elapsed_time": "2:13:13", "remaining_time": "13:31:39"} +{"current_steps": 610, "total_steps": 4291, "loss": 0.2689, "lr": 3.978824272627262e-05, "epoch": 0.9959183673469387, "percentage": 14.22, "elapsed_time": "2:14:24", "remaining_time": "13:31:02"} +{"current_steps": 615, "total_steps": 4291, "loss": 0.2645, "lr": 3.977626990028535e-05, "epoch": 1.003265306122449, "percentage": 14.33, "elapsed_time": "2:15:21", "remaining_time": "13:29:01"} +{"current_steps": 620, "total_steps": 4291, "loss": 0.2485, "lr": 3.976396974571364e-05, "epoch": 1.0114285714285713, "percentage": 14.45, "elapsed_time": "2:16:25", "remaining_time": "13:27:46"} +{"current_steps": 625, "total_steps": 4291, "loss": 0.2686, "lr": 3.9751342466144526e-05, "epoch": 1.019591836734694, "percentage": 14.57, "elapsed_time": "2:17:28", "remaining_time": "13:26:23"} +{"current_steps": 630, "total_steps": 4291, "loss": 0.2776, "lr": 3.9738388270579484e-05, "epoch": 1.0277551020408162, "percentage": 14.68, "elapsed_time": "2:18:33", "remaining_time": "13:25:09"} +{"current_steps": 635, "total_steps": 4291, "loss": 0.2474, "lr": 3.972510737343097e-05, "epoch": 1.0359183673469388, "percentage": 14.8, "elapsed_time": "2:19:49", "remaining_time": "13:25:04"} +{"current_steps": 640, "total_steps": 4291, "loss": 0.2585, "lr": 3.971149999451886e-05, "epoch": 1.0440816326530613, "percentage": 14.91, "elapsed_time": "2:20:58", "remaining_time": "13:24:14"} +{"current_steps": 645, "total_steps": 4291, "loss": 0.2661, "lr": 3.9697566359066805e-05, "epoch": 1.0522448979591836, "percentage": 15.03, "elapsed_time": "2:22:05", "remaining_time": "13:23:14"} +{"current_steps": 650, "total_steps": 4291, "loss": 0.2544, "lr": 3.968330669769857e-05, "epoch": 1.0604081632653062, "percentage": 15.15, "elapsed_time": "2:23:15", "remaining_time": "13:22:26"} +{"current_steps": 655, "total_steps": 4291, "loss": 0.2703, "lr": 3.966872124643411e-05, "epoch": 1.0685714285714285, "percentage": 15.26, "elapsed_time": "2:24:19", "remaining_time": "13:21:09"} +{"current_steps": 660, "total_steps": 4291, "loss": 0.2701, "lr": 3.9653810246685746e-05, "epoch": 1.076734693877551, "percentage": 15.38, "elapsed_time": "2:25:26", "remaining_time": "13:20:11"} +{"current_steps": 665, "total_steps": 4291, "loss": 0.2726, "lr": 3.963857394525413e-05, "epoch": 1.0848979591836734, "percentage": 15.5, "elapsed_time": "2:26:35", "remaining_time": "13:19:16"} +{"current_steps": 670, "total_steps": 4291, "loss": 0.2559, "lr": 3.96230125943242e-05, "epoch": 1.093061224489796, "percentage": 15.61, "elapsed_time": "2:27:42", "remaining_time": "13:18:19"} +{"current_steps": 675, "total_steps": 4291, "loss": 0.2717, "lr": 3.960712645146094e-05, "epoch": 1.1012244897959185, "percentage": 15.73, "elapsed_time": "2:28:39", "remaining_time": "13:16:22"} +{"current_steps": 680, "total_steps": 4291, "loss": 0.267, "lr": 3.959091577960517e-05, "epoch": 1.1093877551020408, "percentage": 15.85, "elapsed_time": "2:29:38", "remaining_time": "13:14:38"} +{"current_steps": 685, "total_steps": 4291, "loss": 0.2567, "lr": 3.957438084706919e-05, "epoch": 1.1175510204081633, "percentage": 15.96, "elapsed_time": "2:30:42", "remaining_time": "13:13:22"} +{"current_steps": 690, "total_steps": 4291, "loss": 0.2669, "lr": 3.955752192753231e-05, "epoch": 1.1257142857142857, "percentage": 16.08, "elapsed_time": "2:31:33", "remaining_time": "13:10:59"} +{"current_steps": 695, "total_steps": 4291, "loss": 0.2782, "lr": 3.954033930003634e-05, "epoch": 1.1338775510204082, "percentage": 16.2, "elapsed_time": "2:32:32", "remaining_time": "13:09:18"} +{"current_steps": 700, "total_steps": 4291, "loss": 0.2636, "lr": 3.9522833248980995e-05, "epoch": 1.1420408163265305, "percentage": 16.31, "elapsed_time": "2:33:32", "remaining_time": "13:07:39"} +{"current_steps": 705, "total_steps": 4291, "loss": 0.2559, "lr": 3.950500406411913e-05, "epoch": 1.150204081632653, "percentage": 16.43, "elapsed_time": "2:34:34", "remaining_time": "13:06:16"} +{"current_steps": 710, "total_steps": 4291, "loss": 0.2745, "lr": 3.948685204055199e-05, "epoch": 1.1583673469387756, "percentage": 16.55, "elapsed_time": "2:35:35", "remaining_time": "13:04:43"} +{"current_steps": 715, "total_steps": 4291, "loss": 0.2773, "lr": 3.946837747872431e-05, "epoch": 1.166530612244898, "percentage": 16.66, "elapsed_time": "2:36:45", "remaining_time": "13:03:59"} +{"current_steps": 720, "total_steps": 4291, "loss": 0.2403, "lr": 3.9449580684419346e-05, "epoch": 1.1746938775510205, "percentage": 16.78, "elapsed_time": "2:37:48", "remaining_time": "13:02:43"} +{"current_steps": 725, "total_steps": 4291, "loss": 0.2483, "lr": 3.943046196875381e-05, "epoch": 1.1828571428571428, "percentage": 16.9, "elapsed_time": "2:38:57", "remaining_time": "13:01:50"} +{"current_steps": 730, "total_steps": 4291, "loss": 0.2741, "lr": 3.941102164817272e-05, "epoch": 1.1910204081632654, "percentage": 17.01, "elapsed_time": "2:40:00", "remaining_time": "13:00:30"} +{"current_steps": 735, "total_steps": 4291, "loss": 0.2581, "lr": 3.939126004444417e-05, "epoch": 1.1991836734693877, "percentage": 17.13, "elapsed_time": "2:41:11", "remaining_time": "12:59:49"} +{"current_steps": 740, "total_steps": 4291, "loss": 0.2621, "lr": 3.9371177484654e-05, "epoch": 1.2073469387755102, "percentage": 17.25, "elapsed_time": "2:42:22", "remaining_time": "12:59:11"} +{"current_steps": 745, "total_steps": 4291, "loss": 0.256, "lr": 3.9350774301200374e-05, "epoch": 1.2155102040816326, "percentage": 17.36, "elapsed_time": "2:43:32", "remaining_time": "12:58:25"} +{"current_steps": 750, "total_steps": 4291, "loss": 0.2486, "lr": 3.933005083178828e-05, "epoch": 1.2236734693877551, "percentage": 17.48, "elapsed_time": "2:44:38", "remaining_time": "12:57:19"} +{"current_steps": 755, "total_steps": 4291, "loss": 0.2696, "lr": 3.930900741942397e-05, "epoch": 1.2318367346938777, "percentage": 17.59, "elapsed_time": "2:45:35", "remaining_time": "12:55:30"} +{"current_steps": 760, "total_steps": 4291, "loss": 0.27, "lr": 3.928764441240922e-05, "epoch": 1.24, "percentage": 17.71, "elapsed_time": "2:46:44", "remaining_time": "12:54:39"} +{"current_steps": 765, "total_steps": 4291, "loss": 0.2669, "lr": 3.9265962164335656e-05, "epoch": 1.2481632653061225, "percentage": 17.83, "elapsed_time": "2:47:46", "remaining_time": "12:53:16"} +{"current_steps": 770, "total_steps": 4291, "loss": 0.2558, "lr": 3.92439610340788e-05, "epoch": 1.2563265306122449, "percentage": 17.94, "elapsed_time": "2:48:51", "remaining_time": "12:52:10"} +{"current_steps": 775, "total_steps": 4291, "loss": 0.2397, "lr": 3.92216413857922e-05, "epoch": 1.2644897959183674, "percentage": 18.06, "elapsed_time": "2:49:57", "remaining_time": "12:51:02"} +{"current_steps": 780, "total_steps": 4291, "loss": 0.2744, "lr": 3.9199003588901375e-05, "epoch": 1.2726530612244897, "percentage": 18.18, "elapsed_time": "2:51:08", "remaining_time": "12:50:22"} +{"current_steps": 785, "total_steps": 4291, "loss": 0.2691, "lr": 3.917604801809774e-05, "epoch": 1.2808163265306123, "percentage": 18.29, "elapsed_time": "2:52:03", "remaining_time": "12:48:27"} +{"current_steps": 790, "total_steps": 4291, "loss": 0.2611, "lr": 3.9152775053332305e-05, "epoch": 1.2889795918367346, "percentage": 18.41, "elapsed_time": "2:53:11", "remaining_time": "12:47:32"} +{"current_steps": 795, "total_steps": 4291, "loss": 0.2589, "lr": 3.912918507980952e-05, "epoch": 1.2971428571428572, "percentage": 18.53, "elapsed_time": "2:54:14", "remaining_time": "12:46:11"} +{"current_steps": 800, "total_steps": 4291, "loss": 0.2609, "lr": 3.91052784879808e-05, "epoch": 1.3053061224489797, "percentage": 18.64, "elapsed_time": "2:55:22", "remaining_time": "12:45:17"} +{"current_steps": 805, "total_steps": 4291, "loss": 0.2711, "lr": 3.9081055673538093e-05, "epoch": 1.313469387755102, "percentage": 18.76, "elapsed_time": "2:57:04", "remaining_time": "12:46:47"} +{"current_steps": 810, "total_steps": 4291, "loss": 0.2571, "lr": 3.9056517037407334e-05, "epoch": 1.3216326530612246, "percentage": 18.88, "elapsed_time": "2:58:02", "remaining_time": "12:45:07"} +{"current_steps": 815, "total_steps": 4291, "loss": 0.2583, "lr": 3.903166298574181e-05, "epoch": 1.329795918367347, "percentage": 18.99, "elapsed_time": "2:59:02", "remaining_time": "12:43:38"} +{"current_steps": 820, "total_steps": 4291, "loss": 0.2646, "lr": 3.900649392991545e-05, "epoch": 1.3379591836734694, "percentage": 19.11, "elapsed_time": "3:00:11", "remaining_time": "12:42:46"} +{"current_steps": 825, "total_steps": 4291, "loss": 0.2741, "lr": 3.898101028651597e-05, "epoch": 1.3461224489795918, "percentage": 19.23, "elapsed_time": "3:01:18", "remaining_time": "12:41:44"} +{"current_steps": 830, "total_steps": 4291, "loss": 0.2617, "lr": 3.895521247733803e-05, "epoch": 1.3542857142857143, "percentage": 19.34, "elapsed_time": "3:02:29", "remaining_time": "12:40:56"} +{"current_steps": 835, "total_steps": 4291, "loss": 0.249, "lr": 3.892910092937622e-05, "epoch": 1.3624489795918366, "percentage": 19.46, "elapsed_time": "3:03:31", "remaining_time": "12:39:36"} +{"current_steps": 840, "total_steps": 4291, "loss": 0.2603, "lr": 3.890267607481801e-05, "epoch": 1.3706122448979592, "percentage": 19.58, "elapsed_time": "3:04:37", "remaining_time": "12:38:31"} +{"current_steps": 845, "total_steps": 4291, "loss": 0.2741, "lr": 3.887593835103659e-05, "epoch": 1.3787755102040817, "percentage": 19.69, "elapsed_time": "3:05:41", "remaining_time": "12:37:15"} +{"current_steps": 850, "total_steps": 4291, "loss": 0.2703, "lr": 3.8848888200583637e-05, "epoch": 1.386938775510204, "percentage": 19.81, "elapsed_time": "3:06:46", "remaining_time": "12:36:07"} +{"current_steps": 855, "total_steps": 4291, "loss": 0.2695, "lr": 3.882152607118197e-05, "epoch": 1.3951020408163266, "percentage": 19.93, "elapsed_time": "3:08:02", "remaining_time": "12:35:40"} +{"current_steps": 860, "total_steps": 4291, "loss": 0.2583, "lr": 3.879385241571817e-05, "epoch": 1.403265306122449, "percentage": 20.04, "elapsed_time": "3:09:11", "remaining_time": "12:34:45"} +{"current_steps": 865, "total_steps": 4291, "loss": 0.249, "lr": 3.876586769223505e-05, "epoch": 1.4114285714285715, "percentage": 20.16, "elapsed_time": "3:10:11", "remaining_time": "12:33:18"} +{"current_steps": 870, "total_steps": 4291, "loss": 0.2586, "lr": 3.87375723639241e-05, "epoch": 1.4195918367346938, "percentage": 20.27, "elapsed_time": "3:11:17", "remaining_time": "12:32:12"} +{"current_steps": 875, "total_steps": 4291, "loss": 0.2525, "lr": 3.8708966899117815e-05, "epoch": 1.4277551020408163, "percentage": 20.39, "elapsed_time": "3:12:20", "remaining_time": "12:30:53"} +{"current_steps": 880, "total_steps": 4291, "loss": 0.2422, "lr": 3.868005177128193e-05, "epoch": 1.4359183673469387, "percentage": 20.51, "elapsed_time": "3:13:29", "remaining_time": "12:29:59"} +{"current_steps": 885, "total_steps": 4291, "loss": 0.274, "lr": 3.8650827459007615e-05, "epoch": 1.4440816326530612, "percentage": 20.62, "elapsed_time": "3:14:36", "remaining_time": "12:28:56"} +{"current_steps": 890, "total_steps": 4291, "loss": 0.2664, "lr": 3.8621294446003504e-05, "epoch": 1.4522448979591838, "percentage": 20.74, "elapsed_time": "3:15:36", "remaining_time": "12:27:27"} +{"current_steps": 895, "total_steps": 4291, "loss": 0.2375, "lr": 3.859145322108774e-05, "epoch": 1.460408163265306, "percentage": 20.86, "elapsed_time": "3:16:38", "remaining_time": "12:26:07"} +{"current_steps": 900, "total_steps": 4291, "loss": 0.2744, "lr": 3.856130427817984e-05, "epoch": 1.4685714285714286, "percentage": 20.97, "elapsed_time": "3:17:31", "remaining_time": "12:24:12"} +{"current_steps": 905, "total_steps": 4291, "loss": 0.2561, "lr": 3.853084811629257e-05, "epoch": 1.476734693877551, "percentage": 21.09, "elapsed_time": "3:18:45", "remaining_time": "12:23:38"} +{"current_steps": 910, "total_steps": 4291, "loss": 0.2484, "lr": 3.8500085239523624e-05, "epoch": 1.4848979591836735, "percentage": 21.21, "elapsed_time": "3:20:01", "remaining_time": "12:23:08"} +{"current_steps": 915, "total_steps": 4291, "loss": 0.2398, "lr": 3.846901615704734e-05, "epoch": 1.4930612244897958, "percentage": 21.32, "elapsed_time": "3:21:12", "remaining_time": "12:22:23"} +{"current_steps": 920, "total_steps": 4291, "loss": 0.2615, "lr": 3.843764138310622e-05, "epoch": 1.5012244897959184, "percentage": 21.44, "elapsed_time": "3:22:21", "remaining_time": "12:21:28"} +{"current_steps": 925, "total_steps": 4291, "loss": 0.2483, "lr": 3.840596143700246e-05, "epoch": 1.5093877551020407, "percentage": 21.56, "elapsed_time": "3:23:29", "remaining_time": "12:20:27"} +{"current_steps": 930, "total_steps": 4291, "loss": 0.2527, "lr": 3.8373976843089355e-05, "epoch": 1.5175510204081633, "percentage": 21.67, "elapsed_time": "3:24:38", "remaining_time": "12:19:34"} +{"current_steps": 935, "total_steps": 4291, "loss": 0.2682, "lr": 3.8341688130762574e-05, "epoch": 1.5257142857142858, "percentage": 21.79, "elapsed_time": "3:25:44", "remaining_time": "12:18:26"} +{"current_steps": 940, "total_steps": 4291, "loss": 0.257, "lr": 3.830909583445143e-05, "epoch": 1.5338775510204081, "percentage": 21.91, "elapsed_time": "3:26:47", "remaining_time": "12:17:12"} +{"current_steps": 945, "total_steps": 4291, "loss": 0.2681, "lr": 3.827620049361004e-05, "epoch": 1.5420408163265305, "percentage": 22.02, "elapsed_time": "3:27:47", "remaining_time": "12:15:42"} +{"current_steps": 950, "total_steps": 4291, "loss": 0.2613, "lr": 3.82430026527084e-05, "epoch": 1.550204081632653, "percentage": 22.14, "elapsed_time": "3:28:49", "remaining_time": "12:14:25"} +{"current_steps": 955, "total_steps": 4291, "loss": 0.2737, "lr": 3.8209502861223344e-05, "epoch": 1.5583673469387755, "percentage": 22.26, "elapsed_time": "3:29:57", "remaining_time": "12:13:25"} +{"current_steps": 960, "total_steps": 4291, "loss": 0.2552, "lr": 3.817570167362947e-05, "epoch": 1.566530612244898, "percentage": 22.37, "elapsed_time": "3:31:04", "remaining_time": "12:12:24"} +{"current_steps": 965, "total_steps": 4291, "loss": 0.2476, "lr": 3.814159964938996e-05, "epoch": 1.5746938775510204, "percentage": 22.49, "elapsed_time": "3:32:14", "remaining_time": "12:11:31"} +{"current_steps": 970, "total_steps": 4291, "loss": 0.2548, "lr": 3.810719735294731e-05, "epoch": 1.5828571428571427, "percentage": 22.61, "elapsed_time": "3:33:22", "remaining_time": "12:10:30"} +{"current_steps": 975, "total_steps": 4291, "loss": 0.2542, "lr": 3.8072495353714025e-05, "epoch": 1.5910204081632653, "percentage": 22.72, "elapsed_time": "3:34:16", "remaining_time": "12:08:44"} +{"current_steps": 980, "total_steps": 4291, "loss": 0.2603, "lr": 3.8037494226063126e-05, "epoch": 1.5991836734693878, "percentage": 22.84, "elapsed_time": "3:35:14", "remaining_time": "12:07:13"} +{"current_steps": 985, "total_steps": 4291, "loss": 0.271, "lr": 3.800219454931871e-05, "epoch": 1.6073469387755102, "percentage": 22.96, "elapsed_time": "3:36:08", "remaining_time": "12:05:26"} +{"current_steps": 990, "total_steps": 4291, "loss": 0.2595, "lr": 3.7966596907746344e-05, "epoch": 1.6155102040816327, "percentage": 23.07, "elapsed_time": "3:36:54", "remaining_time": "12:03:14"} +{"current_steps": 995, "total_steps": 4291, "loss": 0.2548, "lr": 3.793070189054335e-05, "epoch": 1.623673469387755, "percentage": 23.19, "elapsed_time": "3:37:52", "remaining_time": "12:01:42"} +{"current_steps": 1000, "total_steps": 4291, "loss": 0.2688, "lr": 3.789451009182911e-05, "epoch": 1.6318367346938776, "percentage": 23.3, "elapsed_time": "3:38:46", "remaining_time": "11:59:58"} +{"current_steps": 1005, "total_steps": 4291, "loss": 0.241, "lr": 3.78580221106352e-05, "epoch": 1.6400000000000001, "percentage": 23.42, "elapsed_time": "3:40:35", "remaining_time": "12:01:14"} +{"current_steps": 1010, "total_steps": 4291, "loss": 0.2569, "lr": 3.78212385508955e-05, "epoch": 1.6481632653061224, "percentage": 23.54, "elapsed_time": "3:41:45", "remaining_time": "12:00:22"} +{"current_steps": 1015, "total_steps": 4291, "loss": 0.2678, "lr": 3.778416002143617e-05, "epoch": 1.6563265306122448, "percentage": 23.65, "elapsed_time": "3:42:44", "remaining_time": "11:58:54"} +{"current_steps": 1020, "total_steps": 4291, "loss": 0.2459, "lr": 3.774678713596559e-05, "epoch": 1.6644897959183673, "percentage": 23.77, "elapsed_time": "3:43:47", "remaining_time": "11:57:41"} +{"current_steps": 1025, "total_steps": 4291, "loss": 0.2655, "lr": 3.7709120513064196e-05, "epoch": 1.6726530612244899, "percentage": 23.89, "elapsed_time": "3:44:48", "remaining_time": "11:56:19"} +{"current_steps": 1030, "total_steps": 4291, "loss": 0.2589, "lr": 3.767116077617426e-05, "epoch": 1.6808163265306122, "percentage": 24.0, "elapsed_time": "3:45:58", "remaining_time": "11:55:27"} +{"current_steps": 1035, "total_steps": 4291, "loss": 0.2527, "lr": 3.763290855358952e-05, "epoch": 1.6889795918367347, "percentage": 24.12, "elapsed_time": "3:47:12", "remaining_time": "11:54:46"} +{"current_steps": 1040, "total_steps": 4291, "loss": 0.2496, "lr": 3.759436447844485e-05, "epoch": 1.697142857142857, "percentage": 24.24, "elapsed_time": "3:48:29", "remaining_time": "11:54:15"} +{"current_steps": 1045, "total_steps": 4291, "loss": 0.2622, "lr": 3.7555529188705716e-05, "epoch": 1.7053061224489796, "percentage": 24.35, "elapsed_time": "3:49:26", "remaining_time": "11:52:42"} +{"current_steps": 1050, "total_steps": 4291, "loss": 0.2394, "lr": 3.751640332715768e-05, "epoch": 1.7134693877551022, "percentage": 24.47, "elapsed_time": "3:50:34", "remaining_time": "11:51:41"} +{"current_steps": 1055, "total_steps": 4291, "loss": 0.2453, "lr": 3.74769875413957e-05, "epoch": 1.7216326530612245, "percentage": 24.59, "elapsed_time": "3:51:34", "remaining_time": "11:50:18"} +{"current_steps": 1060, "total_steps": 4291, "loss": 0.2516, "lr": 3.7437282483813455e-05, "epoch": 1.7297959183673468, "percentage": 24.7, "elapsed_time": "3:52:35", "remaining_time": "11:48:56"} +{"current_steps": 1065, "total_steps": 4291, "loss": 0.2542, "lr": 3.739728881159252e-05, "epoch": 1.7379591836734694, "percentage": 24.82, "elapsed_time": "3:53:28", "remaining_time": "11:47:11"} +{"current_steps": 1070, "total_steps": 4291, "loss": 0.2481, "lr": 3.735700718669151e-05, "epoch": 1.746122448979592, "percentage": 24.94, "elapsed_time": "3:54:38", "remaining_time": "11:46:20"} +{"current_steps": 1075, "total_steps": 4291, "loss": 0.2649, "lr": 3.73164382758351e-05, "epoch": 1.7542857142857144, "percentage": 25.05, "elapsed_time": "3:55:40", "remaining_time": "11:45:03"} +{"current_steps": 1080, "total_steps": 4291, "loss": 0.2657, "lr": 3.727558275050301e-05, "epoch": 1.7624489795918368, "percentage": 25.17, "elapsed_time": "3:56:48", "remaining_time": "11:44:02"} +{"current_steps": 1085, "total_steps": 4291, "loss": 0.2575, "lr": 3.723444128691888e-05, "epoch": 1.770612244897959, "percentage": 25.29, "elapsed_time": "3:57:56", "remaining_time": "11:43:04"} +{"current_steps": 1090, "total_steps": 4291, "loss": 0.2479, "lr": 3.71930145660391e-05, "epoch": 1.7787755102040816, "percentage": 25.4, "elapsed_time": "3:59:05", "remaining_time": "11:42:09"} +{"current_steps": 1095, "total_steps": 4291, "loss": 0.2459, "lr": 3.715130327354148e-05, "epoch": 1.7869387755102042, "percentage": 25.52, "elapsed_time": "4:00:08", "remaining_time": "11:40:54"} +{"current_steps": 1100, "total_steps": 4291, "loss": 0.2477, "lr": 3.7109308099813986e-05, "epoch": 1.7951020408163265, "percentage": 25.64, "elapsed_time": "4:01:10", "remaining_time": "11:39:37"} +{"current_steps": 1105, "total_steps": 4291, "loss": 0.2567, "lr": 3.706702973994323e-05, "epoch": 1.8032653061224488, "percentage": 25.75, "elapsed_time": "4:02:17", "remaining_time": "11:38:35"} +{"current_steps": 1110, "total_steps": 4291, "loss": 0.2476, "lr": 3.702446889370303e-05, "epoch": 1.8114285714285714, "percentage": 25.87, "elapsed_time": "4:03:26", "remaining_time": "11:37:39"} +{"current_steps": 1115, "total_steps": 4291, "loss": 0.2478, "lr": 3.698162626554279e-05, "epoch": 1.819591836734694, "percentage": 25.98, "elapsed_time": "4:04:31", "remaining_time": "11:36:31"} +{"current_steps": 1120, "total_steps": 4291, "loss": 0.2385, "lr": 3.693850256457585e-05, "epoch": 1.8277551020408165, "percentage": 26.1, "elapsed_time": "4:05:36", "remaining_time": "11:35:22"} +{"current_steps": 1125, "total_steps": 4291, "loss": 0.2372, "lr": 3.689509850456774e-05, "epoch": 1.8359183673469388, "percentage": 26.22, "elapsed_time": "4:06:32", "remaining_time": "11:33:48"} +{"current_steps": 1130, "total_steps": 4291, "loss": 0.2538, "lr": 3.685141480392441e-05, "epoch": 1.8440816326530611, "percentage": 26.33, "elapsed_time": "4:07:45", "remaining_time": "11:33:03"} +{"current_steps": 1135, "total_steps": 4291, "loss": 0.2475, "lr": 3.680745218568026e-05, "epoch": 1.8522448979591837, "percentage": 26.45, "elapsed_time": "4:08:53", "remaining_time": "11:32:03"} +{"current_steps": 1140, "total_steps": 4291, "loss": 0.25, "lr": 3.676321137748626e-05, "epoch": 1.8604081632653062, "percentage": 26.57, "elapsed_time": "4:09:58", "remaining_time": "11:30:55"} +{"current_steps": 1145, "total_steps": 4291, "loss": 0.2548, "lr": 3.671869311159783e-05, "epoch": 1.8685714285714285, "percentage": 26.68, "elapsed_time": "4:11:06", "remaining_time": "11:29:55"} +{"current_steps": 1150, "total_steps": 4291, "loss": 0.2519, "lr": 3.6673898124862776e-05, "epoch": 1.8767346938775509, "percentage": 26.8, "elapsed_time": "4:12:01", "remaining_time": "11:28:20"} +{"current_steps": 1155, "total_steps": 4291, "loss": 0.2687, "lr": 3.662882715870907e-05, "epoch": 1.8848979591836734, "percentage": 26.92, "elapsed_time": "4:13:04", "remaining_time": "11:27:08"} +{"current_steps": 1160, "total_steps": 4291, "loss": 0.2469, "lr": 3.658348095913257e-05, "epoch": 1.893061224489796, "percentage": 27.03, "elapsed_time": "4:14:16", "remaining_time": "11:26:18"} +{"current_steps": 1165, "total_steps": 4291, "loss": 0.2402, "lr": 3.6537860276684686e-05, "epoch": 1.9012244897959185, "percentage": 27.15, "elapsed_time": "4:15:20", "remaining_time": "11:25:08"} +{"current_steps": 1170, "total_steps": 4291, "loss": 0.2605, "lr": 3.6491965866459966e-05, "epoch": 1.9093877551020408, "percentage": 27.27, "elapsed_time": "4:16:25", "remaining_time": "11:24:01"} +{"current_steps": 1175, "total_steps": 4291, "loss": 0.2224, "lr": 3.6445798488083576e-05, "epoch": 1.9175510204081632, "percentage": 27.38, "elapsed_time": "4:17:34", "remaining_time": "11:23:03"} +{"current_steps": 1180, "total_steps": 4291, "loss": 0.2504, "lr": 3.639935890569873e-05, "epoch": 1.9257142857142857, "percentage": 27.5, "elapsed_time": "4:18:43", "remaining_time": "11:22:07"} +{"current_steps": 1185, "total_steps": 4291, "loss": 0.2548, "lr": 3.635264788795407e-05, "epoch": 1.9338775510204083, "percentage": 27.62, "elapsed_time": "4:19:46", "remaining_time": "11:20:54"} +{"current_steps": 1190, "total_steps": 4291, "loss": 0.2537, "lr": 3.6305666207990886e-05, "epoch": 1.9420408163265306, "percentage": 27.73, "elapsed_time": "4:20:50", "remaining_time": "11:19:43"} +{"current_steps": 1195, "total_steps": 4291, "loss": 0.2597, "lr": 3.625841464343042e-05, "epoch": 1.950204081632653, "percentage": 27.85, "elapsed_time": "4:21:58", "remaining_time": "11:18:43"} +{"current_steps": 1200, "total_steps": 4291, "loss": 0.2587, "lr": 3.621089397636086e-05, "epoch": 1.9583673469387755, "percentage": 27.97, "elapsed_time": "4:23:00", "remaining_time": "11:17:28"} +{"current_steps": 1205, "total_steps": 4291, "loss": 0.2514, "lr": 3.616310499332451e-05, "epoch": 1.966530612244898, "percentage": 28.08, "elapsed_time": "4:24:47", "remaining_time": "11:18:08"} +{"current_steps": 1210, "total_steps": 4291, "loss": 0.2492, "lr": 3.6115048485304736e-05, "epoch": 1.9746938775510205, "percentage": 28.2, "elapsed_time": "4:25:53", "remaining_time": "11:17:01"} +{"current_steps": 1215, "total_steps": 4291, "loss": 0.2538, "lr": 3.606672524771281e-05, "epoch": 1.9828571428571429, "percentage": 28.32, "elapsed_time": "4:26:59", "remaining_time": "11:15:57"} +{"current_steps": 1220, "total_steps": 4291, "loss": 0.2525, "lr": 3.601813608037488e-05, "epoch": 1.9910204081632652, "percentage": 28.43, "elapsed_time": "4:28:07", "remaining_time": "11:14:56"} +{"current_steps": 1225, "total_steps": 4291, "loss": 0.2563, "lr": 3.596928178751858e-05, "epoch": 1.9991836734693877, "percentage": 28.55, "elapsed_time": "4:29:10", "remaining_time": "11:13:42"} +{"current_steps": 1230, "total_steps": 4291, "loss": 0.2295, "lr": 3.592016317775984e-05, "epoch": 2.006530612244898, "percentage": 28.66, "elapsed_time": "4:30:10", "remaining_time": "11:12:20"} +{"current_steps": 1235, "total_steps": 4291, "loss": 0.2196, "lr": 3.587078106408942e-05, "epoch": 2.0146938775510206, "percentage": 28.78, "elapsed_time": "4:31:12", "remaining_time": "11:11:05"} +{"current_steps": 1240, "total_steps": 4291, "loss": 0.2343, "lr": 3.582113626385951e-05, "epoch": 2.0228571428571427, "percentage": 28.9, "elapsed_time": "4:32:16", "remaining_time": "11:09:55"} +{"current_steps": 1245, "total_steps": 4291, "loss": 0.2393, "lr": 3.577122959877017e-05, "epoch": 2.0310204081632652, "percentage": 29.01, "elapsed_time": "4:33:28", "remaining_time": "11:09:05"} +{"current_steps": 1250, "total_steps": 4291, "loss": 0.2212, "lr": 3.5721061894855756e-05, "epoch": 2.039183673469388, "percentage": 29.13, "elapsed_time": "4:34:38", "remaining_time": "11:08:08"} +{"current_steps": 1255, "total_steps": 4291, "loss": 0.236, "lr": 3.5670633982471185e-05, "epoch": 2.0473469387755103, "percentage": 29.25, "elapsed_time": "4:35:34", "remaining_time": "11:06:39"} +{"current_steps": 1260, "total_steps": 4291, "loss": 0.2375, "lr": 3.561994669627828e-05, "epoch": 2.0555102040816324, "percentage": 29.36, "elapsed_time": "4:36:44", "remaining_time": "11:05:43"} +{"current_steps": 1265, "total_steps": 4291, "loss": 0.2315, "lr": 3.55690008752319e-05, "epoch": 2.063673469387755, "percentage": 29.48, "elapsed_time": "4:37:44", "remaining_time": "11:04:22"} +{"current_steps": 1270, "total_steps": 4291, "loss": 0.2266, "lr": 3.551779736256604e-05, "epoch": 2.0718367346938775, "percentage": 29.6, "elapsed_time": "4:38:42", "remaining_time": "11:02:58"} +{"current_steps": 1275, "total_steps": 4291, "loss": 0.2335, "lr": 3.546633700577994e-05, "epoch": 2.08, "percentage": 29.71, "elapsed_time": "4:39:42", "remaining_time": "11:01:38"} +{"current_steps": 1280, "total_steps": 4291, "loss": 0.2299, "lr": 3.541462065662399e-05, "epoch": 2.0881632653061226, "percentage": 29.83, "elapsed_time": "4:40:47", "remaining_time": "11:00:31"} +{"current_steps": 1285, "total_steps": 4291, "loss": 0.225, "lr": 3.536264917108567e-05, "epoch": 2.0963265306122447, "percentage": 29.95, "elapsed_time": "4:41:54", "remaining_time": "10:59:28"} +{"current_steps": 1290, "total_steps": 4291, "loss": 0.2234, "lr": 3.531042340937537e-05, "epoch": 2.1044897959183673, "percentage": 30.06, "elapsed_time": "4:42:52", "remaining_time": "10:58:04"} +{"current_steps": 1295, "total_steps": 4291, "loss": 0.2397, "lr": 3.525794423591214e-05, "epoch": 2.11265306122449, "percentage": 30.18, "elapsed_time": "4:43:55", "remaining_time": "10:56:52"} +{"current_steps": 1300, "total_steps": 4291, "loss": 0.2111, "lr": 3.520521251930941e-05, "epoch": 2.1208163265306124, "percentage": 30.3, "elapsed_time": "4:44:59", "remaining_time": "10:55:41"} +{"current_steps": 1305, "total_steps": 4291, "loss": 0.2318, "lr": 3.51522291323606e-05, "epoch": 2.1289795918367345, "percentage": 30.41, "elapsed_time": "4:46:02", "remaining_time": "10:54:29"} +{"current_steps": 1310, "total_steps": 4291, "loss": 0.2425, "lr": 3.509899495202467e-05, "epoch": 2.137142857142857, "percentage": 30.53, "elapsed_time": "4:47:07", "remaining_time": "10:53:22"} +{"current_steps": 1315, "total_steps": 4291, "loss": 0.2292, "lr": 3.50455108594116e-05, "epoch": 2.1453061224489796, "percentage": 30.65, "elapsed_time": "4:48:04", "remaining_time": "10:51:57"} +{"current_steps": 1320, "total_steps": 4291, "loss": 0.2274, "lr": 3.499177773976781e-05, "epoch": 2.153469387755102, "percentage": 30.76, "elapsed_time": "4:49:09", "remaining_time": "10:50:49"} +{"current_steps": 1325, "total_steps": 4291, "loss": 0.2266, "lr": 3.493779648246154e-05, "epoch": 2.1616326530612247, "percentage": 30.88, "elapsed_time": "4:50:07", "remaining_time": "10:49:27"} +{"current_steps": 1330, "total_steps": 4291, "loss": 0.2326, "lr": 3.488356798096808e-05, "epoch": 2.1697959183673468, "percentage": 31.0, "elapsed_time": "4:51:05", "remaining_time": "10:48:04"} +{"current_steps": 1335, "total_steps": 4291, "loss": 0.2218, "lr": 3.482909313285499e-05, "epoch": 2.1779591836734693, "percentage": 31.11, "elapsed_time": "4:52:14", "remaining_time": "10:47:04"} +{"current_steps": 1340, "total_steps": 4291, "loss": 0.2284, "lr": 3.47743728397673e-05, "epoch": 2.186122448979592, "percentage": 31.23, "elapsed_time": "4:53:25", "remaining_time": "10:46:10"} +{"current_steps": 1345, "total_steps": 4291, "loss": 0.2197, "lr": 3.4719408007412494e-05, "epoch": 2.1942857142857144, "percentage": 31.34, "elapsed_time": "4:54:38", "remaining_time": "10:45:22"} +{"current_steps": 1350, "total_steps": 4291, "loss": 0.2252, "lr": 3.4664199545545586e-05, "epoch": 2.202448979591837, "percentage": 31.46, "elapsed_time": "4:55:45", "remaining_time": "10:44:18"} +{"current_steps": 1355, "total_steps": 4291, "loss": 0.2474, "lr": 3.4608748367954064e-05, "epoch": 2.210612244897959, "percentage": 31.58, "elapsed_time": "4:56:37", "remaining_time": "10:42:43"} +{"current_steps": 1360, "total_steps": 4291, "loss": 0.2397, "lr": 3.455305539244272e-05, "epoch": 2.2187755102040816, "percentage": 31.69, "elapsed_time": "4:57:43", "remaining_time": "10:41:38"} +{"current_steps": 1365, "total_steps": 4291, "loss": 0.2319, "lr": 3.44971215408185e-05, "epoch": 2.226938775510204, "percentage": 31.81, "elapsed_time": "4:58:47", "remaining_time": "10:40:29"} +{"current_steps": 1370, "total_steps": 4291, "loss": 0.2322, "lr": 3.4440947738875215e-05, "epoch": 2.2351020408163267, "percentage": 31.93, "elapsed_time": "4:59:53", "remaining_time": "10:39:24"} +{"current_steps": 1375, "total_steps": 4291, "loss": 0.212, "lr": 3.438453491637825e-05, "epoch": 2.243265306122449, "percentage": 32.04, "elapsed_time": "5:01:07", "remaining_time": "10:38:36"} +{"current_steps": 1380, "total_steps": 4291, "loss": 0.2289, "lr": 3.432788400704915e-05, "epoch": 2.2514285714285713, "percentage": 32.16, "elapsed_time": "5:02:17", "remaining_time": "10:37:40"} +{"current_steps": 1385, "total_steps": 4291, "loss": 0.2238, "lr": 3.4270995948550175e-05, "epoch": 2.259591836734694, "percentage": 32.28, "elapsed_time": "5:03:18", "remaining_time": "10:36:23"} +{"current_steps": 1390, "total_steps": 4291, "loss": 0.235, "lr": 3.421387168246879e-05, "epoch": 2.2677551020408164, "percentage": 32.39, "elapsed_time": "5:04:17", "remaining_time": "10:35:04"} +{"current_steps": 1395, "total_steps": 4291, "loss": 0.2203, "lr": 3.415651215430205e-05, "epoch": 2.2759183673469385, "percentage": 32.51, "elapsed_time": "5:05:13", "remaining_time": "10:33:37"} +{"current_steps": 1400, "total_steps": 4291, "loss": 0.2295, "lr": 3.409891831344099e-05, "epoch": 2.284081632653061, "percentage": 32.63, "elapsed_time": "5:06:11", "remaining_time": "10:32:16"} +{"current_steps": 1405, "total_steps": 4291, "loss": 0.2443, "lr": 3.404109111315487e-05, "epoch": 2.2922448979591836, "percentage": 32.74, "elapsed_time": "5:08:10", "remaining_time": "10:33:00"} +{"current_steps": 1410, "total_steps": 4291, "loss": 0.2353, "lr": 3.398303151057543e-05, "epoch": 2.300408163265306, "percentage": 32.86, "elapsed_time": "5:09:10", "remaining_time": "10:31:43"} +{"current_steps": 1415, "total_steps": 4291, "loss": 0.2354, "lr": 3.392474046668105e-05, "epoch": 2.3085714285714287, "percentage": 32.98, "elapsed_time": "5:10:17", "remaining_time": "10:30:40"} +{"current_steps": 1420, "total_steps": 4291, "loss": 0.2396, "lr": 3.38662189462808e-05, "epoch": 2.3167346938775513, "percentage": 33.09, "elapsed_time": "5:11:20", "remaining_time": "10:29:28"} +{"current_steps": 1425, "total_steps": 4291, "loss": 0.2294, "lr": 3.3807467917998534e-05, "epoch": 2.3248979591836734, "percentage": 33.21, "elapsed_time": "5:12:26", "remaining_time": "10:28:23"} +{"current_steps": 1430, "total_steps": 4291, "loss": 0.2148, "lr": 3.374848835425679e-05, "epoch": 2.333061224489796, "percentage": 33.33, "elapsed_time": "5:13:37", "remaining_time": "10:27:27"} +{"current_steps": 1435, "total_steps": 4291, "loss": 0.2216, "lr": 3.3689281231260755e-05, "epoch": 2.3412244897959185, "percentage": 33.44, "elapsed_time": "5:14:49", "remaining_time": "10:26:33"} +{"current_steps": 1440, "total_steps": 4291, "loss": 0.228, "lr": 3.362984752898208e-05, "epoch": 2.349387755102041, "percentage": 33.56, "elapsed_time": "5:15:55", "remaining_time": "10:25:29"} +{"current_steps": 1445, "total_steps": 4291, "loss": 0.2438, "lr": 3.3570188231142647e-05, "epoch": 2.357551020408163, "percentage": 33.68, "elapsed_time": "5:16:55", "remaining_time": "10:24:12"} +{"current_steps": 1450, "total_steps": 4291, "loss": 0.222, "lr": 3.351030432519832e-05, "epoch": 2.3657142857142857, "percentage": 33.79, "elapsed_time": "5:18:00", "remaining_time": "10:23:04"} +{"current_steps": 1455, "total_steps": 4291, "loss": 0.2175, "lr": 3.345019680232256e-05, "epoch": 2.373877551020408, "percentage": 33.91, "elapsed_time": "5:19:04", "remaining_time": "10:21:54"} +{"current_steps": 1460, "total_steps": 4291, "loss": 0.239, "lr": 3.3389866657390074e-05, "epoch": 2.3820408163265308, "percentage": 34.02, "elapsed_time": "5:20:08", "remaining_time": "10:20:46"} +{"current_steps": 1465, "total_steps": 4291, "loss": 0.2384, "lr": 3.332931488896029e-05, "epoch": 2.3902040816326533, "percentage": 34.14, "elapsed_time": "5:21:13", "remaining_time": "10:19:37"} +{"current_steps": 1470, "total_steps": 4291, "loss": 0.2354, "lr": 3.3268542499260874e-05, "epoch": 2.3983673469387754, "percentage": 34.26, "elapsed_time": "5:22:13", "remaining_time": "10:18:21"} +{"current_steps": 1475, "total_steps": 4291, "loss": 0.218, "lr": 3.32075504941711e-05, "epoch": 2.406530612244898, "percentage": 34.37, "elapsed_time": "5:23:17", "remaining_time": "10:17:12"} +{"current_steps": 1480, "total_steps": 4291, "loss": 0.2342, "lr": 3.3146339883205243e-05, "epoch": 2.4146938775510205, "percentage": 34.49, "elapsed_time": "5:24:20", "remaining_time": "10:16:01"} +{"current_steps": 1485, "total_steps": 4291, "loss": 0.2275, "lr": 3.308491167949585e-05, "epoch": 2.422857142857143, "percentage": 34.61, "elapsed_time": "5:25:32", "remaining_time": "10:15:08"} +{"current_steps": 1490, "total_steps": 4291, "loss": 0.2258, "lr": 3.302326689977697e-05, "epoch": 2.431020408163265, "percentage": 34.72, "elapsed_time": "5:26:42", "remaining_time": "10:14:10"} +{"current_steps": 1495, "total_steps": 4291, "loss": 0.2247, "lr": 3.2961406564367335e-05, "epoch": 2.4391836734693877, "percentage": 34.84, "elapsed_time": "5:27:46", "remaining_time": "10:13:00"} +{"current_steps": 1500, "total_steps": 4291, "loss": 0.2415, "lr": 3.2899331697153445e-05, "epoch": 2.4473469387755102, "percentage": 34.96, "elapsed_time": "5:28:58", "remaining_time": "10:12:07"} +{"current_steps": 1505, "total_steps": 4291, "loss": 0.2351, "lr": 3.283704332557266e-05, "epoch": 2.455510204081633, "percentage": 35.07, "elapsed_time": "5:30:02", "remaining_time": "10:10:57"} +{"current_steps": 1510, "total_steps": 4291, "loss": 0.2223, "lr": 3.2774542480596165e-05, "epoch": 2.4636734693877553, "percentage": 35.19, "elapsed_time": "5:31:11", "remaining_time": "10:09:57"} +{"current_steps": 1515, "total_steps": 4291, "loss": 0.2221, "lr": 3.2711830196711926e-05, "epoch": 2.4718367346938774, "percentage": 35.31, "elapsed_time": "5:32:09", "remaining_time": "10:08:38"} +{"current_steps": 1520, "total_steps": 4291, "loss": 0.2231, "lr": 3.2648907511907544e-05, "epoch": 2.48, "percentage": 35.42, "elapsed_time": "5:33:18", "remaining_time": "10:07:38"} +{"current_steps": 1525, "total_steps": 4291, "loss": 0.2274, "lr": 3.258577546765311e-05, "epoch": 2.4881632653061225, "percentage": 35.54, "elapsed_time": "5:34:31", "remaining_time": "10:06:44"} +{"current_steps": 1530, "total_steps": 4291, "loss": 0.2246, "lr": 3.2522435108883925e-05, "epoch": 2.496326530612245, "percentage": 35.66, "elapsed_time": "5:35:38", "remaining_time": "10:05:42"} +{"current_steps": 1535, "total_steps": 4291, "loss": 0.2305, "lr": 3.245888748398322e-05, "epoch": 2.504489795918367, "percentage": 35.77, "elapsed_time": "5:36:46", "remaining_time": "10:04:39"} +{"current_steps": 1540, "total_steps": 4291, "loss": 0.2247, "lr": 3.239513364476484e-05, "epoch": 2.5126530612244897, "percentage": 35.89, "elapsed_time": "5:37:53", "remaining_time": "10:03:36"} +{"current_steps": 1545, "total_steps": 4291, "loss": 0.2272, "lr": 3.233117464645577e-05, "epoch": 2.5208163265306123, "percentage": 36.01, "elapsed_time": "5:38:50", "remaining_time": "10:02:14"} +{"current_steps": 1550, "total_steps": 4291, "loss": 0.2399, "lr": 3.2267011547678735e-05, "epoch": 2.528979591836735, "percentage": 36.12, "elapsed_time": "5:39:56", "remaining_time": "10:01:09"} +{"current_steps": 1555, "total_steps": 4291, "loss": 0.2248, "lr": 3.220264541043461e-05, "epoch": 2.5371428571428574, "percentage": 36.24, "elapsed_time": "5:40:55", "remaining_time": "9:59:51"} +{"current_steps": 1560, "total_steps": 4291, "loss": 0.2327, "lr": 3.21380773000849e-05, "epoch": 2.5453061224489795, "percentage": 36.36, "elapsed_time": "5:42:01", "remaining_time": "9:58:45"} +{"current_steps": 1565, "total_steps": 4291, "loss": 0.2059, "lr": 3.2073308285334085e-05, "epoch": 2.553469387755102, "percentage": 36.47, "elapsed_time": "5:43:11", "remaining_time": "9:57:47"} +{"current_steps": 1570, "total_steps": 4291, "loss": 0.2271, "lr": 3.200833943821192e-05, "epoch": 2.5616326530612246, "percentage": 36.59, "elapsed_time": "5:44:15", "remaining_time": "9:56:38"} +{"current_steps": 1575, "total_steps": 4291, "loss": 0.2249, "lr": 3.194317183405573e-05, "epoch": 2.5697959183673467, "percentage": 36.7, "elapsed_time": "5:45:17", "remaining_time": "9:55:25"} +{"current_steps": 1580, "total_steps": 4291, "loss": 0.2208, "lr": 3.187780655149253e-05, "epoch": 2.577959183673469, "percentage": 36.82, "elapsed_time": "5:46:25", "remaining_time": "9:54:24"} +{"current_steps": 1585, "total_steps": 4291, "loss": 0.2358, "lr": 3.1812244672421293e-05, "epoch": 2.5861224489795918, "percentage": 36.94, "elapsed_time": "5:47:34", "remaining_time": "9:53:23"} +{"current_steps": 1590, "total_steps": 4291, "loss": 0.2257, "lr": 3.174648728199491e-05, "epoch": 2.5942857142857143, "percentage": 37.05, "elapsed_time": "5:48:42", "remaining_time": "9:52:21"} +{"current_steps": 1595, "total_steps": 4291, "loss": 0.2274, "lr": 3.168053546860234e-05, "epoch": 2.602448979591837, "percentage": 37.17, "elapsed_time": "5:49:41", "remaining_time": "9:51:04"} +{"current_steps": 1600, "total_steps": 4291, "loss": 0.2171, "lr": 3.1614390323850545e-05, "epoch": 2.6106122448979594, "percentage": 37.29, "elapsed_time": "5:50:42", "remaining_time": "9:49:51"} +{"current_steps": 1605, "total_steps": 4291, "loss": 0.227, "lr": 3.15480529425464e-05, "epoch": 2.6187755102040815, "percentage": 37.4, "elapsed_time": "5:52:37", "remaining_time": "9:50:08"} +{"current_steps": 1610, "total_steps": 4291, "loss": 0.2199, "lr": 3.148152442267863e-05, "epoch": 2.626938775510204, "percentage": 37.52, "elapsed_time": "5:53:41", "remaining_time": "9:48:57"} +{"current_steps": 1615, "total_steps": 4291, "loss": 0.23, "lr": 3.141480586539958e-05, "epoch": 2.6351020408163266, "percentage": 37.64, "elapsed_time": "5:54:41", "remaining_time": "9:47:43"} +{"current_steps": 1620, "total_steps": 4291, "loss": 0.2286, "lr": 3.1347898375007046e-05, "epoch": 2.643265306122449, "percentage": 37.75, "elapsed_time": "5:55:52", "remaining_time": "9:46:45"} +{"current_steps": 1625, "total_steps": 4291, "loss": 0.2354, "lr": 3.128080305892593e-05, "epoch": 2.6514285714285712, "percentage": 37.87, "elapsed_time": "5:56:58", "remaining_time": "9:45:40"} +{"current_steps": 1630, "total_steps": 4291, "loss": 0.2344, "lr": 3.121352102768998e-05, "epoch": 2.659591836734694, "percentage": 37.99, "elapsed_time": "5:57:57", "remaining_time": "9:44:22"} +{"current_steps": 1635, "total_steps": 4291, "loss": 0.2179, "lr": 3.114605339492335e-05, "epoch": 2.6677551020408163, "percentage": 38.1, "elapsed_time": "5:59:04", "remaining_time": "9:43:18"} +{"current_steps": 1640, "total_steps": 4291, "loss": 0.2297, "lr": 3.107840127732221e-05, "epoch": 2.675918367346939, "percentage": 38.22, "elapsed_time": "6:00:10", "remaining_time": "9:42:13"} +{"current_steps": 1645, "total_steps": 4291, "loss": 0.2196, "lr": 3.1010565794636236e-05, "epoch": 2.6840816326530614, "percentage": 38.34, "elapsed_time": "6:01:20", "remaining_time": "9:41:12"} +{"current_steps": 1650, "total_steps": 4291, "loss": 0.2262, "lr": 3.0942548069650085e-05, "epoch": 2.6922448979591835, "percentage": 38.45, "elapsed_time": "6:02:29", "remaining_time": "9:40:12"} +{"current_steps": 1655, "total_steps": 4291, "loss": 0.2254, "lr": 3.0874349228164836e-05, "epoch": 2.700408163265306, "percentage": 38.57, "elapsed_time": "6:03:42", "remaining_time": "9:39:17"} +{"current_steps": 1660, "total_steps": 4291, "loss": 0.2378, "lr": 3.08059703989793e-05, "epoch": 2.7085714285714286, "percentage": 38.69, "elapsed_time": "6:04:55", "remaining_time": "9:38:22"} +{"current_steps": 1665, "total_steps": 4291, "loss": 0.2153, "lr": 3.073741271387138e-05, "epoch": 2.716734693877551, "percentage": 38.8, "elapsed_time": "6:05:52", "remaining_time": "9:37:03"} +{"current_steps": 1670, "total_steps": 4291, "loss": 0.2284, "lr": 3.0668677307579344e-05, "epoch": 2.7248979591836733, "percentage": 38.92, "elapsed_time": "6:06:55", "remaining_time": "9:35:52"} +{"current_steps": 1675, "total_steps": 4291, "loss": 0.235, "lr": 3.059976531778301e-05, "epoch": 2.733061224489796, "percentage": 39.04, "elapsed_time": "6:07:58", "remaining_time": "9:34:41"} +{"current_steps": 1680, "total_steps": 4291, "loss": 0.2242, "lr": 3.0530677885084946e-05, "epoch": 2.7412244897959184, "percentage": 39.15, "elapsed_time": "6:09:03", "remaining_time": "9:33:35"} +{"current_steps": 1685, "total_steps": 4291, "loss": 0.214, "lr": 3.0461416152991555e-05, "epoch": 2.749387755102041, "percentage": 39.27, "elapsed_time": "6:10:14", "remaining_time": "9:32:37"} +{"current_steps": 1690, "total_steps": 4291, "loss": 0.2332, "lr": 3.0391981267894195e-05, "epoch": 2.7575510204081635, "percentage": 39.38, "elapsed_time": "6:11:24", "remaining_time": "9:31:37"} +{"current_steps": 1695, "total_steps": 4291, "loss": 0.234, "lr": 3.0322374379050166e-05, "epoch": 2.7657142857142856, "percentage": 39.5, "elapsed_time": "6:12:27", "remaining_time": "9:30:26"} +{"current_steps": 1700, "total_steps": 4291, "loss": 0.2336, "lr": 3.0252596638563717e-05, "epoch": 2.773877551020408, "percentage": 39.62, "elapsed_time": "6:13:32", "remaining_time": "9:29:19"} +{"current_steps": 1705, "total_steps": 4291, "loss": 0.2263, "lr": 3.0182649201366944e-05, "epoch": 2.7820408163265307, "percentage": 39.73, "elapsed_time": "6:14:41", "remaining_time": "9:28:18"} +{"current_steps": 1710, "total_steps": 4291, "loss": 0.2095, "lr": 3.0112533225200694e-05, "epoch": 2.790204081632653, "percentage": 39.85, "elapsed_time": "6:15:39", "remaining_time": "9:27:00"} +{"current_steps": 1715, "total_steps": 4291, "loss": 0.2282, "lr": 3.0042249870595424e-05, "epoch": 2.7983673469387753, "percentage": 39.97, "elapsed_time": "6:16:33", "remaining_time": "9:25:36"} +{"current_steps": 1720, "total_steps": 4291, "loss": 0.2358, "lr": 2.9971800300851947e-05, "epoch": 2.806530612244898, "percentage": 40.08, "elapsed_time": "6:17:32", "remaining_time": "9:24:19"} +{"current_steps": 1725, "total_steps": 4291, "loss": 0.2224, "lr": 2.9901185682022208e-05, "epoch": 2.8146938775510204, "percentage": 40.2, "elapsed_time": "6:18:42", "remaining_time": "9:23:19"} +{"current_steps": 1730, "total_steps": 4291, "loss": 0.2267, "lr": 2.9830407182889966e-05, "epoch": 2.822857142857143, "percentage": 40.32, "elapsed_time": "6:19:43", "remaining_time": "9:22:06"} +{"current_steps": 1735, "total_steps": 4291, "loss": 0.2304, "lr": 2.9759465974951473e-05, "epoch": 2.8310204081632655, "percentage": 40.43, "elapsed_time": "6:20:56", "remaining_time": "9:21:11"} +{"current_steps": 1740, "total_steps": 4291, "loss": 0.2322, "lr": 2.9688363232396056e-05, "epoch": 2.8391836734693876, "percentage": 40.55, "elapsed_time": "6:21:58", "remaining_time": "9:20:00"} +{"current_steps": 1745, "total_steps": 4291, "loss": 0.2453, "lr": 2.9617100132086705e-05, "epoch": 2.84734693877551, "percentage": 40.67, "elapsed_time": "6:23:01", "remaining_time": "9:18:50"} +{"current_steps": 1750, "total_steps": 4291, "loss": 0.219, "lr": 2.9545677853540584e-05, "epoch": 2.8555102040816327, "percentage": 40.78, "elapsed_time": "6:24:10", "remaining_time": "9:17:49"} +{"current_steps": 1755, "total_steps": 4291, "loss": 0.2268, "lr": 2.9474097578909507e-05, "epoch": 2.8636734693877552, "percentage": 40.9, "elapsed_time": "6:25:22", "remaining_time": "9:16:51"} +{"current_steps": 1760, "total_steps": 4291, "loss": 0.2207, "lr": 2.9402360492960377e-05, "epoch": 2.8718367346938773, "percentage": 41.02, "elapsed_time": "6:26:30", "remaining_time": "9:15:49"} +{"current_steps": 1765, "total_steps": 4291, "loss": 0.2261, "lr": 2.933046778305556e-05, "epoch": 2.88, "percentage": 41.13, "elapsed_time": "6:27:37", "remaining_time": "9:14:45"} +{"current_steps": 1770, "total_steps": 4291, "loss": 0.2329, "lr": 2.9258420639133253e-05, "epoch": 2.8881632653061224, "percentage": 41.25, "elapsed_time": "6:28:42", "remaining_time": "9:13:38"} +{"current_steps": 1775, "total_steps": 4291, "loss": 0.2305, "lr": 2.918622025368779e-05, "epoch": 2.896326530612245, "percentage": 41.37, "elapsed_time": "6:29:53", "remaining_time": "9:12:39"} +{"current_steps": 1780, "total_steps": 4291, "loss": 0.2261, "lr": 2.9113867821749883e-05, "epoch": 2.9044897959183675, "percentage": 41.48, "elapsed_time": "6:31:05", "remaining_time": "9:11:41"} +{"current_steps": 1785, "total_steps": 4291, "loss": 0.2202, "lr": 2.9041364540866855e-05, "epoch": 2.9126530612244896, "percentage": 41.6, "elapsed_time": "6:32:04", "remaining_time": "9:10:26"} +{"current_steps": 1790, "total_steps": 4291, "loss": 0.2316, "lr": 2.8968711611082825e-05, "epoch": 2.920816326530612, "percentage": 41.72, "elapsed_time": "6:33:12", "remaining_time": "9:09:22"} +{"current_steps": 1795, "total_steps": 4291, "loss": 0.231, "lr": 2.8895910234918828e-05, "epoch": 2.9289795918367347, "percentage": 41.83, "elapsed_time": "6:34:03", "remaining_time": "9:07:57"} +{"current_steps": 1800, "total_steps": 4291, "loss": 0.2268, "lr": 2.882296161735293e-05, "epoch": 2.9371428571428573, "percentage": 41.95, "elapsed_time": "6:35:06", "remaining_time": "9:06:46"} +{"current_steps": 1805, "total_steps": 4291, "loss": 0.2238, "lr": 2.874986696580028e-05, "epoch": 2.9453061224489794, "percentage": 42.06, "elapsed_time": "6:37:13", "remaining_time": "9:07:05"} +{"current_steps": 1810, "total_steps": 4291, "loss": 0.2331, "lr": 2.86766274900931e-05, "epoch": 2.953469387755102, "percentage": 42.18, "elapsed_time": "6:38:11", "remaining_time": "9:05:48"} +{"current_steps": 1815, "total_steps": 4291, "loss": 0.228, "lr": 2.8603244402460703e-05, "epoch": 2.9616326530612245, "percentage": 42.3, "elapsed_time": "6:39:15", "remaining_time": "9:04:40"} +{"current_steps": 1820, "total_steps": 4291, "loss": 0.2218, "lr": 2.8529718917509393e-05, "epoch": 2.969795918367347, "percentage": 42.41, "elapsed_time": "6:40:24", "remaining_time": "9:03:38"} +{"current_steps": 1825, "total_steps": 4291, "loss": 0.2212, "lr": 2.8456052252202385e-05, "epoch": 2.9779591836734696, "percentage": 42.53, "elapsed_time": "6:41:28", "remaining_time": "9:02:29"} +{"current_steps": 1830, "total_steps": 4291, "loss": 0.2297, "lr": 2.838224562583965e-05, "epoch": 2.9861224489795917, "percentage": 42.65, "elapsed_time": "6:42:36", "remaining_time": "9:01:25"} +{"current_steps": 1835, "total_steps": 4291, "loss": 0.2321, "lr": 2.8308300260037734e-05, "epoch": 2.994285714285714, "percentage": 42.76, "elapsed_time": "6:43:33", "remaining_time": "9:00:07"} +{"current_steps": 1840, "total_steps": 4291, "loss": 0.2193, "lr": 2.8234217378709533e-05, "epoch": 3.0016326530612245, "percentage": 42.88, "elapsed_time": "6:44:35", "remaining_time": "8:58:56"} +{"current_steps": 1845, "total_steps": 4291, "loss": 0.21, "lr": 2.8159998208044054e-05, "epoch": 3.009795918367347, "percentage": 43.0, "elapsed_time": "6:45:38", "remaining_time": "8:57:47"} +{"current_steps": 1850, "total_steps": 4291, "loss": 0.2094, "lr": 2.80856439764861e-05, "epoch": 3.0179591836734696, "percentage": 43.11, "elapsed_time": "6:46:52", "remaining_time": "8:56:51"} +{"current_steps": 1855, "total_steps": 4291, "loss": 0.1975, "lr": 2.8011155914715966e-05, "epoch": 3.0261224489795917, "percentage": 43.23, "elapsed_time": "6:47:53", "remaining_time": "8:55:39"} +{"current_steps": 1860, "total_steps": 4291, "loss": 0.2058, "lr": 2.7936535255629012e-05, "epoch": 3.0342857142857143, "percentage": 43.35, "elapsed_time": "6:48:52", "remaining_time": "8:54:23"} +{"current_steps": 1865, "total_steps": 4291, "loss": 0.2012, "lr": 2.786178323431534e-05, "epoch": 3.042448979591837, "percentage": 43.46, "elapsed_time": "6:49:46", "remaining_time": "8:53:02"} +{"current_steps": 1870, "total_steps": 4291, "loss": 0.2073, "lr": 2.7786901088039252e-05, "epoch": 3.0506122448979593, "percentage": 43.58, "elapsed_time": "6:50:52", "remaining_time": "8:51:56"} +{"current_steps": 1875, "total_steps": 4291, "loss": 0.2011, "lr": 2.771189005621887e-05, "epoch": 3.0587755102040814, "percentage": 43.7, "elapsed_time": "6:51:59", "remaining_time": "8:50:51"} +{"current_steps": 1880, "total_steps": 4291, "loss": 0.2192, "lr": 2.7636751380405548e-05, "epoch": 3.066938775510204, "percentage": 43.81, "elapsed_time": "6:53:03", "remaining_time": "8:49:43"} +{"current_steps": 1885, "total_steps": 4291, "loss": 0.2113, "lr": 2.7561486304263368e-05, "epoch": 3.0751020408163265, "percentage": 43.93, "elapsed_time": "6:54:13", "remaining_time": "8:48:42"} +{"current_steps": 1890, "total_steps": 4291, "loss": 0.2043, "lr": 2.7486096073548525e-05, "epoch": 3.083265306122449, "percentage": 44.05, "elapsed_time": "6:55:11", "remaining_time": "8:47:26"} +{"current_steps": 1895, "total_steps": 4291, "loss": 0.2124, "lr": 2.7410581936088724e-05, "epoch": 3.0914285714285716, "percentage": 44.16, "elapsed_time": "6:56:14", "remaining_time": "8:46:17"} +{"current_steps": 1900, "total_steps": 4291, "loss": 0.2096, "lr": 2.733494514176253e-05, "epoch": 3.0995918367346937, "percentage": 44.28, "elapsed_time": "6:57:22", "remaining_time": "8:45:14"} +{"current_steps": 1905, "total_steps": 4291, "loss": 0.2004, "lr": 2.7259186942478656e-05, "epoch": 3.1077551020408163, "percentage": 44.4, "elapsed_time": "6:58:29", "remaining_time": "8:44:09"} +{"current_steps": 1910, "total_steps": 4291, "loss": 0.2042, "lr": 2.718330859215528e-05, "epoch": 3.115918367346939, "percentage": 44.51, "elapsed_time": "6:59:29", "remaining_time": "8:42:56"} +{"current_steps": 1915, "total_steps": 4291, "loss": 0.2063, "lr": 2.710731134669926e-05, "epoch": 3.1240816326530614, "percentage": 44.63, "elapsed_time": "7:00:33", "remaining_time": "8:41:47"} +{"current_steps": 1920, "total_steps": 4291, "loss": 0.2146, "lr": 2.7031196463985355e-05, "epoch": 3.1322448979591835, "percentage": 44.74, "elapsed_time": "7:01:33", "remaining_time": "8:40:34"} +{"current_steps": 1925, "total_steps": 4291, "loss": 0.2118, "lr": 2.69549652038354e-05, "epoch": 3.140408163265306, "percentage": 44.86, "elapsed_time": "7:02:45", "remaining_time": "8:39:36"} +{"current_steps": 1930, "total_steps": 4291, "loss": 0.201, "lr": 2.687861882799748e-05, "epoch": 3.1485714285714286, "percentage": 44.98, "elapsed_time": "7:04:00", "remaining_time": "8:38:41"} +{"current_steps": 1935, "total_steps": 4291, "loss": 0.2129, "lr": 2.6802158600125022e-05, "epoch": 3.156734693877551, "percentage": 45.09, "elapsed_time": "7:05:05", "remaining_time": "8:37:34"} +{"current_steps": 1940, "total_steps": 4291, "loss": 0.2002, "lr": 2.6725585785755873e-05, "epoch": 3.1648979591836737, "percentage": 45.21, "elapsed_time": "7:06:12", "remaining_time": "8:36:29"} +{"current_steps": 1945, "total_steps": 4291, "loss": 0.2059, "lr": 2.6648901652291366e-05, "epoch": 3.1730612244897958, "percentage": 45.33, "elapsed_time": "7:07:22", "remaining_time": "8:35:29"} +{"current_steps": 1950, "total_steps": 4291, "loss": 0.1962, "lr": 2.657210746897535e-05, "epoch": 3.1812244897959183, "percentage": 45.44, "elapsed_time": "7:08:29", "remaining_time": "8:34:25"} +{"current_steps": 1955, "total_steps": 4291, "loss": 0.2031, "lr": 2.649520450687317e-05, "epoch": 3.189387755102041, "percentage": 45.56, "elapsed_time": "7:09:35", "remaining_time": "8:33:19"} +{"current_steps": 1960, "total_steps": 4291, "loss": 0.2058, "lr": 2.6418194038850634e-05, "epoch": 3.1975510204081634, "percentage": 45.68, "elapsed_time": "7:10:40", "remaining_time": "8:32:11"} +{"current_steps": 1965, "total_steps": 4291, "loss": 0.1986, "lr": 2.6341077339552938e-05, "epoch": 3.2057142857142855, "percentage": 45.79, "elapsed_time": "7:11:36", "remaining_time": "8:30:54"} +{"current_steps": 1970, "total_steps": 4291, "loss": 0.1889, "lr": 2.6263855685383585e-05, "epoch": 3.213877551020408, "percentage": 45.91, "elapsed_time": "7:12:39", "remaining_time": "8:29:44"} +{"current_steps": 1975, "total_steps": 4291, "loss": 0.2088, "lr": 2.6186530354483224e-05, "epoch": 3.2220408163265306, "percentage": 46.03, "elapsed_time": "7:13:41", "remaining_time": "8:28:34"} +{"current_steps": 1980, "total_steps": 4291, "loss": 0.2026, "lr": 2.6109102626708547e-05, "epoch": 3.230204081632653, "percentage": 46.14, "elapsed_time": "7:14:46", "remaining_time": "8:27:27"} +{"current_steps": 1985, "total_steps": 4291, "loss": 0.203, "lr": 2.6031573783611053e-05, "epoch": 3.2383673469387757, "percentage": 46.26, "elapsed_time": "7:16:02", "remaining_time": "8:26:33"} +{"current_steps": 1990, "total_steps": 4291, "loss": 0.2118, "lr": 2.5953945108415883e-05, "epoch": 3.246530612244898, "percentage": 46.38, "elapsed_time": "7:17:01", "remaining_time": "8:25:19"} +{"current_steps": 1995, "total_steps": 4291, "loss": 0.2243, "lr": 2.5876217886000526e-05, "epoch": 3.2546938775510204, "percentage": 46.49, "elapsed_time": "7:18:01", "remaining_time": "8:24:07"} +{"current_steps": 2000, "total_steps": 4291, "loss": 0.2172, "lr": 2.5798393402873614e-05, "epoch": 3.262857142857143, "percentage": 46.61, "elapsed_time": "7:19:00", "remaining_time": "8:22:52"} +{"current_steps": 2005, "total_steps": 4291, "loss": 0.201, "lr": 2.5720472947153582e-05, "epoch": 3.2710204081632654, "percentage": 46.73, "elapsed_time": "7:20:58", "remaining_time": "8:22:46"} +{"current_steps": 2010, "total_steps": 4291, "loss": 0.2076, "lr": 2.5642457808547365e-05, "epoch": 3.2791836734693875, "percentage": 46.84, "elapsed_time": "7:22:04", "remaining_time": "8:21:40"} +{"current_steps": 2015, "total_steps": 4291, "loss": 0.2099, "lr": 2.5564349278329056e-05, "epoch": 3.28734693877551, "percentage": 46.96, "elapsed_time": "7:23:02", "remaining_time": "8:20:26"} +{"current_steps": 2020, "total_steps": 4291, "loss": 0.2014, "lr": 2.5486148649318514e-05, "epoch": 3.2955102040816326, "percentage": 47.08, "elapsed_time": "7:24:09", "remaining_time": "8:19:21"} +{"current_steps": 2025, "total_steps": 4291, "loss": 0.1946, "lr": 2.540785721586e-05, "epoch": 3.303673469387755, "percentage": 47.19, "elapsed_time": "7:25:06", "remaining_time": "8:18:04"} +{"current_steps": 2030, "total_steps": 4291, "loss": 0.2151, "lr": 2.53294762738007e-05, "epoch": 3.3118367346938777, "percentage": 47.31, "elapsed_time": "7:26:12", "remaining_time": "8:16:58"} +{"current_steps": 2035, "total_steps": 4291, "loss": 0.2101, "lr": 2.5251007120469347e-05, "epoch": 3.32, "percentage": 47.42, "elapsed_time": "7:27:21", "remaining_time": "8:15:56"} +{"current_steps": 2040, "total_steps": 4291, "loss": 0.2121, "lr": 2.5172451054654685e-05, "epoch": 3.3281632653061224, "percentage": 47.54, "elapsed_time": "7:28:23", "remaining_time": "8:14:46"} +{"current_steps": 2045, "total_steps": 4291, "loss": 0.2177, "lr": 2.5093809376584006e-05, "epoch": 3.336326530612245, "percentage": 47.66, "elapsed_time": "7:29:22", "remaining_time": "8:13:32"} +{"current_steps": 2050, "total_steps": 4291, "loss": 0.2069, "lr": 2.501508338790162e-05, "epoch": 3.3444897959183675, "percentage": 47.77, "elapsed_time": "7:30:27", "remaining_time": "8:12:26"} +{"current_steps": 2055, "total_steps": 4291, "loss": 0.2135, "lr": 2.49362743916473e-05, "epoch": 3.3526530612244896, "percentage": 47.89, "elapsed_time": "7:31:24", "remaining_time": "8:11:09"} +{"current_steps": 2060, "total_steps": 4291, "loss": 0.2021, "lr": 2.4857383692234742e-05, "epoch": 3.360816326530612, "percentage": 48.01, "elapsed_time": "7:32:30", "remaining_time": "8:10:04"} +{"current_steps": 2065, "total_steps": 4291, "loss": 0.2085, "lr": 2.4778412595429955e-05, "epoch": 3.3689795918367347, "percentage": 48.12, "elapsed_time": "7:33:28", "remaining_time": "8:08:50"} +{"current_steps": 2070, "total_steps": 4291, "loss": 0.2, "lr": 2.4699362408329646e-05, "epoch": 3.3771428571428572, "percentage": 48.24, "elapsed_time": "7:34:35", "remaining_time": "8:07:45"} +{"current_steps": 2075, "total_steps": 4291, "loss": 0.2168, "lr": 2.46202344393396e-05, "epoch": 3.3853061224489798, "percentage": 48.36, "elapsed_time": "7:35:31", "remaining_time": "8:06:29"} +{"current_steps": 2080, "total_steps": 4291, "loss": 0.1992, "lr": 2.4541029998153e-05, "epoch": 3.393469387755102, "percentage": 48.47, "elapsed_time": "7:36:37", "remaining_time": "8:05:22"} +{"current_steps": 2085, "total_steps": 4291, "loss": 0.2018, "lr": 2.446175039572878e-05, "epoch": 3.4016326530612244, "percentage": 48.59, "elapsed_time": "7:37:44", "remaining_time": "8:04:18"} +{"current_steps": 2090, "total_steps": 4291, "loss": 0.2076, "lr": 2.4382396944269912e-05, "epoch": 3.409795918367347, "percentage": 48.71, "elapsed_time": "7:38:42", "remaining_time": "8:03:04"} +{"current_steps": 2095, "total_steps": 4291, "loss": 0.2141, "lr": 2.430297095720168e-05, "epoch": 3.4179591836734695, "percentage": 48.82, "elapsed_time": "7:39:45", "remaining_time": "8:01:55"} +{"current_steps": 2100, "total_steps": 4291, "loss": 0.1987, "lr": 2.4223473749149937e-05, "epoch": 3.4261224489795916, "percentage": 48.94, "elapsed_time": "7:40:59", "remaining_time": "8:00:58"} +{"current_steps": 2105, "total_steps": 4291, "loss": 0.2079, "lr": 2.4143906635919385e-05, "epoch": 3.434285714285714, "percentage": 49.06, "elapsed_time": "7:42:00", "remaining_time": "7:59:47"} +{"current_steps": 2110, "total_steps": 4291, "loss": 0.2105, "lr": 2.4064270934471735e-05, "epoch": 3.4424489795918367, "percentage": 49.17, "elapsed_time": "7:43:00", "remaining_time": "7:58:35"} +{"current_steps": 2115, "total_steps": 4291, "loss": 0.2021, "lr": 2.3984567962903968e-05, "epoch": 3.4506122448979593, "percentage": 49.29, "elapsed_time": "7:44:15", "remaining_time": "7:57:38"} +{"current_steps": 2120, "total_steps": 4291, "loss": 0.2074, "lr": 2.3904799040426476e-05, "epoch": 3.458775510204082, "percentage": 49.41, "elapsed_time": "7:45:18", "remaining_time": "7:56:29"} +{"current_steps": 2125, "total_steps": 4291, "loss": 0.2102, "lr": 2.3824965487341247e-05, "epoch": 3.466938775510204, "percentage": 49.52, "elapsed_time": "7:46:29", "remaining_time": "7:55:29"} +{"current_steps": 2130, "total_steps": 4291, "loss": 0.2144, "lr": 2.3745068625020008e-05, "epoch": 3.4751020408163265, "percentage": 49.64, "elapsed_time": "7:47:30", "remaining_time": "7:54:18"} +{"current_steps": 2135, "total_steps": 4291, "loss": 0.2071, "lr": 2.3665109775882352e-05, "epoch": 3.483265306122449, "percentage": 49.76, "elapsed_time": "7:48:33", "remaining_time": "7:53:10"} +{"current_steps": 2140, "total_steps": 4291, "loss": 0.2106, "lr": 2.3585090263373868e-05, "epoch": 3.4914285714285715, "percentage": 49.87, "elapsed_time": "7:49:40", "remaining_time": "7:52:05"} +{"current_steps": 2145, "total_steps": 4291, "loss": 0.2066, "lr": 2.350501141194419e-05, "epoch": 3.4995918367346937, "percentage": 49.99, "elapsed_time": "7:50:35", "remaining_time": "7:50:48"} +{"current_steps": 2150, "total_steps": 4291, "loss": 0.2127, "lr": 2.3424874547025125e-05, "epoch": 3.507755102040816, "percentage": 50.1, "elapsed_time": "7:51:32", "remaining_time": "7:49:34"} +{"current_steps": 2155, "total_steps": 4291, "loss": 0.2134, "lr": 2.33446809950087e-05, "epoch": 3.5159183673469387, "percentage": 50.22, "elapsed_time": "7:52:37", "remaining_time": "7:48:27"} +{"current_steps": 2160, "total_steps": 4291, "loss": 0.2091, "lr": 2.326443208322519e-05, "epoch": 3.5240816326530613, "percentage": 50.34, "elapsed_time": "7:53:45", "remaining_time": "7:47:23"} +{"current_steps": 2165, "total_steps": 4291, "loss": 0.2054, "lr": 2.3184129139921162e-05, "epoch": 3.532244897959184, "percentage": 50.45, "elapsed_time": "7:54:49", "remaining_time": "7:46:16"} +{"current_steps": 2170, "total_steps": 4291, "loss": 0.1912, "lr": 2.3103773494237504e-05, "epoch": 3.5404081632653064, "percentage": 50.57, "elapsed_time": "7:55:58", "remaining_time": "7:45:13"} +{"current_steps": 2175, "total_steps": 4291, "loss": 0.2069, "lr": 2.3023366476187402e-05, "epoch": 3.5485714285714285, "percentage": 50.69, "elapsed_time": "7:57:13", "remaining_time": "7:44:16"} +{"current_steps": 2180, "total_steps": 4291, "loss": 0.1996, "lr": 2.2942909416634326e-05, "epoch": 3.556734693877551, "percentage": 50.8, "elapsed_time": "7:58:19", "remaining_time": "7:43:11"} +{"current_steps": 2185, "total_steps": 4291, "loss": 0.2169, "lr": 2.2862403647270033e-05, "epoch": 3.5648979591836736, "percentage": 50.92, "elapsed_time": "7:59:29", "remaining_time": "7:42:09"} +{"current_steps": 2190, "total_steps": 4291, "loss": 0.2071, "lr": 2.2781850500592487e-05, "epoch": 3.5730612244897957, "percentage": 51.04, "elapsed_time": "8:00:44", "remaining_time": "7:41:11"} +{"current_steps": 2195, "total_steps": 4291, "loss": 0.2055, "lr": 2.2701251309883837e-05, "epoch": 3.5812244897959182, "percentage": 51.15, "elapsed_time": "8:01:41", "remaining_time": "7:39:58"} +{"current_steps": 2200, "total_steps": 4291, "loss": 0.2104, "lr": 2.2620607409188326e-05, "epoch": 3.5893877551020408, "percentage": 51.27, "elapsed_time": "8:02:34", "remaining_time": "7:38:40"} +{"current_steps": 2205, "total_steps": 4291, "loss": 0.2077, "lr": 2.2539920133290208e-05, "epoch": 3.5975510204081633, "percentage": 51.39, "elapsed_time": "8:04:44", "remaining_time": "7:38:35"} +{"current_steps": 2210, "total_steps": 4291, "loss": 0.2064, "lr": 2.245919081769168e-05, "epoch": 3.605714285714286, "percentage": 51.5, "elapsed_time": "8:05:54", "remaining_time": "7:37:32"} +{"current_steps": 2215, "total_steps": 4291, "loss": 0.2085, "lr": 2.2378420798590764e-05, "epoch": 3.6138775510204084, "percentage": 51.62, "elapsed_time": "8:06:59", "remaining_time": "7:36:25"} +{"current_steps": 2220, "total_steps": 4291, "loss": 0.213, "lr": 2.2297611412859178e-05, "epoch": 3.6220408163265305, "percentage": 51.74, "elapsed_time": "8:08:08", "remaining_time": "7:35:23"} +{"current_steps": 2225, "total_steps": 4291, "loss": 0.2065, "lr": 2.2216763998020222e-05, "epoch": 3.630204081632653, "percentage": 51.85, "elapsed_time": "8:09:19", "remaining_time": "7:34:21"} +{"current_steps": 2230, "total_steps": 4291, "loss": 0.1884, "lr": 2.2135879892226643e-05, "epoch": 3.6383673469387756, "percentage": 51.97, "elapsed_time": "8:10:16", "remaining_time": "7:33:07"} +{"current_steps": 2235, "total_steps": 4291, "loss": 0.2129, "lr": 2.205496043423849e-05, "epoch": 3.6465306122448977, "percentage": 52.09, "elapsed_time": "8:11:22", "remaining_time": "7:32:01"} +{"current_steps": 2240, "total_steps": 4291, "loss": 0.2045, "lr": 2.197400696340091e-05, "epoch": 3.6546938775510203, "percentage": 52.2, "elapsed_time": "8:12:22", "remaining_time": "7:30:50"} +{"current_steps": 2245, "total_steps": 4291, "loss": 0.2147, "lr": 2.189302081962207e-05, "epoch": 3.662857142857143, "percentage": 52.32, "elapsed_time": "8:13:27", "remaining_time": "7:29:42"} +{"current_steps": 2250, "total_steps": 4291, "loss": 0.2095, "lr": 2.1812003343350886e-05, "epoch": 3.6710204081632654, "percentage": 52.44, "elapsed_time": "8:14:29", "remaining_time": "7:28:33"} +{"current_steps": 2255, "total_steps": 4291, "loss": 0.1989, "lr": 2.1730955875554894e-05, "epoch": 3.679183673469388, "percentage": 52.55, "elapsed_time": "8:15:27", "remaining_time": "7:27:20"} +{"current_steps": 2260, "total_steps": 4291, "loss": 0.2041, "lr": 2.164987975769803e-05, "epoch": 3.6873469387755105, "percentage": 52.67, "elapsed_time": "8:16:28", "remaining_time": "7:26:10"} +{"current_steps": 2265, "total_steps": 4291, "loss": 0.2108, "lr": 2.1568776331718428e-05, "epoch": 3.6955102040816326, "percentage": 52.78, "elapsed_time": "8:17:39", "remaining_time": "7:25:08"} +{"current_steps": 2270, "total_steps": 4291, "loss": 0.2105, "lr": 2.148764694000624e-05, "epoch": 3.703673469387755, "percentage": 52.9, "elapsed_time": "8:18:32", "remaining_time": "7:23:51"} +{"current_steps": 2275, "total_steps": 4291, "loss": 0.1967, "lr": 2.1406492925381358e-05, "epoch": 3.7118367346938776, "percentage": 53.02, "elapsed_time": "8:19:34", "remaining_time": "7:22:42"} +{"current_steps": 2280, "total_steps": 4291, "loss": 0.1972, "lr": 2.1325315631071257e-05, "epoch": 3.7199999999999998, "percentage": 53.13, "elapsed_time": "8:20:43", "remaining_time": "7:21:38"} +{"current_steps": 2285, "total_steps": 4291, "loss": 0.2078, "lr": 2.124411640068869e-05, "epoch": 3.7281632653061223, "percentage": 53.25, "elapsed_time": "8:21:40", "remaining_time": "7:20:24"} +{"current_steps": 2290, "total_steps": 4291, "loss": 0.219, "lr": 2.1162896578209517e-05, "epoch": 3.736326530612245, "percentage": 53.37, "elapsed_time": "8:22:38", "remaining_time": "7:19:12"} +{"current_steps": 2295, "total_steps": 4291, "loss": 0.2048, "lr": 2.1081657507950423e-05, "epoch": 3.7444897959183674, "percentage": 53.48, "elapsed_time": "8:23:50", "remaining_time": "7:18:11"} +{"current_steps": 2300, "total_steps": 4291, "loss": 0.2145, "lr": 2.100040053454666e-05, "epoch": 3.75265306122449, "percentage": 53.6, "elapsed_time": "8:24:45", "remaining_time": "7:16:56"} +{"current_steps": 2305, "total_steps": 4291, "loss": 0.2033, "lr": 2.091912700292983e-05, "epoch": 3.7608163265306125, "percentage": 53.72, "elapsed_time": "8:25:50", "remaining_time": "7:15:50"} +{"current_steps": 2310, "total_steps": 4291, "loss": 0.2076, "lr": 2.083783825830557e-05, "epoch": 3.7689795918367346, "percentage": 53.83, "elapsed_time": "8:26:50", "remaining_time": "7:14:39"} +{"current_steps": 2315, "total_steps": 4291, "loss": 0.2055, "lr": 2.075653564613134e-05, "epoch": 3.777142857142857, "percentage": 53.95, "elapsed_time": "8:27:59", "remaining_time": "7:13:36"} +{"current_steps": 2320, "total_steps": 4291, "loss": 0.2042, "lr": 2.0675220512094124e-05, "epoch": 3.7853061224489797, "percentage": 54.07, "elapsed_time": "8:29:06", "remaining_time": "7:12:31"} +{"current_steps": 2325, "total_steps": 4291, "loss": 0.2156, "lr": 2.059389420208816e-05, "epoch": 3.793469387755102, "percentage": 54.18, "elapsed_time": "8:30:16", "remaining_time": "7:11:28"} +{"current_steps": 2330, "total_steps": 4291, "loss": 0.2117, "lr": 2.0512558062192655e-05, "epoch": 3.8016326530612243, "percentage": 54.3, "elapsed_time": "8:31:29", "remaining_time": "7:10:29"} +{"current_steps": 2335, "total_steps": 4291, "loss": 0.209, "lr": 2.043121343864954e-05, "epoch": 3.809795918367347, "percentage": 54.42, "elapsed_time": "8:32:37", "remaining_time": "7:09:24"} +{"current_steps": 2340, "total_steps": 4291, "loss": 0.2123, "lr": 2.0349861677841148e-05, "epoch": 3.8179591836734694, "percentage": 54.53, "elapsed_time": "8:33:33", "remaining_time": "7:08:10"} +{"current_steps": 2345, "total_steps": 4291, "loss": 0.2062, "lr": 2.0268504126267952e-05, "epoch": 3.826122448979592, "percentage": 54.65, "elapsed_time": "8:34:52", "remaining_time": "7:07:16"} +{"current_steps": 2350, "total_steps": 4291, "loss": 0.206, "lr": 2.0187142130526266e-05, "epoch": 3.8342857142857145, "percentage": 54.77, "elapsed_time": "8:35:59", "remaining_time": "7:06:11"} +{"current_steps": 2355, "total_steps": 4291, "loss": 0.2108, "lr": 2.0105777037285957e-05, "epoch": 3.8424489795918366, "percentage": 54.88, "elapsed_time": "8:37:00", "remaining_time": "7:05:01"} +{"current_steps": 2360, "total_steps": 4291, "loss": 0.2055, "lr": 2.0024410193268177e-05, "epoch": 3.850612244897959, "percentage": 55.0, "elapsed_time": "8:38:02", "remaining_time": "7:03:51"} +{"current_steps": 2365, "total_steps": 4291, "loss": 0.201, "lr": 1.994304294522304e-05, "epoch": 3.8587755102040817, "percentage": 55.12, "elapsed_time": "8:39:12", "remaining_time": "7:02:49"} +{"current_steps": 2370, "total_steps": 4291, "loss": 0.2133, "lr": 1.9861676639907358e-05, "epoch": 3.866938775510204, "percentage": 55.23, "elapsed_time": "8:40:17", "remaining_time": "7:01:43"} +{"current_steps": 2375, "total_steps": 4291, "loss": 0.2151, "lr": 1.978031262406233e-05, "epoch": 3.8751020408163264, "percentage": 55.35, "elapsed_time": "8:41:14", "remaining_time": "7:00:30"} +{"current_steps": 2380, "total_steps": 4291, "loss": 0.2147, "lr": 1.9698952244391273e-05, "epoch": 3.883265306122449, "percentage": 55.46, "elapsed_time": "8:42:13", "remaining_time": "6:59:18"} +{"current_steps": 2385, "total_steps": 4291, "loss": 0.2015, "lr": 1.9617596847537314e-05, "epoch": 3.8914285714285715, "percentage": 55.58, "elapsed_time": "8:43:24", "remaining_time": "6:58:17"} +{"current_steps": 2390, "total_steps": 4291, "loss": 0.216, "lr": 1.953624778006109e-05, "epoch": 3.899591836734694, "percentage": 55.7, "elapsed_time": "8:44:26", "remaining_time": "6:57:08"} +{"current_steps": 2395, "total_steps": 4291, "loss": 0.2018, "lr": 1.9454906388418507e-05, "epoch": 3.9077551020408166, "percentage": 55.81, "elapsed_time": "8:45:28", "remaining_time": "6:55:59"} +{"current_steps": 2400, "total_steps": 4291, "loss": 0.1993, "lr": 1.93735740189384e-05, "epoch": 3.9159183673469387, "percentage": 55.93, "elapsed_time": "8:46:33", "remaining_time": "6:54:52"} +{"current_steps": 2405, "total_steps": 4291, "loss": 0.1962, "lr": 1.929225201780029e-05, "epoch": 3.924081632653061, "percentage": 56.05, "elapsed_time": "8:48:47", "remaining_time": "6:54:40"} +{"current_steps": 2410, "total_steps": 4291, "loss": 0.2034, "lr": 1.9210941731012058e-05, "epoch": 3.9322448979591837, "percentage": 56.16, "elapsed_time": "8:49:49", "remaining_time": "6:53:31"} +{"current_steps": 2415, "total_steps": 4291, "loss": 0.2118, "lr": 1.9129644504387738e-05, "epoch": 3.940408163265306, "percentage": 56.28, "elapsed_time": "8:50:40", "remaining_time": "6:52:13"} +{"current_steps": 2420, "total_steps": 4291, "loss": 0.1873, "lr": 1.9048361683525155e-05, "epoch": 3.9485714285714284, "percentage": 56.4, "elapsed_time": "8:51:51", "remaining_time": "6:51:11"} +{"current_steps": 2425, "total_steps": 4291, "loss": 0.1998, "lr": 1.8967094613783726e-05, "epoch": 3.956734693877551, "percentage": 56.51, "elapsed_time": "8:53:04", "remaining_time": "6:50:11"} +{"current_steps": 2430, "total_steps": 4291, "loss": 0.1969, "lr": 1.888584464026213e-05, "epoch": 3.9648979591836735, "percentage": 56.63, "elapsed_time": "8:54:07", "remaining_time": "6:49:03"} +{"current_steps": 2435, "total_steps": 4291, "loss": 0.201, "lr": 1.8804613107776118e-05, "epoch": 3.973061224489796, "percentage": 56.75, "elapsed_time": "8:55:17", "remaining_time": "6:48:00"} +{"current_steps": 2440, "total_steps": 4291, "loss": 0.2127, "lr": 1.8723401360836178e-05, "epoch": 3.9812244897959186, "percentage": 56.86, "elapsed_time": "8:56:22", "remaining_time": "6:46:53"} +{"current_steps": 2445, "total_steps": 4291, "loss": 0.212, "lr": 1.8642210743625325e-05, "epoch": 3.9893877551020407, "percentage": 56.98, "elapsed_time": "8:57:30", "remaining_time": "6:45:49"} +{"current_steps": 2450, "total_steps": 4291, "loss": 0.2127, "lr": 1.8561042599976854e-05, "epoch": 3.9975510204081632, "percentage": 57.1, "elapsed_time": "8:58:32", "remaining_time": "6:44:40"} +{"current_steps": 2455, "total_steps": 4291, "loss": 0.1961, "lr": 1.8479898273352084e-05, "epoch": 4.0048979591836735, "percentage": 57.21, "elapsed_time": "8:59:36", "remaining_time": "6:43:32"} +{"current_steps": 2460, "total_steps": 4291, "loss": 0.1964, "lr": 1.8398779106818115e-05, "epoch": 4.013061224489796, "percentage": 57.33, "elapsed_time": "9:00:31", "remaining_time": "6:42:18"} +{"current_steps": 2465, "total_steps": 4291, "loss": 0.1966, "lr": 1.83176864430256e-05, "epoch": 4.021224489795919, "percentage": 57.45, "elapsed_time": "9:01:31", "remaining_time": "6:41:08"} +{"current_steps": 2470, "total_steps": 4291, "loss": 0.1912, "lr": 1.823662162418657e-05, "epoch": 4.029387755102041, "percentage": 57.56, "elapsed_time": "9:02:32", "remaining_time": "6:39:59"} +{"current_steps": 2475, "total_steps": 4291, "loss": 0.1975, "lr": 1.815558599205213e-05, "epoch": 4.037551020408163, "percentage": 57.68, "elapsed_time": "9:03:41", "remaining_time": "6:38:55"} +{"current_steps": 2480, "total_steps": 4291, "loss": 0.2029, "lr": 1.8074580887890317e-05, "epoch": 4.045714285714285, "percentage": 57.8, "elapsed_time": "9:04:47", "remaining_time": "6:37:49"} +{"current_steps": 2485, "total_steps": 4291, "loss": 0.1926, "lr": 1.79936076524639e-05, "epoch": 4.053877551020408, "percentage": 57.91, "elapsed_time": "9:05:46", "remaining_time": "6:36:38"} +{"current_steps": 2490, "total_steps": 4291, "loss": 0.1995, "lr": 1.791266762600815e-05, "epoch": 4.0620408163265305, "percentage": 58.03, "elapsed_time": "9:06:41", "remaining_time": "6:35:25"} +{"current_steps": 2495, "total_steps": 4291, "loss": 0.1856, "lr": 1.7831762148208672e-05, "epoch": 4.070204081632653, "percentage": 58.14, "elapsed_time": "9:07:36", "remaining_time": "6:34:11"} +{"current_steps": 2500, "total_steps": 4291, "loss": 0.1975, "lr": 1.775089255817925e-05, "epoch": 4.078367346938776, "percentage": 58.26, "elapsed_time": "9:08:50", "remaining_time": "6:33:11"} +{"current_steps": 2505, "total_steps": 4291, "loss": 0.2024, "lr": 1.7670060194439665e-05, "epoch": 4.086530612244898, "percentage": 58.38, "elapsed_time": "9:09:50", "remaining_time": "6:32:01"} +{"current_steps": 2510, "total_steps": 4291, "loss": 0.1854, "lr": 1.758926639489354e-05, "epoch": 4.094693877551021, "percentage": 58.49, "elapsed_time": "9:11:00", "remaining_time": "6:30:58"} +{"current_steps": 2515, "total_steps": 4291, "loss": 0.1932, "lr": 1.7508512496806202e-05, "epoch": 4.102857142857143, "percentage": 58.61, "elapsed_time": "9:11:57", "remaining_time": "6:29:46"} +{"current_steps": 2520, "total_steps": 4291, "loss": 0.1941, "lr": 1.7427799836782554e-05, "epoch": 4.111020408163265, "percentage": 58.73, "elapsed_time": "9:13:01", "remaining_time": "6:28:39"} +{"current_steps": 2525, "total_steps": 4291, "loss": 0.2046, "lr": 1.7347129750744923e-05, "epoch": 4.119183673469387, "percentage": 58.84, "elapsed_time": "9:14:07", "remaining_time": "6:27:33"} +{"current_steps": 2530, "total_steps": 4291, "loss": 0.1975, "lr": 1.7266503573910995e-05, "epoch": 4.12734693877551, "percentage": 58.96, "elapsed_time": "9:15:07", "remaining_time": "6:26:23"} +{"current_steps": 2535, "total_steps": 4291, "loss": 0.1773, "lr": 1.7185922640771657e-05, "epoch": 4.1355102040816325, "percentage": 59.08, "elapsed_time": "9:16:05", "remaining_time": "6:25:12"} +{"current_steps": 2540, "total_steps": 4291, "loss": 0.1867, "lr": 1.710538828506898e-05, "epoch": 4.143673469387755, "percentage": 59.19, "elapsed_time": "9:17:15", "remaining_time": "6:24:09"} +{"current_steps": 2545, "total_steps": 4291, "loss": 0.1936, "lr": 1.7024901839774074e-05, "epoch": 4.151836734693878, "percentage": 59.31, "elapsed_time": "9:18:21", "remaining_time": "6:23:03"} +{"current_steps": 2550, "total_steps": 4291, "loss": 0.1796, "lr": 1.6944464637065062e-05, "epoch": 4.16, "percentage": 59.43, "elapsed_time": "9:19:26", "remaining_time": "6:21:57"} +{"current_steps": 2555, "total_steps": 4291, "loss": 0.1828, "lr": 1.686407800830504e-05, "epoch": 4.168163265306123, "percentage": 59.54, "elapsed_time": "9:20:31", "remaining_time": "6:20:51"} +{"current_steps": 2560, "total_steps": 4291, "loss": 0.194, "lr": 1.678374328402001e-05, "epoch": 4.176326530612245, "percentage": 59.66, "elapsed_time": "9:21:32", "remaining_time": "6:19:41"} +{"current_steps": 2565, "total_steps": 4291, "loss": 0.1864, "lr": 1.6703461793876876e-05, "epoch": 4.184489795918367, "percentage": 59.78, "elapsed_time": "9:22:43", "remaining_time": "6:18:39"} +{"current_steps": 2570, "total_steps": 4291, "loss": 0.1895, "lr": 1.6623234866661425e-05, "epoch": 4.192653061224489, "percentage": 59.89, "elapsed_time": "9:23:57", "remaining_time": "6:17:38"} +{"current_steps": 2575, "total_steps": 4291, "loss": 0.191, "lr": 1.6543063830256365e-05, "epoch": 4.200816326530612, "percentage": 60.01, "elapsed_time": "9:25:04", "remaining_time": "6:16:34"} +{"current_steps": 2580, "total_steps": 4291, "loss": 0.1888, "lr": 1.6462950011619294e-05, "epoch": 4.2089795918367345, "percentage": 60.13, "elapsed_time": "9:26:08", "remaining_time": "6:15:27"} +{"current_steps": 2585, "total_steps": 4291, "loss": 0.1819, "lr": 1.638289473676079e-05, "epoch": 4.217142857142857, "percentage": 60.24, "elapsed_time": "9:27:12", "remaining_time": "6:14:19"} +{"current_steps": 2590, "total_steps": 4291, "loss": 0.1759, "lr": 1.6302899330722408e-05, "epoch": 4.22530612244898, "percentage": 60.36, "elapsed_time": "9:28:14", "remaining_time": "6:13:12"} +{"current_steps": 2595, "total_steps": 4291, "loss": 0.182, "lr": 1.622296511755482e-05, "epoch": 4.233469387755102, "percentage": 60.48, "elapsed_time": "9:29:21", "remaining_time": "6:12:06"} +{"current_steps": 2600, "total_steps": 4291, "loss": 0.1902, "lr": 1.6143093420295823e-05, "epoch": 4.241632653061225, "percentage": 60.59, "elapsed_time": "9:30:30", "remaining_time": "6:11:03"} +{"current_steps": 2605, "total_steps": 4291, "loss": 0.1892, "lr": 1.6063285560948492e-05, "epoch": 4.249795918367347, "percentage": 60.71, "elapsed_time": "9:32:38", "remaining_time": "6:10:37"} +{"current_steps": 2610, "total_steps": 4291, "loss": 0.1899, "lr": 1.5983542860459292e-05, "epoch": 4.257959183673469, "percentage": 60.82, "elapsed_time": "9:33:42", "remaining_time": "6:09:30"} +{"current_steps": 2615, "total_steps": 4291, "loss": 0.1843, "lr": 1.590386663869619e-05, "epoch": 4.2661224489795915, "percentage": 60.94, "elapsed_time": "9:34:46", "remaining_time": "6:08:23"} +{"current_steps": 2620, "total_steps": 4291, "loss": 0.1874, "lr": 1.5824258214426833e-05, "epoch": 4.274285714285714, "percentage": 61.06, "elapsed_time": "9:35:57", "remaining_time": "6:07:20"} +{"current_steps": 2625, "total_steps": 4291, "loss": 0.1972, "lr": 1.57447189052967e-05, "epoch": 4.282448979591837, "percentage": 61.17, "elapsed_time": "9:36:55", "remaining_time": "6:06:09"} +{"current_steps": 2630, "total_steps": 4291, "loss": 0.1807, "lr": 1.5665250027807326e-05, "epoch": 4.290612244897959, "percentage": 61.29, "elapsed_time": "9:37:58", "remaining_time": "6:05:01"} +{"current_steps": 2635, "total_steps": 4291, "loss": 0.1886, "lr": 1.558585289729448e-05, "epoch": 4.298775510204082, "percentage": 61.41, "elapsed_time": "9:39:03", "remaining_time": "6:03:54"} +{"current_steps": 2640, "total_steps": 4291, "loss": 0.1931, "lr": 1.55065288279064e-05, "epoch": 4.306938775510204, "percentage": 61.52, "elapsed_time": "9:40:05", "remaining_time": "6:02:46"} +{"current_steps": 2645, "total_steps": 4291, "loss": 0.195, "lr": 1.5427279132582055e-05, "epoch": 4.315102040816327, "percentage": 61.64, "elapsed_time": "9:41:22", "remaining_time": "6:01:47"} +{"current_steps": 2650, "total_steps": 4291, "loss": 0.1861, "lr": 1.5348105123029404e-05, "epoch": 4.323265306122449, "percentage": 61.76, "elapsed_time": "9:42:24", "remaining_time": "6:00:39"} +{"current_steps": 2655, "total_steps": 4291, "loss": 0.1949, "lr": 1.5269008109703676e-05, "epoch": 4.331428571428571, "percentage": 61.87, "elapsed_time": "9:43:17", "remaining_time": "5:59:25"} +{"current_steps": 2660, "total_steps": 4291, "loss": 0.1928, "lr": 1.5189989401785698e-05, "epoch": 4.3395918367346935, "percentage": 61.99, "elapsed_time": "9:44:21", "remaining_time": "5:58:18"} +{"current_steps": 2665, "total_steps": 4291, "loss": 0.1893, "lr": 1.5111050307160237e-05, "epoch": 4.347755102040816, "percentage": 62.11, "elapsed_time": "9:45:25", "remaining_time": "5:57:11"} +{"current_steps": 2670, "total_steps": 4291, "loss": 0.1784, "lr": 1.5032192132394305e-05, "epoch": 4.355918367346939, "percentage": 62.22, "elapsed_time": "9:46:32", "remaining_time": "5:56:05"} +{"current_steps": 2675, "total_steps": 4291, "loss": 0.1923, "lr": 1.4953416182715566e-05, "epoch": 4.364081632653061, "percentage": 62.34, "elapsed_time": "9:47:30", "remaining_time": "5:54:55"} +{"current_steps": 2680, "total_steps": 4291, "loss": 0.1946, "lr": 1.4874723761990751e-05, "epoch": 4.372244897959184, "percentage": 62.46, "elapsed_time": "9:48:39", "remaining_time": "5:53:50"} +{"current_steps": 2685, "total_steps": 4291, "loss": 0.1991, "lr": 1.4796116172704032e-05, "epoch": 4.380408163265306, "percentage": 62.57, "elapsed_time": "9:49:48", "remaining_time": "5:52:46"} +{"current_steps": 2690, "total_steps": 4291, "loss": 0.197, "lr": 1.4717594715935498e-05, "epoch": 4.388571428571429, "percentage": 62.69, "elapsed_time": "9:50:48", "remaining_time": "5:51:37"} +{"current_steps": 2695, "total_steps": 4291, "loss": 0.1954, "lr": 1.4639160691339593e-05, "epoch": 4.396734693877551, "percentage": 62.81, "elapsed_time": "9:51:50", "remaining_time": "5:50:29"} +{"current_steps": 2700, "total_steps": 4291, "loss": 0.1885, "lr": 1.456081539712365e-05, "epoch": 4.404897959183674, "percentage": 62.92, "elapsed_time": "9:52:56", "remaining_time": "5:49:23"} +{"current_steps": 2705, "total_steps": 4291, "loss": 0.187, "lr": 1.4482560130026342e-05, "epoch": 4.4130612244897955, "percentage": 63.04, "elapsed_time": "9:54:01", "remaining_time": "5:48:17"} +{"current_steps": 2710, "total_steps": 4291, "loss": 0.1916, "lr": 1.4404396185296268e-05, "epoch": 4.421224489795918, "percentage": 63.16, "elapsed_time": "9:55:07", "remaining_time": "5:47:11"} +{"current_steps": 2715, "total_steps": 4291, "loss": 0.1904, "lr": 1.4326324856670494e-05, "epoch": 4.429387755102041, "percentage": 63.27, "elapsed_time": "9:56:14", "remaining_time": "5:46:06"} +{"current_steps": 2720, "total_steps": 4291, "loss": 0.1908, "lr": 1.4248347436353147e-05, "epoch": 4.437551020408163, "percentage": 63.39, "elapsed_time": "9:57:24", "remaining_time": "5:45:02"} +{"current_steps": 2725, "total_steps": 4291, "loss": 0.1885, "lr": 1.4170465214994011e-05, "epoch": 4.445714285714286, "percentage": 63.51, "elapsed_time": "9:58:30", "remaining_time": "5:43:57"} +{"current_steps": 2730, "total_steps": 4291, "loss": 0.192, "lr": 1.409267948166718e-05, "epoch": 4.453877551020408, "percentage": 63.62, "elapsed_time": "9:59:35", "remaining_time": "5:42:50"} +{"current_steps": 2735, "total_steps": 4291, "loss": 0.1984, "lr": 1.4014991523849726e-05, "epoch": 4.462040816326531, "percentage": 63.74, "elapsed_time": "10:00:30", "remaining_time": "5:41:38"} +{"current_steps": 2740, "total_steps": 4291, "loss": 0.1971, "lr": 1.3937402627400384e-05, "epoch": 4.470204081632653, "percentage": 63.85, "elapsed_time": "10:01:30", "remaining_time": "5:40:29"} +{"current_steps": 2745, "total_steps": 4291, "loss": 0.1944, "lr": 1.3859914076538245e-05, "epoch": 4.478367346938776, "percentage": 63.97, "elapsed_time": "10:02:37", "remaining_time": "5:39:24"} +{"current_steps": 2750, "total_steps": 4291, "loss": 0.1797, "lr": 1.3782527153821554e-05, "epoch": 4.486530612244898, "percentage": 64.09, "elapsed_time": "10:03:39", "remaining_time": "5:38:15"} +{"current_steps": 2755, "total_steps": 4291, "loss": 0.1829, "lr": 1.370524314012642e-05, "epoch": 4.49469387755102, "percentage": 64.2, "elapsed_time": "10:04:34", "remaining_time": "5:37:03"} +{"current_steps": 2760, "total_steps": 4291, "loss": 0.1785, "lr": 1.3628063314625655e-05, "epoch": 4.502857142857143, "percentage": 64.32, "elapsed_time": "10:05:41", "remaining_time": "5:35:58"} +{"current_steps": 2765, "total_steps": 4291, "loss": 0.1913, "lr": 1.355098895476759e-05, "epoch": 4.511020408163265, "percentage": 64.44, "elapsed_time": "10:06:47", "remaining_time": "5:34:53"} +{"current_steps": 2770, "total_steps": 4291, "loss": 0.1762, "lr": 1.347402133625495e-05, "epoch": 4.519183673469388, "percentage": 64.55, "elapsed_time": "10:07:53", "remaining_time": "5:33:47"} +{"current_steps": 2775, "total_steps": 4291, "loss": 0.1892, "lr": 1.3397161733023698e-05, "epoch": 4.52734693877551, "percentage": 64.67, "elapsed_time": "10:09:01", "remaining_time": "5:32:42"} +{"current_steps": 2780, "total_steps": 4291, "loss": 0.1967, "lr": 1.3320411417221974e-05, "epoch": 4.535510204081633, "percentage": 64.79, "elapsed_time": "10:10:02", "remaining_time": "5:31:34"} +{"current_steps": 2785, "total_steps": 4291, "loss": 0.1875, "lr": 1.324377165918906e-05, "epoch": 4.543673469387755, "percentage": 64.9, "elapsed_time": "10:11:07", "remaining_time": "5:30:27"} +{"current_steps": 2790, "total_steps": 4291, "loss": 0.1974, "lr": 1.3167243727434307e-05, "epoch": 4.551836734693877, "percentage": 65.02, "elapsed_time": "10:12:09", "remaining_time": "5:29:20"} +{"current_steps": 2795, "total_steps": 4291, "loss": 0.1794, "lr": 1.3090828888616183e-05, "epoch": 4.5600000000000005, "percentage": 65.14, "elapsed_time": "10:13:15", "remaining_time": "5:28:14"} +{"current_steps": 2800, "total_steps": 4291, "loss": 0.1902, "lr": 1.301452840752127e-05, "epoch": 4.568163265306122, "percentage": 65.25, "elapsed_time": "10:14:21", "remaining_time": "5:27:08"} +{"current_steps": 2805, "total_steps": 4291, "loss": 0.1879, "lr": 1.293834354704337e-05, "epoch": 4.576326530612245, "percentage": 65.37, "elapsed_time": "10:16:21", "remaining_time": "5:26:31"} +{"current_steps": 2810, "total_steps": 4291, "loss": 0.189, "lr": 1.2862275568162566e-05, "epoch": 4.584489795918367, "percentage": 65.49, "elapsed_time": "10:17:29", "remaining_time": "5:25:26"} +{"current_steps": 2815, "total_steps": 4291, "loss": 0.1857, "lr": 1.2786325729924362e-05, "epoch": 4.59265306122449, "percentage": 65.6, "elapsed_time": "10:18:26", "remaining_time": "5:24:16"} +{"current_steps": 2820, "total_steps": 4291, "loss": 0.1898, "lr": 1.2710495289418869e-05, "epoch": 4.600816326530612, "percentage": 65.72, "elapsed_time": "10:19:39", "remaining_time": "5:23:14"} +{"current_steps": 2825, "total_steps": 4291, "loss": 0.1881, "lr": 1.2634785501759957e-05, "epoch": 4.608979591836735, "percentage": 65.84, "elapsed_time": "10:20:55", "remaining_time": "5:22:13"} +{"current_steps": 2830, "total_steps": 4291, "loss": 0.1804, "lr": 1.2559197620064501e-05, "epoch": 4.617142857142857, "percentage": 65.95, "elapsed_time": "10:21:57", "remaining_time": "5:21:05"} +{"current_steps": 2835, "total_steps": 4291, "loss": 0.2024, "lr": 1.2483732895431637e-05, "epoch": 4.625306122448979, "percentage": 66.07, "elapsed_time": "10:23:02", "remaining_time": "5:19:58"} +{"current_steps": 2840, "total_steps": 4291, "loss": 0.1959, "lr": 1.2408392576922075e-05, "epoch": 4.6334693877551025, "percentage": 66.19, "elapsed_time": "10:24:08", "remaining_time": "5:18:52"} +{"current_steps": 2845, "total_steps": 4291, "loss": 0.1953, "lr": 1.2333177911537396e-05, "epoch": 4.641632653061224, "percentage": 66.3, "elapsed_time": "10:25:15", "remaining_time": "5:17:47"} +{"current_steps": 2850, "total_steps": 4291, "loss": 0.1869, "lr": 1.2258090144199409e-05, "epoch": 4.649795918367347, "percentage": 66.42, "elapsed_time": "10:26:17", "remaining_time": "5:16:39"} +{"current_steps": 2855, "total_steps": 4291, "loss": 0.1831, "lr": 1.2183130517729586e-05, "epoch": 4.657959183673469, "percentage": 66.53, "elapsed_time": "10:27:23", "remaining_time": "5:15:33"} +{"current_steps": 2860, "total_steps": 4291, "loss": 0.1939, "lr": 1.2108300272828446e-05, "epoch": 4.666122448979592, "percentage": 66.65, "elapsed_time": "10:28:38", "remaining_time": "5:14:32"} +{"current_steps": 2865, "total_steps": 4291, "loss": 0.1966, "lr": 1.2033600648055038e-05, "epoch": 4.674285714285714, "percentage": 66.77, "elapsed_time": "10:29:33", "remaining_time": "5:13:21"} +{"current_steps": 2870, "total_steps": 4291, "loss": 0.1918, "lr": 1.1959032879806442e-05, "epoch": 4.682448979591837, "percentage": 66.88, "elapsed_time": "10:30:39", "remaining_time": "5:12:15"} +{"current_steps": 2875, "total_steps": 4291, "loss": 0.1908, "lr": 1.1884598202297327e-05, "epoch": 4.6906122448979595, "percentage": 67.0, "elapsed_time": "10:31:40", "remaining_time": "5:11:06"} +{"current_steps": 2880, "total_steps": 4291, "loss": 0.2062, "lr": 1.1810297847539464e-05, "epoch": 4.698775510204082, "percentage": 67.12, "elapsed_time": "10:32:55", "remaining_time": "5:10:05"} +{"current_steps": 2885, "total_steps": 4291, "loss": 0.1886, "lr": 1.1736133045321387e-05, "epoch": 4.706938775510205, "percentage": 67.23, "elapsed_time": "10:34:04", "remaining_time": "5:09:01"} +{"current_steps": 2890, "total_steps": 4291, "loss": 0.1797, "lr": 1.1662105023188027e-05, "epoch": 4.715102040816326, "percentage": 67.35, "elapsed_time": "10:35:07", "remaining_time": "5:07:53"} +{"current_steps": 2895, "total_steps": 4291, "loss": 0.1928, "lr": 1.1588215006420374e-05, "epoch": 4.723265306122449, "percentage": 67.47, "elapsed_time": "10:36:08", "remaining_time": "5:06:45"} +{"current_steps": 2900, "total_steps": 4291, "loss": 0.1862, "lr": 1.1514464218015218e-05, "epoch": 4.731428571428571, "percentage": 67.58, "elapsed_time": "10:37:05", "remaining_time": "5:05:35"} +{"current_steps": 2905, "total_steps": 4291, "loss": 0.1928, "lr": 1.1440853878664902e-05, "epoch": 4.739591836734694, "percentage": 67.7, "elapsed_time": "10:38:10", "remaining_time": "5:04:28"} +{"current_steps": 2910, "total_steps": 4291, "loss": 0.1846, "lr": 1.1367385206737106e-05, "epoch": 4.747755102040816, "percentage": 67.82, "elapsed_time": "10:39:15", "remaining_time": "5:03:22"} +{"current_steps": 2915, "total_steps": 4291, "loss": 0.1934, "lr": 1.129405941825471e-05, "epoch": 4.755918367346939, "percentage": 67.93, "elapsed_time": "10:40:28", "remaining_time": "5:02:19"} +{"current_steps": 2920, "total_steps": 4291, "loss": 0.1896, "lr": 1.1220877726875603e-05, "epoch": 4.7640816326530615, "percentage": 68.05, "elapsed_time": "10:41:32", "remaining_time": "5:01:13"} +{"current_steps": 2925, "total_steps": 4291, "loss": 0.1815, "lr": 1.1147841343872699e-05, "epoch": 4.772244897959184, "percentage": 68.17, "elapsed_time": "10:42:34", "remaining_time": "5:00:05"} +{"current_steps": 2930, "total_steps": 4291, "loss": 0.2038, "lr": 1.1074951478113774e-05, "epoch": 4.780408163265307, "percentage": 68.28, "elapsed_time": "10:43:35", "remaining_time": "4:58:57"} +{"current_steps": 2935, "total_steps": 4291, "loss": 0.184, "lr": 1.1002209336041542e-05, "epoch": 4.788571428571428, "percentage": 68.4, "elapsed_time": "10:44:37", "remaining_time": "4:57:49"} +{"current_steps": 2940, "total_steps": 4291, "loss": 0.1815, "lr": 1.092961612165365e-05, "epoch": 4.796734693877551, "percentage": 68.52, "elapsed_time": "10:45:46", "remaining_time": "4:56:44"} +{"current_steps": 2945, "total_steps": 4291, "loss": 0.1966, "lr": 1.085717303648275e-05, "epoch": 4.804897959183673, "percentage": 68.63, "elapsed_time": "10:46:45", "remaining_time": "4:55:35"} +{"current_steps": 2950, "total_steps": 4291, "loss": 0.1917, "lr": 1.0784881279576635e-05, "epoch": 4.813061224489796, "percentage": 68.75, "elapsed_time": "10:47:46", "remaining_time": "4:54:27"} +{"current_steps": 2955, "total_steps": 4291, "loss": 0.1937, "lr": 1.0712742047478346e-05, "epoch": 4.8212244897959184, "percentage": 68.87, "elapsed_time": "10:48:45", "remaining_time": "4:53:18"} +{"current_steps": 2960, "total_steps": 4291, "loss": 0.1837, "lr": 1.0640756534206412e-05, "epoch": 4.829387755102041, "percentage": 68.98, "elapsed_time": "10:49:53", "remaining_time": "4:52:13"} +{"current_steps": 2965, "total_steps": 4291, "loss": 0.1898, "lr": 1.0568925931235095e-05, "epoch": 4.8375510204081635, "percentage": 69.1, "elapsed_time": "10:50:46", "remaining_time": "4:51:02"} +{"current_steps": 2970, "total_steps": 4291, "loss": 0.1896, "lr": 1.0497251427474602e-05, "epoch": 4.845714285714286, "percentage": 69.21, "elapsed_time": "10:51:58", "remaining_time": "4:49:59"} +{"current_steps": 2975, "total_steps": 4291, "loss": 0.2019, "lr": 1.0425734209251486e-05, "epoch": 4.853877551020409, "percentage": 69.33, "elapsed_time": "10:52:54", "remaining_time": "4:48:48"} +{"current_steps": 2980, "total_steps": 4291, "loss": 0.1796, "lr": 1.035437546028896e-05, "epoch": 4.86204081632653, "percentage": 69.45, "elapsed_time": "10:54:04", "remaining_time": "4:47:44"} +{"current_steps": 2985, "total_steps": 4291, "loss": 0.1872, "lr": 1.028317636168732e-05, "epoch": 4.870204081632653, "percentage": 69.56, "elapsed_time": "10:55:09", "remaining_time": "4:46:38"} +{"current_steps": 2990, "total_steps": 4291, "loss": 0.1945, "lr": 1.0212138091904409e-05, "epoch": 4.878367346938775, "percentage": 69.68, "elapsed_time": "10:56:14", "remaining_time": "4:45:32"} +{"current_steps": 2995, "total_steps": 4291, "loss": 0.1889, "lr": 1.0141261826736067e-05, "epoch": 4.886530612244898, "percentage": 69.8, "elapsed_time": "10:57:16", "remaining_time": "4:44:25"} +{"current_steps": 3000, "total_steps": 4291, "loss": 0.1766, "lr": 1.0070548739296746e-05, "epoch": 4.8946938775510205, "percentage": 69.91, "elapsed_time": "10:58:24", "remaining_time": "4:43:20"} +{"current_steps": 3005, "total_steps": 4291, "loss": 0.1967, "lr": 1.0000000000000006e-05, "epoch": 4.902857142857143, "percentage": 70.03, "elapsed_time": "11:00:24", "remaining_time": "4:42:37"} +{"current_steps": 3010, "total_steps": 4291, "loss": 0.1966, "lr": 9.929616776539199e-06, "epoch": 4.911020408163266, "percentage": 70.15, "elapsed_time": "11:01:22", "remaining_time": "4:41:28"} +{"current_steps": 3015, "total_steps": 4291, "loss": 0.1829, "lr": 9.859400233868159e-06, "epoch": 4.919183673469388, "percentage": 70.26, "elapsed_time": "11:02:22", "remaining_time": "4:40:19"} +{"current_steps": 3020, "total_steps": 4291, "loss": 0.187, "lr": 9.789351534181837e-06, "epoch": 4.927346938775511, "percentage": 70.38, "elapsed_time": "11:03:32", "remaining_time": "4:39:15"} +{"current_steps": 3025, "total_steps": 4291, "loss": 0.1794, "lr": 9.719471836897156e-06, "epoch": 4.935510204081632, "percentage": 70.5, "elapsed_time": "11:04:38", "remaining_time": "4:38:09"} +{"current_steps": 3030, "total_steps": 4291, "loss": 0.2045, "lr": 9.649762298633747e-06, "epoch": 4.943673469387755, "percentage": 70.61, "elapsed_time": "11:05:51", "remaining_time": "4:37:06"} +{"current_steps": 3035, "total_steps": 4291, "loss": 0.192, "lr": 9.58022407319488e-06, "epoch": 4.951836734693877, "percentage": 70.73, "elapsed_time": "11:06:55", "remaining_time": "4:35:59"} +{"current_steps": 3040, "total_steps": 4291, "loss": 0.1919, "lr": 9.510858311548306e-06, "epoch": 4.96, "percentage": 70.85, "elapsed_time": "11:07:59", "remaining_time": "4:34:53"} +{"current_steps": 3045, "total_steps": 4291, "loss": 0.1943, "lr": 9.441666161807206e-06, "epoch": 4.9681632653061225, "percentage": 70.96, "elapsed_time": "11:09:02", "remaining_time": "4:33:46"} +{"current_steps": 3050, "total_steps": 4291, "loss": 0.1918, "lr": 9.372648769211259e-06, "epoch": 4.976326530612245, "percentage": 71.08, "elapsed_time": "11:10:01", "remaining_time": "4:32:37"} +{"current_steps": 3055, "total_steps": 4291, "loss": 0.1999, "lr": 9.30380727610759e-06, "epoch": 4.984489795918368, "percentage": 71.2, "elapsed_time": "11:11:03", "remaining_time": "4:31:29"} +{"current_steps": 3060, "total_steps": 4291, "loss": 0.1861, "lr": 9.235142821931928e-06, "epoch": 4.99265306122449, "percentage": 71.31, "elapsed_time": "11:12:05", "remaining_time": "4:30:22"} +{"current_steps": 3065, "total_steps": 4291, "loss": 0.184, "lr": 9.166656543189733e-06, "epoch": 5.0, "percentage": 71.43, "elapsed_time": "11:13:03", "remaining_time": "4:29:13"} +{"current_steps": 3070, "total_steps": 4291, "loss": 0.1672, "lr": 9.098349573437368e-06, "epoch": 5.0081632653061225, "percentage": 71.55, "elapsed_time": "11:14:15", "remaining_time": "4:28:09"} +{"current_steps": 3075, "total_steps": 4291, "loss": 0.1836, "lr": 9.03022304326336e-06, "epoch": 5.016326530612245, "percentage": 71.66, "elapsed_time": "11:15:25", "remaining_time": "4:27:05"} +{"current_steps": 3080, "total_steps": 4291, "loss": 0.1868, "lr": 8.962278080269649e-06, "epoch": 5.024489795918368, "percentage": 71.78, "elapsed_time": "11:16:24", "remaining_time": "4:25:56"} +{"current_steps": 3085, "total_steps": 4291, "loss": 0.1822, "lr": 8.89451580905299e-06, "epoch": 5.03265306122449, "percentage": 71.89, "elapsed_time": "11:17:19", "remaining_time": "4:24:47"} +{"current_steps": 3090, "total_steps": 4291, "loss": 0.1642, "lr": 8.82693735118628e-06, "epoch": 5.040816326530612, "percentage": 72.01, "elapsed_time": "11:18:22", "remaining_time": "4:23:39"} +{"current_steps": 3095, "total_steps": 4291, "loss": 0.1879, "lr": 8.759543825200001e-06, "epoch": 5.048979591836734, "percentage": 72.13, "elapsed_time": "11:19:34", "remaining_time": "4:22:36"} +{"current_steps": 3100, "total_steps": 4291, "loss": 0.1739, "lr": 8.692336346563735e-06, "epoch": 5.057142857142857, "percentage": 72.24, "elapsed_time": "11:20:39", "remaining_time": "4:21:30"} +{"current_steps": 3105, "total_steps": 4291, "loss": 0.1794, "lr": 8.625316027667685e-06, "epoch": 5.0653061224489795, "percentage": 72.36, "elapsed_time": "11:21:40", "remaining_time": "4:20:22"} +{"current_steps": 3110, "total_steps": 4291, "loss": 0.1792, "lr": 8.558483977804265e-06, "epoch": 5.073469387755102, "percentage": 72.48, "elapsed_time": "11:22:52", "remaining_time": "4:19:19"} +{"current_steps": 3115, "total_steps": 4291, "loss": 0.1831, "lr": 8.491841303149728e-06, "epoch": 5.081632653061225, "percentage": 72.59, "elapsed_time": "11:23:55", "remaining_time": "4:18:12"} +{"current_steps": 3120, "total_steps": 4291, "loss": 0.1739, "lr": 8.425389106745881e-06, "epoch": 5.089795918367347, "percentage": 72.71, "elapsed_time": "11:24:54", "remaining_time": "4:17:03"} +{"current_steps": 3125, "total_steps": 4291, "loss": 0.1748, "lr": 8.35912848848181e-06, "epoch": 5.09795918367347, "percentage": 72.83, "elapsed_time": "11:25:53", "remaining_time": "4:15:55"} +{"current_steps": 3130, "total_steps": 4291, "loss": 0.1844, "lr": 8.293060545075667e-06, "epoch": 5.106122448979592, "percentage": 72.94, "elapsed_time": "11:27:04", "remaining_time": "4:14:51"} +{"current_steps": 3135, "total_steps": 4291, "loss": 0.1803, "lr": 8.227186370056546e-06, "epoch": 5.114285714285714, "percentage": 73.06, "elapsed_time": "11:28:18", "remaining_time": "4:13:48"} +{"current_steps": 3140, "total_steps": 4291, "loss": 0.1826, "lr": 8.16150705374636e-06, "epoch": 5.122448979591836, "percentage": 73.18, "elapsed_time": "11:29:22", "remaining_time": "4:12:41"} +{"current_steps": 3145, "total_steps": 4291, "loss": 0.1794, "lr": 8.096023683241805e-06, "epoch": 5.130612244897959, "percentage": 73.29, "elapsed_time": "11:30:30", "remaining_time": "4:11:36"} +{"current_steps": 3150, "total_steps": 4291, "loss": 0.1816, "lr": 8.030737342396362e-06, "epoch": 5.1387755102040815, "percentage": 73.41, "elapsed_time": "11:31:40", "remaining_time": "4:10:32"} +{"current_steps": 3155, "total_steps": 4291, "loss": 0.178, "lr": 7.965649111802358e-06, "epoch": 5.146938775510204, "percentage": 73.53, "elapsed_time": "11:32:39", "remaining_time": "4:09:24"} +{"current_steps": 3160, "total_steps": 4291, "loss": 0.1818, "lr": 7.900760068773083e-06, "epoch": 5.155102040816327, "percentage": 73.64, "elapsed_time": "11:33:40", "remaining_time": "4:08:16"} +{"current_steps": 3165, "total_steps": 4291, "loss": 0.1878, "lr": 7.836071287324965e-06, "epoch": 5.163265306122449, "percentage": 73.76, "elapsed_time": "11:34:52", "remaining_time": "4:07:12"} +{"current_steps": 3170, "total_steps": 4291, "loss": 0.171, "lr": 7.771583838159756e-06, "epoch": 5.171428571428572, "percentage": 73.88, "elapsed_time": "11:36:04", "remaining_time": "4:06:09"} +{"current_steps": 3175, "total_steps": 4291, "loss": 0.1774, "lr": 7.70729878864689e-06, "epoch": 5.179591836734694, "percentage": 73.99, "elapsed_time": "11:37:11", "remaining_time": "4:05:03"} +{"current_steps": 3180, "total_steps": 4291, "loss": 0.175, "lr": 7.643217202805715e-06, "epoch": 5.187755102040816, "percentage": 74.11, "elapsed_time": "11:38:10", "remaining_time": "4:03:55"} +{"current_steps": 3185, "total_steps": 4291, "loss": 0.1806, "lr": 7.579340141287965e-06, "epoch": 5.1959183673469385, "percentage": 74.23, "elapsed_time": "11:39:16", "remaining_time": "4:02:49"} +{"current_steps": 3190, "total_steps": 4291, "loss": 0.1885, "lr": 7.515668661360162e-06, "epoch": 5.204081632653061, "percentage": 74.34, "elapsed_time": "11:40:13", "remaining_time": "4:01:40"} +{"current_steps": 3195, "total_steps": 4291, "loss": 0.182, "lr": 7.452203816886132e-06, "epoch": 5.2122448979591836, "percentage": 74.46, "elapsed_time": "11:41:09", "remaining_time": "4:00:31"} +{"current_steps": 3200, "total_steps": 4291, "loss": 0.1725, "lr": 7.388946658309557e-06, "epoch": 5.220408163265306, "percentage": 74.57, "elapsed_time": "11:42:19", "remaining_time": "3:59:26"} +{"current_steps": 3205, "total_steps": 4291, "loss": 0.1797, "lr": 7.3258982326365656e-06, "epoch": 5.228571428571429, "percentage": 74.69, "elapsed_time": "11:44:10", "remaining_time": "3:58:36"} +{"current_steps": 3210, "total_steps": 4291, "loss": 0.1807, "lr": 7.263059583418473e-06, "epoch": 5.236734693877551, "percentage": 74.81, "elapsed_time": "11:45:11", "remaining_time": "3:57:28"} +{"current_steps": 3215, "total_steps": 4291, "loss": 0.1887, "lr": 7.200431750734414e-06, "epoch": 5.244897959183674, "percentage": 74.92, "elapsed_time": "11:46:12", "remaining_time": "3:56:21"} +{"current_steps": 3220, "total_steps": 4291, "loss": 0.1709, "lr": 7.138015771174196e-06, "epoch": 5.253061224489796, "percentage": 75.04, "elapsed_time": "11:47:12", "remaining_time": "3:55:13"} +{"current_steps": 3225, "total_steps": 4291, "loss": 0.1807, "lr": 7.075812677821145e-06, "epoch": 5.261224489795918, "percentage": 75.16, "elapsed_time": "11:48:19", "remaining_time": "3:54:07"} +{"current_steps": 3230, "total_steps": 4291, "loss": 0.1757, "lr": 7.013823500234938e-06, "epoch": 5.2693877551020405, "percentage": 75.27, "elapsed_time": "11:49:26", "remaining_time": "3:53:02"} +{"current_steps": 3235, "total_steps": 4291, "loss": 0.1711, "lr": 6.952049264434644e-06, "epoch": 5.277551020408163, "percentage": 75.39, "elapsed_time": "11:50:35", "remaining_time": "3:51:57"} +{"current_steps": 3240, "total_steps": 4291, "loss": 0.1656, "lr": 6.890490992881671e-06, "epoch": 5.285714285714286, "percentage": 75.51, "elapsed_time": "11:51:45", "remaining_time": "3:50:53"} +{"current_steps": 3245, "total_steps": 4291, "loss": 0.1784, "lr": 6.829149704462918e-06, "epoch": 5.293877551020408, "percentage": 75.62, "elapsed_time": "11:52:59", "remaining_time": "3:49:49"} +{"current_steps": 3250, "total_steps": 4291, "loss": 0.1806, "lr": 6.768026414473845e-06, "epoch": 5.302040816326531, "percentage": 75.74, "elapsed_time": "11:54:05", "remaining_time": "3:48:43"} +{"current_steps": 3255, "total_steps": 4291, "loss": 0.1732, "lr": 6.707122134601676e-06, "epoch": 5.310204081632653, "percentage": 75.86, "elapsed_time": "11:55:14", "remaining_time": "3:47:38"} +{"current_steps": 3260, "total_steps": 4291, "loss": 0.1698, "lr": 6.6464378729087196e-06, "epoch": 5.318367346938776, "percentage": 75.97, "elapsed_time": "11:56:16", "remaining_time": "3:46:31"} +{"current_steps": 3265, "total_steps": 4291, "loss": 0.1833, "lr": 6.585974633815584e-06, "epoch": 5.326530612244898, "percentage": 76.09, "elapsed_time": "11:57:20", "remaining_time": "3:45:25"} +{"current_steps": 3270, "total_steps": 4291, "loss": 0.1849, "lr": 6.525733418084635e-06, "epoch": 5.33469387755102, "percentage": 76.21, "elapsed_time": "11:58:31", "remaining_time": "3:44:20"} +{"current_steps": 3275, "total_steps": 4291, "loss": 0.178, "lr": 6.465715222803386e-06, "epoch": 5.3428571428571425, "percentage": 76.32, "elapsed_time": "11:59:35", "remaining_time": "3:43:14"} +{"current_steps": 3280, "total_steps": 4291, "loss": 0.1827, "lr": 6.4059210413680175e-06, "epoch": 5.351020408163265, "percentage": 76.44, "elapsed_time": "12:00:39", "remaining_time": "3:42:07"} +{"current_steps": 3285, "total_steps": 4291, "loss": 0.1723, "lr": 6.346351863466926e-06, "epoch": 5.359183673469388, "percentage": 76.56, "elapsed_time": "12:01:50", "remaining_time": "3:41:03"} +{"current_steps": 3290, "total_steps": 4291, "loss": 0.1831, "lr": 6.2870086750643215e-06, "epoch": 5.36734693877551, "percentage": 76.67, "elapsed_time": "12:02:52", "remaining_time": "3:39:56"} +{"current_steps": 3295, "total_steps": 4291, "loss": 0.1777, "lr": 6.2278924583839665e-06, "epoch": 5.375510204081633, "percentage": 76.79, "elapsed_time": "12:04:01", "remaining_time": "3:38:51"} +{"current_steps": 3300, "total_steps": 4291, "loss": 0.1859, "lr": 6.16900419189286e-06, "epoch": 5.383673469387755, "percentage": 76.91, "elapsed_time": "12:04:58", "remaining_time": "3:37:42"} +{"current_steps": 3305, "total_steps": 4291, "loss": 0.1834, "lr": 6.110344850285059e-06, "epoch": 5.391836734693878, "percentage": 77.02, "elapsed_time": "12:06:02", "remaining_time": "3:36:36"} +{"current_steps": 3310, "total_steps": 4291, "loss": 0.1813, "lr": 6.0519154044655674e-06, "epoch": 5.4, "percentage": 77.14, "elapsed_time": "12:07:04", "remaining_time": "3:35:29"} +{"current_steps": 3315, "total_steps": 4291, "loss": 0.182, "lr": 5.9937168215342345e-06, "epoch": 5.408163265306122, "percentage": 77.25, "elapsed_time": "12:08:15", "remaining_time": "3:34:24"} +{"current_steps": 3320, "total_steps": 4291, "loss": 0.1711, "lr": 5.9357500647697786e-06, "epoch": 5.416326530612245, "percentage": 77.37, "elapsed_time": "12:09:28", "remaining_time": "3:33:21"} +{"current_steps": 3325, "total_steps": 4291, "loss": 0.1936, "lr": 5.878016093613816e-06, "epoch": 5.424489795918367, "percentage": 77.49, "elapsed_time": "12:10:32", "remaining_time": "3:32:14"} +{"current_steps": 3330, "total_steps": 4291, "loss": 0.1745, "lr": 5.820515863654999e-06, "epoch": 5.43265306122449, "percentage": 77.6, "elapsed_time": "12:11:33", "remaining_time": "3:31:07"} +{"current_steps": 3335, "total_steps": 4291, "loss": 0.1734, "lr": 5.7632503266131925e-06, "epoch": 5.440816326530612, "percentage": 77.72, "elapsed_time": "12:12:40", "remaining_time": "3:30:01"} +{"current_steps": 3340, "total_steps": 4291, "loss": 0.1784, "lr": 5.70622043032371e-06, "epoch": 5.448979591836735, "percentage": 77.84, "elapsed_time": "12:13:43", "remaining_time": "3:28:54"} +{"current_steps": 3345, "total_steps": 4291, "loss": 0.1809, "lr": 5.649427118721655e-06, "epoch": 5.457142857142857, "percentage": 77.95, "elapsed_time": "12:14:35", "remaining_time": "3:27:45"} +{"current_steps": 3350, "total_steps": 4291, "loss": 0.1803, "lr": 5.592871331826271e-06, "epoch": 5.46530612244898, "percentage": 78.07, "elapsed_time": "12:15:41", "remaining_time": "3:26:39"} +{"current_steps": 3355, "total_steps": 4291, "loss": 0.1687, "lr": 5.536554005725392e-06, "epoch": 5.473469387755102, "percentage": 78.19, "elapsed_time": "12:16:40", "remaining_time": "3:25:31"} +{"current_steps": 3360, "total_steps": 4291, "loss": 0.1772, "lr": 5.480476072559955e-06, "epoch": 5.481632653061224, "percentage": 78.3, "elapsed_time": "12:17:50", "remaining_time": "3:24:26"} +{"current_steps": 3365, "total_steps": 4291, "loss": 0.1792, "lr": 5.42463846050854e-06, "epoch": 5.489795918367347, "percentage": 78.42, "elapsed_time": "12:18:55", "remaining_time": "3:23:20"} +{"current_steps": 3370, "total_steps": 4291, "loss": 0.1864, "lr": 5.369042093772068e-06, "epoch": 5.497959183673469, "percentage": 78.54, "elapsed_time": "12:20:02", "remaining_time": "3:22:14"} +{"current_steps": 3375, "total_steps": 4291, "loss": 0.1797, "lr": 5.313687892558455e-06, "epoch": 5.506122448979592, "percentage": 78.65, "elapsed_time": "12:21:05", "remaining_time": "3:21:08"} +{"current_steps": 3380, "total_steps": 4291, "loss": 0.1731, "lr": 5.258576773067366e-06, "epoch": 5.514285714285714, "percentage": 78.77, "elapsed_time": "12:22:13", "remaining_time": "3:20:02"} +{"current_steps": 3385, "total_steps": 4291, "loss": 0.177, "lr": 5.203709647475128e-06, "epoch": 5.522448979591837, "percentage": 78.89, "elapsed_time": "12:23:19", "remaining_time": "3:18:57"} +{"current_steps": 3390, "total_steps": 4291, "loss": 0.1638, "lr": 5.149087423919541e-06, "epoch": 5.530612244897959, "percentage": 79.0, "elapsed_time": "12:24:32", "remaining_time": "3:17:53"} +{"current_steps": 3395, "total_steps": 4291, "loss": 0.175, "lr": 5.094711006484907e-06, "epoch": 5.538775510204082, "percentage": 79.12, "elapsed_time": "12:25:42", "remaining_time": "3:16:48"} +{"current_steps": 3400, "total_steps": 4291, "loss": 0.1736, "lr": 5.040581295187048e-06, "epoch": 5.546938775510204, "percentage": 79.24, "elapsed_time": "12:26:45", "remaining_time": "3:15:41"} +{"current_steps": 3405, "total_steps": 4291, "loss": 0.1745, "lr": 4.986699185958403e-06, "epoch": 5.555102040816326, "percentage": 79.35, "elapsed_time": "12:28:48", "remaining_time": "3:14:50"} +{"current_steps": 3410, "total_steps": 4291, "loss": 0.1768, "lr": 4.933065570633215e-06, "epoch": 5.563265306122449, "percentage": 79.47, "elapsed_time": "12:29:57", "remaining_time": "3:13:45"} +{"current_steps": 3415, "total_steps": 4291, "loss": 0.1804, "lr": 4.879681336932733e-06, "epoch": 5.571428571428571, "percentage": 79.59, "elapsed_time": "12:31:11", "remaining_time": "3:12:41"} +{"current_steps": 3420, "total_steps": 4291, "loss": 0.1799, "lr": 4.8265473684505844e-06, "epoch": 5.579591836734694, "percentage": 79.7, "elapsed_time": "12:32:19", "remaining_time": "3:11:36"} +{"current_steps": 3425, "total_steps": 4291, "loss": 0.1751, "lr": 4.773664544638079e-06, "epoch": 5.587755102040816, "percentage": 79.82, "elapsed_time": "12:33:23", "remaining_time": "3:10:29"} +{"current_steps": 3430, "total_steps": 4291, "loss": 0.1845, "lr": 4.721033740789696e-06, "epoch": 5.595918367346939, "percentage": 79.93, "elapsed_time": "12:34:27", "remaining_time": "3:09:22"} +{"current_steps": 3435, "total_steps": 4291, "loss": 0.1838, "lr": 4.6686558280285875e-06, "epoch": 5.604081632653061, "percentage": 80.05, "elapsed_time": "12:35:21", "remaining_time": "3:08:13"} +{"current_steps": 3440, "total_steps": 4291, "loss": 0.1797, "lr": 4.6165316732921525e-06, "epoch": 5.612244897959184, "percentage": 80.17, "elapsed_time": "12:36:29", "remaining_time": "3:07:08"} +{"current_steps": 3445, "total_steps": 4291, "loss": 0.1865, "lr": 4.5646621393177e-06, "epoch": 5.6204081632653065, "percentage": 80.28, "elapsed_time": "12:37:41", "remaining_time": "3:06:04"} +{"current_steps": 3450, "total_steps": 4291, "loss": 0.1765, "lr": 4.513048084628153e-06, "epoch": 5.628571428571428, "percentage": 80.4, "elapsed_time": "12:38:47", "remaining_time": "3:04:58"} +{"current_steps": 3455, "total_steps": 4291, "loss": 0.1793, "lr": 4.461690363517856e-06, "epoch": 5.636734693877551, "percentage": 80.52, "elapsed_time": "12:39:43", "remaining_time": "3:03:49"} +{"current_steps": 3460, "total_steps": 4291, "loss": 0.1835, "lr": 4.410589826038427e-06, "epoch": 5.644897959183673, "percentage": 80.63, "elapsed_time": "12:40:45", "remaining_time": "3:02:42"} +{"current_steps": 3465, "total_steps": 4291, "loss": 0.17, "lr": 4.35974731798467e-06, "epoch": 5.653061224489796, "percentage": 80.75, "elapsed_time": "12:41:50", "remaining_time": "3:01:36"} +{"current_steps": 3470, "total_steps": 4291, "loss": 0.1857, "lr": 4.309163680880619e-06, "epoch": 5.661224489795918, "percentage": 80.87, "elapsed_time": "12:42:45", "remaining_time": "3:00:28"} +{"current_steps": 3475, "total_steps": 4291, "loss": 0.1803, "lr": 4.258839751965566e-06, "epoch": 5.669387755102041, "percentage": 80.98, "elapsed_time": "12:43:44", "remaining_time": "2:59:20"} +{"current_steps": 3480, "total_steps": 4291, "loss": 0.1775, "lr": 4.208776364180231e-06, "epoch": 5.677551020408163, "percentage": 81.1, "elapsed_time": "12:44:49", "remaining_time": "2:58:14"} +{"current_steps": 3485, "total_steps": 4291, "loss": 0.1768, "lr": 4.15897434615296e-06, "epoch": 5.685714285714286, "percentage": 81.22, "elapsed_time": "12:45:46", "remaining_time": "2:57:06"} +{"current_steps": 3490, "total_steps": 4291, "loss": 0.1922, "lr": 4.10943452218602e-06, "epoch": 5.6938775510204085, "percentage": 81.33, "elapsed_time": "12:46:59", "remaining_time": "2:56:02"} +{"current_steps": 3495, "total_steps": 4291, "loss": 0.1863, "lr": 4.06015771224195e-06, "epoch": 5.70204081632653, "percentage": 81.45, "elapsed_time": "12:48:06", "remaining_time": "2:54:56"} +{"current_steps": 3500, "total_steps": 4291, "loss": 0.1909, "lr": 4.011144731929981e-06, "epoch": 5.710204081632653, "percentage": 81.57, "elapsed_time": "12:49:00", "remaining_time": "2:53:47"} +{"current_steps": 3505, "total_steps": 4291, "loss": 0.1835, "lr": 3.962396392492554e-06, "epoch": 5.718367346938775, "percentage": 81.68, "elapsed_time": "12:50:01", "remaining_time": "2:52:40"} +{"current_steps": 3510, "total_steps": 4291, "loss": 0.1815, "lr": 3.9139135007919e-06, "epoch": 5.726530612244898, "percentage": 81.8, "elapsed_time": "12:50:57", "remaining_time": "2:51:32"} +{"current_steps": 3515, "total_steps": 4291, "loss": 0.1815, "lr": 3.865696859296635e-06, "epoch": 5.73469387755102, "percentage": 81.92, "elapsed_time": "12:52:04", "remaining_time": "2:50:26"} +{"current_steps": 3520, "total_steps": 4291, "loss": 0.1756, "lr": 3.817747266068541e-06, "epoch": 5.742857142857143, "percentage": 82.03, "elapsed_time": "12:53:10", "remaining_time": "2:49:20"} +{"current_steps": 3525, "total_steps": 4291, "loss": 0.1731, "lr": 3.770065514749317e-06, "epoch": 5.751020408163265, "percentage": 82.15, "elapsed_time": "12:54:05", "remaining_time": "2:48:12"} +{"current_steps": 3530, "total_steps": 4291, "loss": 0.1839, "lr": 3.7226523945474567e-06, "epoch": 5.759183673469388, "percentage": 82.27, "elapsed_time": "12:55:09", "remaining_time": "2:47:06"} +{"current_steps": 3535, "total_steps": 4291, "loss": 0.1838, "lr": 3.6755086902251847e-06, "epoch": 5.7673469387755105, "percentage": 82.38, "elapsed_time": "12:56:07", "remaining_time": "2:45:58"} +{"current_steps": 3540, "total_steps": 4291, "loss": 0.1856, "lr": 3.6286351820854514e-06, "epoch": 5.775510204081632, "percentage": 82.5, "elapsed_time": "12:57:12", "remaining_time": "2:44:52"} +{"current_steps": 3545, "total_steps": 4291, "loss": 0.1729, "lr": 3.582032645959066e-06, "epoch": 5.783673469387755, "percentage": 82.61, "elapsed_time": "12:58:13", "remaining_time": "2:43:46"} +{"current_steps": 3550, "total_steps": 4291, "loss": 0.1854, "lr": 3.535701853191784e-06, "epoch": 5.791836734693877, "percentage": 82.73, "elapsed_time": "12:59:16", "remaining_time": "2:42:39"} +{"current_steps": 3555, "total_steps": 4291, "loss": 0.1783, "lr": 3.4896435706316e-06, "epoch": 5.8, "percentage": 82.85, "elapsed_time": "13:00:14", "remaining_time": "2:41:32"} +{"current_steps": 3560, "total_steps": 4291, "loss": 0.1716, "lr": 3.443858560616029e-06, "epoch": 5.808163265306122, "percentage": 82.96, "elapsed_time": "13:01:18", "remaining_time": "2:40:25"} +{"current_steps": 3565, "total_steps": 4291, "loss": 0.1762, "lr": 3.3983475809594892e-06, "epoch": 5.816326530612245, "percentage": 83.08, "elapsed_time": "13:02:18", "remaining_time": "2:39:18"} +{"current_steps": 3570, "total_steps": 4291, "loss": 0.1751, "lr": 3.353111384940775e-06, "epoch": 5.8244897959183675, "percentage": 83.2, "elapsed_time": "13:03:20", "remaining_time": "2:38:12"} +{"current_steps": 3575, "total_steps": 4291, "loss": 0.1829, "lr": 3.308150721290546e-06, "epoch": 5.83265306122449, "percentage": 83.31, "elapsed_time": "13:04:18", "remaining_time": "2:37:04"} +{"current_steps": 3580, "total_steps": 4291, "loss": 0.1783, "lr": 3.263466334179002e-06, "epoch": 5.840816326530613, "percentage": 83.43, "elapsed_time": "13:05:25", "remaining_time": "2:35:59"} +{"current_steps": 3585, "total_steps": 4291, "loss": 0.181, "lr": 3.2190589632035117e-06, "epoch": 5.848979591836734, "percentage": 83.55, "elapsed_time": "13:06:41", "remaining_time": "2:34:55"} +{"current_steps": 3590, "total_steps": 4291, "loss": 0.1877, "lr": 3.174929343376374e-06, "epoch": 5.857142857142857, "percentage": 83.66, "elapsed_time": "13:07:44", "remaining_time": "2:33:49"} +{"current_steps": 3595, "total_steps": 4291, "loss": 0.1859, "lr": 3.1310782051126986e-06, "epoch": 5.865306122448979, "percentage": 83.78, "elapsed_time": "13:08:57", "remaining_time": "2:32:44"} +{"current_steps": 3600, "total_steps": 4291, "loss": 0.1735, "lr": 3.0875062742182526e-06, "epoch": 5.873469387755102, "percentage": 83.9, "elapsed_time": "13:10:03", "remaining_time": "2:31:38"} +{"current_steps": 3605, "total_steps": 4291, "loss": 0.1804, "lr": 3.0442142718774992e-06, "epoch": 5.881632653061224, "percentage": 84.01, "elapsed_time": "13:12:06", "remaining_time": "2:30:43"} +{"current_steps": 3610, "total_steps": 4291, "loss": 0.1786, "lr": 3.001202914641628e-06, "epoch": 5.889795918367347, "percentage": 84.13, "elapsed_time": "13:13:09", "remaining_time": "2:29:37"} +{"current_steps": 3615, "total_steps": 4291, "loss": 0.1814, "lr": 2.958472914416721e-06, "epoch": 5.8979591836734695, "percentage": 84.25, "elapsed_time": "13:14:16", "remaining_time": "2:28:31"} +{"current_steps": 3620, "total_steps": 4291, "loss": 0.1914, "lr": 2.916024978451948e-06, "epoch": 5.906122448979592, "percentage": 84.36, "elapsed_time": "13:15:16", "remaining_time": "2:27:24"} +{"current_steps": 3625, "total_steps": 4291, "loss": 0.1772, "lr": 2.8738598093278593e-06, "epoch": 5.914285714285715, "percentage": 84.48, "elapsed_time": "13:16:12", "remaining_time": "2:26:16"} +{"current_steps": 3630, "total_steps": 4291, "loss": 0.1755, "lr": 2.8319781049447923e-06, "epoch": 5.922448979591836, "percentage": 84.6, "elapsed_time": "13:17:17", "remaining_time": "2:25:10"} +{"current_steps": 3635, "total_steps": 4291, "loss": 0.1751, "lr": 2.7903805585112674e-06, "epoch": 5.930612244897959, "percentage": 84.71, "elapsed_time": "13:18:20", "remaining_time": "2:24:04"} +{"current_steps": 3640, "total_steps": 4291, "loss": 0.1699, "lr": 2.7490678585325526e-06, "epoch": 5.938775510204081, "percentage": 84.83, "elapsed_time": "13:19:32", "remaining_time": "2:22:59"} +{"current_steps": 3645, "total_steps": 4291, "loss": 0.1806, "lr": 2.708040688799258e-06, "epoch": 5.946938775510204, "percentage": 84.95, "elapsed_time": "13:20:35", "remaining_time": "2:21:53"} +{"current_steps": 3650, "total_steps": 4291, "loss": 0.1825, "lr": 2.6672997283760137e-06, "epoch": 5.955102040816326, "percentage": 85.06, "elapsed_time": "13:21:31", "remaining_time": "2:20:45"} +{"current_steps": 3655, "total_steps": 4291, "loss": 0.1801, "lr": 2.6268456515902283e-06, "epoch": 5.963265306122449, "percentage": 85.18, "elapsed_time": "13:22:44", "remaining_time": "2:19:41"} +{"current_steps": 3660, "total_steps": 4291, "loss": 0.1735, "lr": 2.5866791280209368e-06, "epoch": 5.9714285714285715, "percentage": 85.29, "elapsed_time": "13:23:50", "remaining_time": "2:18:35"} +{"current_steps": 3665, "total_steps": 4291, "loss": 0.1768, "lr": 2.546800822487714e-06, "epoch": 5.979591836734694, "percentage": 85.41, "elapsed_time": "13:24:52", "remaining_time": "2:17:28"} +{"current_steps": 3670, "total_steps": 4291, "loss": 0.1783, "lr": 2.507211395039668e-06, "epoch": 5.987755102040817, "percentage": 85.53, "elapsed_time": "13:25:58", "remaining_time": "2:16:22"} +{"current_steps": 3675, "total_steps": 4291, "loss": 0.1636, "lr": 2.4679115009445132e-06, "epoch": 5.995918367346938, "percentage": 85.64, "elapsed_time": "13:27:04", "remaining_time": "2:15:16"} +{"current_steps": 3680, "total_steps": 4291, "loss": 0.1667, "lr": 2.428901790677731e-06, "epoch": 6.003265306122449, "percentage": 85.76, "elapsed_time": "13:28:05", "remaining_time": "2:14:10"} +{"current_steps": 3685, "total_steps": 4291, "loss": 0.1717, "lr": 2.3901829099118044e-06, "epoch": 6.011428571428572, "percentage": 85.88, "elapsed_time": "13:29:11", "remaining_time": "2:13:04"} +{"current_steps": 3690, "total_steps": 4291, "loss": 0.1726, "lr": 2.351755499505526e-06, "epoch": 6.019591836734694, "percentage": 85.99, "elapsed_time": "13:30:15", "remaining_time": "2:11:58"} +{"current_steps": 3695, "total_steps": 4291, "loss": 0.1731, "lr": 2.313620195493389e-06, "epoch": 6.027755102040817, "percentage": 86.11, "elapsed_time": "13:31:18", "remaining_time": "2:10:51"} +{"current_steps": 3700, "total_steps": 4291, "loss": 0.1757, "lr": 2.2757776290750687e-06, "epoch": 6.035918367346939, "percentage": 86.23, "elapsed_time": "13:32:20", "remaining_time": "2:09:45"} +{"current_steps": 3705, "total_steps": 4291, "loss": 0.1753, "lr": 2.238228426604967e-06, "epoch": 6.044081632653061, "percentage": 86.34, "elapsed_time": "13:33:22", "remaining_time": "2:08:38"} +{"current_steps": 3710, "total_steps": 4291, "loss": 0.176, "lr": 2.2009732095818425e-06, "epoch": 6.052244897959183, "percentage": 86.46, "elapsed_time": "13:34:27", "remaining_time": "2:07:32"} +{"current_steps": 3715, "total_steps": 4291, "loss": 0.1603, "lr": 2.1640125946385317e-06, "epoch": 6.060408163265306, "percentage": 86.58, "elapsed_time": "13:35:34", "remaining_time": "2:06:27"} +{"current_steps": 3720, "total_steps": 4291, "loss": 0.1647, "lr": 2.127347193531757e-06, "epoch": 6.0685714285714285, "percentage": 86.69, "elapsed_time": "13:36:35", "remaining_time": "2:05:20"} +{"current_steps": 3725, "total_steps": 4291, "loss": 0.1756, "lr": 2.090977613131955e-06, "epoch": 6.076734693877551, "percentage": 86.81, "elapsed_time": "13:37:40", "remaining_time": "2:04:14"} +{"current_steps": 3730, "total_steps": 4291, "loss": 0.169, "lr": 2.0549044554132824e-06, "epoch": 6.084897959183674, "percentage": 86.93, "elapsed_time": "13:38:46", "remaining_time": "2:03:08"} +{"current_steps": 3735, "total_steps": 4291, "loss": 0.1807, "lr": 2.019128317443626e-06, "epoch": 6.093061224489796, "percentage": 87.04, "elapsed_time": "13:39:41", "remaining_time": "2:02:01"} +{"current_steps": 3740, "total_steps": 4291, "loss": 0.1757, "lr": 1.98364979137472e-06, "epoch": 6.101224489795919, "percentage": 87.16, "elapsed_time": "13:40:41", "remaining_time": "2:00:54"} +{"current_steps": 3745, "total_steps": 4291, "loss": 0.176, "lr": 1.948469464432359e-06, "epoch": 6.109387755102041, "percentage": 87.28, "elapsed_time": "13:41:48", "remaining_time": "1:59:48"} +{"current_steps": 3750, "total_steps": 4291, "loss": 0.1708, "lr": 1.9135879189066562e-06, "epoch": 6.117551020408163, "percentage": 87.39, "elapsed_time": "13:42:52", "remaining_time": "1:58:42"} +{"current_steps": 3755, "total_steps": 4291, "loss": 0.1768, "lr": 1.8790057321424427e-06, "epoch": 6.1257142857142854, "percentage": 87.51, "elapsed_time": "13:43:52", "remaining_time": "1:57:36"} +{"current_steps": 3760, "total_steps": 4291, "loss": 0.1755, "lr": 1.8447234765296617e-06, "epoch": 6.133877551020408, "percentage": 87.63, "elapsed_time": "13:44:55", "remaining_time": "1:56:29"} +{"current_steps": 3765, "total_steps": 4291, "loss": 0.1696, "lr": 1.810741719493936e-06, "epoch": 6.1420408163265305, "percentage": 87.74, "elapsed_time": "13:46:02", "remaining_time": "1:55:24"} +{"current_steps": 3770, "total_steps": 4291, "loss": 0.1654, "lr": 1.7770610234871588e-06, "epoch": 6.150204081632653, "percentage": 87.86, "elapsed_time": "13:47:03", "remaining_time": "1:54:17"} +{"current_steps": 3775, "total_steps": 4291, "loss": 0.1786, "lr": 1.743681945978184e-06, "epoch": 6.158367346938776, "percentage": 87.97, "elapsed_time": "13:48:05", "remaining_time": "1:53:11"} +{"current_steps": 3780, "total_steps": 4291, "loss": 0.1759, "lr": 1.710605039443609e-06, "epoch": 6.166530612244898, "percentage": 88.09, "elapsed_time": "13:49:11", "remaining_time": "1:52:05"} +{"current_steps": 3785, "total_steps": 4291, "loss": 0.1647, "lr": 1.6778308513586084e-06, "epoch": 6.174693877551021, "percentage": 88.21, "elapsed_time": "13:50:18", "remaining_time": "1:51:00"} +{"current_steps": 3790, "total_steps": 4291, "loss": 0.1634, "lr": 1.6453599241879014e-06, "epoch": 6.182857142857143, "percentage": 88.32, "elapsed_time": "13:51:18", "remaining_time": "1:49:53"} +{"current_steps": 3795, "total_steps": 4291, "loss": 0.1674, "lr": 1.6131927953767567e-06, "epoch": 6.191020408163265, "percentage": 88.44, "elapsed_time": "13:52:29", "remaining_time": "1:48:48"} +{"current_steps": 3800, "total_steps": 4291, "loss": 0.1682, "lr": 1.581329997342087e-06, "epoch": 6.1991836734693875, "percentage": 88.56, "elapsed_time": "13:53:32", "remaining_time": "1:47:42"} +{"current_steps": 3805, "total_steps": 4291, "loss": 0.1778, "lr": 1.5497720574636632e-06, "epoch": 6.20734693877551, "percentage": 88.67, "elapsed_time": "13:55:31", "remaining_time": "1:46:43"} +{"current_steps": 3810, "total_steps": 4291, "loss": 0.1696, "lr": 1.518519498075366e-06, "epoch": 6.215510204081633, "percentage": 88.79, "elapsed_time": "13:56:33", "remaining_time": "1:45:36"} +{"current_steps": 3815, "total_steps": 4291, "loss": 0.1652, "lr": 1.4875728364565411e-06, "epoch": 6.223673469387755, "percentage": 88.91, "elapsed_time": "13:57:41", "remaining_time": "1:44:31"} +{"current_steps": 3820, "total_steps": 4291, "loss": 0.1755, "lr": 1.4569325848234473e-06, "epoch": 6.231836734693878, "percentage": 89.02, "elapsed_time": "13:58:49", "remaining_time": "1:43:25"} +{"current_steps": 3825, "total_steps": 4291, "loss": 0.1671, "lr": 1.4265992503207704e-06, "epoch": 6.24, "percentage": 89.14, "elapsed_time": "13:59:51", "remaining_time": "1:42:19"} +{"current_steps": 3830, "total_steps": 4291, "loss": 0.1831, "lr": 1.396573335013236e-06, "epoch": 6.248163265306123, "percentage": 89.26, "elapsed_time": "14:00:49", "remaining_time": "1:41:12"} +{"current_steps": 3835, "total_steps": 4291, "loss": 0.163, "lr": 1.3668553358772863e-06, "epoch": 6.256326530612245, "percentage": 89.37, "elapsed_time": "14:01:56", "remaining_time": "1:40:06"} +{"current_steps": 3840, "total_steps": 4291, "loss": 0.1776, "lr": 1.337445744792867e-06, "epoch": 6.264489795918367, "percentage": 89.49, "elapsed_time": "14:02:57", "remaining_time": "1:39:00"} +{"current_steps": 3845, "total_steps": 4291, "loss": 0.168, "lr": 1.3083450485352867e-06, "epoch": 6.2726530612244895, "percentage": 89.61, "elapsed_time": "14:04:08", "remaining_time": "1:37:54"} +{"current_steps": 3850, "total_steps": 4291, "loss": 0.1706, "lr": 1.279553728767151e-06, "epoch": 6.280816326530612, "percentage": 89.72, "elapsed_time": "14:05:07", "remaining_time": "1:36:48"} +{"current_steps": 3855, "total_steps": 4291, "loss": 0.1733, "lr": 1.2510722620303972e-06, "epoch": 6.288979591836735, "percentage": 89.84, "elapsed_time": "14:06:08", "remaining_time": "1:35:41"} +{"current_steps": 3860, "total_steps": 4291, "loss": 0.1695, "lr": 1.2229011197384021e-06, "epoch": 6.297142857142857, "percentage": 89.96, "elapsed_time": "14:07:15", "remaining_time": "1:34:36"} +{"current_steps": 3865, "total_steps": 4291, "loss": 0.1707, "lr": 1.1950407681681831e-06, "epoch": 6.30530612244898, "percentage": 90.07, "elapsed_time": "14:08:14", "remaining_time": "1:33:29"} +{"current_steps": 3870, "total_steps": 4291, "loss": 0.1785, "lr": 1.1674916684526782e-06, "epoch": 6.313469387755102, "percentage": 90.19, "elapsed_time": "14:09:15", "remaining_time": "1:32:23"} +{"current_steps": 3875, "total_steps": 4291, "loss": 0.1683, "lr": 1.1402542765731072e-06, "epoch": 6.321632653061225, "percentage": 90.31, "elapsed_time": "14:10:14", "remaining_time": "1:31:16"} +{"current_steps": 3880, "total_steps": 4291, "loss": 0.1691, "lr": 1.1133290433514522e-06, "epoch": 6.329795918367347, "percentage": 90.42, "elapsed_time": "14:11:26", "remaining_time": "1:30:11"} +{"current_steps": 3885, "total_steps": 4291, "loss": 0.1736, "lr": 1.086716414442952e-06, "epoch": 6.337959183673469, "percentage": 90.54, "elapsed_time": "14:12:33", "remaining_time": "1:29:05"} +{"current_steps": 3890, "total_steps": 4291, "loss": 0.1711, "lr": 1.0604168303287588e-06, "epoch": 6.3461224489795915, "percentage": 90.65, "elapsed_time": "14:13:41", "remaining_time": "1:28:00"} +{"current_steps": 3895, "total_steps": 4291, "loss": 0.1744, "lr": 1.0344307263086417e-06, "epoch": 6.354285714285714, "percentage": 90.77, "elapsed_time": "14:14:49", "remaining_time": "1:26:54"} +{"current_steps": 3900, "total_steps": 4291, "loss": 0.1734, "lr": 1.008758532493772e-06, "epoch": 6.362448979591837, "percentage": 90.89, "elapsed_time": "14:15:53", "remaining_time": "1:25:48"} +{"current_steps": 3905, "total_steps": 4291, "loss": 0.1625, "lr": 9.834006737996126e-07, "epoch": 6.370612244897959, "percentage": 91.0, "elapsed_time": "14:16:53", "remaining_time": "1:24:42"} +{"current_steps": 3910, "total_steps": 4291, "loss": 0.1777, "lr": 9.583575699388748e-07, "epoch": 6.378775510204082, "percentage": 91.12, "elapsed_time": "14:17:52", "remaining_time": "1:23:35"} +{"current_steps": 3915, "total_steps": 4291, "loss": 0.17, "lr": 9.336296354145969e-07, "epoch": 6.386938775510204, "percentage": 91.24, "elapsed_time": "14:18:59", "remaining_time": "1:22:29"} +{"current_steps": 3920, "total_steps": 4291, "loss": 0.1727, "lr": 9.092172795132459e-07, "epoch": 6.395102040816327, "percentage": 91.35, "elapsed_time": "14:20:10", "remaining_time": "1:21:24"} +{"current_steps": 3925, "total_steps": 4291, "loss": 0.1805, "lr": 8.851209062979715e-07, "epoch": 6.403265306122449, "percentage": 91.47, "elapsed_time": "14:21:17", "remaining_time": "1:20:18"} +{"current_steps": 3930, "total_steps": 4291, "loss": 0.1786, "lr": 8.613409146019158e-07, "epoch": 6.411428571428571, "percentage": 91.59, "elapsed_time": "14:22:18", "remaining_time": "1:19:12"} +{"current_steps": 3935, "total_steps": 4291, "loss": 0.1779, "lr": 8.378776980215919e-07, "epoch": 6.419591836734694, "percentage": 91.7, "elapsed_time": "14:23:24", "remaining_time": "1:18:06"} +{"current_steps": 3940, "total_steps": 4291, "loss": 0.1741, "lr": 8.147316449103959e-07, "epoch": 6.427755102040816, "percentage": 91.82, "elapsed_time": "14:24:31", "remaining_time": "1:17:01"} +{"current_steps": 3945, "total_steps": 4291, "loss": 0.1793, "lr": 7.919031383721565e-07, "epoch": 6.435918367346939, "percentage": 91.94, "elapsed_time": "14:25:40", "remaining_time": "1:15:55"} +{"current_steps": 3950, "total_steps": 4291, "loss": 0.166, "lr": 7.69392556254811e-07, "epoch": 6.444081632653061, "percentage": 92.05, "elapsed_time": "14:26:43", "remaining_time": "1:14:49"} +{"current_steps": 3955, "total_steps": 4291, "loss": 0.1782, "lr": 7.472002711441418e-07, "epoch": 6.452244897959184, "percentage": 92.17, "elapsed_time": "14:27:56", "remaining_time": "1:13:44"} +{"current_steps": 3960, "total_steps": 4291, "loss": 0.1787, "lr": 7.253266503576051e-07, "epoch": 6.460408163265306, "percentage": 92.29, "elapsed_time": "14:28:51", "remaining_time": "1:12:37"} +{"current_steps": 3965, "total_steps": 4291, "loss": 0.1743, "lr": 7.037720559382699e-07, "epoch": 6.468571428571429, "percentage": 92.4, "elapsed_time": "14:29:54", "remaining_time": "1:11:31"} +{"current_steps": 3970, "total_steps": 4291, "loss": 0.1759, "lr": 6.825368446488068e-07, "epoch": 6.476734693877551, "percentage": 92.52, "elapsed_time": "14:30:58", "remaining_time": "1:10:25"} +{"current_steps": 3975, "total_steps": 4291, "loss": 0.1727, "lr": 6.616213679655881e-07, "epoch": 6.484897959183673, "percentage": 92.64, "elapsed_time": "14:32:03", "remaining_time": "1:09:19"} +{"current_steps": 3980, "total_steps": 4291, "loss": 0.17, "lr": 6.410259720728751e-07, "epoch": 6.493061224489796, "percentage": 92.75, "elapsed_time": "14:33:10", "remaining_time": "1:08:13"} +{"current_steps": 3985, "total_steps": 4291, "loss": 0.1648, "lr": 6.207509978570848e-07, "epoch": 6.501224489795918, "percentage": 92.87, "elapsed_time": "14:34:14", "remaining_time": "1:07:07"} +{"current_steps": 3990, "total_steps": 4291, "loss": 0.1694, "lr": 6.007967809011517e-07, "epoch": 6.509387755102041, "percentage": 92.99, "elapsed_time": "14:35:14", "remaining_time": "1:06:01"} +{"current_steps": 3995, "total_steps": 4291, "loss": 0.181, "lr": 5.811636514789598e-07, "epoch": 6.517551020408163, "percentage": 93.1, "elapsed_time": "14:36:15", "remaining_time": "1:04:55"} +{"current_steps": 4000, "total_steps": 4291, "loss": 0.1637, "lr": 5.618519345498952e-07, "epoch": 6.525714285714286, "percentage": 93.22, "elapsed_time": "14:37:23", "remaining_time": "1:03:49"} +{"current_steps": 4005, "total_steps": 4291, "loss": 0.1779, "lr": 5.428619497534548e-07, "epoch": 6.533877551020408, "percentage": 93.33, "elapsed_time": "14:39:28", "remaining_time": "1:02:48"} +{"current_steps": 4010, "total_steps": 4291, "loss": 0.17, "lr": 5.2419401140396e-07, "epoch": 6.542040816326531, "percentage": 93.45, "elapsed_time": "14:40:34", "remaining_time": "1:01:42"} +{"current_steps": 4015, "total_steps": 4291, "loss": 0.1625, "lr": 5.058484284853494e-07, "epoch": 6.550204081632653, "percentage": 93.57, "elapsed_time": "14:41:42", "remaining_time": "1:00:36"} +{"current_steps": 4020, "total_steps": 4291, "loss": 0.179, "lr": 4.878255046460755e-07, "epoch": 6.558367346938775, "percentage": 93.68, "elapsed_time": "14:42:46", "remaining_time": "0:59:30"} +{"current_steps": 4025, "total_steps": 4291, "loss": 0.1809, "lr": 4.701255381940684e-07, "epoch": 6.566530612244898, "percentage": 93.8, "elapsed_time": "14:43:59", "remaining_time": "0:58:25"} +{"current_steps": 4030, "total_steps": 4291, "loss": 0.1795, "lr": 4.5274882209180195e-07, "epoch": 6.57469387755102, "percentage": 93.92, "elapsed_time": "14:45:02", "remaining_time": "0:57:19"} +{"current_steps": 4035, "total_steps": 4291, "loss": 0.1757, "lr": 4.3569564395144595e-07, "epoch": 6.582857142857143, "percentage": 94.03, "elapsed_time": "14:46:00", "remaining_time": "0:56:12"} +{"current_steps": 4040, "total_steps": 4291, "loss": 0.1832, "lr": 4.189662860301069e-07, "epoch": 6.591020408163265, "percentage": 94.15, "elapsed_time": "14:47:07", "remaining_time": "0:55:06"} +{"current_steps": 4045, "total_steps": 4291, "loss": 0.1747, "lr": 4.025610252251433e-07, "epoch": 6.599183673469388, "percentage": 94.27, "elapsed_time": "14:48:12", "remaining_time": "0:54:01"} +{"current_steps": 4050, "total_steps": 4291, "loss": 0.1721, "lr": 3.8648013306960664e-07, "epoch": 6.60734693877551, "percentage": 94.38, "elapsed_time": "14:49:22", "remaining_time": "0:52:55"} +{"current_steps": 4055, "total_steps": 4291, "loss": 0.1798, "lr": 3.7072387572772317e-07, "epoch": 6.615510204081633, "percentage": 94.5, "elapsed_time": "14:50:23", "remaining_time": "0:51:49"} +{"current_steps": 4060, "total_steps": 4291, "loss": 0.1815, "lr": 3.552925139905083e-07, "epoch": 6.6236734693877555, "percentage": 94.62, "elapsed_time": "14:51:26", "remaining_time": "0:50:43"} +{"current_steps": 4065, "total_steps": 4291, "loss": 0.1734, "lr": 3.4018630327143474e-07, "epoch": 6.631836734693877, "percentage": 94.73, "elapsed_time": "14:52:37", "remaining_time": "0:49:37"} +{"current_steps": 4070, "total_steps": 4291, "loss": 0.176, "lr": 3.2540549360221105e-07, "epoch": 6.64, "percentage": 94.85, "elapsed_time": "14:53:44", "remaining_time": "0:48:31"} +{"current_steps": 4075, "total_steps": 4291, "loss": 0.1665, "lr": 3.109503296286431e-07, "epoch": 6.648163265306122, "percentage": 94.97, "elapsed_time": "14:54:52", "remaining_time": "0:47:26"} +{"current_steps": 4080, "total_steps": 4291, "loss": 0.1687, "lr": 2.968210506065883e-07, "epoch": 6.656326530612245, "percentage": 95.08, "elapsed_time": "14:56:00", "remaining_time": "0:46:20"} +{"current_steps": 4085, "total_steps": 4291, "loss": 0.1722, "lr": 2.8301789039798523e-07, "epoch": 6.664489795918367, "percentage": 95.2, "elapsed_time": "14:57:00", "remaining_time": "0:45:14"} +{"current_steps": 4090, "total_steps": 4291, "loss": 0.1724, "lr": 2.695410774669993e-07, "epoch": 6.67265306122449, "percentage": 95.32, "elapsed_time": "14:58:11", "remaining_time": "0:44:08"} +{"current_steps": 4095, "total_steps": 4291, "loss": 0.1777, "lr": 2.5639083487622116e-07, "epoch": 6.680816326530612, "percentage": 95.43, "elapsed_time": "14:59:21", "remaining_time": "0:43:02"} +{"current_steps": 4100, "total_steps": 4291, "loss": 0.1822, "lr": 2.435673802829941e-07, "epoch": 6.688979591836735, "percentage": 95.55, "elapsed_time": "15:00:34", "remaining_time": "0:41:57"} +{"current_steps": 4105, "total_steps": 4291, "loss": 0.1735, "lr": 2.3107092593579905e-07, "epoch": 6.6971428571428575, "percentage": 95.67, "elapsed_time": "15:01:30", "remaining_time": "0:40:50"} +{"current_steps": 4110, "total_steps": 4291, "loss": 0.1664, "lr": 2.1890167867074208e-07, "epoch": 6.705306122448979, "percentage": 95.78, "elapsed_time": "15:02:37", "remaining_time": "0:39:45"} +{"current_steps": 4115, "total_steps": 4291, "loss": 0.187, "lr": 2.0705983990813915e-07, "epoch": 6.713469387755102, "percentage": 95.9, "elapsed_time": "15:03:41", "remaining_time": "0:38:39"} +{"current_steps": 4120, "total_steps": 4291, "loss": 0.1713, "lr": 1.9554560564917001e-07, "epoch": 6.721632653061224, "percentage": 96.01, "elapsed_time": "15:04:44", "remaining_time": "0:37:33"} +{"current_steps": 4125, "total_steps": 4291, "loss": 0.1804, "lr": 1.843591664726496e-07, "epoch": 6.729795918367347, "percentage": 96.13, "elapsed_time": "15:05:48", "remaining_time": "0:36:27"} +{"current_steps": 4130, "total_steps": 4291, "loss": 0.175, "lr": 1.7350070753186176e-07, "epoch": 6.737959183673469, "percentage": 96.25, "elapsed_time": "15:06:51", "remaining_time": "0:35:21"} +{"current_steps": 4135, "total_steps": 4291, "loss": 0.1785, "lr": 1.6297040855149493e-07, "epoch": 6.746122448979592, "percentage": 96.36, "elapsed_time": "15:07:47", "remaining_time": "0:34:14"} +{"current_steps": 4140, "total_steps": 4291, "loss": 0.1704, "lr": 1.5276844382467125e-07, "epoch": 6.7542857142857144, "percentage": 96.48, "elapsed_time": "15:08:51", "remaining_time": "0:33:08"} +{"current_steps": 4145, "total_steps": 4291, "loss": 0.1775, "lr": 1.4289498221006893e-07, "epoch": 6.762448979591837, "percentage": 96.6, "elapsed_time": "15:10:02", "remaining_time": "0:32:03"} +{"current_steps": 4150, "total_steps": 4291, "loss": 0.1598, "lr": 1.333501871291065e-07, "epoch": 6.7706122448979595, "percentage": 96.71, "elapsed_time": "15:11:10", "remaining_time": "0:30:57"} +{"current_steps": 4155, "total_steps": 4291, "loss": 0.1788, "lr": 1.2413421656326307e-07, "epoch": 6.778775510204081, "percentage": 96.83, "elapsed_time": "15:12:20", "remaining_time": "0:29:51"} +{"current_steps": 4160, "total_steps": 4291, "loss": 0.1791, "lr": 1.1524722305144231e-07, "epoch": 6.786938775510204, "percentage": 96.95, "elapsed_time": "15:13:24", "remaining_time": "0:28:45"} +{"current_steps": 4165, "total_steps": 4291, "loss": 0.1801, "lr": 1.0668935368746136e-07, "epoch": 6.795102040816326, "percentage": 97.06, "elapsed_time": "15:14:42", "remaining_time": "0:27:40"} +{"current_steps": 4170, "total_steps": 4291, "loss": 0.1712, "lr": 9.846075011761048e-08, "epoch": 6.803265306122449, "percentage": 97.18, "elapsed_time": "15:15:42", "remaining_time": "0:26:34"} +{"current_steps": 4175, "total_steps": 4291, "loss": 0.1786, "lr": 9.056154853830823e-08, "epoch": 6.811428571428571, "percentage": 97.3, "elapsed_time": "15:16:38", "remaining_time": "0:25:28"} +{"current_steps": 4180, "total_steps": 4291, "loss": 0.1685, "lr": 8.299187969385226e-08, "epoch": 6.819591836734694, "percentage": 97.41, "elapsed_time": "15:17:41", "remaining_time": "0:24:22"} +{"current_steps": 4185, "total_steps": 4291, "loss": 0.1696, "lr": 7.575186887424757e-08, "epoch": 6.8277551020408165, "percentage": 97.53, "elapsed_time": "15:18:46", "remaining_time": "0:23:16"} +{"current_steps": 4190, "total_steps": 4291, "loss": 0.162, "lr": 6.884163591313941e-08, "epoch": 6.835918367346939, "percentage": 97.65, "elapsed_time": "15:19:55", "remaining_time": "0:22:10"} +{"current_steps": 4195, "total_steps": 4291, "loss": 0.1722, "lr": 6.226129518582369e-08, "epoch": 6.844081632653062, "percentage": 97.76, "elapsed_time": "15:21:00", "remaining_time": "0:21:04"} +{"current_steps": 4200, "total_steps": 4291, "loss": 0.1681, "lr": 5.6010955607366245e-08, "epoch": 6.852244897959183, "percentage": 97.88, "elapsed_time": "15:22:01", "remaining_time": "0:19:58"} +{"current_steps": 4205, "total_steps": 4291, "loss": 0.185, "lr": 5.009072063077769e-08, "epoch": 6.860408163265306, "percentage": 98.0, "elapsed_time": "15:23:59", "remaining_time": "0:18:53"} +{"current_steps": 4210, "total_steps": 4291, "loss": 0.1779, "lr": 4.450068824532583e-08, "epoch": 6.868571428571428, "percentage": 98.11, "elapsed_time": "15:24:56", "remaining_time": "0:17:47"} +{"current_steps": 4215, "total_steps": 4291, "loss": 0.173, "lr": 3.924095097489922e-08, "epoch": 6.876734693877551, "percentage": 98.23, "elapsed_time": "15:26:01", "remaining_time": "0:16:41"} +{"current_steps": 4220, "total_steps": 4291, "loss": 0.1761, "lr": 3.4311595876479475e-08, "epoch": 6.884897959183673, "percentage": 98.35, "elapsed_time": "15:27:07", "remaining_time": "0:15:35"} +{"current_steps": 4225, "total_steps": 4291, "loss": 0.1733, "lr": 2.971270453869579e-08, "epoch": 6.893061224489796, "percentage": 98.46, "elapsed_time": "15:28:04", "remaining_time": "0:14:29"} +{"current_steps": 4230, "total_steps": 4291, "loss": 0.1784, "lr": 2.5444353080485983e-08, "epoch": 6.9012244897959185, "percentage": 98.58, "elapsed_time": "15:29:14", "remaining_time": "0:13:24"} +{"current_steps": 4235, "total_steps": 4291, "loss": 0.1679, "lr": 2.150661214982641e-08, "epoch": 6.909387755102041, "percentage": 98.69, "elapsed_time": "15:30:11", "remaining_time": "0:12:18"} +{"current_steps": 4240, "total_steps": 4291, "loss": 0.182, "lr": 1.7899546922564016e-08, "epoch": 6.917551020408164, "percentage": 98.81, "elapsed_time": "15:31:18", "remaining_time": "0:11:12"} +{"current_steps": 4245, "total_steps": 4291, "loss": 0.1655, "lr": 1.4623217101339405e-08, "epoch": 6.925714285714285, "percentage": 98.93, "elapsed_time": "15:32:18", "remaining_time": "0:10:06"} +{"current_steps": 4250, "total_steps": 4291, "loss": 0.1645, "lr": 1.1677676914603197e-08, "epoch": 6.933877551020408, "percentage": 99.04, "elapsed_time": "15:33:26", "remaining_time": "0:09:00"} +{"current_steps": 4255, "total_steps": 4291, "loss": 0.178, "lr": 9.06297511570564e-09, "epoch": 6.94204081632653, "percentage": 99.16, "elapsed_time": "15:34:30", "remaining_time": "0:07:54"} +{"current_steps": 4260, "total_steps": 4291, "loss": 0.1617, "lr": 6.779154982103908e-09, "epoch": 6.950204081632653, "percentage": 99.28, "elapsed_time": "15:35:30", "remaining_time": "0:06:48"} +{"current_steps": 4265, "total_steps": 4291, "loss": 0.1806, "lr": 4.8262543146382345e-09, "epoch": 6.9583673469387755, "percentage": 99.39, "elapsed_time": "15:36:38", "remaining_time": "0:05:42"} +{"current_steps": 4270, "total_steps": 4291, "loss": 0.1783, "lr": 3.2043054369057523e-09, "epoch": 6.966530612244898, "percentage": 99.51, "elapsed_time": "15:37:44", "remaining_time": "0:04:36"} +{"current_steps": 4275, "total_steps": 4291, "loss": 0.1772, "lr": 1.9133351947320246e-09, "epoch": 6.9746938775510205, "percentage": 99.63, "elapsed_time": "15:38:51", "remaining_time": "0:03:30"} +{"current_steps": 4280, "total_steps": 4291, "loss": 0.1724, "lr": 9.53364955715852e-10, "epoch": 6.982857142857143, "percentage": 99.74, "elapsed_time": "15:39:52", "remaining_time": "0:02:24"} +{"current_steps": 4285, "total_steps": 4291, "loss": 0.1788, "lr": 3.244106088873267e-10, "epoch": 6.991020408163266, "percentage": 99.86, "elapsed_time": "15:41:01", "remaining_time": "0:01:19"} +{"current_steps": 4290, "total_steps": 4291, "loss": 0.1759, "lr": 2.648256443471553e-11, "epoch": 6.999183673469387, "percentage": 99.98, "elapsed_time": "15:42:04", "remaining_time": "0:00:13"} +{"current_steps": 4291, "total_steps": 4291, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "15:43:04", "remaining_time": "0:00:00"} diff --git a/trainer_state.json b/trainer_state.json new file mode 100644 index 0000000..ed7fe65 --- /dev/null +++ b/trainer_state.json @@ -0,0 +1,9485 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 7.0, + "eval_steps": 500, + "global_step": 4291, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.00816326530612245, + "grad_norm": 12.33506016365843, + "learning_rate": 3.7209302325581396e-07, + "loss": 0.75, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.31827449798583984, + "step": 5, + "valid_targets_mean": 3888.0, + "valid_targets_min": 2300 + }, + { + "epoch": 0.0163265306122449, + "grad_norm": 14.06557282235133, + "learning_rate": 8.372093023255814e-07, + "loss": 0.7413, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4188452959060669, + "step": 10, + "valid_targets_mean": 3943.4, + "valid_targets_min": 1645 + }, + { + "epoch": 0.024489795918367346, + "grad_norm": 12.64563346356677, + "learning_rate": 1.302325581395349e-06, + "loss": 0.7263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.4356209933757782, + "step": 15, + "valid_targets_mean": 5323.4, + "valid_targets_min": 2731 + }, + { + "epoch": 0.0326530612244898, + "grad_norm": 7.5861629954645045, + "learning_rate": 1.7674418604651164e-06, + "loss": 0.7189, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3566373586654663, + "step": 20, + "valid_targets_mean": 4771.5, + "valid_targets_min": 231 + }, + { + "epoch": 0.04081632653061224, + "grad_norm": 5.884180511617331, + "learning_rate": 2.232558139534884e-06, + "loss": 0.6782, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2910090386867523, + "step": 25, + "valid_targets_mean": 2981.0, + "valid_targets_min": 697 + }, + { + "epoch": 0.04897959183673469, + "grad_norm": 4.048322161590491, + "learning_rate": 2.6976744186046517e-06, + "loss": 0.5857, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2788548767566681, + "step": 30, + "valid_targets_mean": 4749.9, + "valid_targets_min": 2045 + }, + { + "epoch": 0.05714285714285714, + "grad_norm": 3.8721871492073894, + "learning_rate": 3.1627906976744187e-06, + "loss": 0.5584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2676370143890381, + "step": 35, + "valid_targets_mean": 4417.5, + "valid_targets_min": 1155 + }, + { + "epoch": 0.0653061224489796, + "grad_norm": 2.1663266935097294, + "learning_rate": 3.6279069767441866e-06, + "loss": 0.5306, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28581827878952026, + "step": 40, + "valid_targets_mean": 4260.9, + "valid_targets_min": 2287 + }, + { + "epoch": 0.07346938775510205, + "grad_norm": 1.4335626902963403, + "learning_rate": 4.0930232558139536e-06, + "loss": 0.5073, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28230318427085876, + "step": 45, + "valid_targets_mean": 4514.6, + "valid_targets_min": 1947 + }, + { + "epoch": 0.08163265306122448, + "grad_norm": 0.8187625274651436, + "learning_rate": 4.558139534883721e-06, + "loss": 0.463, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1898300051689148, + "step": 50, + "valid_targets_mean": 6367.4, + "valid_targets_min": 1883 + }, + { + "epoch": 0.08979591836734693, + "grad_norm": 0.9581011591667019, + "learning_rate": 5.023255813953489e-06, + "loss": 0.4861, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22689494490623474, + "step": 55, + "valid_targets_mean": 3721.6, + "valid_targets_min": 2174 + }, + { + "epoch": 0.09795918367346938, + "grad_norm": 0.688459144076369, + "learning_rate": 5.488372093023256e-06, + "loss": 0.4585, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21300674974918365, + "step": 60, + "valid_targets_mean": 5311.5, + "valid_targets_min": 179 + }, + { + "epoch": 0.10612244897959183, + "grad_norm": 0.7211400009988276, + "learning_rate": 5.953488372093023e-06, + "loss": 0.4672, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2803425192832947, + "step": 65, + "valid_targets_mean": 5541.2, + "valid_targets_min": 225 + }, + { + "epoch": 0.11428571428571428, + "grad_norm": 0.6534824892302817, + "learning_rate": 6.418604651162791e-06, + "loss": 0.44, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2042653113603592, + "step": 70, + "valid_targets_mean": 3739.4, + "valid_targets_min": 1544 + }, + { + "epoch": 0.12244897959183673, + "grad_norm": 0.7238400984773256, + "learning_rate": 6.883720930232559e-06, + "loss": 0.4608, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2297825664281845, + "step": 75, + "valid_targets_mean": 3874.8, + "valid_targets_min": 2090 + }, + { + "epoch": 0.1306122448979592, + "grad_norm": 0.6261187440459337, + "learning_rate": 7.348837209302326e-06, + "loss": 0.4456, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20637601613998413, + "step": 80, + "valid_targets_mean": 4730.2, + "valid_targets_min": 2225 + }, + { + "epoch": 0.13877551020408163, + "grad_norm": 0.6083383898632471, + "learning_rate": 7.813953488372094e-06, + "loss": 0.4337, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21255721151828766, + "step": 85, + "valid_targets_mean": 4527.2, + "valid_targets_min": 2565 + }, + { + "epoch": 0.1469387755102041, + "grad_norm": 0.49647727411141274, + "learning_rate": 8.279069767441861e-06, + "loss": 0.3833, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18475192785263062, + "step": 90, + "valid_targets_mean": 5922.8, + "valid_targets_min": 2468 + }, + { + "epoch": 0.15510204081632653, + "grad_norm": 0.6578636753926649, + "learning_rate": 8.74418604651163e-06, + "loss": 0.4174, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19449181854724884, + "step": 95, + "valid_targets_mean": 4025.9, + "valid_targets_min": 2181 + }, + { + "epoch": 0.16326530612244897, + "grad_norm": 0.5554069864757591, + "learning_rate": 9.209302325581397e-06, + "loss": 0.4086, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2075560986995697, + "step": 100, + "valid_targets_mean": 5168.0, + "valid_targets_min": 2191 + }, + { + "epoch": 0.17142857142857143, + "grad_norm": 0.5442721349991868, + "learning_rate": 9.674418604651164e-06, + "loss": 0.3998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17283964157104492, + "step": 105, + "valid_targets_mean": 4438.6, + "valid_targets_min": 1285 + }, + { + "epoch": 0.17959183673469387, + "grad_norm": 0.5267931849085247, + "learning_rate": 1.0139534883720932e-05, + "loss": 0.3843, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20762407779693604, + "step": 110, + "valid_targets_mean": 6056.4, + "valid_targets_min": 2910 + }, + { + "epoch": 0.18775510204081633, + "grad_norm": 0.5536849451475337, + "learning_rate": 1.0604651162790698e-05, + "loss": 0.3954, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20768246054649353, + "step": 115, + "valid_targets_mean": 4792.0, + "valid_targets_min": 864 + }, + { + "epoch": 0.19591836734693877, + "grad_norm": 0.5236424454710094, + "learning_rate": 1.1069767441860466e-05, + "loss": 0.3875, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21655716001987457, + "step": 120, + "valid_targets_mean": 6482.5, + "valid_targets_min": 1500 + }, + { + "epoch": 0.20408163265306123, + "grad_norm": 0.5025306927483865, + "learning_rate": 1.1534883720930235e-05, + "loss": 0.3829, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17163898050785065, + "step": 125, + "valid_targets_mean": 4536.0, + "valid_targets_min": 2532 + }, + { + "epoch": 0.21224489795918366, + "grad_norm": 0.4284847923366298, + "learning_rate": 1.2e-05, + "loss": 0.3573, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13200239837169647, + "step": 130, + "valid_targets_mean": 4378.4, + "valid_targets_min": 1834 + }, + { + "epoch": 0.22040816326530613, + "grad_norm": 0.5724886335853256, + "learning_rate": 1.2465116279069769e-05, + "loss": 0.3791, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17680321633815765, + "step": 135, + "valid_targets_mean": 4725.5, + "valid_targets_min": 2184 + }, + { + "epoch": 0.22857142857142856, + "grad_norm": 0.5262974819415349, + "learning_rate": 1.2930232558139534e-05, + "loss": 0.3689, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1830161213874817, + "step": 140, + "valid_targets_mean": 5207.1, + "valid_targets_min": 1525 + }, + { + "epoch": 0.23673469387755103, + "grad_norm": 0.5943237872366759, + "learning_rate": 1.3395348837209303e-05, + "loss": 0.3841, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16007442772388458, + "step": 145, + "valid_targets_mean": 4077.4, + "valid_targets_min": 3143 + }, + { + "epoch": 0.24489795918367346, + "grad_norm": 0.5011629694901417, + "learning_rate": 1.3860465116279072e-05, + "loss": 0.3655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21169030666351318, + "step": 150, + "valid_targets_mean": 7955.6, + "valid_targets_min": 2451 + }, + { + "epoch": 0.2530612244897959, + "grad_norm": 0.5412588141865855, + "learning_rate": 1.4325581395348837e-05, + "loss": 0.3682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1689492166042328, + "step": 155, + "valid_targets_mean": 5123.8, + "valid_targets_min": 2164 + }, + { + "epoch": 0.2612244897959184, + "grad_norm": 0.6643295450246909, + "learning_rate": 1.4790697674418606e-05, + "loss": 0.3562, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1859186887741089, + "step": 160, + "valid_targets_mean": 6049.8, + "valid_targets_min": 3021 + }, + { + "epoch": 0.2693877551020408, + "grad_norm": 0.597344608075607, + "learning_rate": 1.5255813953488374e-05, + "loss": 0.3692, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15211470425128937, + "step": 165, + "valid_targets_mean": 3266.6, + "valid_targets_min": 1135 + }, + { + "epoch": 0.27755102040816326, + "grad_norm": 0.6202606312284262, + "learning_rate": 1.572093023255814e-05, + "loss": 0.3689, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16817247867584229, + "step": 170, + "valid_targets_mean": 3775.8, + "valid_targets_min": 2136 + }, + { + "epoch": 0.2857142857142857, + "grad_norm": 0.5839399121332536, + "learning_rate": 1.618604651162791e-05, + "loss": 0.3431, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15334473550319672, + "step": 175, + "valid_targets_mean": 4572.5, + "valid_targets_min": 2188 + }, + { + "epoch": 0.2938775510204082, + "grad_norm": 0.46689156867921305, + "learning_rate": 1.6651162790697674e-05, + "loss": 0.3154, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1432073712348938, + "step": 180, + "valid_targets_mean": 5915.1, + "valid_targets_min": 2132 + }, + { + "epoch": 0.3020408163265306, + "grad_norm": 0.5309548508364261, + "learning_rate": 1.7116279069767444e-05, + "loss": 0.3333, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20056375861167908, + "step": 185, + "valid_targets_mean": 6440.8, + "valid_targets_min": 2163 + }, + { + "epoch": 0.31020408163265306, + "grad_norm": 0.5604620864213024, + "learning_rate": 1.758139534883721e-05, + "loss": 0.3352, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1434939056634903, + "step": 190, + "valid_targets_mean": 4429.0, + "valid_targets_min": 2288 + }, + { + "epoch": 0.3183673469387755, + "grad_norm": 0.5450998899560288, + "learning_rate": 1.8046511627906978e-05, + "loss": 0.345, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17527905106544495, + "step": 195, + "valid_targets_mean": 5860.5, + "valid_targets_min": 1828 + }, + { + "epoch": 0.32653061224489793, + "grad_norm": 0.5852906718193894, + "learning_rate": 1.8511627906976745e-05, + "loss": 0.3586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15143322944641113, + "step": 200, + "valid_targets_mean": 4496.5, + "valid_targets_min": 1416 + }, + { + "epoch": 0.3346938775510204, + "grad_norm": 0.6271764625641036, + "learning_rate": 1.8976744186046516e-05, + "loss": 0.3368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14210546016693115, + "step": 205, + "valid_targets_mean": 3699.0, + "valid_targets_min": 2282 + }, + { + "epoch": 0.34285714285714286, + "grad_norm": 0.5616628504866027, + "learning_rate": 1.944186046511628e-05, + "loss": 0.3404, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2095171958208084, + "step": 210, + "valid_targets_mean": 5037.4, + "valid_targets_min": 1231 + }, + { + "epoch": 0.3510204081632653, + "grad_norm": 0.5615575210169799, + "learning_rate": 1.990697674418605e-05, + "loss": 0.3455, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1549168974161148, + "step": 215, + "valid_targets_mean": 4484.1, + "valid_targets_min": 665 + }, + { + "epoch": 0.35918367346938773, + "grad_norm": 0.507550652268599, + "learning_rate": 2.0372093023255813e-05, + "loss": 0.3256, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1841312050819397, + "step": 220, + "valid_targets_mean": 6250.6, + "valid_targets_min": 3309 + }, + { + "epoch": 0.3673469387755102, + "grad_norm": 0.5094360914962326, + "learning_rate": 2.0837209302325584e-05, + "loss": 0.2986, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1224166676402092, + "step": 225, + "valid_targets_mean": 3814.6, + "valid_targets_min": 2198 + }, + { + "epoch": 0.37551020408163266, + "grad_norm": 0.6246439139516935, + "learning_rate": 2.130232558139535e-05, + "loss": 0.3275, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1374599039554596, + "step": 230, + "valid_targets_mean": 3553.8, + "valid_targets_min": 217 + }, + { + "epoch": 0.3836734693877551, + "grad_norm": 0.5951069221562814, + "learning_rate": 2.176744186046512e-05, + "loss": 0.3489, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18446165323257446, + "step": 235, + "valid_targets_mean": 4747.2, + "valid_targets_min": 1414 + }, + { + "epoch": 0.39183673469387753, + "grad_norm": 0.6523853794671483, + "learning_rate": 2.2232558139534888e-05, + "loss": 0.333, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15122082829475403, + "step": 240, + "valid_targets_mean": 3160.5, + "valid_targets_min": 1638 + }, + { + "epoch": 0.4, + "grad_norm": 0.4725934988884912, + "learning_rate": 2.269767441860465e-05, + "loss": 0.3198, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13313765823841095, + "step": 245, + "valid_targets_mean": 6339.0, + "valid_targets_min": 1913 + }, + { + "epoch": 0.40816326530612246, + "grad_norm": 0.49423205201340153, + "learning_rate": 2.316279069767442e-05, + "loss": 0.3368, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16928038001060486, + "step": 250, + "valid_targets_mean": 7214.2, + "valid_targets_min": 3139 + }, + { + "epoch": 0.4163265306122449, + "grad_norm": 0.4974690927027136, + "learning_rate": 2.362790697674419e-05, + "loss": 0.3345, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16831715404987335, + "step": 255, + "valid_targets_mean": 5773.4, + "valid_targets_min": 2589 + }, + { + "epoch": 0.42448979591836733, + "grad_norm": 0.5602390366587728, + "learning_rate": 2.4093023255813956e-05, + "loss": 0.3444, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17655014991760254, + "step": 260, + "valid_targets_mean": 5242.8, + "valid_targets_min": 2248 + }, + { + "epoch": 0.4326530612244898, + "grad_norm": 0.6143492193154322, + "learning_rate": 2.4558139534883726e-05, + "loss": 0.3311, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16245931386947632, + "step": 265, + "valid_targets_mean": 4020.8, + "valid_targets_min": 1845 + }, + { + "epoch": 0.44081632653061226, + "grad_norm": 0.5185982721162242, + "learning_rate": 2.502325581395349e-05, + "loss": 0.3395, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.169978067278862, + "step": 270, + "valid_targets_mean": 5750.5, + "valid_targets_min": 1634 + }, + { + "epoch": 0.4489795918367347, + "grad_norm": 0.5510287595838287, + "learning_rate": 2.5488372093023257e-05, + "loss": 0.3273, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15716497600078583, + "step": 275, + "valid_targets_mean": 6445.8, + "valid_targets_min": 1772 + }, + { + "epoch": 0.45714285714285713, + "grad_norm": 0.5410082300355423, + "learning_rate": 2.5953488372093024e-05, + "loss": 0.3111, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19414874911308289, + "step": 280, + "valid_targets_mean": 5556.1, + "valid_targets_min": 1816 + }, + { + "epoch": 0.46530612244897956, + "grad_norm": 0.5707835911746898, + "learning_rate": 2.6418604651162794e-05, + "loss": 0.2959, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16951584815979004, + "step": 285, + "valid_targets_mean": 6885.9, + "valid_targets_min": 1210 + }, + { + "epoch": 0.47346938775510206, + "grad_norm": 0.6040168533013804, + "learning_rate": 2.688372093023256e-05, + "loss": 0.3089, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1755496859550476, + "step": 290, + "valid_targets_mean": 4725.0, + "valid_targets_min": 1232 + }, + { + "epoch": 0.4816326530612245, + "grad_norm": 0.5954085643873348, + "learning_rate": 2.7348837209302325e-05, + "loss": 0.3217, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13682439923286438, + "step": 295, + "valid_targets_mean": 4376.2, + "valid_targets_min": 2137 + }, + { + "epoch": 0.4897959183673469, + "grad_norm": 0.6431457999458166, + "learning_rate": 2.7813953488372095e-05, + "loss": 0.3085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1722913682460785, + "step": 300, + "valid_targets_mean": 4122.9, + "valid_targets_min": 2335 + }, + { + "epoch": 0.49795918367346936, + "grad_norm": 0.5475282840206034, + "learning_rate": 2.8279069767441862e-05, + "loss": 0.3217, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14570221304893494, + "step": 305, + "valid_targets_mean": 3578.1, + "valid_targets_min": 2075 + }, + { + "epoch": 0.5061224489795918, + "grad_norm": 0.5977501757018006, + "learning_rate": 2.874418604651163e-05, + "loss": 0.3227, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13097640872001648, + "step": 310, + "valid_targets_mean": 3568.1, + "valid_targets_min": 1950 + }, + { + "epoch": 0.5142857142857142, + "grad_norm": 0.6119511908814466, + "learning_rate": 2.92093023255814e-05, + "loss": 0.3193, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1389477699995041, + "step": 315, + "valid_targets_mean": 3503.9, + "valid_targets_min": 1251 + }, + { + "epoch": 0.5224489795918368, + "grad_norm": 0.5252404379494299, + "learning_rate": 2.9674418604651167e-05, + "loss": 0.3192, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1406920701265335, + "step": 320, + "valid_targets_mean": 4817.9, + "valid_targets_min": 2302 + }, + { + "epoch": 0.5306122448979592, + "grad_norm": 0.5672213393499427, + "learning_rate": 3.013953488372093e-05, + "loss": 0.3121, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17177443206310272, + "step": 325, + "valid_targets_mean": 5428.4, + "valid_targets_min": 1917 + }, + { + "epoch": 0.5387755102040817, + "grad_norm": 0.5873908296751622, + "learning_rate": 3.06046511627907e-05, + "loss": 0.3149, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1694718897342682, + "step": 330, + "valid_targets_mean": 4601.1, + "valid_targets_min": 2799 + }, + { + "epoch": 0.5469387755102041, + "grad_norm": 0.5200193608156581, + "learning_rate": 3.1069767441860465e-05, + "loss": 0.3143, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1358920931816101, + "step": 335, + "valid_targets_mean": 5289.6, + "valid_targets_min": 3964 + }, + { + "epoch": 0.5551020408163265, + "grad_norm": 0.6481598582680852, + "learning_rate": 3.1534883720930235e-05, + "loss": 0.3014, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17161871492862701, + "step": 340, + "valid_targets_mean": 4119.6, + "valid_targets_min": 1883 + }, + { + "epoch": 0.563265306122449, + "grad_norm": 0.5109022013798562, + "learning_rate": 3.2000000000000005e-05, + "loss": 0.313, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13416041433811188, + "step": 345, + "valid_targets_mean": 4991.4, + "valid_targets_min": 657 + }, + { + "epoch": 0.5714285714285714, + "grad_norm": 0.5339499139390554, + "learning_rate": 3.246511627906977e-05, + "loss": 0.2967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14657235145568848, + "step": 350, + "valid_targets_mean": 5172.1, + "valid_targets_min": 2580 + }, + { + "epoch": 0.5795918367346938, + "grad_norm": 0.6023496113208678, + "learning_rate": 3.293023255813954e-05, + "loss": 0.3015, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1426255851984024, + "step": 355, + "valid_targets_mean": 5163.6, + "valid_targets_min": 2359 + }, + { + "epoch": 0.5877551020408164, + "grad_norm": 0.6011243129435078, + "learning_rate": 3.33953488372093e-05, + "loss": 0.3067, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19121070206165314, + "step": 360, + "valid_targets_mean": 4767.0, + "valid_targets_min": 2148 + }, + { + "epoch": 0.5959183673469388, + "grad_norm": 0.6749077948841615, + "learning_rate": 3.386046511627907e-05, + "loss": 0.3107, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18281207978725433, + "step": 365, + "valid_targets_mean": 3520.6, + "valid_targets_min": 1283 + }, + { + "epoch": 0.6040816326530613, + "grad_norm": 0.5445436044445798, + "learning_rate": 3.4325581395348844e-05, + "loss": 0.2927, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12598849833011627, + "step": 370, + "valid_targets_mean": 5079.1, + "valid_targets_min": 1151 + }, + { + "epoch": 0.6122448979591837, + "grad_norm": 0.698927419975243, + "learning_rate": 3.479069767441861e-05, + "loss": 0.306, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1887175440788269, + "step": 375, + "valid_targets_mean": 5366.5, + "valid_targets_min": 1548 + }, + { + "epoch": 0.6204081632653061, + "grad_norm": 0.5846609611760529, + "learning_rate": 3.525581395348837e-05, + "loss": 0.3136, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15284685790538788, + "step": 380, + "valid_targets_mean": 3979.0, + "valid_targets_min": 1392 + }, + { + "epoch": 0.6285714285714286, + "grad_norm": 0.5376221163617384, + "learning_rate": 3.572093023255814e-05, + "loss": 0.2908, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1378687173128128, + "step": 385, + "valid_targets_mean": 4058.5, + "valid_targets_min": 1270 + }, + { + "epoch": 0.636734693877551, + "grad_norm": 0.5530069475016968, + "learning_rate": 3.618604651162791e-05, + "loss": 0.3002, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13155528903007507, + "step": 390, + "valid_targets_mean": 4221.2, + "valid_targets_min": 1822 + }, + { + "epoch": 0.6448979591836734, + "grad_norm": 0.4978881241002247, + "learning_rate": 3.6651162790697675e-05, + "loss": 0.3036, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11640087515115738, + "step": 395, + "valid_targets_mean": 3738.6, + "valid_targets_min": 2168 + }, + { + "epoch": 0.6530612244897959, + "grad_norm": 0.5430912992210294, + "learning_rate": 3.7116279069767446e-05, + "loss": 0.2896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.155136376619339, + "step": 400, + "valid_targets_mean": 5601.9, + "valid_targets_min": 3271 + }, + { + "epoch": 0.6612244897959184, + "grad_norm": 0.49528034996711595, + "learning_rate": 3.758139534883721e-05, + "loss": 0.3105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13687986135482788, + "step": 405, + "valid_targets_mean": 5079.8, + "valid_targets_min": 2260 + }, + { + "epoch": 0.6693877551020408, + "grad_norm": 0.5921551637945828, + "learning_rate": 3.804651162790698e-05, + "loss": 0.3186, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11712108552455902, + "step": 410, + "valid_targets_mean": 2763.8, + "valid_targets_min": 1469 + }, + { + "epoch": 0.6775510204081633, + "grad_norm": 0.43782810606243927, + "learning_rate": 3.851162790697675e-05, + "loss": 0.29, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09977801144123077, + "step": 415, + "valid_targets_mean": 5919.1, + "valid_targets_min": 2181 + }, + { + "epoch": 0.6857142857142857, + "grad_norm": 0.4890700657801529, + "learning_rate": 3.8976744186046514e-05, + "loss": 0.2986, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11042569577693939, + "step": 420, + "valid_targets_mean": 3841.8, + "valid_targets_min": 1487 + }, + { + "epoch": 0.6938775510204082, + "grad_norm": 0.5529585097612127, + "learning_rate": 3.9441860465116284e-05, + "loss": 0.2987, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14143112301826477, + "step": 425, + "valid_targets_mean": 4780.0, + "valid_targets_min": 1625 + }, + { + "epoch": 0.7020408163265306, + "grad_norm": 0.5508262572064092, + "learning_rate": 3.990697674418605e-05, + "loss": 0.2894, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1426255702972412, + "step": 430, + "valid_targets_mean": 5180.5, + "valid_targets_min": 2789 + }, + { + "epoch": 0.710204081632653, + "grad_norm": 0.5066137206557315, + "learning_rate": 3.99998940698124e-05, + "loss": 0.3196, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16803446412086487, + "step": 435, + "valid_targets_mean": 5961.6, + "valid_targets_min": 3011 + }, + { + "epoch": 0.7183673469387755, + "grad_norm": 0.5025096770678751, + "learning_rate": 3.999946373034843e-05, + "loss": 0.2733, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15143230557441711, + "step": 440, + "valid_targets_mean": 6720.8, + "valid_targets_min": 3227 + }, + { + "epoch": 0.726530612244898, + "grad_norm": 0.6002164385003982, + "learning_rate": 3.999870236808868e-05, + "loss": 0.2884, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1483706831932068, + "step": 445, + "valid_targets_mean": 4109.9, + "valid_targets_min": 215 + }, + { + "epoch": 0.7346938775510204, + "grad_norm": 0.7700094830439378, + "learning_rate": 3.999760999563491e-05, + "loss": 0.2935, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14717252552509308, + "step": 450, + "valid_targets_mean": 4119.0, + "valid_targets_min": 2148 + }, + { + "epoch": 0.7428571428571429, + "grad_norm": 0.46482998773287904, + "learning_rate": 3.999618663106761e-05, + "loss": 0.2933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1267111450433731, + "step": 455, + "valid_targets_mean": 7414.5, + "valid_targets_min": 1461 + }, + { + "epoch": 0.7510204081632653, + "grad_norm": 0.5950339538598457, + "learning_rate": 3.999443229794572e-05, + "loss": 0.2951, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14380259811878204, + "step": 460, + "valid_targets_mean": 3798.4, + "valid_targets_min": 1922 + }, + { + "epoch": 0.7591836734693878, + "grad_norm": 0.5308130390938092, + "learning_rate": 3.9992347025306236e-05, + "loss": 0.2931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1623890995979309, + "step": 465, + "valid_targets_mean": 5759.2, + "valid_targets_min": 2835 + }, + { + "epoch": 0.7673469387755102, + "grad_norm": 0.48879962344124256, + "learning_rate": 3.9989930847663706e-05, + "loss": 0.2778, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13271786272525787, + "step": 470, + "valid_targets_mean": 5688.9, + "valid_targets_min": 1937 + }, + { + "epoch": 0.7755102040816326, + "grad_norm": 0.5934438944710013, + "learning_rate": 3.998718380500971e-05, + "loss": 0.292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1767587959766388, + "step": 475, + "valid_targets_mean": 4823.4, + "valid_targets_min": 2594 + }, + { + "epoch": 0.7836734693877551, + "grad_norm": 0.5617745043664201, + "learning_rate": 3.998410594281215e-05, + "loss": 0.2923, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14815029501914978, + "step": 480, + "valid_targets_mean": 4843.0, + "valid_targets_min": 2702 + }, + { + "epoch": 0.7918367346938775, + "grad_norm": 0.5280138408881649, + "learning_rate": 3.998069731201452e-05, + "loss": 0.2956, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15829509496688843, + "step": 485, + "valid_targets_mean": 5016.4, + "valid_targets_min": 1744 + }, + { + "epoch": 0.8, + "grad_norm": 0.5374828402758218, + "learning_rate": 3.997695796903506e-05, + "loss": 0.2764, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14506173133850098, + "step": 490, + "valid_targets_mean": 5647.2, + "valid_targets_min": 556 + }, + { + "epoch": 0.8081632653061225, + "grad_norm": 0.5363773773173921, + "learning_rate": 3.997288797576581e-05, + "loss": 0.2857, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13211554288864136, + "step": 495, + "valid_targets_mean": 5134.4, + "valid_targets_min": 2075 + }, + { + "epoch": 0.8163265306122449, + "grad_norm": 0.5371091766380987, + "learning_rate": 3.996848739957162e-05, + "loss": 0.2943, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14844098687171936, + "step": 500, + "valid_targets_mean": 4823.0, + "valid_targets_min": 2984 + }, + { + "epoch": 0.8244897959183674, + "grad_norm": 0.5553565489257202, + "learning_rate": 3.996375631328898e-05, + "loss": 0.3024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.174277663230896, + "step": 505, + "valid_targets_mean": 5254.9, + "valid_targets_min": 2011 + }, + { + "epoch": 0.8326530612244898, + "grad_norm": 0.6386966753151678, + "learning_rate": 3.995869479522486e-05, + "loss": 0.3079, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14175033569335938, + "step": 510, + "valid_targets_mean": 3835.4, + "valid_targets_min": 2278 + }, + { + "epoch": 0.8408163265306122, + "grad_norm": 0.5488250127766011, + "learning_rate": 3.99533029291554e-05, + "loss": 0.2807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14564122259616852, + "step": 515, + "valid_targets_mean": 3774.8, + "valid_targets_min": 1732 + }, + { + "epoch": 0.8489795918367347, + "grad_norm": 0.47254565574720186, + "learning_rate": 3.994758080432453e-05, + "loss": 0.2915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1435089409351349, + "step": 520, + "valid_targets_mean": 4876.8, + "valid_targets_min": 1882 + }, + { + "epoch": 0.8571428571428571, + "grad_norm": 0.5007153163086048, + "learning_rate": 3.994152851544248e-05, + "loss": 0.287, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1679384410381317, + "step": 525, + "valid_targets_mean": 5747.9, + "valid_targets_min": 1080 + }, + { + "epoch": 0.8653061224489796, + "grad_norm": 1.2081947619323428, + "learning_rate": 3.9935146162684206e-05, + "loss": 0.2786, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12986889481544495, + "step": 530, + "valid_targets_mean": 5772.0, + "valid_targets_min": 1397 + }, + { + "epoch": 0.8734693877551021, + "grad_norm": 0.4993470220190206, + "learning_rate": 3.992843385168775e-05, + "loss": 0.2746, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1556703746318817, + "step": 535, + "valid_targets_mean": 6776.5, + "valid_targets_min": 1956 + }, + { + "epoch": 0.8816326530612245, + "grad_norm": 0.5776232745040066, + "learning_rate": 3.99213916935525e-05, + "loss": 0.2898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13914015889167786, + "step": 540, + "valid_targets_mean": 4373.4, + "valid_targets_min": 703 + }, + { + "epoch": 0.889795918367347, + "grad_norm": 0.4812680988740483, + "learning_rate": 3.991401980483731e-05, + "loss": 0.2792, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1518934965133667, + "step": 545, + "valid_targets_mean": 6895.9, + "valid_targets_min": 2243 + }, + { + "epoch": 0.8979591836734694, + "grad_norm": 0.5738308130244382, + "learning_rate": 3.990631830755863e-05, + "loss": 0.2863, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1788262128829956, + "step": 550, + "valid_targets_mean": 4816.4, + "valid_targets_min": 2138 + }, + { + "epoch": 0.9061224489795918, + "grad_norm": 0.5224458825908416, + "learning_rate": 3.9898287329188426e-05, + "loss": 0.292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14131759107112885, + "step": 555, + "valid_targets_mean": 4659.0, + "valid_targets_min": 1993 + }, + { + "epoch": 0.9142857142857143, + "grad_norm": 0.5698616740709329, + "learning_rate": 3.988992700265211e-05, + "loss": 0.2751, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14461690187454224, + "step": 560, + "valid_targets_mean": 3929.5, + "valid_targets_min": 1063 + }, + { + "epoch": 0.9224489795918367, + "grad_norm": 0.543077032885466, + "learning_rate": 3.988123746632632e-05, + "loss": 0.2664, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13961417973041534, + "step": 565, + "valid_targets_mean": 4097.1, + "valid_targets_min": 1526 + }, + { + "epoch": 0.9306122448979591, + "grad_norm": 0.5536871068210806, + "learning_rate": 3.9872218864036656e-05, + "loss": 0.2805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12596586346626282, + "step": 570, + "valid_targets_mean": 4359.4, + "valid_targets_min": 1569 + }, + { + "epoch": 0.9387755102040817, + "grad_norm": 0.5691487205064515, + "learning_rate": 3.986287134505525e-05, + "loss": 0.2866, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11891940236091614, + "step": 575, + "valid_targets_mean": 4227.9, + "valid_targets_min": 2328 + }, + { + "epoch": 0.9469387755102041, + "grad_norm": 0.5175770297779654, + "learning_rate": 3.985319506409836e-05, + "loss": 0.2864, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14197929203510284, + "step": 580, + "valid_targets_mean": 4714.2, + "valid_targets_min": 2851 + }, + { + "epoch": 0.9551020408163265, + "grad_norm": 0.4841626589697453, + "learning_rate": 3.9843190181323744e-05, + "loss": 0.2805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1410512924194336, + "step": 585, + "valid_targets_mean": 6033.8, + "valid_targets_min": 2286 + }, + { + "epoch": 0.963265306122449, + "grad_norm": 0.5070780388430243, + "learning_rate": 3.9832856862328074e-05, + "loss": 0.2883, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14377368986606598, + "step": 590, + "valid_targets_mean": 5524.2, + "valid_targets_min": 2156 + }, + { + "epoch": 0.9714285714285714, + "grad_norm": 0.5420150728069583, + "learning_rate": 3.982219527814413e-05, + "loss": 0.2788, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16339737176895142, + "step": 595, + "valid_targets_mean": 5497.6, + "valid_targets_min": 2958 + }, + { + "epoch": 0.9795918367346939, + "grad_norm": 0.4456547760032798, + "learning_rate": 3.981120560523802e-05, + "loss": 0.2731, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13854515552520752, + "step": 600, + "valid_targets_mean": 6230.0, + "valid_targets_min": 2322 + }, + { + "epoch": 0.9877551020408163, + "grad_norm": 0.5810811547000726, + "learning_rate": 3.979988802550622e-05, + "loss": 0.2832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1590469479560852, + "step": 605, + "valid_targets_mean": 4077.9, + "valid_targets_min": 2857 + }, + { + "epoch": 0.9959183673469387, + "grad_norm": 0.4891857085078456, + "learning_rate": 3.978824272627262e-05, + "loss": 0.2689, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11531509459018707, + "step": 610, + "valid_targets_mean": 4370.0, + "valid_targets_min": 1733 + }, + { + "epoch": 1.003265306122449, + "grad_norm": 0.5532155013235863, + "learning_rate": 3.977626990028535e-05, + "loss": 0.2645, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11366681754589081, + "step": 615, + "valid_targets_mean": 3601.6, + "valid_targets_min": 2097 + }, + { + "epoch": 1.0114285714285713, + "grad_norm": 0.45978091826871326, + "learning_rate": 3.976396974571364e-05, + "loss": 0.2485, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11303171515464783, + "step": 620, + "valid_targets_mean": 6528.6, + "valid_targets_min": 2089 + }, + { + "epoch": 1.019591836734694, + "grad_norm": 0.4523914061946981, + "learning_rate": 3.9751342466144526e-05, + "loss": 0.2686, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1494063138961792, + "step": 625, + "valid_targets_mean": 7086.5, + "valid_targets_min": 2442 + }, + { + "epoch": 1.0277551020408162, + "grad_norm": 0.544225989094849, + "learning_rate": 3.9738388270579484e-05, + "loss": 0.2776, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14501337707042694, + "step": 630, + "valid_targets_mean": 4636.4, + "valid_targets_min": 2459 + }, + { + "epoch": 1.0359183673469388, + "grad_norm": 0.4867306661490911, + "learning_rate": 3.972510737343097e-05, + "loss": 0.2474, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10430733859539032, + "step": 635, + "valid_targets_mean": 4400.8, + "valid_targets_min": 2464 + }, + { + "epoch": 1.0440816326530613, + "grad_norm": 0.4444263422457097, + "learning_rate": 3.971149999451886e-05, + "loss": 0.2585, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11613980680704117, + "step": 640, + "valid_targets_mean": 5829.5, + "valid_targets_min": 1063 + }, + { + "epoch": 1.0522448979591836, + "grad_norm": 0.5230816472689035, + "learning_rate": 3.9697566359066805e-05, + "loss": 0.2661, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.146030992269516, + "step": 645, + "valid_targets_mean": 5094.1, + "valid_targets_min": 2429 + }, + { + "epoch": 1.0604081632653062, + "grad_norm": 0.47487568623337845, + "learning_rate": 3.968330669769857e-05, + "loss": 0.2544, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1315501481294632, + "step": 650, + "valid_targets_mean": 5665.8, + "valid_targets_min": 2710 + }, + { + "epoch": 1.0685714285714285, + "grad_norm": 0.49799195804012064, + "learning_rate": 3.966872124643411e-05, + "loss": 0.2703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1552223563194275, + "step": 655, + "valid_targets_mean": 5938.4, + "valid_targets_min": 1687 + }, + { + "epoch": 1.076734693877551, + "grad_norm": 0.5297556483893657, + "learning_rate": 3.9653810246685746e-05, + "loss": 0.2701, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13517478108406067, + "step": 660, + "valid_targets_mean": 4919.1, + "valid_targets_min": 1791 + }, + { + "epoch": 1.0848979591836734, + "grad_norm": 0.5601908106981581, + "learning_rate": 3.963857394525413e-05, + "loss": 0.2726, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1491164267063141, + "step": 665, + "valid_targets_mean": 5137.8, + "valid_targets_min": 2264 + }, + { + "epoch": 1.093061224489796, + "grad_norm": 0.4525683784590435, + "learning_rate": 3.96230125943242e-05, + "loss": 0.2559, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.113084577023983, + "step": 670, + "valid_targets_mean": 5150.8, + "valid_targets_min": 2168 + }, + { + "epoch": 1.1012244897959185, + "grad_norm": 0.519392037612358, + "learning_rate": 3.960712645146094e-05, + "loss": 0.2717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13259488344192505, + "step": 675, + "valid_targets_mean": 4962.0, + "valid_targets_min": 2038 + }, + { + "epoch": 1.1093877551020408, + "grad_norm": 0.5539538809708434, + "learning_rate": 3.959091577960517e-05, + "loss": 0.267, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14472344517707825, + "step": 680, + "valid_targets_mean": 4567.8, + "valid_targets_min": 2960 + }, + { + "epoch": 1.1175510204081633, + "grad_norm": 0.4887317566930953, + "learning_rate": 3.957438084706919e-05, + "loss": 0.2567, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10747165977954865, + "step": 685, + "valid_targets_mean": 4679.0, + "valid_targets_min": 2378 + }, + { + "epoch": 1.1257142857142857, + "grad_norm": 0.555310737006963, + "learning_rate": 3.955752192753231e-05, + "loss": 0.2669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12046556919813156, + "step": 690, + "valid_targets_mean": 4157.1, + "valid_targets_min": 2181 + }, + { + "epoch": 1.1338775510204082, + "grad_norm": 0.5121166772677435, + "learning_rate": 3.954033930003634e-05, + "loss": 0.2782, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1369376927614212, + "step": 695, + "valid_targets_mean": 5442.1, + "valid_targets_min": 3151 + }, + { + "epoch": 1.1420408163265305, + "grad_norm": 0.531503784823624, + "learning_rate": 3.9522833248980995e-05, + "loss": 0.2636, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11154456436634064, + "step": 700, + "valid_targets_mean": 4064.0, + "valid_targets_min": 2285 + }, + { + "epoch": 1.150204081632653, + "grad_norm": 0.5489557045500898, + "learning_rate": 3.950500406411913e-05, + "loss": 0.2559, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12069879472255707, + "step": 705, + "valid_targets_mean": 3922.4, + "valid_targets_min": 160 + }, + { + "epoch": 1.1583673469387756, + "grad_norm": 0.5440676931905495, + "learning_rate": 3.948685204055199e-05, + "loss": 0.2745, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15134495496749878, + "step": 710, + "valid_targets_mean": 4761.1, + "valid_targets_min": 1830 + }, + { + "epoch": 1.166530612244898, + "grad_norm": 0.4377127575255182, + "learning_rate": 3.946837747872431e-05, + "loss": 0.2773, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12781107425689697, + "step": 715, + "valid_targets_mean": 6668.2, + "valid_targets_min": 2050 + }, + { + "epoch": 1.1746938775510205, + "grad_norm": 0.46747115033774317, + "learning_rate": 3.9449580684419346e-05, + "loss": 0.2403, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1349138617515564, + "step": 720, + "valid_targets_mean": 7172.1, + "valid_targets_min": 2796 + }, + { + "epoch": 1.1828571428571428, + "grad_norm": 0.46477551009936313, + "learning_rate": 3.943046196875381e-05, + "loss": 0.2483, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09906481206417084, + "step": 725, + "valid_targets_mean": 3456.6, + "valid_targets_min": 971 + }, + { + "epoch": 1.1910204081632654, + "grad_norm": 0.5174344184504699, + "learning_rate": 3.941102164817272e-05, + "loss": 0.2741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1233099177479744, + "step": 730, + "valid_targets_mean": 4068.8, + "valid_targets_min": 2132 + }, + { + "epoch": 1.1991836734693877, + "grad_norm": 0.4559852424556182, + "learning_rate": 3.939126004444417e-05, + "loss": 0.2581, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12599629163742065, + "step": 735, + "valid_targets_mean": 5618.4, + "valid_targets_min": 1192 + }, + { + "epoch": 1.2073469387755102, + "grad_norm": 0.6236070648619523, + "learning_rate": 3.9371177484654e-05, + "loss": 0.2621, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1476593315601349, + "step": 740, + "valid_targets_mean": 4005.5, + "valid_targets_min": 1569 + }, + { + "epoch": 1.2155102040816326, + "grad_norm": 0.4113790501215956, + "learning_rate": 3.9350774301200374e-05, + "loss": 0.256, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10646657645702362, + "step": 745, + "valid_targets_mean": 5235.4, + "valid_targets_min": 2214 + }, + { + "epoch": 1.2236734693877551, + "grad_norm": 0.4400441587321837, + "learning_rate": 3.933005083178828e-05, + "loss": 0.2486, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09953781217336655, + "step": 750, + "valid_targets_mean": 4607.5, + "valid_targets_min": 1155 + }, + { + "epoch": 1.2318367346938777, + "grad_norm": 0.49581897488480603, + "learning_rate": 3.930900741942397e-05, + "loss": 0.2696, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12676215171813965, + "step": 755, + "valid_targets_mean": 4551.0, + "valid_targets_min": 845 + }, + { + "epoch": 1.24, + "grad_norm": 0.44448319184762836, + "learning_rate": 3.928764441240922e-05, + "loss": 0.27, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13121074438095093, + "step": 760, + "valid_targets_mean": 6703.2, + "valid_targets_min": 3792 + }, + { + "epoch": 1.2481632653061225, + "grad_norm": 0.6007678431285308, + "learning_rate": 3.9265962164335656e-05, + "loss": 0.2669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12405592948198318, + "step": 765, + "valid_targets_mean": 3433.2, + "valid_targets_min": 574 + }, + { + "epoch": 1.2563265306122449, + "grad_norm": 0.5103975132542746, + "learning_rate": 3.92439610340788e-05, + "loss": 0.2558, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12031939625740051, + "step": 770, + "valid_targets_mean": 4843.1, + "valid_targets_min": 2665 + }, + { + "epoch": 1.2644897959183674, + "grad_norm": 0.5233070874477446, + "learning_rate": 3.92216413857922e-05, + "loss": 0.2397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1319732964038849, + "step": 775, + "valid_targets_mean": 4303.9, + "valid_targets_min": 1636 + }, + { + "epoch": 1.2726530612244897, + "grad_norm": 0.45711695489790305, + "learning_rate": 3.9199003588901375e-05, + "loss": 0.2744, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12586191296577454, + "step": 780, + "valid_targets_mean": 5050.6, + "valid_targets_min": 1917 + }, + { + "epoch": 1.2808163265306123, + "grad_norm": 0.4912001376959002, + "learning_rate": 3.917604801809774e-05, + "loss": 0.2691, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13443773984909058, + "step": 785, + "valid_targets_mean": 4459.2, + "valid_targets_min": 2587 + }, + { + "epoch": 1.2889795918367346, + "grad_norm": 0.5195769905666631, + "learning_rate": 3.9152775053332305e-05, + "loss": 0.2611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1408383846282959, + "step": 790, + "valid_targets_mean": 4617.0, + "valid_targets_min": 1967 + }, + { + "epoch": 1.2971428571428572, + "grad_norm": 0.5616503318329411, + "learning_rate": 3.912918507980952e-05, + "loss": 0.2589, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14663545787334442, + "step": 795, + "valid_targets_mean": 4577.2, + "valid_targets_min": 917 + }, + { + "epoch": 1.3053061224489797, + "grad_norm": 0.5302531747449998, + "learning_rate": 3.91052784879808e-05, + "loss": 0.2609, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12137632071971893, + "step": 800, + "valid_targets_mean": 3798.2, + "valid_targets_min": 1148 + }, + { + "epoch": 1.313469387755102, + "grad_norm": 0.5111215932189485, + "learning_rate": 3.9081055673538093e-05, + "loss": 0.2711, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12059567123651505, + "step": 805, + "valid_targets_mean": 4635.2, + "valid_targets_min": 385 + }, + { + "epoch": 1.3216326530612246, + "grad_norm": 0.48693696302052697, + "learning_rate": 3.9056517037407334e-05, + "loss": 0.2571, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1377648413181305, + "step": 810, + "valid_targets_mean": 5086.8, + "valid_targets_min": 3186 + }, + { + "epoch": 1.329795918367347, + "grad_norm": 0.5020090791110657, + "learning_rate": 3.903166298574181e-05, + "loss": 0.2583, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12303146719932556, + "step": 815, + "valid_targets_mean": 4312.8, + "valid_targets_min": 1772 + }, + { + "epoch": 1.3379591836734694, + "grad_norm": 0.5428165968206224, + "learning_rate": 3.900649392991545e-05, + "loss": 0.2646, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13852190971374512, + "step": 820, + "valid_targets_mean": 4192.0, + "valid_targets_min": 736 + }, + { + "epoch": 1.3461224489795918, + "grad_norm": 0.507038409712324, + "learning_rate": 3.898101028651597e-05, + "loss": 0.2741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11246784031391144, + "step": 825, + "valid_targets_mean": 4226.5, + "valid_targets_min": 1760 + }, + { + "epoch": 1.3542857142857143, + "grad_norm": 0.4846864352304685, + "learning_rate": 3.895521247733803e-05, + "loss": 0.2617, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15112227201461792, + "step": 830, + "valid_targets_mean": 4990.0, + "valid_targets_min": 2134 + }, + { + "epoch": 1.3624489795918366, + "grad_norm": 0.5498149576955207, + "learning_rate": 3.892910092937622e-05, + "loss": 0.249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1215394139289856, + "step": 835, + "valid_targets_mean": 4482.6, + "valid_targets_min": 3359 + }, + { + "epoch": 1.3706122448979592, + "grad_norm": 0.506582687397877, + "learning_rate": 3.890267607481801e-05, + "loss": 0.2603, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14072394371032715, + "step": 840, + "valid_targets_mean": 4253.4, + "valid_targets_min": 2900 + }, + { + "epoch": 1.3787755102040817, + "grad_norm": 0.5310004085942591, + "learning_rate": 3.887593835103659e-05, + "loss": 0.2741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11263677477836609, + "step": 845, + "valid_targets_mean": 3621.0, + "valid_targets_min": 2181 + }, + { + "epoch": 1.386938775510204, + "grad_norm": 0.4339415251534723, + "learning_rate": 3.8848888200583637e-05, + "loss": 0.2703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10042256116867065, + "step": 850, + "valid_targets_mean": 4672.8, + "valid_targets_min": 2327 + }, + { + "epoch": 1.3951020408163266, + "grad_norm": 0.5441893667906397, + "learning_rate": 3.882152607118197e-05, + "loss": 0.2695, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12515777349472046, + "step": 855, + "valid_targets_mean": 3982.5, + "valid_targets_min": 2314 + }, + { + "epoch": 1.403265306122449, + "grad_norm": 0.503757452633956, + "learning_rate": 3.879385241571817e-05, + "loss": 0.2583, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12633231282234192, + "step": 860, + "valid_targets_mean": 5279.4, + "valid_targets_min": 1544 + }, + { + "epoch": 1.4114285714285715, + "grad_norm": 0.4894760287198677, + "learning_rate": 3.876586769223505e-05, + "loss": 0.249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12801289558410645, + "step": 865, + "valid_targets_mean": 4574.9, + "valid_targets_min": 2731 + }, + { + "epoch": 1.4195918367346938, + "grad_norm": 0.5146349691454388, + "learning_rate": 3.87375723639241e-05, + "loss": 0.2586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10573422908782959, + "step": 870, + "valid_targets_mean": 5117.6, + "valid_targets_min": 726 + }, + { + "epoch": 1.4277551020408163, + "grad_norm": 0.41242003835171465, + "learning_rate": 3.8708966899117815e-05, + "loss": 0.2525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15078118443489075, + "step": 875, + "valid_targets_mean": 8537.8, + "valid_targets_min": 4366 + }, + { + "epoch": 1.4359183673469387, + "grad_norm": 0.45629679864975425, + "learning_rate": 3.868005177128193e-05, + "loss": 0.2422, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10194846987724304, + "step": 880, + "valid_targets_mean": 3460.2, + "valid_targets_min": 469 + }, + { + "epoch": 1.4440816326530612, + "grad_norm": 0.5171964688688151, + "learning_rate": 3.8650827459007615e-05, + "loss": 0.274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14588257670402527, + "step": 885, + "valid_targets_mean": 5561.6, + "valid_targets_min": 3511 + }, + { + "epoch": 1.4522448979591838, + "grad_norm": 0.4084420445366524, + "learning_rate": 3.8621294446003504e-05, + "loss": 0.2664, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15651914477348328, + "step": 890, + "valid_targets_mean": 7917.4, + "valid_targets_min": 3301 + }, + { + "epoch": 1.460408163265306, + "grad_norm": 0.5083640639561413, + "learning_rate": 3.859145322108774e-05, + "loss": 0.2375, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1252627670764923, + "step": 895, + "valid_targets_mean": 3932.4, + "valid_targets_min": 1273 + }, + { + "epoch": 1.4685714285714286, + "grad_norm": 0.5443678273198022, + "learning_rate": 3.856130427817984e-05, + "loss": 0.2744, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13401755690574646, + "step": 900, + "valid_targets_mean": 4187.8, + "valid_targets_min": 1703 + }, + { + "epoch": 1.476734693877551, + "grad_norm": 0.46848277024034957, + "learning_rate": 3.853084811629257e-05, + "loss": 0.2561, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13730978965759277, + "step": 905, + "valid_targets_mean": 5213.6, + "valid_targets_min": 1923 + }, + { + "epoch": 1.4848979591836735, + "grad_norm": 0.5269763713171643, + "learning_rate": 3.8500085239523624e-05, + "loss": 0.2484, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14742831885814667, + "step": 910, + "valid_targets_mean": 5609.5, + "valid_targets_min": 1810 + }, + { + "epoch": 1.4930612244897958, + "grad_norm": 0.4918260517625046, + "learning_rate": 3.846901615704734e-05, + "loss": 0.2398, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12141484022140503, + "step": 915, + "valid_targets_mean": 4149.9, + "valid_targets_min": 755 + }, + { + "epoch": 1.5012244897959184, + "grad_norm": 0.48992607672168265, + "learning_rate": 3.843764138310622e-05, + "loss": 0.2615, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16299812495708466, + "step": 920, + "valid_targets_mean": 5494.4, + "valid_targets_min": 1927 + }, + { + "epoch": 1.5093877551020407, + "grad_norm": 0.47445400914699815, + "learning_rate": 3.840596143700246e-05, + "loss": 0.2483, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13037662208080292, + "step": 925, + "valid_targets_mean": 5328.8, + "valid_targets_min": 2703 + }, + { + "epoch": 1.5175510204081633, + "grad_norm": 0.5085959632245128, + "learning_rate": 3.8373976843089355e-05, + "loss": 0.2527, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14229968190193176, + "step": 930, + "valid_targets_mean": 4230.6, + "valid_targets_min": 503 + }, + { + "epoch": 1.5257142857142858, + "grad_norm": 0.4982028747496671, + "learning_rate": 3.8341688130762574e-05, + "loss": 0.2682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14528988301753998, + "step": 935, + "valid_targets_mean": 5524.4, + "valid_targets_min": 1399 + }, + { + "epoch": 1.5338775510204081, + "grad_norm": 0.5614943316039491, + "learning_rate": 3.830909583445143e-05, + "loss": 0.257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14664828777313232, + "step": 940, + "valid_targets_mean": 3903.5, + "valid_targets_min": 720 + }, + { + "epoch": 1.5420408163265305, + "grad_norm": 0.4217324955618739, + "learning_rate": 3.827620049361004e-05, + "loss": 0.2681, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13761885464191437, + "step": 945, + "valid_targets_mean": 4984.2, + "valid_targets_min": 2669 + }, + { + "epoch": 1.550204081632653, + "grad_norm": 0.5031297388057706, + "learning_rate": 3.82430026527084e-05, + "loss": 0.2613, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11942078918218613, + "step": 950, + "valid_targets_mean": 4265.5, + "valid_targets_min": 2414 + }, + { + "epoch": 1.5583673469387755, + "grad_norm": 0.594624788084472, + "learning_rate": 3.8209502861223344e-05, + "loss": 0.2737, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14711017906665802, + "step": 955, + "valid_targets_mean": 4126.1, + "valid_targets_min": 1653 + }, + { + "epoch": 1.566530612244898, + "grad_norm": 0.5147712518084654, + "learning_rate": 3.817570167362947e-05, + "loss": 0.2552, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11636977642774582, + "step": 960, + "valid_targets_mean": 4158.8, + "valid_targets_min": 533 + }, + { + "epoch": 1.5746938775510204, + "grad_norm": 0.49811050757421693, + "learning_rate": 3.814159964938996e-05, + "loss": 0.2476, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09341903030872345, + "step": 965, + "valid_targets_mean": 4023.4, + "valid_targets_min": 1942 + }, + { + "epoch": 1.5828571428571427, + "grad_norm": 0.4865312743365694, + "learning_rate": 3.810719735294731e-05, + "loss": 0.2548, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11299189925193787, + "step": 970, + "valid_targets_mean": 4597.5, + "valid_targets_min": 1141 + }, + { + "epoch": 1.5910204081632653, + "grad_norm": 0.5076603149396371, + "learning_rate": 3.8072495353714025e-05, + "loss": 0.2542, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16264669597148895, + "step": 975, + "valid_targets_mean": 5229.2, + "valid_targets_min": 1843 + }, + { + "epoch": 1.5991836734693878, + "grad_norm": 0.5194180616689766, + "learning_rate": 3.8037494226063126e-05, + "loss": 0.2603, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15082842111587524, + "step": 980, + "valid_targets_mean": 4919.4, + "valid_targets_min": 2627 + }, + { + "epoch": 1.6073469387755102, + "grad_norm": 0.596887240652063, + "learning_rate": 3.800219454931871e-05, + "loss": 0.271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1672152429819107, + "step": 985, + "valid_targets_mean": 4447.2, + "valid_targets_min": 2166 + }, + { + "epoch": 1.6155102040816327, + "grad_norm": 0.5350107457946114, + "learning_rate": 3.7966596907746344e-05, + "loss": 0.2595, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1205388754606247, + "step": 990, + "valid_targets_mean": 3446.4, + "valid_targets_min": 1205 + }, + { + "epoch": 1.623673469387755, + "grad_norm": 0.5081950112157592, + "learning_rate": 3.793070189054335e-05, + "loss": 0.2548, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12550537288188934, + "step": 995, + "valid_targets_mean": 4020.2, + "valid_targets_min": 2696 + }, + { + "epoch": 1.6318367346938776, + "grad_norm": 0.5318057562656638, + "learning_rate": 3.789451009182911e-05, + "loss": 0.2688, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13900066912174225, + "step": 1000, + "valid_targets_mean": 4368.8, + "valid_targets_min": 2759 + }, + { + "epoch": 1.6400000000000001, + "grad_norm": 0.510742602562625, + "learning_rate": 3.78580221106352e-05, + "loss": 0.241, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1156846210360527, + "step": 1005, + "valid_targets_mean": 3944.9, + "valid_targets_min": 1110 + }, + { + "epoch": 1.6481632653061224, + "grad_norm": 0.4750127711790795, + "learning_rate": 3.78212385508955e-05, + "loss": 0.2569, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16291871666908264, + "step": 1010, + "valid_targets_mean": 6779.4, + "valid_targets_min": 3330 + }, + { + "epoch": 1.6563265306122448, + "grad_norm": 0.478006168965345, + "learning_rate": 3.778416002143617e-05, + "loss": 0.2678, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1426718533039093, + "step": 1015, + "valid_targets_mean": 6106.1, + "valid_targets_min": 2608 + }, + { + "epoch": 1.6644897959183673, + "grad_norm": 0.47169805607015175, + "learning_rate": 3.774678713596559e-05, + "loss": 0.2459, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11533847451210022, + "step": 1020, + "valid_targets_mean": 5635.5, + "valid_targets_min": 2118 + }, + { + "epoch": 1.6726530612244899, + "grad_norm": 0.5401353076658538, + "learning_rate": 3.7709120513064196e-05, + "loss": 0.2655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.148171067237854, + "step": 1025, + "valid_targets_mean": 5144.6, + "valid_targets_min": 1841 + }, + { + "epoch": 1.6808163265306122, + "grad_norm": 0.5251628231233804, + "learning_rate": 3.767116077617426e-05, + "loss": 0.2589, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12557116150856018, + "step": 1030, + "valid_targets_mean": 4335.6, + "valid_targets_min": 2394 + }, + { + "epoch": 1.6889795918367347, + "grad_norm": 0.4650011640177102, + "learning_rate": 3.763290855358952e-05, + "loss": 0.2527, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15282702445983887, + "step": 1035, + "valid_targets_mean": 6148.4, + "valid_targets_min": 552 + }, + { + "epoch": 1.697142857142857, + "grad_norm": 0.47198000057835904, + "learning_rate": 3.759436447844485e-05, + "loss": 0.2496, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09859016537666321, + "step": 1040, + "valid_targets_mean": 4540.5, + "valid_targets_min": 2642 + }, + { + "epoch": 1.7053061224489796, + "grad_norm": 0.550851013931262, + "learning_rate": 3.7555529188705716e-05, + "loss": 0.2622, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1216202899813652, + "step": 1045, + "valid_targets_mean": 3473.2, + "valid_targets_min": 2111 + }, + { + "epoch": 1.7134693877551022, + "grad_norm": 0.37040303534509933, + "learning_rate": 3.751640332715768e-05, + "loss": 0.2394, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10425127297639847, + "step": 1050, + "valid_targets_mean": 6519.0, + "valid_targets_min": 3227 + }, + { + "epoch": 1.7216326530612245, + "grad_norm": 0.53785718338026, + "learning_rate": 3.74769875413957e-05, + "loss": 0.2453, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13041691482067108, + "step": 1055, + "valid_targets_mean": 4362.1, + "valid_targets_min": 2065 + }, + { + "epoch": 1.7297959183673468, + "grad_norm": 0.440296077925039, + "learning_rate": 3.7437282483813455e-05, + "loss": 0.2516, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1237720474600792, + "step": 1060, + "valid_targets_mean": 5525.8, + "valid_targets_min": 2969 + }, + { + "epoch": 1.7379591836734694, + "grad_norm": 0.4894636219664195, + "learning_rate": 3.739728881159252e-05, + "loss": 0.2542, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1411234736442566, + "step": 1065, + "valid_targets_mean": 4698.5, + "valid_targets_min": 2461 + }, + { + "epoch": 1.746122448979592, + "grad_norm": 0.44036148587883767, + "learning_rate": 3.735700718669151e-05, + "loss": 0.2481, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13445939123630524, + "step": 1070, + "valid_targets_mean": 6161.1, + "valid_targets_min": 2236 + }, + { + "epoch": 1.7542857142857144, + "grad_norm": 0.5008003842103482, + "learning_rate": 3.73164382758351e-05, + "loss": 0.2649, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14120852947235107, + "step": 1075, + "valid_targets_mean": 5492.5, + "valid_targets_min": 3082 + }, + { + "epoch": 1.7624489795918368, + "grad_norm": 0.6470980556874778, + "learning_rate": 3.727558275050301e-05, + "loss": 0.2657, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15491387248039246, + "step": 1080, + "valid_targets_mean": 6985.4, + "valid_targets_min": 1928 + }, + { + "epoch": 1.770612244897959, + "grad_norm": 0.4683575217814649, + "learning_rate": 3.723444128691888e-05, + "loss": 0.2575, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13208508491516113, + "step": 1085, + "valid_targets_mean": 4950.8, + "valid_targets_min": 2287 + }, + { + "epoch": 1.7787755102040816, + "grad_norm": 0.36384695238620846, + "learning_rate": 3.71930145660391e-05, + "loss": 0.2479, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06327690929174423, + "step": 1090, + "valid_targets_mean": 3406.9, + "valid_targets_min": 2404 + }, + { + "epoch": 1.7869387755102042, + "grad_norm": 0.46447438386840667, + "learning_rate": 3.715130327354148e-05, + "loss": 0.2459, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12327778339385986, + "step": 1095, + "valid_targets_mean": 4916.1, + "valid_targets_min": 1133 + }, + { + "epoch": 1.7951020408163265, + "grad_norm": 0.4993865990758053, + "learning_rate": 3.7109308099813986e-05, + "loss": 0.2477, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13568077981472015, + "step": 1100, + "valid_targets_mean": 5312.5, + "valid_targets_min": 2093 + }, + { + "epoch": 1.8032653061224488, + "grad_norm": 0.45257355005085736, + "learning_rate": 3.706702973994323e-05, + "loss": 0.2567, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1562289148569107, + "step": 1105, + "valid_targets_mean": 7640.9, + "valid_targets_min": 5005 + }, + { + "epoch": 1.8114285714285714, + "grad_norm": 0.4722207457039624, + "learning_rate": 3.702446889370303e-05, + "loss": 0.2476, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11828452348709106, + "step": 1110, + "valid_targets_mean": 4442.2, + "valid_targets_min": 1883 + }, + { + "epoch": 1.819591836734694, + "grad_norm": 0.4484385738226761, + "learning_rate": 3.698162626554279e-05, + "loss": 0.2478, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13094788789749146, + "step": 1115, + "valid_targets_mean": 6618.0, + "valid_targets_min": 2405 + }, + { + "epoch": 1.8277551020408165, + "grad_norm": 0.4955177645979594, + "learning_rate": 3.693850256457585e-05, + "loss": 0.2385, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16945160925388336, + "step": 1120, + "valid_targets_mean": 6220.6, + "valid_targets_min": 3866 + }, + { + "epoch": 1.8359183673469388, + "grad_norm": 0.5870278446195637, + "learning_rate": 3.689509850456774e-05, + "loss": 0.2372, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11503128707408905, + "step": 1125, + "valid_targets_mean": 4211.6, + "valid_targets_min": 2709 + }, + { + "epoch": 1.8440816326530611, + "grad_norm": 0.5875782256134718, + "learning_rate": 3.685141480392441e-05, + "loss": 0.2538, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10766669362783432, + "step": 1130, + "valid_targets_mean": 3444.6, + "valid_targets_min": 1349 + }, + { + "epoch": 1.8522448979591837, + "grad_norm": 0.43013135938747526, + "learning_rate": 3.680745218568026e-05, + "loss": 0.2475, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1484908163547516, + "step": 1135, + "valid_targets_mean": 7711.4, + "valid_targets_min": 1330 + }, + { + "epoch": 1.8604081632653062, + "grad_norm": 0.4031243286403073, + "learning_rate": 3.676321137748626e-05, + "loss": 0.25, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11534357070922852, + "step": 1140, + "valid_targets_mean": 6376.8, + "valid_targets_min": 1852 + }, + { + "epoch": 1.8685714285714285, + "grad_norm": 0.4706254490578847, + "learning_rate": 3.671869311159783e-05, + "loss": 0.2548, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1212092861533165, + "step": 1145, + "valid_targets_mean": 5040.8, + "valid_targets_min": 1542 + }, + { + "epoch": 1.8767346938775509, + "grad_norm": 0.3795817061920972, + "learning_rate": 3.6673898124862776e-05, + "loss": 0.2519, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13475094735622406, + "step": 1150, + "valid_targets_mean": 8024.8, + "valid_targets_min": 2130 + }, + { + "epoch": 1.8848979591836734, + "grad_norm": 0.5252621439889926, + "learning_rate": 3.662882715870907e-05, + "loss": 0.2687, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12451595067977905, + "step": 1155, + "valid_targets_mean": 4560.5, + "valid_targets_min": 3147 + }, + { + "epoch": 1.893061224489796, + "grad_norm": 0.5148704957978237, + "learning_rate": 3.658348095913257e-05, + "loss": 0.2469, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1368659883737564, + "step": 1160, + "valid_targets_mean": 4340.2, + "valid_targets_min": 1934 + }, + { + "epoch": 1.9012244897959185, + "grad_norm": 0.6121916844321755, + "learning_rate": 3.6537860276684686e-05, + "loss": 0.2402, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12297011911869049, + "step": 1165, + "valid_targets_mean": 5257.2, + "valid_targets_min": 2234 + }, + { + "epoch": 1.9093877551020408, + "grad_norm": 0.5724642448538444, + "learning_rate": 3.6491965866459966e-05, + "loss": 0.2605, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12487930059432983, + "step": 1170, + "valid_targets_mean": 3325.8, + "valid_targets_min": 1434 + }, + { + "epoch": 1.9175510204081632, + "grad_norm": 0.46933023905049503, + "learning_rate": 3.6445798488083576e-05, + "loss": 0.2224, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11442330479621887, + "step": 1175, + "valid_targets_mean": 4419.6, + "valid_targets_min": 1587 + }, + { + "epoch": 1.9257142857142857, + "grad_norm": 0.5036546733874333, + "learning_rate": 3.639935890569873e-05, + "loss": 0.2504, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12574106454849243, + "step": 1180, + "valid_targets_mean": 4680.1, + "valid_targets_min": 1697 + }, + { + "epoch": 1.9338775510204083, + "grad_norm": 0.43369263982057915, + "learning_rate": 3.635264788795407e-05, + "loss": 0.2548, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11851448565721512, + "step": 1185, + "valid_targets_mean": 5754.1, + "valid_targets_min": 2418 + }, + { + "epoch": 1.9420408163265306, + "grad_norm": 0.5419617720778334, + "learning_rate": 3.6305666207990886e-05, + "loss": 0.2537, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11305820941925049, + "step": 1190, + "valid_targets_mean": 3559.0, + "valid_targets_min": 197 + }, + { + "epoch": 1.950204081632653, + "grad_norm": 0.5314154428545551, + "learning_rate": 3.625841464343042e-05, + "loss": 0.2597, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13911515474319458, + "step": 1195, + "valid_targets_mean": 3929.6, + "valid_targets_min": 1821 + }, + { + "epoch": 1.9583673469387755, + "grad_norm": 0.4896698812099476, + "learning_rate": 3.621089397636086e-05, + "loss": 0.2587, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14098498225212097, + "step": 1200, + "valid_targets_mean": 4840.2, + "valid_targets_min": 2400 + }, + { + "epoch": 1.966530612244898, + "grad_norm": 0.5856569580560823, + "learning_rate": 3.616310499332451e-05, + "loss": 0.2514, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11941181123256683, + "step": 1205, + "valid_targets_mean": 3408.5, + "valid_targets_min": 1840 + }, + { + "epoch": 1.9746938775510205, + "grad_norm": 0.4982878645027437, + "learning_rate": 3.6115048485304736e-05, + "loss": 0.2492, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15822862088680267, + "step": 1210, + "valid_targets_mean": 4985.1, + "valid_targets_min": 1426 + }, + { + "epoch": 1.9828571428571429, + "grad_norm": 0.5618752039712213, + "learning_rate": 3.606672524771281e-05, + "loss": 0.2538, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1203514039516449, + "step": 1215, + "valid_targets_mean": 3678.4, + "valid_targets_min": 1868 + }, + { + "epoch": 1.9910204081632652, + "grad_norm": 0.5291588003096855, + "learning_rate": 3.601813608037488e-05, + "loss": 0.2525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12971103191375732, + "step": 1220, + "valid_targets_mean": 4032.5, + "valid_targets_min": 2465 + }, + { + "epoch": 1.9991836734693877, + "grad_norm": 0.45963462836895563, + "learning_rate": 3.596928178751858e-05, + "loss": 0.2563, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16318891942501068, + "step": 1225, + "valid_targets_mean": 7103.4, + "valid_targets_min": 2453 + }, + { + "epoch": 2.006530612244898, + "grad_norm": 0.43056684768932413, + "learning_rate": 3.592016317775984e-05, + "loss": 0.2295, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1186676174402237, + "step": 1230, + "valid_targets_mean": 8096.0, + "valid_targets_min": 3368 + }, + { + "epoch": 2.0146938775510206, + "grad_norm": 0.48079760996523624, + "learning_rate": 3.587078106408942e-05, + "loss": 0.2196, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10643745213747025, + "step": 1235, + "valid_targets_mean": 4640.6, + "valid_targets_min": 2642 + }, + { + "epoch": 2.0228571428571427, + "grad_norm": 0.5739864792377497, + "learning_rate": 3.582113626385951e-05, + "loss": 0.2343, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12379193305969238, + "step": 1240, + "valid_targets_mean": 3463.2, + "valid_targets_min": 1642 + }, + { + "epoch": 2.0310204081632652, + "grad_norm": 0.4995233604930039, + "learning_rate": 3.577122959877017e-05, + "loss": 0.2393, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13157770037651062, + "step": 1245, + "valid_targets_mean": 4963.4, + "valid_targets_min": 2360 + }, + { + "epoch": 2.039183673469388, + "grad_norm": 0.5292051470490394, + "learning_rate": 3.5721061894855756e-05, + "loss": 0.2212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1389136165380478, + "step": 1250, + "valid_targets_mean": 5193.1, + "valid_targets_min": 2899 + }, + { + "epoch": 2.0473469387755103, + "grad_norm": 0.4888907093559062, + "learning_rate": 3.5670633982471185e-05, + "loss": 0.236, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15732908248901367, + "step": 1255, + "valid_targets_mean": 6197.2, + "valid_targets_min": 922 + }, + { + "epoch": 2.0555102040816324, + "grad_norm": 0.46552539231145457, + "learning_rate": 3.561994669627828e-05, + "loss": 0.2375, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10380949079990387, + "step": 1260, + "valid_targets_mean": 4741.5, + "valid_targets_min": 1876 + }, + { + "epoch": 2.063673469387755, + "grad_norm": 0.502350861185688, + "learning_rate": 3.55690008752319e-05, + "loss": 0.2315, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1098252683877945, + "step": 1265, + "valid_targets_mean": 4458.0, + "valid_targets_min": 426 + }, + { + "epoch": 2.0718367346938775, + "grad_norm": 0.5119241705569415, + "learning_rate": 3.551779736256604e-05, + "loss": 0.2266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10564523935317993, + "step": 1270, + "valid_targets_mean": 4026.6, + "valid_targets_min": 1932 + }, + { + "epoch": 2.08, + "grad_norm": 0.5445752890676937, + "learning_rate": 3.546633700577994e-05, + "loss": 0.2335, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13011673092842102, + "step": 1275, + "valid_targets_mean": 4863.9, + "valid_targets_min": 2948 + }, + { + "epoch": 2.0881632653061226, + "grad_norm": 0.503922154527504, + "learning_rate": 3.541462065662399e-05, + "loss": 0.2299, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11520066112279892, + "step": 1280, + "valid_targets_mean": 5264.2, + "valid_targets_min": 699 + }, + { + "epoch": 2.0963265306122447, + "grad_norm": 0.5078588427716441, + "learning_rate": 3.536264917108567e-05, + "loss": 0.225, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13688494265079498, + "step": 1285, + "valid_targets_mean": 5433.9, + "valid_targets_min": 428 + }, + { + "epoch": 2.1044897959183673, + "grad_norm": 0.5346239759699418, + "learning_rate": 3.531042340937537e-05, + "loss": 0.2234, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09234683960676193, + "step": 1290, + "valid_targets_mean": 3514.4, + "valid_targets_min": 1542 + }, + { + "epoch": 2.11265306122449, + "grad_norm": 0.48695969669957334, + "learning_rate": 3.525794423591214e-05, + "loss": 0.2397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10513603687286377, + "step": 1295, + "valid_targets_mean": 4443.8, + "valid_targets_min": 2414 + }, + { + "epoch": 2.1208163265306124, + "grad_norm": 0.3494135233315135, + "learning_rate": 3.520521251930941e-05, + "loss": 0.2111, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08017104119062424, + "step": 1300, + "valid_targets_mean": 4831.8, + "valid_targets_min": 1849 + }, + { + "epoch": 2.1289795918367345, + "grad_norm": 0.51369999487573, + "learning_rate": 3.51522291323606e-05, + "loss": 0.2318, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12431921064853668, + "step": 1305, + "valid_targets_mean": 4205.5, + "valid_targets_min": 1934 + }, + { + "epoch": 2.137142857142857, + "grad_norm": 0.48865254132392544, + "learning_rate": 3.509899495202467e-05, + "loss": 0.2425, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11693074554204941, + "step": 1310, + "valid_targets_mean": 4546.9, + "valid_targets_min": 1284 + }, + { + "epoch": 2.1453061224489796, + "grad_norm": 0.44405836098264856, + "learning_rate": 3.50455108594116e-05, + "loss": 0.2292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09972664713859558, + "step": 1315, + "valid_targets_mean": 5387.9, + "valid_targets_min": 2268 + }, + { + "epoch": 2.153469387755102, + "grad_norm": 0.5265273408182132, + "learning_rate": 3.499177773976781e-05, + "loss": 0.2274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13111922144889832, + "step": 1320, + "valid_targets_mean": 4415.8, + "valid_targets_min": 2872 + }, + { + "epoch": 2.1616326530612247, + "grad_norm": 0.5205910134073096, + "learning_rate": 3.493779648246154e-05, + "loss": 0.2266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1394486427307129, + "step": 1325, + "valid_targets_mean": 5611.8, + "valid_targets_min": 1843 + }, + { + "epoch": 2.1697959183673468, + "grad_norm": 0.5319230026891569, + "learning_rate": 3.488356798096808e-05, + "loss": 0.2326, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09613202512264252, + "step": 1330, + "valid_targets_mean": 3297.8, + "valid_targets_min": 2093 + }, + { + "epoch": 2.1779591836734693, + "grad_norm": 0.4979347957927288, + "learning_rate": 3.482909313285499e-05, + "loss": 0.2218, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10724276304244995, + "step": 1335, + "valid_targets_mean": 3485.8, + "valid_targets_min": 1248 + }, + { + "epoch": 2.186122448979592, + "grad_norm": 0.4937495740699081, + "learning_rate": 3.47743728397673e-05, + "loss": 0.2284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12111664563417435, + "step": 1340, + "valid_targets_mean": 4159.6, + "valid_targets_min": 1846 + }, + { + "epoch": 2.1942857142857144, + "grad_norm": 0.4324928163179517, + "learning_rate": 3.4719408007412494e-05, + "loss": 0.2197, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08307406306266785, + "step": 1345, + "valid_targets_mean": 3331.8, + "valid_targets_min": 1356 + }, + { + "epoch": 2.202448979591837, + "grad_norm": 0.5400878779239833, + "learning_rate": 3.4664199545545586e-05, + "loss": 0.2252, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11346576362848282, + "step": 1350, + "valid_targets_mean": 4346.8, + "valid_targets_min": 1594 + }, + { + "epoch": 2.210612244897959, + "grad_norm": 0.5751437287217012, + "learning_rate": 3.4608748367954064e-05, + "loss": 0.2474, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12245942652225494, + "step": 1355, + "valid_targets_mean": 3260.6, + "valid_targets_min": 2321 + }, + { + "epoch": 2.2187755102040816, + "grad_norm": 0.47449273447519114, + "learning_rate": 3.455305539244272e-05, + "loss": 0.2397, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11719304323196411, + "step": 1360, + "valid_targets_mean": 5256.8, + "valid_targets_min": 1937 + }, + { + "epoch": 2.226938775510204, + "grad_norm": 0.5472839028591866, + "learning_rate": 3.44971215408185e-05, + "loss": 0.2319, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14066477119922638, + "step": 1365, + "valid_targets_mean": 4968.9, + "valid_targets_min": 2880 + }, + { + "epoch": 2.2351020408163267, + "grad_norm": 0.46090788127961696, + "learning_rate": 3.4440947738875215e-05, + "loss": 0.2322, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13428354263305664, + "step": 1370, + "valid_targets_mean": 5852.8, + "valid_targets_min": 2568 + }, + { + "epoch": 2.243265306122449, + "grad_norm": 0.5012503306811049, + "learning_rate": 3.438453491637825e-05, + "loss": 0.212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12726007401943207, + "step": 1375, + "valid_targets_mean": 4681.1, + "valid_targets_min": 2095 + }, + { + "epoch": 2.2514285714285713, + "grad_norm": 0.5273115750782535, + "learning_rate": 3.432788400704915e-05, + "loss": 0.2289, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13001614809036255, + "step": 1380, + "valid_targets_mean": 5218.6, + "valid_targets_min": 2828 + }, + { + "epoch": 2.259591836734694, + "grad_norm": 0.4808418482598062, + "learning_rate": 3.4270995948550175e-05, + "loss": 0.2238, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11392636597156525, + "step": 1385, + "valid_targets_mean": 4804.0, + "valid_targets_min": 1839 + }, + { + "epoch": 2.2677551020408164, + "grad_norm": 0.49225698613286484, + "learning_rate": 3.421387168246879e-05, + "loss": 0.235, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13683432340621948, + "step": 1390, + "valid_targets_mean": 5497.4, + "valid_targets_min": 2835 + }, + { + "epoch": 2.2759183673469385, + "grad_norm": 0.4707613798201057, + "learning_rate": 3.415651215430205e-05, + "loss": 0.2203, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11409764736890793, + "step": 1395, + "valid_targets_mean": 4437.4, + "valid_targets_min": 2262 + }, + { + "epoch": 2.284081632653061, + "grad_norm": 0.5457878781003122, + "learning_rate": 3.409891831344099e-05, + "loss": 0.2295, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15483315289020538, + "step": 1400, + "valid_targets_mean": 4922.2, + "valid_targets_min": 2454 + }, + { + "epoch": 2.2922448979591836, + "grad_norm": 0.526618791139005, + "learning_rate": 3.404109111315487e-05, + "loss": 0.2443, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1386374831199646, + "step": 1405, + "valid_targets_mean": 5216.4, + "valid_targets_min": 2441 + }, + { + "epoch": 2.300408163265306, + "grad_norm": 0.5133511190962673, + "learning_rate": 3.398303151057543e-05, + "loss": 0.2353, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1506226360797882, + "step": 1410, + "valid_targets_mean": 5939.0, + "valid_targets_min": 2086 + }, + { + "epoch": 2.3085714285714287, + "grad_norm": 0.5246066974685838, + "learning_rate": 3.392474046668105e-05, + "loss": 0.2354, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1148165687918663, + "step": 1415, + "valid_targets_mean": 3559.0, + "valid_targets_min": 319 + }, + { + "epoch": 2.3167346938775513, + "grad_norm": 0.45674426779962213, + "learning_rate": 3.38662189462808e-05, + "loss": 0.2396, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10377395153045654, + "step": 1420, + "valid_targets_mean": 4903.6, + "valid_targets_min": 1524 + }, + { + "epoch": 2.3248979591836734, + "grad_norm": 0.5549953930504047, + "learning_rate": 3.3807467917998534e-05, + "loss": 0.2294, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10187067836523056, + "step": 1425, + "valid_targets_mean": 5022.4, + "valid_targets_min": 179 + }, + { + "epoch": 2.333061224489796, + "grad_norm": 0.5667679888700494, + "learning_rate": 3.374848835425679e-05, + "loss": 0.2148, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10153408348560333, + "step": 1430, + "valid_targets_mean": 4056.2, + "valid_targets_min": 514 + }, + { + "epoch": 2.3412244897959185, + "grad_norm": 0.38849681119607615, + "learning_rate": 3.3689281231260755e-05, + "loss": 0.2216, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09394226968288422, + "step": 1435, + "valid_targets_mean": 7920.5, + "valid_targets_min": 2036 + }, + { + "epoch": 2.349387755102041, + "grad_norm": 0.4849228280895079, + "learning_rate": 3.362984752898208e-05, + "loss": 0.228, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10231403261423111, + "step": 1440, + "valid_targets_mean": 5006.1, + "valid_targets_min": 1883 + }, + { + "epoch": 2.357551020408163, + "grad_norm": 0.507994637708824, + "learning_rate": 3.3570188231142647e-05, + "loss": 0.2438, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11404062807559967, + "step": 1445, + "valid_targets_mean": 4506.0, + "valid_targets_min": 2781 + }, + { + "epoch": 2.3657142857142857, + "grad_norm": 0.47807993839798824, + "learning_rate": 3.351030432519832e-05, + "loss": 0.222, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11850126087665558, + "step": 1450, + "valid_targets_mean": 5020.4, + "valid_targets_min": 3208 + }, + { + "epoch": 2.373877551020408, + "grad_norm": 0.48000969417395906, + "learning_rate": 3.345019680232256e-05, + "loss": 0.2175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1052422970533371, + "step": 1455, + "valid_targets_mean": 4933.5, + "valid_targets_min": 2331 + }, + { + "epoch": 2.3820408163265308, + "grad_norm": 0.7487096129240244, + "learning_rate": 3.3389866657390074e-05, + "loss": 0.239, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11607220768928528, + "step": 1460, + "valid_targets_mean": 4892.9, + "valid_targets_min": 1966 + }, + { + "epoch": 2.3902040816326533, + "grad_norm": 0.48365221971953487, + "learning_rate": 3.332931488896029e-05, + "loss": 0.2384, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11191040277481079, + "step": 1465, + "valid_targets_mean": 4724.5, + "valid_targets_min": 274 + }, + { + "epoch": 2.3983673469387754, + "grad_norm": 0.4928050884387631, + "learning_rate": 3.3268542499260874e-05, + "loss": 0.2354, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11471347510814667, + "step": 1470, + "valid_targets_mean": 4724.8, + "valid_targets_min": 729 + }, + { + "epoch": 2.406530612244898, + "grad_norm": 0.5035592274557454, + "learning_rate": 3.32075504941711e-05, + "loss": 0.218, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1193801611661911, + "step": 1475, + "valid_targets_mean": 4907.6, + "valid_targets_min": 2960 + }, + { + "epoch": 2.4146938775510205, + "grad_norm": 0.5363290738713465, + "learning_rate": 3.3146339883205243e-05, + "loss": 0.2342, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09244148433208466, + "step": 1480, + "valid_targets_mean": 4252.2, + "valid_targets_min": 1957 + }, + { + "epoch": 2.422857142857143, + "grad_norm": 0.5780978434020073, + "learning_rate": 3.308491167949585e-05, + "loss": 0.2275, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09925157576799393, + "step": 1485, + "valid_targets_mean": 4348.6, + "valid_targets_min": 1251 + }, + { + "epoch": 2.431020408163265, + "grad_norm": 0.42097104143162, + "learning_rate": 3.302326689977697e-05, + "loss": 0.2258, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10086309164762497, + "step": 1490, + "valid_targets_mean": 4883.4, + "valid_targets_min": 1399 + }, + { + "epoch": 2.4391836734693877, + "grad_norm": 0.5932340797258087, + "learning_rate": 3.2961406564367335e-05, + "loss": 0.2247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11674109101295471, + "step": 1495, + "valid_targets_mean": 3973.2, + "valid_targets_min": 1920 + }, + { + "epoch": 2.4473469387755102, + "grad_norm": 0.4972235138038265, + "learning_rate": 3.2899331697153445e-05, + "loss": 0.2415, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1610492318868637, + "step": 1500, + "valid_targets_mean": 7308.4, + "valid_targets_min": 3836 + }, + { + "epoch": 2.455510204081633, + "grad_norm": 0.5172073645695415, + "learning_rate": 3.283704332557266e-05, + "loss": 0.2351, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10869985818862915, + "step": 1505, + "valid_targets_mean": 4040.6, + "valid_targets_min": 755 + }, + { + "epoch": 2.4636734693877553, + "grad_norm": 0.5477449704840818, + "learning_rate": 3.2774542480596165e-05, + "loss": 0.2223, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10822654515504837, + "step": 1510, + "valid_targets_mean": 3207.1, + "valid_targets_min": 580 + }, + { + "epoch": 2.4718367346938774, + "grad_norm": 0.5657022689186633, + "learning_rate": 3.2711830196711926e-05, + "loss": 0.2221, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11191504448652267, + "step": 1515, + "valid_targets_mean": 4077.8, + "valid_targets_min": 2333 + }, + { + "epoch": 2.48, + "grad_norm": 0.4381042129056105, + "learning_rate": 3.2648907511907544e-05, + "loss": 0.2231, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10943110287189484, + "step": 1520, + "valid_targets_mean": 6497.8, + "valid_targets_min": 2968 + }, + { + "epoch": 2.4881632653061225, + "grad_norm": 0.4131104301772404, + "learning_rate": 3.258577546765311e-05, + "loss": 0.2274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10738101601600647, + "step": 1525, + "valid_targets_mean": 7473.2, + "valid_targets_min": 2792 + }, + { + "epoch": 2.496326530612245, + "grad_norm": 0.4381525165508198, + "learning_rate": 3.2522435108883925e-05, + "loss": 0.2246, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09731510281562805, + "step": 1530, + "valid_targets_mean": 4936.4, + "valid_targets_min": 1337 + }, + { + "epoch": 2.504489795918367, + "grad_norm": 0.4643511154489369, + "learning_rate": 3.245888748398322e-05, + "loss": 0.2305, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12030883133411407, + "step": 1535, + "valid_targets_mean": 4852.2, + "valid_targets_min": 1923 + }, + { + "epoch": 2.5126530612244897, + "grad_norm": 0.5157394733487625, + "learning_rate": 3.239513364476484e-05, + "loss": 0.2247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10848742723464966, + "step": 1540, + "valid_targets_mean": 3704.1, + "valid_targets_min": 2390 + }, + { + "epoch": 2.5208163265306123, + "grad_norm": 0.431398191953788, + "learning_rate": 3.233117464645577e-05, + "loss": 0.2272, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10247580707073212, + "step": 1545, + "valid_targets_mean": 4868.9, + "valid_targets_min": 2738 + }, + { + "epoch": 2.528979591836735, + "grad_norm": 0.49163079122173614, + "learning_rate": 3.2267011547678735e-05, + "loss": 0.2399, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10695581138134003, + "step": 1550, + "valid_targets_mean": 4697.6, + "valid_targets_min": 1898 + }, + { + "epoch": 2.5371428571428574, + "grad_norm": 0.5377249490301479, + "learning_rate": 3.220264541043461e-05, + "loss": 0.2248, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14529483020305634, + "step": 1555, + "valid_targets_mean": 4557.8, + "valid_targets_min": 2132 + }, + { + "epoch": 2.5453061224489795, + "grad_norm": 0.4657234274209734, + "learning_rate": 3.21380773000849e-05, + "loss": 0.2327, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07366371154785156, + "step": 1560, + "valid_targets_mean": 3585.4, + "valid_targets_min": 172 + }, + { + "epoch": 2.553469387755102, + "grad_norm": 0.5015836388954693, + "learning_rate": 3.2073308285334085e-05, + "loss": 0.2059, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08729308098554611, + "step": 1565, + "valid_targets_mean": 3744.1, + "valid_targets_min": 503 + }, + { + "epoch": 2.5616326530612246, + "grad_norm": 0.4722130178319186, + "learning_rate": 3.200833943821192e-05, + "loss": 0.2271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11033571511507034, + "step": 1570, + "valid_targets_mean": 5072.1, + "valid_targets_min": 2320 + }, + { + "epoch": 2.5697959183673467, + "grad_norm": 0.5032636329867763, + "learning_rate": 3.194317183405573e-05, + "loss": 0.2249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10128222405910492, + "step": 1575, + "valid_targets_mean": 4459.9, + "valid_targets_min": 1882 + }, + { + "epoch": 2.577959183673469, + "grad_norm": 0.44825229490940477, + "learning_rate": 3.187780655149253e-05, + "loss": 0.2208, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10649105161428452, + "step": 1580, + "valid_targets_mean": 5630.6, + "valid_targets_min": 2670 + }, + { + "epoch": 2.5861224489795918, + "grad_norm": 0.4744318297919026, + "learning_rate": 3.1812244672421293e-05, + "loss": 0.2358, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12488362193107605, + "step": 1585, + "valid_targets_mean": 5510.8, + "valid_targets_min": 1727 + }, + { + "epoch": 2.5942857142857143, + "grad_norm": 0.46273596039967474, + "learning_rate": 3.174648728199491e-05, + "loss": 0.2257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09472452104091644, + "step": 1590, + "valid_targets_mean": 4258.9, + "valid_targets_min": 1055 + }, + { + "epoch": 2.602448979591837, + "grad_norm": 0.6295062662437589, + "learning_rate": 3.168053546860234e-05, + "loss": 0.2274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13179591298103333, + "step": 1595, + "valid_targets_mean": 4865.5, + "valid_targets_min": 991 + }, + { + "epoch": 2.6106122448979594, + "grad_norm": 0.47096472069028195, + "learning_rate": 3.1614390323850545e-05, + "loss": 0.2171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08240728825330734, + "step": 1600, + "valid_targets_mean": 3505.2, + "valid_targets_min": 2442 + }, + { + "epoch": 2.6187755102040815, + "grad_norm": 0.46338642405080044, + "learning_rate": 3.15480529425464e-05, + "loss": 0.227, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17011001706123352, + "step": 1605, + "valid_targets_mean": 6040.1, + "valid_targets_min": 2068 + }, + { + "epoch": 2.626938775510204, + "grad_norm": 0.4699772712282333, + "learning_rate": 3.148152442267863e-05, + "loss": 0.2199, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07595792412757874, + "step": 1610, + "valid_targets_mean": 2887.2, + "valid_targets_min": 580 + }, + { + "epoch": 2.6351020408163266, + "grad_norm": 0.5128860468493535, + "learning_rate": 3.141480586539958e-05, + "loss": 0.23, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11134334653615952, + "step": 1615, + "valid_targets_mean": 3978.0, + "valid_targets_min": 1984 + }, + { + "epoch": 2.643265306122449, + "grad_norm": 0.48807682019439635, + "learning_rate": 3.1347898375007046e-05, + "loss": 0.2286, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10528915375471115, + "step": 1620, + "valid_targets_mean": 4671.4, + "valid_targets_min": 2219 + }, + { + "epoch": 2.6514285714285712, + "grad_norm": 0.5321071877691989, + "learning_rate": 3.128080305892593e-05, + "loss": 0.2354, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11503913253545761, + "step": 1625, + "valid_targets_mean": 4098.6, + "valid_targets_min": 2062 + }, + { + "epoch": 2.659591836734694, + "grad_norm": 0.47764688293351737, + "learning_rate": 3.121352102768998e-05, + "loss": 0.2344, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13357749581336975, + "step": 1630, + "valid_targets_mean": 5889.1, + "valid_targets_min": 1806 + }, + { + "epoch": 2.6677551020408163, + "grad_norm": 0.43425010363034666, + "learning_rate": 3.114605339492335e-05, + "loss": 0.2179, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08906761556863785, + "step": 1635, + "valid_targets_mean": 4261.2, + "valid_targets_min": 1482 + }, + { + "epoch": 2.675918367346939, + "grad_norm": 0.4790146606325519, + "learning_rate": 3.107840127732221e-05, + "loss": 0.2297, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09749139845371246, + "step": 1640, + "valid_targets_mean": 3916.4, + "valid_targets_min": 1664 + }, + { + "epoch": 2.6840816326530614, + "grad_norm": 0.47802026551929605, + "learning_rate": 3.1010565794636236e-05, + "loss": 0.2196, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0704977810382843, + "step": 1645, + "valid_targets_mean": 2540.2, + "valid_targets_min": 408 + }, + { + "epoch": 2.6922448979591835, + "grad_norm": 0.4139517698882833, + "learning_rate": 3.0942548069650085e-05, + "loss": 0.2262, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08792874217033386, + "step": 1650, + "valid_targets_mean": 4686.0, + "valid_targets_min": 385 + }, + { + "epoch": 2.700408163265306, + "grad_norm": 0.49595022031258346, + "learning_rate": 3.0874349228164836e-05, + "loss": 0.2254, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11520551890134811, + "step": 1655, + "valid_targets_mean": 4993.1, + "valid_targets_min": 2719 + }, + { + "epoch": 2.7085714285714286, + "grad_norm": 0.4840889074246445, + "learning_rate": 3.08059703989793e-05, + "loss": 0.2378, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09348023682832718, + "step": 1660, + "valid_targets_mean": 3658.9, + "valid_targets_min": 1500 + }, + { + "epoch": 2.716734693877551, + "grad_norm": 0.46075186086680664, + "learning_rate": 3.073741271387138e-05, + "loss": 0.2153, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09797358512878418, + "step": 1665, + "valid_targets_mean": 4534.5, + "valid_targets_min": 182 + }, + { + "epoch": 2.7248979591836733, + "grad_norm": 0.44333047710393725, + "learning_rate": 3.0668677307579344e-05, + "loss": 0.2284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11998357623815536, + "step": 1670, + "valid_targets_mean": 5285.8, + "valid_targets_min": 1452 + }, + { + "epoch": 2.733061224489796, + "grad_norm": 0.48610888044017014, + "learning_rate": 3.059976531778301e-05, + "loss": 0.235, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12351099401712418, + "step": 1675, + "valid_targets_mean": 5781.0, + "valid_targets_min": 2358 + }, + { + "epoch": 2.7412244897959184, + "grad_norm": 0.5038017541644728, + "learning_rate": 3.0530677885084946e-05, + "loss": 0.2242, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11074423789978027, + "step": 1680, + "valid_targets_mean": 3891.0, + "valid_targets_min": 231 + }, + { + "epoch": 2.749387755102041, + "grad_norm": 0.39086467925287205, + "learning_rate": 3.0461416152991555e-05, + "loss": 0.214, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10134381055831909, + "step": 1685, + "valid_targets_mean": 5829.8, + "valid_targets_min": 2469 + }, + { + "epoch": 2.7575510204081635, + "grad_norm": 0.5035516790420694, + "learning_rate": 3.0391981267894195e-05, + "loss": 0.2332, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13245612382888794, + "step": 1690, + "valid_targets_mean": 5677.1, + "valid_targets_min": 2756 + }, + { + "epoch": 2.7657142857142856, + "grad_norm": 0.534397662500551, + "learning_rate": 3.0322374379050166e-05, + "loss": 0.234, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10749916732311249, + "step": 1695, + "valid_targets_mean": 3503.0, + "valid_targets_min": 612 + }, + { + "epoch": 2.773877551020408, + "grad_norm": 0.5164717311339068, + "learning_rate": 3.0252596638563717e-05, + "loss": 0.2336, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09786826372146606, + "step": 1700, + "valid_targets_mean": 4356.8, + "valid_targets_min": 2568 + }, + { + "epoch": 2.7820408163265307, + "grad_norm": 0.4634558853870556, + "learning_rate": 3.0182649201366944e-05, + "loss": 0.2263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12021247297525406, + "step": 1705, + "valid_targets_mean": 5664.5, + "valid_targets_min": 2675 + }, + { + "epoch": 2.790204081632653, + "grad_norm": 0.4967608998125744, + "learning_rate": 3.0112533225200694e-05, + "loss": 0.2095, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07486874610185623, + "step": 1710, + "valid_targets_mean": 3141.6, + "valid_targets_min": 1362 + }, + { + "epoch": 2.7983673469387753, + "grad_norm": 0.4872167497296232, + "learning_rate": 3.0042249870595424e-05, + "loss": 0.2282, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10342056304216385, + "step": 1715, + "valid_targets_mean": 4081.1, + "valid_targets_min": 497 + }, + { + "epoch": 2.806530612244898, + "grad_norm": 0.5608970204261299, + "learning_rate": 2.9971800300851947e-05, + "loss": 0.2358, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09813547134399414, + "step": 1720, + "valid_targets_mean": 3014.8, + "valid_targets_min": 842 + }, + { + "epoch": 2.8146938775510204, + "grad_norm": 0.4983776943603883, + "learning_rate": 2.9901185682022208e-05, + "loss": 0.2224, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1242903396487236, + "step": 1725, + "valid_targets_mean": 4135.0, + "valid_targets_min": 2853 + }, + { + "epoch": 2.822857142857143, + "grad_norm": 0.4616169148945666, + "learning_rate": 2.9830407182889966e-05, + "loss": 0.2267, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13832835853099823, + "step": 1730, + "valid_targets_mean": 5644.0, + "valid_targets_min": 2770 + }, + { + "epoch": 2.8310204081632655, + "grad_norm": 0.43901300519214903, + "learning_rate": 2.9759465974951473e-05, + "loss": 0.2304, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1145050972700119, + "step": 1735, + "valid_targets_mean": 5897.6, + "valid_targets_min": 2242 + }, + { + "epoch": 2.8391836734693876, + "grad_norm": 0.37515014808594804, + "learning_rate": 2.9688363232396056e-05, + "loss": 0.2322, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10348987579345703, + "step": 1740, + "valid_targets_mean": 6573.2, + "valid_targets_min": 3306 + }, + { + "epoch": 2.84734693877551, + "grad_norm": 0.44557975300400043, + "learning_rate": 2.9617100132086705e-05, + "loss": 0.2453, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1028202623128891, + "step": 1745, + "valid_targets_mean": 4791.2, + "valid_targets_min": 2091 + }, + { + "epoch": 2.8555102040816327, + "grad_norm": 0.4618656993079806, + "learning_rate": 2.9545677853540584e-05, + "loss": 0.219, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12122482061386108, + "step": 1750, + "valid_targets_mean": 5584.4, + "valid_targets_min": 2143 + }, + { + "epoch": 2.8636734693877552, + "grad_norm": 0.5201277039799321, + "learning_rate": 2.9474097578909507e-05, + "loss": 0.2268, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12581226229667664, + "step": 1755, + "valid_targets_mean": 4061.0, + "valid_targets_min": 871 + }, + { + "epoch": 2.8718367346938773, + "grad_norm": 0.3861661290634797, + "learning_rate": 2.9402360492960377e-05, + "loss": 0.2207, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09558389335870743, + "step": 1760, + "valid_targets_mean": 7256.2, + "valid_targets_min": 1539 + }, + { + "epoch": 2.88, + "grad_norm": 0.46182446582565007, + "learning_rate": 2.933046778305556e-05, + "loss": 0.2261, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11080855131149292, + "step": 1765, + "valid_targets_mean": 4730.5, + "valid_targets_min": 1340 + }, + { + "epoch": 2.8881632653061224, + "grad_norm": 0.5463925298521676, + "learning_rate": 2.9258420639133253e-05, + "loss": 0.2329, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14682868123054504, + "step": 1770, + "valid_targets_mean": 4417.8, + "valid_targets_min": 1779 + }, + { + "epoch": 2.896326530612245, + "grad_norm": 0.4948282687111177, + "learning_rate": 2.918622025368779e-05, + "loss": 0.2305, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10451407730579376, + "step": 1775, + "valid_targets_mean": 4030.2, + "valid_targets_min": 1111 + }, + { + "epoch": 2.9044897959183675, + "grad_norm": 0.4208329318988087, + "learning_rate": 2.9113867821749883e-05, + "loss": 0.2261, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10942228138446808, + "step": 1780, + "valid_targets_mean": 6234.6, + "valid_targets_min": 2505 + }, + { + "epoch": 2.9126530612244896, + "grad_norm": 0.46262665189609575, + "learning_rate": 2.9041364540866855e-05, + "loss": 0.2202, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1118980124592781, + "step": 1785, + "valid_targets_mean": 5584.4, + "valid_targets_min": 2249 + }, + { + "epoch": 2.920816326530612, + "grad_norm": 0.4653192043672795, + "learning_rate": 2.8968711611082825e-05, + "loss": 0.2316, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11189016699790955, + "step": 1790, + "valid_targets_mean": 4856.0, + "valid_targets_min": 1638 + }, + { + "epoch": 2.9289795918367347, + "grad_norm": 0.5381032783967711, + "learning_rate": 2.8895910234918828e-05, + "loss": 0.231, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12157237529754639, + "step": 1795, + "valid_targets_mean": 4281.2, + "valid_targets_min": 1729 + }, + { + "epoch": 2.9371428571428573, + "grad_norm": 0.45749805411420635, + "learning_rate": 2.882296161735293e-05, + "loss": 0.2268, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17637786269187927, + "step": 1800, + "valid_targets_mean": 7033.2, + "valid_targets_min": 3103 + }, + { + "epoch": 2.9453061224489794, + "grad_norm": 0.5185768036987373, + "learning_rate": 2.874986696580028e-05, + "loss": 0.2238, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1181664690375328, + "step": 1805, + "valid_targets_mean": 4420.5, + "valid_targets_min": 1700 + }, + { + "epoch": 2.953469387755102, + "grad_norm": 0.530425221551735, + "learning_rate": 2.86766274900931e-05, + "loss": 0.2331, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0907604768872261, + "step": 1810, + "valid_targets_mean": 3195.0, + "valid_targets_min": 868 + }, + { + "epoch": 2.9616326530612245, + "grad_norm": 0.463810084198026, + "learning_rate": 2.8603244402460703e-05, + "loss": 0.228, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09564442187547684, + "step": 1815, + "valid_targets_mean": 4673.5, + "valid_targets_min": 1035 + }, + { + "epoch": 2.969795918367347, + "grad_norm": 0.4874888396808933, + "learning_rate": 2.8529718917509393e-05, + "loss": 0.2218, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14039471745491028, + "step": 1820, + "valid_targets_mean": 5813.9, + "valid_targets_min": 2879 + }, + { + "epoch": 2.9779591836734696, + "grad_norm": 0.5728968555651159, + "learning_rate": 2.8456052252202385e-05, + "loss": 0.2212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1152806431055069, + "step": 1825, + "valid_targets_mean": 3103.2, + "valid_targets_min": 394 + }, + { + "epoch": 2.9861224489795917, + "grad_norm": 0.4835541798438136, + "learning_rate": 2.838224562583965e-05, + "loss": 0.2297, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09890221059322357, + "step": 1830, + "valid_targets_mean": 3995.4, + "valid_targets_min": 1916 + }, + { + "epoch": 2.994285714285714, + "grad_norm": 0.5015938870952967, + "learning_rate": 2.8308300260037734e-05, + "loss": 0.2321, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11656036227941513, + "step": 1835, + "valid_targets_mean": 4335.5, + "valid_targets_min": 2135 + }, + { + "epoch": 3.0016326530612245, + "grad_norm": 0.5164883314650555, + "learning_rate": 2.8234217378709533e-05, + "loss": 0.2193, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1306457817554474, + "step": 1840, + "valid_targets_mean": 5073.4, + "valid_targets_min": 3300 + }, + { + "epoch": 3.009795918367347, + "grad_norm": 0.4295219668494872, + "learning_rate": 2.8159998208044054e-05, + "loss": 0.21, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09993826597929001, + "step": 1845, + "valid_targets_mean": 5941.2, + "valid_targets_min": 2309 + }, + { + "epoch": 3.0179591836734696, + "grad_norm": 0.47919659318523444, + "learning_rate": 2.80856439764861e-05, + "loss": 0.2094, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09795888513326645, + "step": 1850, + "valid_targets_mean": 4683.8, + "valid_targets_min": 1422 + }, + { + "epoch": 3.0261224489795917, + "grad_norm": 0.40561110927902666, + "learning_rate": 2.8011155914715966e-05, + "loss": 0.1975, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08308486640453339, + "step": 1855, + "valid_targets_mean": 4387.9, + "valid_targets_min": 1836 + }, + { + "epoch": 3.0342857142857143, + "grad_norm": 0.4522047892147258, + "learning_rate": 2.7936535255629012e-05, + "loss": 0.2058, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07871349155902863, + "step": 1860, + "valid_targets_mean": 4495.1, + "valid_targets_min": 494 + }, + { + "epoch": 3.042448979591837, + "grad_norm": 0.5334853924211195, + "learning_rate": 2.786178323431534e-05, + "loss": 0.2012, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08362503349781036, + "step": 1865, + "valid_targets_mean": 3050.2, + "valid_targets_min": 228 + }, + { + "epoch": 3.0506122448979593, + "grad_norm": 0.5357442255015453, + "learning_rate": 2.7786901088039252e-05, + "loss": 0.2073, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1317649483680725, + "step": 1870, + "valid_targets_mean": 5041.1, + "valid_targets_min": 2540 + }, + { + "epoch": 3.0587755102040814, + "grad_norm": 0.3913265665588566, + "learning_rate": 2.771189005621887e-05, + "loss": 0.2011, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09924866259098053, + "step": 1875, + "valid_targets_mean": 7752.0, + "valid_targets_min": 2987 + }, + { + "epoch": 3.066938775510204, + "grad_norm": 0.5177456280517249, + "learning_rate": 2.7636751380405548e-05, + "loss": 0.2192, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08077879995107651, + "step": 1880, + "valid_targets_mean": 3772.6, + "valid_targets_min": 428 + }, + { + "epoch": 3.0751020408163265, + "grad_norm": 0.554211366794004, + "learning_rate": 2.7561486304263368e-05, + "loss": 0.2113, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0937132015824318, + "step": 1885, + "valid_targets_mean": 3425.9, + "valid_targets_min": 1557 + }, + { + "epoch": 3.083265306122449, + "grad_norm": 0.5273556149528048, + "learning_rate": 2.7486096073548525e-05, + "loss": 0.2043, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10002836585044861, + "step": 1890, + "valid_targets_mean": 3580.9, + "valid_targets_min": 347 + }, + { + "epoch": 3.0914285714285716, + "grad_norm": 0.4568666684236704, + "learning_rate": 2.7410581936088724e-05, + "loss": 0.2124, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12599283456802368, + "step": 1895, + "valid_targets_mean": 5725.8, + "valid_targets_min": 1676 + }, + { + "epoch": 3.0995918367346937, + "grad_norm": 0.5178725047067185, + "learning_rate": 2.733494514176253e-05, + "loss": 0.2096, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14653289318084717, + "step": 1900, + "valid_targets_mean": 5193.2, + "valid_targets_min": 2352 + }, + { + "epoch": 3.1077551020408163, + "grad_norm": 0.5266790142927706, + "learning_rate": 2.7259186942478656e-05, + "loss": 0.2004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1208992749452591, + "step": 1905, + "valid_targets_mean": 5259.6, + "valid_targets_min": 3002 + }, + { + "epoch": 3.115918367346939, + "grad_norm": 0.45599128475629863, + "learning_rate": 2.718330859215528e-05, + "loss": 0.2042, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09286023676395416, + "step": 1910, + "valid_targets_mean": 4556.6, + "valid_targets_min": 1514 + }, + { + "epoch": 3.1240816326530614, + "grad_norm": 0.4647475863679506, + "learning_rate": 2.710731134669926e-05, + "loss": 0.2063, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1082848533987999, + "step": 1915, + "valid_targets_mean": 5435.9, + "valid_targets_min": 2714 + }, + { + "epoch": 3.1322448979591835, + "grad_norm": 0.5485809884441133, + "learning_rate": 2.7031196463985355e-05, + "loss": 0.2146, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1032031774520874, + "step": 1920, + "valid_targets_mean": 4039.2, + "valid_targets_min": 2406 + }, + { + "epoch": 3.140408163265306, + "grad_norm": 0.4890342953844648, + "learning_rate": 2.69549652038354e-05, + "loss": 0.2118, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09243686497211456, + "step": 1925, + "valid_targets_mean": 4101.8, + "valid_targets_min": 1504 + }, + { + "epoch": 3.1485714285714286, + "grad_norm": 0.550436781586566, + "learning_rate": 2.687861882799748e-05, + "loss": 0.201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09199976921081543, + "step": 1930, + "valid_targets_mean": 4970.2, + "valid_targets_min": 1889 + }, + { + "epoch": 3.156734693877551, + "grad_norm": 0.5099546687084738, + "learning_rate": 2.6802158600125022e-05, + "loss": 0.2129, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07879088819026947, + "step": 1935, + "valid_targets_mean": 3468.5, + "valid_targets_min": 2111 + }, + { + "epoch": 3.1648979591836737, + "grad_norm": 0.555886818946411, + "learning_rate": 2.6725585785755873e-05, + "loss": 0.2002, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.109479621052742, + "step": 1940, + "valid_targets_mean": 5166.8, + "valid_targets_min": 1744 + }, + { + "epoch": 3.1730612244897958, + "grad_norm": 0.4420617318565779, + "learning_rate": 2.6648901652291366e-05, + "loss": 0.2059, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10851933062076569, + "step": 1945, + "valid_targets_mean": 6564.5, + "valid_targets_min": 755 + }, + { + "epoch": 3.1812244897959183, + "grad_norm": 0.4094155215148865, + "learning_rate": 2.657210746897535e-05, + "loss": 0.1962, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08861730992794037, + "step": 1950, + "valid_targets_mean": 4720.9, + "valid_targets_min": 2427 + }, + { + "epoch": 3.189387755102041, + "grad_norm": 0.47205587038214847, + "learning_rate": 2.649520450687317e-05, + "loss": 0.2031, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0953550860285759, + "step": 1955, + "valid_targets_mean": 4718.2, + "valid_targets_min": 1569 + }, + { + "epoch": 3.1975510204081634, + "grad_norm": 0.5460699888433964, + "learning_rate": 2.6418194038850634e-05, + "loss": 0.2058, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12797680497169495, + "step": 1960, + "valid_targets_mean": 5194.0, + "valid_targets_min": 2231 + }, + { + "epoch": 3.2057142857142855, + "grad_norm": 0.5232868644903416, + "learning_rate": 2.6341077339552938e-05, + "loss": 0.1986, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0944906622171402, + "step": 1965, + "valid_targets_mean": 4287.2, + "valid_targets_min": 2086 + }, + { + "epoch": 3.213877551020408, + "grad_norm": 0.4393077695920715, + "learning_rate": 2.6263855685383585e-05, + "loss": 0.1889, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0835411548614502, + "step": 1970, + "valid_targets_mean": 4544.9, + "valid_targets_min": 2044 + }, + { + "epoch": 3.2220408163265306, + "grad_norm": 0.5802825664153698, + "learning_rate": 2.6186530354483224e-05, + "loss": 0.2088, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07232633233070374, + "step": 1975, + "valid_targets_mean": 3895.1, + "valid_targets_min": 1513 + }, + { + "epoch": 3.230204081632653, + "grad_norm": 0.7167557743459537, + "learning_rate": 2.6109102626708547e-05, + "loss": 0.2026, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10113707929849625, + "step": 1980, + "valid_targets_mean": 3797.6, + "valid_targets_min": 875 + }, + { + "epoch": 3.2383673469387757, + "grad_norm": 0.4045153165661452, + "learning_rate": 2.6031573783611053e-05, + "loss": 0.203, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.131624236702919, + "step": 1985, + "valid_targets_mean": 7943.2, + "valid_targets_min": 2568 + }, + { + "epoch": 3.246530612244898, + "grad_norm": 0.5833282836855167, + "learning_rate": 2.5953945108415883e-05, + "loss": 0.2118, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10405785590410233, + "step": 1990, + "valid_targets_mean": 3333.1, + "valid_targets_min": 1429 + }, + { + "epoch": 3.2546938775510204, + "grad_norm": 0.508945720636517, + "learning_rate": 2.5876217886000526e-05, + "loss": 0.2243, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11219657957553864, + "step": 1995, + "valid_targets_mean": 4835.9, + "valid_targets_min": 629 + }, + { + "epoch": 3.262857142857143, + "grad_norm": 0.5076274988600784, + "learning_rate": 2.5798393402873614e-05, + "loss": 0.2172, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0813908651471138, + "step": 2000, + "valid_targets_mean": 3677.1, + "valid_targets_min": 1703 + }, + { + "epoch": 3.2710204081632654, + "grad_norm": 0.5434822373214382, + "learning_rate": 2.5720472947153582e-05, + "loss": 0.201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12669488787651062, + "step": 2005, + "valid_targets_mean": 4550.0, + "valid_targets_min": 1585 + }, + { + "epoch": 3.2791836734693875, + "grad_norm": 0.451823896304527, + "learning_rate": 2.5642457808547365e-05, + "loss": 0.2076, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09495006501674652, + "step": 2010, + "valid_targets_mean": 5470.9, + "valid_targets_min": 2327 + }, + { + "epoch": 3.28734693877551, + "grad_norm": 0.6430875308815419, + "learning_rate": 2.5564349278329056e-05, + "loss": 0.2099, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08256812393665314, + "step": 2015, + "valid_targets_mean": 2062.4, + "valid_targets_min": 998 + }, + { + "epoch": 3.2955102040816326, + "grad_norm": 0.43098561769347765, + "learning_rate": 2.5486148649318514e-05, + "loss": 0.2014, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08821088820695877, + "step": 2020, + "valid_targets_mean": 5786.9, + "valid_targets_min": 2578 + }, + { + "epoch": 3.303673469387755, + "grad_norm": 0.6356272149997206, + "learning_rate": 2.540785721586e-05, + "loss": 0.1946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09246645122766495, + "step": 2025, + "valid_targets_mean": 3517.8, + "valid_targets_min": 2197 + }, + { + "epoch": 3.3118367346938777, + "grad_norm": 0.4454137575172535, + "learning_rate": 2.53294762738007e-05, + "loss": 0.2151, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10923437029123306, + "step": 2030, + "valid_targets_mean": 6211.1, + "valid_targets_min": 1175 + }, + { + "epoch": 3.32, + "grad_norm": 0.5257535142881744, + "learning_rate": 2.5251007120469347e-05, + "loss": 0.2101, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13460084795951843, + "step": 2035, + "valid_targets_mean": 5227.9, + "valid_targets_min": 2360 + }, + { + "epoch": 3.3281632653061224, + "grad_norm": 0.5088064453423816, + "learning_rate": 2.5172451054654685e-05, + "loss": 0.2121, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10537538677453995, + "step": 2040, + "valid_targets_mean": 4792.5, + "valid_targets_min": 1346 + }, + { + "epoch": 3.336326530612245, + "grad_norm": 0.49456466651480946, + "learning_rate": 2.5093809376584006e-05, + "loss": 0.2177, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08632659167051315, + "step": 2045, + "valid_targets_mean": 4183.0, + "valid_targets_min": 2064 + }, + { + "epoch": 3.3444897959183675, + "grad_norm": 0.5409869448697774, + "learning_rate": 2.501508338790162e-05, + "loss": 0.2069, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12297812104225159, + "step": 2050, + "valid_targets_mean": 5444.6, + "valid_targets_min": 2788 + }, + { + "epoch": 3.3526530612244896, + "grad_norm": 0.5093822050798533, + "learning_rate": 2.49362743916473e-05, + "loss": 0.2135, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09434787929058075, + "step": 2055, + "valid_targets_mean": 3569.2, + "valid_targets_min": 2053 + }, + { + "epoch": 3.360816326530612, + "grad_norm": 0.47603529261568117, + "learning_rate": 2.4857383692234742e-05, + "loss": 0.2021, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12311601638793945, + "step": 2060, + "valid_targets_mean": 6423.6, + "valid_targets_min": 2096 + }, + { + "epoch": 3.3689795918367347, + "grad_norm": 0.534211960665438, + "learning_rate": 2.4778412595429955e-05, + "loss": 0.2085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10466711223125458, + "step": 2065, + "valid_targets_mean": 4221.9, + "valid_targets_min": 1956 + }, + { + "epoch": 3.3771428571428572, + "grad_norm": 0.5472416262428637, + "learning_rate": 2.4699362408329646e-05, + "loss": 0.2, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10589661449193954, + "step": 2070, + "valid_targets_mean": 3920.5, + "valid_targets_min": 620 + }, + { + "epoch": 3.3853061224489798, + "grad_norm": 0.547677700443028, + "learning_rate": 2.46202344393396e-05, + "loss": 0.2168, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1077793538570404, + "step": 2075, + "valid_targets_mean": 3886.6, + "valid_targets_min": 684 + }, + { + "epoch": 3.393469387755102, + "grad_norm": 0.5831345321921154, + "learning_rate": 2.4541029998153e-05, + "loss": 0.1992, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1060214415192604, + "step": 2080, + "valid_targets_mean": 3710.5, + "valid_targets_min": 1638 + }, + { + "epoch": 3.4016326530612244, + "grad_norm": 0.493306213537092, + "learning_rate": 2.446175039572878e-05, + "loss": 0.2018, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11661600321531296, + "step": 2085, + "valid_targets_mean": 5181.2, + "valid_targets_min": 1399 + }, + { + "epoch": 3.409795918367347, + "grad_norm": 0.5871415626775445, + "learning_rate": 2.4382396944269912e-05, + "loss": 0.2076, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09259402751922607, + "step": 2090, + "valid_targets_mean": 3459.8, + "valid_targets_min": 1845 + }, + { + "epoch": 3.4179591836734695, + "grad_norm": 0.5745344129842763, + "learning_rate": 2.430297095720168e-05, + "loss": 0.2141, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09636922925710678, + "step": 2095, + "valid_targets_mean": 3819.6, + "valid_targets_min": 186 + }, + { + "epoch": 3.4261224489795916, + "grad_norm": 0.4721493101920305, + "learning_rate": 2.4223473749149937e-05, + "loss": 0.1987, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09481912851333618, + "step": 2100, + "valid_targets_mean": 4979.5, + "valid_targets_min": 1675 + }, + { + "epoch": 3.434285714285714, + "grad_norm": 0.49501466744059974, + "learning_rate": 2.4143906635919385e-05, + "loss": 0.2079, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08848510682582855, + "step": 2105, + "valid_targets_mean": 4279.2, + "valid_targets_min": 839 + }, + { + "epoch": 3.4424489795918367, + "grad_norm": 0.3912763685993033, + "learning_rate": 2.4064270934471735e-05, + "loss": 0.2105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0881570428609848, + "step": 2110, + "valid_targets_mean": 5399.5, + "valid_targets_min": 2613 + }, + { + "epoch": 3.4506122448979593, + "grad_norm": 0.43492407833873475, + "learning_rate": 2.3984567962903968e-05, + "loss": 0.2021, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1029646098613739, + "step": 2115, + "valid_targets_mean": 4683.2, + "valid_targets_min": 3756 + }, + { + "epoch": 3.458775510204082, + "grad_norm": 0.5260597065670592, + "learning_rate": 2.3904799040426476e-05, + "loss": 0.2074, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09860092401504517, + "step": 2120, + "valid_targets_mean": 3959.4, + "valid_targets_min": 2219 + }, + { + "epoch": 3.466938775510204, + "grad_norm": 0.47356375082893604, + "learning_rate": 2.3824965487341247e-05, + "loss": 0.2102, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11778430640697479, + "step": 2125, + "valid_targets_mean": 5864.9, + "valid_targets_min": 3196 + }, + { + "epoch": 3.4751020408163265, + "grad_norm": 0.5462984356735526, + "learning_rate": 2.3745068625020008e-05, + "loss": 0.2144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.119639553129673, + "step": 2130, + "valid_targets_mean": 4661.4, + "valid_targets_min": 2031 + }, + { + "epoch": 3.483265306122449, + "grad_norm": 0.5206942495860799, + "learning_rate": 2.3665109775882352e-05, + "loss": 0.2071, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08757778257131577, + "step": 2135, + "valid_targets_mean": 3686.8, + "valid_targets_min": 1095 + }, + { + "epoch": 3.4914285714285715, + "grad_norm": 0.5018763236389352, + "learning_rate": 2.3585090263373868e-05, + "loss": 0.2106, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11070487648248672, + "step": 2140, + "valid_targets_mean": 4611.9, + "valid_targets_min": 1927 + }, + { + "epoch": 3.4995918367346937, + "grad_norm": 0.47575628659027597, + "learning_rate": 2.350501141194419e-05, + "loss": 0.2066, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10880281031131744, + "step": 2145, + "valid_targets_mean": 5161.1, + "valid_targets_min": 1502 + }, + { + "epoch": 3.507755102040816, + "grad_norm": 0.4928847665812081, + "learning_rate": 2.3424874547025125e-05, + "loss": 0.2127, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11148449778556824, + "step": 2150, + "valid_targets_mean": 5960.8, + "valid_targets_min": 1937 + }, + { + "epoch": 3.5159183673469387, + "grad_norm": 0.4433641303426999, + "learning_rate": 2.33446809950087e-05, + "loss": 0.2134, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1022956520318985, + "step": 2155, + "valid_targets_mean": 5677.5, + "valid_targets_min": 2168 + }, + { + "epoch": 3.5240816326530613, + "grad_norm": 0.46201695625079175, + "learning_rate": 2.326443208322519e-05, + "loss": 0.2091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1019848957657814, + "step": 2160, + "valid_targets_mean": 5709.8, + "valid_targets_min": 2873 + }, + { + "epoch": 3.532244897959184, + "grad_norm": 0.4932375115901645, + "learning_rate": 2.3184129139921162e-05, + "loss": 0.2054, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0880424901843071, + "step": 2165, + "valid_targets_mean": 4456.4, + "valid_targets_min": 754 + }, + { + "epoch": 3.5404081632653064, + "grad_norm": 0.4415593903313715, + "learning_rate": 2.3103773494237504e-05, + "loss": 0.1912, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08950482308864594, + "step": 2170, + "valid_targets_mean": 4576.4, + "valid_targets_min": 2308 + }, + { + "epoch": 3.5485714285714285, + "grad_norm": 0.6079075078157, + "learning_rate": 2.3023366476187402e-05, + "loss": 0.2069, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1052299439907074, + "step": 2175, + "valid_targets_mean": 2974.9, + "valid_targets_min": 1248 + }, + { + "epoch": 3.556734693877551, + "grad_norm": 0.3989429150863732, + "learning_rate": 2.2942909416634326e-05, + "loss": 0.1996, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10140019655227661, + "step": 2180, + "valid_targets_mean": 5633.8, + "valid_targets_min": 3165 + }, + { + "epoch": 3.5648979591836736, + "grad_norm": 0.4904211159678837, + "learning_rate": 2.2862403647270033e-05, + "loss": 0.2169, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10210586339235306, + "step": 2185, + "valid_targets_mean": 4799.4, + "valid_targets_min": 589 + }, + { + "epoch": 3.5730612244897957, + "grad_norm": 0.5484079285629692, + "learning_rate": 2.2781850500592487e-05, + "loss": 0.2071, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11891067028045654, + "step": 2190, + "valid_targets_mean": 5383.9, + "valid_targets_min": 3293 + }, + { + "epoch": 3.5812244897959182, + "grad_norm": 0.6008177442360957, + "learning_rate": 2.2701251309883837e-05, + "loss": 0.2055, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10325556993484497, + "step": 2195, + "valid_targets_mean": 3358.9, + "valid_targets_min": 1192 + }, + { + "epoch": 3.5893877551020408, + "grad_norm": 0.5227749077095252, + "learning_rate": 2.2620607409188326e-05, + "loss": 0.2104, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09844657778739929, + "step": 2200, + "valid_targets_mean": 4752.5, + "valid_targets_min": 2204 + }, + { + "epoch": 3.5975510204081633, + "grad_norm": 0.5258378975483752, + "learning_rate": 2.2539920133290208e-05, + "loss": 0.2077, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11812394112348557, + "step": 2205, + "valid_targets_mean": 5001.0, + "valid_targets_min": 1199 + }, + { + "epoch": 3.605714285714286, + "grad_norm": 0.5339466282169013, + "learning_rate": 2.245919081769168e-05, + "loss": 0.2064, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10490740835666656, + "step": 2210, + "valid_targets_mean": 3964.6, + "valid_targets_min": 1520 + }, + { + "epoch": 3.6138775510204084, + "grad_norm": 0.4045810022832643, + "learning_rate": 2.2378420798590764e-05, + "loss": 0.2085, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08282206952571869, + "step": 2215, + "valid_targets_mean": 6330.6, + "valid_targets_min": 1416 + }, + { + "epoch": 3.6220408163265305, + "grad_norm": 0.4675388517138899, + "learning_rate": 2.2297611412859178e-05, + "loss": 0.213, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12077487260103226, + "step": 2220, + "valid_targets_mean": 6203.0, + "valid_targets_min": 2988 + }, + { + "epoch": 3.630204081632653, + "grad_norm": 0.4974295688370866, + "learning_rate": 2.2216763998020222e-05, + "loss": 0.2065, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08385932445526123, + "step": 2225, + "valid_targets_mean": 3604.4, + "valid_targets_min": 1055 + }, + { + "epoch": 3.6383673469387756, + "grad_norm": 0.40766596357353385, + "learning_rate": 2.2135879892226643e-05, + "loss": 0.1884, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0922253355383873, + "step": 2230, + "valid_targets_mean": 6555.6, + "valid_targets_min": 2428 + }, + { + "epoch": 3.6465306122448977, + "grad_norm": 0.5442667722758696, + "learning_rate": 2.205496043423849e-05, + "loss": 0.2129, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10492303967475891, + "step": 2235, + "valid_targets_mean": 4267.6, + "valid_targets_min": 508 + }, + { + "epoch": 3.6546938775510203, + "grad_norm": 0.47975005964527107, + "learning_rate": 2.197400696340091e-05, + "loss": 0.2045, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13115760684013367, + "step": 2240, + "valid_targets_mean": 6332.2, + "valid_targets_min": 4055 + }, + { + "epoch": 3.662857142857143, + "grad_norm": 0.5520568799725798, + "learning_rate": 2.189302081962207e-05, + "loss": 0.2147, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09570181369781494, + "step": 2245, + "valid_targets_mean": 3627.2, + "valid_targets_min": 2068 + }, + { + "epoch": 3.6710204081632654, + "grad_norm": 0.5183011327920499, + "learning_rate": 2.1812003343350886e-05, + "loss": 0.2095, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12069438397884369, + "step": 2250, + "valid_targets_mean": 5529.8, + "valid_targets_min": 1055 + }, + { + "epoch": 3.679183673469388, + "grad_norm": 0.40494166134398335, + "learning_rate": 2.1730955875554894e-05, + "loss": 0.1989, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11352106928825378, + "step": 2255, + "valid_targets_mean": 7916.5, + "valid_targets_min": 2533 + }, + { + "epoch": 3.6873469387755105, + "grad_norm": 0.5377042271103643, + "learning_rate": 2.164987975769803e-05, + "loss": 0.2041, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08446522057056427, + "step": 2260, + "valid_targets_mean": 3122.2, + "valid_targets_min": 1664 + }, + { + "epoch": 3.6955102040816326, + "grad_norm": 0.4714079493569945, + "learning_rate": 2.1568776331718428e-05, + "loss": 0.2108, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12201312184333801, + "step": 2265, + "valid_targets_mean": 5722.1, + "valid_targets_min": 2148 + }, + { + "epoch": 3.703673469387755, + "grad_norm": 0.6138764873320968, + "learning_rate": 2.148764694000624e-05, + "loss": 0.2105, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12149910628795624, + "step": 2270, + "valid_targets_mean": 3841.4, + "valid_targets_min": 581 + }, + { + "epoch": 3.7118367346938776, + "grad_norm": 0.6152682833954068, + "learning_rate": 2.1406492925381358e-05, + "loss": 0.1967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1131722629070282, + "step": 2275, + "valid_targets_mean": 3731.0, + "valid_targets_min": 231 + }, + { + "epoch": 3.7199999999999998, + "grad_norm": 0.4810706951283212, + "learning_rate": 2.1325315631071257e-05, + "loss": 0.1972, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12919890880584717, + "step": 2280, + "valid_targets_mean": 5619.1, + "valid_targets_min": 2068 + }, + { + "epoch": 3.7281632653061223, + "grad_norm": 0.4109059218621779, + "learning_rate": 2.124411640068869e-05, + "loss": 0.2078, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0750608742237091, + "step": 2285, + "valid_targets_mean": 4065.9, + "valid_targets_min": 186 + }, + { + "epoch": 3.736326530612245, + "grad_norm": 0.5452804250141515, + "learning_rate": 2.1162896578209517e-05, + "loss": 0.219, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13831168413162231, + "step": 2290, + "valid_targets_mean": 4769.5, + "valid_targets_min": 2192 + }, + { + "epoch": 3.7444897959183674, + "grad_norm": 0.46277655460159545, + "learning_rate": 2.1081657507950423e-05, + "loss": 0.2048, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10686130821704865, + "step": 2295, + "valid_targets_mean": 5485.1, + "valid_targets_min": 1555 + }, + { + "epoch": 3.75265306122449, + "grad_norm": 0.4422005430051166, + "learning_rate": 2.100040053454666e-05, + "loss": 0.2145, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12035848200321198, + "step": 2300, + "valid_targets_mean": 6698.6, + "valid_targets_min": 2586 + }, + { + "epoch": 3.7608163265306125, + "grad_norm": 0.44658354686920304, + "learning_rate": 2.091912700292983e-05, + "loss": 0.2033, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10489645600318909, + "step": 2305, + "valid_targets_mean": 5490.9, + "valid_targets_min": 2301 + }, + { + "epoch": 3.7689795918367346, + "grad_norm": 0.5729406847308828, + "learning_rate": 2.083783825830557e-05, + "loss": 0.2076, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09271632134914398, + "step": 2310, + "valid_targets_mean": 3407.8, + "valid_targets_min": 1110 + }, + { + "epoch": 3.777142857142857, + "grad_norm": 0.5408692265981219, + "learning_rate": 2.075653564613134e-05, + "loss": 0.2055, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08484731614589691, + "step": 2315, + "valid_targets_mean": 3749.0, + "valid_targets_min": 2862 + }, + { + "epoch": 3.7853061224489797, + "grad_norm": 0.5281560420325818, + "learning_rate": 2.0675220512094124e-05, + "loss": 0.2042, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12320023775100708, + "step": 2320, + "valid_targets_mean": 5550.1, + "valid_targets_min": 2278 + }, + { + "epoch": 3.793469387755102, + "grad_norm": 0.5655352291536425, + "learning_rate": 2.059389420208816e-05, + "loss": 0.2156, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12888170778751373, + "step": 2325, + "valid_targets_mean": 3960.6, + "valid_targets_min": 2243 + }, + { + "epoch": 3.8016326530612243, + "grad_norm": 0.5643861980349051, + "learning_rate": 2.0512558062192655e-05, + "loss": 0.2117, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14314799010753632, + "step": 2330, + "valid_targets_mean": 4727.9, + "valid_targets_min": 2596 + }, + { + "epoch": 3.809795918367347, + "grad_norm": 0.6586685049239898, + "learning_rate": 2.043121343864954e-05, + "loss": 0.209, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10116365551948547, + "step": 2335, + "valid_targets_mean": 5758.2, + "valid_targets_min": 2438 + }, + { + "epoch": 3.8179591836734694, + "grad_norm": 0.7771486185341817, + "learning_rate": 2.0349861677841148e-05, + "loss": 0.2123, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1095040887594223, + "step": 2340, + "valid_targets_mean": 4051.2, + "valid_targets_min": 1823 + }, + { + "epoch": 3.826122448979592, + "grad_norm": 0.3983669329609547, + "learning_rate": 2.0268504126267952e-05, + "loss": 0.2062, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08110864460468292, + "step": 2345, + "valid_targets_mean": 5468.2, + "valid_targets_min": 1297 + }, + { + "epoch": 3.8342857142857145, + "grad_norm": 0.5066751984407911, + "learning_rate": 2.0187142130526266e-05, + "loss": 0.206, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11485150456428528, + "step": 2350, + "valid_targets_mean": 4955.1, + "valid_targets_min": 865 + }, + { + "epoch": 3.8424489795918366, + "grad_norm": 0.5141538006278502, + "learning_rate": 2.0105777037285957e-05, + "loss": 0.2108, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09995323419570923, + "step": 2355, + "valid_targets_mean": 4502.9, + "valid_targets_min": 1772 + }, + { + "epoch": 3.850612244897959, + "grad_norm": 0.6241468197660187, + "learning_rate": 2.0024410193268177e-05, + "loss": 0.2055, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09824708104133606, + "step": 2360, + "valid_targets_mean": 3143.4, + "valid_targets_min": 1446 + }, + { + "epoch": 3.8587755102040817, + "grad_norm": 0.4833346466208415, + "learning_rate": 1.994304294522304e-05, + "loss": 0.201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08165107667446136, + "step": 2365, + "valid_targets_mean": 4159.0, + "valid_targets_min": 532 + }, + { + "epoch": 3.866938775510204, + "grad_norm": 0.49067284008168804, + "learning_rate": 1.9861676639907358e-05, + "loss": 0.2133, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09489576518535614, + "step": 2370, + "valid_targets_mean": 4256.8, + "valid_targets_min": 2423 + }, + { + "epoch": 3.8751020408163264, + "grad_norm": 0.4899645525965157, + "learning_rate": 1.978031262406233e-05, + "loss": 0.2151, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09761510044336319, + "step": 2375, + "valid_targets_mean": 4272.1, + "valid_targets_min": 2099 + }, + { + "epoch": 3.883265306122449, + "grad_norm": 0.4823414520905191, + "learning_rate": 1.9698952244391273e-05, + "loss": 0.2147, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09190724790096283, + "step": 2380, + "valid_targets_mean": 4303.9, + "valid_targets_min": 1791 + }, + { + "epoch": 3.8914285714285715, + "grad_norm": 0.4036421152941417, + "learning_rate": 1.9617596847537314e-05, + "loss": 0.2015, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09016790986061096, + "step": 2385, + "valid_targets_mean": 6770.0, + "valid_targets_min": 2283 + }, + { + "epoch": 3.899591836734694, + "grad_norm": 0.5304022072294978, + "learning_rate": 1.953624778006109e-05, + "loss": 0.216, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11460654437541962, + "step": 2390, + "valid_targets_mean": 4059.8, + "valid_targets_min": 2678 + }, + { + "epoch": 3.9077551020408166, + "grad_norm": 0.5475140898120625, + "learning_rate": 1.9454906388418507e-05, + "loss": 0.2018, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09916389733552933, + "step": 2395, + "valid_targets_mean": 4083.2, + "valid_targets_min": 1999 + }, + { + "epoch": 3.9159183673469387, + "grad_norm": 0.5272651986791439, + "learning_rate": 1.93735740189384e-05, + "loss": 0.1993, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08801892399787903, + "step": 2400, + "valid_targets_mean": 3487.5, + "valid_targets_min": 2172 + }, + { + "epoch": 3.924081632653061, + "grad_norm": 0.4996257867275813, + "learning_rate": 1.929225201780029e-05, + "loss": 0.1962, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12266403436660767, + "step": 2405, + "valid_targets_mean": 5626.6, + "valid_targets_min": 2293 + }, + { + "epoch": 3.9322448979591837, + "grad_norm": 0.4803588830227563, + "learning_rate": 1.9210941731012058e-05, + "loss": 0.2034, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09042441099882126, + "step": 2410, + "valid_targets_mean": 4183.6, + "valid_targets_min": 657 + }, + { + "epoch": 3.940408163265306, + "grad_norm": 0.49225801562943217, + "learning_rate": 1.9129644504387738e-05, + "loss": 0.2118, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09421433508396149, + "step": 2415, + "valid_targets_mean": 4357.8, + "valid_targets_min": 1251 + }, + { + "epoch": 3.9485714285714284, + "grad_norm": 0.4419635106773158, + "learning_rate": 1.9048361683525155e-05, + "loss": 0.1873, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08827171474695206, + "step": 2420, + "valid_targets_mean": 5476.1, + "valid_targets_min": 2469 + }, + { + "epoch": 3.956734693877551, + "grad_norm": 0.5346905119601936, + "learning_rate": 1.8967094613783726e-05, + "loss": 0.1998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08471651375293732, + "step": 2425, + "valid_targets_mean": 3831.2, + "valid_targets_min": 2011 + }, + { + "epoch": 3.9648979591836735, + "grad_norm": 0.4829207115923738, + "learning_rate": 1.888584464026213e-05, + "loss": 0.1969, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11744900792837143, + "step": 2430, + "valid_targets_mean": 5821.8, + "valid_targets_min": 726 + }, + { + "epoch": 3.973061224489796, + "grad_norm": 0.3872676695587164, + "learning_rate": 1.8804613107776118e-05, + "loss": 0.201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08249861001968384, + "step": 2435, + "valid_targets_mean": 6168.4, + "valid_targets_min": 2748 + }, + { + "epoch": 3.9812244897959186, + "grad_norm": 0.44806085077099195, + "learning_rate": 1.8723401360836178e-05, + "loss": 0.2127, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11369358003139496, + "step": 2440, + "valid_targets_mean": 6784.9, + "valid_targets_min": 752 + }, + { + "epoch": 3.9893877551020407, + "grad_norm": 0.4710951192858675, + "learning_rate": 1.8642210743625325e-05, + "loss": 0.212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12091173231601715, + "step": 2445, + "valid_targets_mean": 6086.6, + "valid_targets_min": 2989 + }, + { + "epoch": 3.9975510204081632, + "grad_norm": 0.5450661144221071, + "learning_rate": 1.8561042599976854e-05, + "loss": 0.2127, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1176106184720993, + "step": 2450, + "valid_targets_mean": 4691.8, + "valid_targets_min": 2349 + }, + { + "epoch": 4.0048979591836735, + "grad_norm": 0.49684305530383777, + "learning_rate": 1.8479898273352084e-05, + "loss": 0.1961, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09347330778837204, + "step": 2455, + "valid_targets_mean": 4379.8, + "valid_targets_min": 2255 + }, + { + "epoch": 4.013061224489796, + "grad_norm": 0.5102719717526412, + "learning_rate": 1.8398779106818115e-05, + "loss": 0.1964, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10384904593229294, + "step": 2460, + "valid_targets_mean": 5453.8, + "valid_targets_min": 2491 + }, + { + "epoch": 4.021224489795919, + "grad_norm": 0.4928662125813874, + "learning_rate": 1.83176864430256e-05, + "loss": 0.1966, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08712229132652283, + "step": 2465, + "valid_targets_mean": 4610.5, + "valid_targets_min": 1697 + }, + { + "epoch": 4.029387755102041, + "grad_norm": 0.5021600618126286, + "learning_rate": 1.823662162418657e-05, + "loss": 0.1912, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09403041005134583, + "step": 2470, + "valid_targets_mean": 5114.4, + "valid_targets_min": 1426 + }, + { + "epoch": 4.037551020408163, + "grad_norm": 0.5398163844157312, + "learning_rate": 1.815558599205213e-05, + "loss": 0.1975, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09044396877288818, + "step": 2475, + "valid_targets_mean": 3718.1, + "valid_targets_min": 1760 + }, + { + "epoch": 4.045714285714285, + "grad_norm": 0.9492538214781632, + "learning_rate": 1.8074580887890317e-05, + "loss": 0.2029, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10314054042100906, + "step": 2480, + "valid_targets_mean": 3802.1, + "valid_targets_min": 1962 + }, + { + "epoch": 4.053877551020408, + "grad_norm": 0.4125298411521795, + "learning_rate": 1.79936076524639e-05, + "loss": 0.1926, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09343983232975006, + "step": 2485, + "valid_targets_mean": 5443.4, + "valid_targets_min": 2814 + }, + { + "epoch": 4.0620408163265305, + "grad_norm": 0.5505051260285522, + "learning_rate": 1.791266762600815e-05, + "loss": 0.1995, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10079911351203918, + "step": 2490, + "valid_targets_mean": 4203.0, + "valid_targets_min": 1870 + }, + { + "epoch": 4.070204081632653, + "grad_norm": 0.4793360527143795, + "learning_rate": 1.7831762148208672e-05, + "loss": 0.1856, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0887790098786354, + "step": 2495, + "valid_targets_mean": 4890.6, + "valid_targets_min": 2429 + }, + { + "epoch": 4.078367346938776, + "grad_norm": 0.5006011398859094, + "learning_rate": 1.775089255817925e-05, + "loss": 0.1975, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08640684187412262, + "step": 2500, + "valid_targets_mean": 4471.8, + "valid_targets_min": 729 + }, + { + "epoch": 4.086530612244898, + "grad_norm": 0.4823608555582478, + "learning_rate": 1.7670060194439665e-05, + "loss": 0.2024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09267100691795349, + "step": 2505, + "valid_targets_mean": 6793.4, + "valid_targets_min": 3312 + }, + { + "epoch": 4.094693877551021, + "grad_norm": 0.4950168084238867, + "learning_rate": 1.758926639489354e-05, + "loss": 0.1854, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09220251441001892, + "step": 2510, + "valid_targets_mean": 4972.6, + "valid_targets_min": 3191 + }, + { + "epoch": 4.102857142857143, + "grad_norm": 0.5397701907428266, + "learning_rate": 1.7508512496806202e-05, + "loss": 0.1932, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11720062792301178, + "step": 2515, + "valid_targets_mean": 5273.1, + "valid_targets_min": 2051 + }, + { + "epoch": 4.111020408163265, + "grad_norm": 0.5151168385770297, + "learning_rate": 1.7427799836782554e-05, + "loss": 0.1941, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07783135771751404, + "step": 2520, + "valid_targets_mean": 3984.2, + "valid_targets_min": 2415 + }, + { + "epoch": 4.119183673469387, + "grad_norm": 0.5475399954363521, + "learning_rate": 1.7347129750744923e-05, + "loss": 0.2046, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1131369099020958, + "step": 2525, + "valid_targets_mean": 5291.9, + "valid_targets_min": 1966 + }, + { + "epoch": 4.12734693877551, + "grad_norm": 0.4256526467582212, + "learning_rate": 1.7266503573910995e-05, + "loss": 0.1975, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09197886288166046, + "step": 2530, + "valid_targets_mean": 4821.0, + "valid_targets_min": 2732 + }, + { + "epoch": 4.1355102040816325, + "grad_norm": 0.572725873451191, + "learning_rate": 1.7185922640771657e-05, + "loss": 0.1773, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09851355850696564, + "step": 2535, + "valid_targets_mean": 3625.0, + "valid_targets_min": 2596 + }, + { + "epoch": 4.143673469387755, + "grad_norm": 0.48901114744309754, + "learning_rate": 1.710538828506898e-05, + "loss": 0.1867, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09258966147899628, + "step": 2540, + "valid_targets_mean": 5980.6, + "valid_targets_min": 2077 + }, + { + "epoch": 4.151836734693878, + "grad_norm": 0.4713249554585069, + "learning_rate": 1.7024901839774074e-05, + "loss": 0.1936, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11058518290519714, + "step": 2545, + "valid_targets_mean": 5908.0, + "valid_targets_min": 2449 + }, + { + "epoch": 4.16, + "grad_norm": 0.46189356108689933, + "learning_rate": 1.6944464637065062e-05, + "loss": 0.1796, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09432432055473328, + "step": 2550, + "valid_targets_mean": 6157.5, + "valid_targets_min": 2201 + }, + { + "epoch": 4.168163265306123, + "grad_norm": 0.49702673341480896, + "learning_rate": 1.686407800830504e-05, + "loss": 0.1828, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07615435123443604, + "step": 2555, + "valid_targets_mean": 4524.0, + "valid_targets_min": 796 + }, + { + "epoch": 4.176326530612245, + "grad_norm": 0.5819775268795271, + "learning_rate": 1.678374328402001e-05, + "loss": 0.194, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11597669124603271, + "step": 2560, + "valid_targets_mean": 4323.1, + "valid_targets_min": 1681 + }, + { + "epoch": 4.184489795918367, + "grad_norm": 0.5634760624657875, + "learning_rate": 1.6703461793876876e-05, + "loss": 0.1864, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07644718140363693, + "step": 2565, + "valid_targets_mean": 3535.8, + "valid_targets_min": 1396 + }, + { + "epoch": 4.192653061224489, + "grad_norm": 0.5663166930782293, + "learning_rate": 1.6623234866661425e-05, + "loss": 0.1895, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09392023831605911, + "step": 2570, + "valid_targets_mean": 4895.1, + "valid_targets_min": 2242 + }, + { + "epoch": 4.200816326530612, + "grad_norm": 0.4898814924321093, + "learning_rate": 1.6543063830256365e-05, + "loss": 0.191, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.074742391705513, + "step": 2575, + "valid_targets_mean": 3640.9, + "valid_targets_min": 302 + }, + { + "epoch": 4.2089795918367345, + "grad_norm": 0.5329704151981115, + "learning_rate": 1.6462950011619294e-05, + "loss": 0.1888, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10386515408754349, + "step": 2580, + "valid_targets_mean": 4789.4, + "valid_targets_min": 1685 + }, + { + "epoch": 4.217142857142857, + "grad_norm": 0.5654217835025043, + "learning_rate": 1.638289473676079e-05, + "loss": 0.1819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10331140458583832, + "step": 2585, + "valid_targets_mean": 4166.1, + "valid_targets_min": 1248 + }, + { + "epoch": 4.22530612244898, + "grad_norm": 0.4477703963173839, + "learning_rate": 1.6302899330722408e-05, + "loss": 0.1759, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07035226374864578, + "step": 2590, + "valid_targets_mean": 4336.6, + "valid_targets_min": 1659 + }, + { + "epoch": 4.233469387755102, + "grad_norm": 0.5146643374207694, + "learning_rate": 1.622296511755482e-05, + "loss": 0.182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10337590426206589, + "step": 2595, + "valid_targets_mean": 5123.9, + "valid_targets_min": 1762 + }, + { + "epoch": 4.241632653061225, + "grad_norm": 0.5283593915985714, + "learning_rate": 1.6143093420295823e-05, + "loss": 0.1902, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11016341298818588, + "step": 2600, + "valid_targets_mean": 5239.8, + "valid_targets_min": 1520 + }, + { + "epoch": 4.249795918367347, + "grad_norm": 0.44759186145233176, + "learning_rate": 1.6063285560948492e-05, + "loss": 0.1892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09216510504484177, + "step": 2605, + "valid_targets_mean": 5240.5, + "valid_targets_min": 1707 + }, + { + "epoch": 4.257959183673469, + "grad_norm": 0.6073616846345756, + "learning_rate": 1.5983542860459292e-05, + "loss": 0.1899, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.065554678440094, + "step": 2610, + "valid_targets_mean": 2690.9, + "valid_targets_min": 556 + }, + { + "epoch": 4.2661224489795915, + "grad_norm": 0.494273576410177, + "learning_rate": 1.590386663869619e-05, + "loss": 0.1843, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11738970875740051, + "step": 2615, + "valid_targets_mean": 5866.9, + "valid_targets_min": 3054 + }, + { + "epoch": 4.274285714285714, + "grad_norm": 0.47365274825672116, + "learning_rate": 1.5824258214426833e-05, + "loss": 0.1874, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08815792948007584, + "step": 2620, + "valid_targets_mean": 4863.2, + "valid_targets_min": 1469 + }, + { + "epoch": 4.282448979591837, + "grad_norm": 0.3951324311720292, + "learning_rate": 1.57447189052967e-05, + "loss": 0.1972, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07140126824378967, + "step": 2625, + "valid_targets_mean": 5357.2, + "valid_targets_min": 2212 + }, + { + "epoch": 4.290612244897959, + "grad_norm": 0.3899464760912804, + "learning_rate": 1.5665250027807326e-05, + "loss": 0.1807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07976418733596802, + "step": 2630, + "valid_targets_mean": 6317.6, + "valid_targets_min": 870 + }, + { + "epoch": 4.298775510204082, + "grad_norm": 0.4477544191318257, + "learning_rate": 1.558585289729448e-05, + "loss": 0.1886, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08820460736751556, + "step": 2635, + "valid_targets_mean": 5996.9, + "valid_targets_min": 1867 + }, + { + "epoch": 4.306938775510204, + "grad_norm": 0.49235787781606305, + "learning_rate": 1.55065288279064e-05, + "loss": 0.1931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07115676999092102, + "step": 2640, + "valid_targets_mean": 4227.9, + "valid_targets_min": 2739 + }, + { + "epoch": 4.315102040816327, + "grad_norm": 0.5161672595037532, + "learning_rate": 1.5427279132582055e-05, + "loss": 0.195, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10170981287956238, + "step": 2645, + "valid_targets_mean": 5293.6, + "valid_targets_min": 2011 + }, + { + "epoch": 4.323265306122449, + "grad_norm": 0.5046151825642934, + "learning_rate": 1.5348105123029404e-05, + "loss": 0.1861, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08673232793807983, + "step": 2650, + "valid_targets_mean": 4813.0, + "valid_targets_min": 1235 + }, + { + "epoch": 4.331428571428571, + "grad_norm": 0.4703623860473787, + "learning_rate": 1.5269008109703676e-05, + "loss": 0.1949, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07027247548103333, + "step": 2655, + "valid_targets_mean": 4028.8, + "valid_targets_min": 1481 + }, + { + "epoch": 4.3395918367346935, + "grad_norm": 0.5046661473253734, + "learning_rate": 1.5189989401785698e-05, + "loss": 0.1928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1191679909825325, + "step": 2660, + "valid_targets_mean": 5938.4, + "valid_targets_min": 3219 + }, + { + "epoch": 4.347755102040816, + "grad_norm": 0.4248241877322586, + "learning_rate": 1.5111050307160237e-05, + "loss": 0.1893, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07479171454906464, + "step": 2665, + "valid_targets_mean": 6721.6, + "valid_targets_min": 424 + }, + { + "epoch": 4.355918367346939, + "grad_norm": 0.3557581340270684, + "learning_rate": 1.5032192132394305e-05, + "loss": 0.1784, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07210604846477509, + "step": 2670, + "valid_targets_mean": 6250.0, + "valid_targets_min": 2950 + }, + { + "epoch": 4.364081632653061, + "grad_norm": 0.5382046936001726, + "learning_rate": 1.4953416182715566e-05, + "loss": 0.1923, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07258506119251251, + "step": 2675, + "valid_targets_mean": 3225.1, + "valid_targets_min": 2163 + }, + { + "epoch": 4.372244897959184, + "grad_norm": 0.4899944797037151, + "learning_rate": 1.4874723761990751e-05, + "loss": 0.1946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07114744931459427, + "step": 2680, + "valid_targets_mean": 4258.1, + "valid_targets_min": 704 + }, + { + "epoch": 4.380408163265306, + "grad_norm": 0.49284944300678446, + "learning_rate": 1.4796116172704032e-05, + "loss": 0.1991, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10904914885759354, + "step": 2685, + "valid_targets_mean": 6167.9, + "valid_targets_min": 838 + }, + { + "epoch": 4.388571428571429, + "grad_norm": 0.5432323972808517, + "learning_rate": 1.4717594715935498e-05, + "loss": 0.197, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09540468454360962, + "step": 2690, + "valid_targets_mean": 3727.9, + "valid_targets_min": 3180 + }, + { + "epoch": 4.396734693877551, + "grad_norm": 0.4854316131867214, + "learning_rate": 1.4639160691339593e-05, + "loss": 0.1954, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09044314920902252, + "step": 2695, + "valid_targets_mean": 4713.0, + "valid_targets_min": 2170 + }, + { + "epoch": 4.404897959183674, + "grad_norm": 0.41399018384818026, + "learning_rate": 1.456081539712365e-05, + "loss": 0.1885, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07750784605741501, + "step": 2700, + "valid_targets_mean": 6339.5, + "valid_targets_min": 988 + }, + { + "epoch": 4.4130612244897955, + "grad_norm": 0.45239236721280934, + "learning_rate": 1.4482560130026342e-05, + "loss": 0.187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09557928144931793, + "step": 2705, + "valid_targets_mean": 5853.2, + "valid_targets_min": 2915 + }, + { + "epoch": 4.421224489795918, + "grad_norm": 0.5584019817797293, + "learning_rate": 1.4404396185296268e-05, + "loss": 0.1916, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08847010135650635, + "step": 2710, + "valid_targets_mean": 4407.9, + "valid_targets_min": 1298 + }, + { + "epoch": 4.429387755102041, + "grad_norm": 0.48880926965662114, + "learning_rate": 1.4326324856670494e-05, + "loss": 0.1904, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07466118037700653, + "step": 2715, + "valid_targets_mean": 4065.1, + "valid_targets_min": 2074 + }, + { + "epoch": 4.437551020408163, + "grad_norm": 0.4277135480825074, + "learning_rate": 1.4248347436353147e-05, + "loss": 0.1908, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0795067697763443, + "step": 2720, + "valid_targets_mean": 5436.0, + "valid_targets_min": 1588 + }, + { + "epoch": 4.445714285714286, + "grad_norm": 0.5413528832361828, + "learning_rate": 1.4170465214994011e-05, + "loss": 0.1885, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10669746994972229, + "step": 2725, + "valid_targets_mean": 4772.5, + "valid_targets_min": 1882 + }, + { + "epoch": 4.453877551020408, + "grad_norm": 0.4731571122632174, + "learning_rate": 1.409267948166718e-05, + "loss": 0.192, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08785238116979599, + "step": 2730, + "valid_targets_mean": 5070.4, + "valid_targets_min": 2187 + }, + { + "epoch": 4.462040816326531, + "grad_norm": 0.5039704029344154, + "learning_rate": 1.4014991523849726e-05, + "loss": 0.1984, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10380153357982635, + "step": 2735, + "valid_targets_mean": 4711.8, + "valid_targets_min": 1346 + }, + { + "epoch": 4.470204081632653, + "grad_norm": 0.542316983338691, + "learning_rate": 1.3937402627400384e-05, + "loss": 0.1971, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13545742630958557, + "step": 2740, + "valid_targets_mean": 6321.9, + "valid_targets_min": 1489 + }, + { + "epoch": 4.478367346938776, + "grad_norm": 0.47877743495837094, + "learning_rate": 1.3859914076538245e-05, + "loss": 0.1944, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09974990785121918, + "step": 2745, + "valid_targets_mean": 5603.5, + "valid_targets_min": 2886 + }, + { + "epoch": 4.486530612244898, + "grad_norm": 0.5437938225021358, + "learning_rate": 1.3782527153821554e-05, + "loss": 0.1797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1074422299861908, + "step": 2750, + "valid_targets_mean": 4773.6, + "valid_targets_min": 2787 + }, + { + "epoch": 4.49469387755102, + "grad_norm": 0.5296213977075995, + "learning_rate": 1.370524314012642e-05, + "loss": 0.1829, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1249004527926445, + "step": 2755, + "valid_targets_mean": 5794.8, + "valid_targets_min": 2462 + }, + { + "epoch": 4.502857142857143, + "grad_norm": 0.4448019181831348, + "learning_rate": 1.3628063314625655e-05, + "loss": 0.1785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07824553549289703, + "step": 2760, + "valid_targets_mean": 5313.8, + "valid_targets_min": 1962 + }, + { + "epoch": 4.511020408163265, + "grad_norm": 0.5684669670820979, + "learning_rate": 1.355098895476759e-05, + "loss": 0.1913, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10887186229228973, + "step": 2765, + "valid_targets_mean": 4395.8, + "valid_targets_min": 2515 + }, + { + "epoch": 4.519183673469388, + "grad_norm": 0.37621169472643107, + "learning_rate": 1.347402133625495e-05, + "loss": 0.1762, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07432250678539276, + "step": 2770, + "valid_targets_mean": 5932.8, + "valid_targets_min": 1810 + }, + { + "epoch": 4.52734693877551, + "grad_norm": 0.5608359554833122, + "learning_rate": 1.3397161733023698e-05, + "loss": 0.1892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0860264003276825, + "step": 2775, + "valid_targets_mean": 3904.0, + "valid_targets_min": 1149 + }, + { + "epoch": 4.535510204081633, + "grad_norm": 0.4795668691745758, + "learning_rate": 1.3320411417221974e-05, + "loss": 0.1967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10846267640590668, + "step": 2780, + "valid_targets_mean": 5842.1, + "valid_targets_min": 2204 + }, + { + "epoch": 4.543673469387755, + "grad_norm": 0.375899081661828, + "learning_rate": 1.324377165918906e-05, + "loss": 0.1875, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.055310845375061035, + "step": 2785, + "valid_targets_mean": 5695.2, + "valid_targets_min": 1003 + }, + { + "epoch": 4.551836734693877, + "grad_norm": 0.48026420303225553, + "learning_rate": 1.3167243727434307e-05, + "loss": 0.1974, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10313354432582855, + "step": 2790, + "valid_targets_mean": 5187.1, + "valid_targets_min": 3179 + }, + { + "epoch": 4.5600000000000005, + "grad_norm": 0.42389844482180566, + "learning_rate": 1.3090828888616183e-05, + "loss": 0.1794, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0987774059176445, + "step": 2795, + "valid_targets_mean": 6063.2, + "valid_targets_min": 1414 + }, + { + "epoch": 4.568163265306122, + "grad_norm": 0.492898294974917, + "learning_rate": 1.301452840752127e-05, + "loss": 0.1902, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09452690184116364, + "step": 2800, + "valid_targets_mean": 5125.5, + "valid_targets_min": 1920 + }, + { + "epoch": 4.576326530612245, + "grad_norm": 0.5922819701493808, + "learning_rate": 1.293834354704337e-05, + "loss": 0.1879, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1072399765253067, + "step": 2805, + "valid_targets_mean": 4184.5, + "valid_targets_min": 2043 + }, + { + "epoch": 4.584489795918367, + "grad_norm": 0.5009712991870972, + "learning_rate": 1.2862275568162566e-05, + "loss": 0.189, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12208345532417297, + "step": 2810, + "valid_targets_mean": 5785.0, + "valid_targets_min": 1694 + }, + { + "epoch": 4.59265306122449, + "grad_norm": 0.5019101438266571, + "learning_rate": 1.2786325729924362e-05, + "loss": 0.1857, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11706788837909698, + "step": 2815, + "valid_targets_mean": 6469.1, + "valid_targets_min": 1605 + }, + { + "epoch": 4.600816326530612, + "grad_norm": 0.5425198767387878, + "learning_rate": 1.2710495289418869e-05, + "loss": 0.1898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08403375744819641, + "step": 2820, + "valid_targets_mean": 3734.6, + "valid_targets_min": 1120 + }, + { + "epoch": 4.608979591836735, + "grad_norm": 0.47797549079237955, + "learning_rate": 1.2634785501759957e-05, + "loss": 0.1881, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08239451050758362, + "step": 2825, + "valid_targets_mean": 4446.0, + "valid_targets_min": 2288 + }, + { + "epoch": 4.617142857142857, + "grad_norm": 0.5237661865631653, + "learning_rate": 1.2559197620064501e-05, + "loss": 0.1804, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11617680639028549, + "step": 2830, + "valid_targets_mean": 5307.2, + "valid_targets_min": 2390 + }, + { + "epoch": 4.625306122448979, + "grad_norm": 0.49168671805845504, + "learning_rate": 1.2483732895431637e-05, + "loss": 0.2024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1150754764676094, + "step": 2835, + "valid_targets_mean": 6200.5, + "valid_targets_min": 2398 + }, + { + "epoch": 4.6334693877551025, + "grad_norm": 0.48627176014261897, + "learning_rate": 1.2408392576922075e-05, + "loss": 0.1959, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08021821081638336, + "step": 2840, + "valid_targets_mean": 5957.8, + "valid_targets_min": 2327 + }, + { + "epoch": 4.641632653061224, + "grad_norm": 0.44639892122473207, + "learning_rate": 1.2333177911537396e-05, + "loss": 0.1953, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08871643245220184, + "step": 2845, + "valid_targets_mean": 5963.0, + "valid_targets_min": 2091 + }, + { + "epoch": 4.649795918367347, + "grad_norm": 0.5226387363165819, + "learning_rate": 1.2258090144199409e-05, + "loss": 0.1869, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08227618038654327, + "step": 2850, + "valid_targets_mean": 3925.1, + "valid_targets_min": 2181 + }, + { + "epoch": 4.657959183673469, + "grad_norm": 0.4332462492586316, + "learning_rate": 1.2183130517729586e-05, + "loss": 0.1831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.081905797123909, + "step": 2855, + "valid_targets_mean": 5561.9, + "valid_targets_min": 2039 + }, + { + "epoch": 4.666122448979592, + "grad_norm": 0.5683336516651115, + "learning_rate": 1.2108300272828446e-05, + "loss": 0.1939, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10651646554470062, + "step": 2860, + "valid_targets_mean": 5148.9, + "valid_targets_min": 2376 + }, + { + "epoch": 4.674285714285714, + "grad_norm": 0.44839763343296324, + "learning_rate": 1.2033600648055038e-05, + "loss": 0.1966, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08386656641960144, + "step": 2865, + "valid_targets_mean": 5665.2, + "valid_targets_min": 228 + }, + { + "epoch": 4.682448979591837, + "grad_norm": 0.5024860369005062, + "learning_rate": 1.1959032879806442e-05, + "loss": 0.1918, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09913516044616699, + "step": 2870, + "valid_targets_mean": 4861.0, + "valid_targets_min": 2413 + }, + { + "epoch": 4.6906122448979595, + "grad_norm": 0.49564273386489915, + "learning_rate": 1.1884598202297327e-05, + "loss": 0.1908, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10049331933259964, + "step": 2875, + "valid_targets_mean": 5382.5, + "valid_targets_min": 1733 + }, + { + "epoch": 4.698775510204082, + "grad_norm": 0.46680473946107015, + "learning_rate": 1.1810297847539464e-05, + "loss": 0.2062, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08834764361381531, + "step": 2880, + "valid_targets_mean": 4403.9, + "valid_targets_min": 1545 + }, + { + "epoch": 4.706938775510205, + "grad_norm": 0.5099284522866282, + "learning_rate": 1.1736133045321387e-05, + "loss": 0.1886, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09604378044605255, + "step": 2885, + "valid_targets_mean": 5398.1, + "valid_targets_min": 2002 + }, + { + "epoch": 4.715102040816326, + "grad_norm": 0.44693345599274104, + "learning_rate": 1.1662105023188027e-05, + "loss": 0.1797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.04823494702577591, + "step": 2890, + "valid_targets_mean": 3021.4, + "valid_targets_min": 350 + }, + { + "epoch": 4.723265306122449, + "grad_norm": 0.5114382893021212, + "learning_rate": 1.1588215006420374e-05, + "loss": 0.1928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08465202152729034, + "step": 2895, + "valid_targets_mean": 3635.6, + "valid_targets_min": 574 + }, + { + "epoch": 4.731428571428571, + "grad_norm": 0.5061445498959004, + "learning_rate": 1.1514464218015218e-05, + "loss": 0.1862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08960931748151779, + "step": 2900, + "valid_targets_mean": 5353.8, + "valid_targets_min": 2572 + }, + { + "epoch": 4.739591836734694, + "grad_norm": 1.103649871943999, + "learning_rate": 1.1440853878664902e-05, + "loss": 0.1928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11720076203346252, + "step": 2905, + "valid_targets_mean": 5939.9, + "valid_targets_min": 1889 + }, + { + "epoch": 4.747755102040816, + "grad_norm": 0.4954748711036458, + "learning_rate": 1.1367385206737106e-05, + "loss": 0.1846, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07366099953651428, + "step": 2910, + "valid_targets_mean": 3961.4, + "valid_targets_min": 2342 + }, + { + "epoch": 4.755918367346939, + "grad_norm": 0.49215147868377257, + "learning_rate": 1.129405941825471e-05, + "loss": 0.1934, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08957966417074203, + "step": 2915, + "valid_targets_mean": 4154.5, + "valid_targets_min": 1411 + }, + { + "epoch": 4.7640816326530615, + "grad_norm": 0.5405766702143414, + "learning_rate": 1.1220877726875603e-05, + "loss": 0.1896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10561898350715637, + "step": 2920, + "valid_targets_mean": 4969.0, + "valid_targets_min": 2225 + }, + { + "epoch": 4.772244897959184, + "grad_norm": 0.460865615051362, + "learning_rate": 1.1147841343872699e-05, + "loss": 0.1815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07561326026916504, + "step": 2925, + "valid_targets_mean": 4488.9, + "valid_targets_min": 1389 + }, + { + "epoch": 4.780408163265307, + "grad_norm": 0.5238153626279427, + "learning_rate": 1.1074951478113774e-05, + "loss": 0.2038, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09728378057479858, + "step": 2930, + "valid_targets_mean": 5040.6, + "valid_targets_min": 2158 + }, + { + "epoch": 4.788571428571428, + "grad_norm": 0.44238272143257545, + "learning_rate": 1.1002209336041542e-05, + "loss": 0.184, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08578670024871826, + "step": 2935, + "valid_targets_mean": 4627.5, + "valid_targets_min": 1124 + }, + { + "epoch": 4.796734693877551, + "grad_norm": 0.43479390970924037, + "learning_rate": 1.092961612165365e-05, + "loss": 0.1815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06755372136831284, + "step": 2940, + "valid_targets_mean": 6472.2, + "valid_targets_min": 1480 + }, + { + "epoch": 4.804897959183673, + "grad_norm": 0.5853935194690773, + "learning_rate": 1.085717303648275e-05, + "loss": 0.1966, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11776626110076904, + "step": 2945, + "valid_targets_mean": 4299.6, + "valid_targets_min": 148 + }, + { + "epoch": 4.813061224489796, + "grad_norm": 0.5131471997499106, + "learning_rate": 1.0784881279576635e-05, + "loss": 0.1917, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.096668541431427, + "step": 2950, + "valid_targets_mean": 4616.1, + "valid_targets_min": 2963 + }, + { + "epoch": 4.8212244897959184, + "grad_norm": 0.5679561757392165, + "learning_rate": 1.0712742047478346e-05, + "loss": 0.1937, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07576416432857513, + "step": 2955, + "valid_targets_mean": 3169.6, + "valid_targets_min": 1363 + }, + { + "epoch": 4.829387755102041, + "grad_norm": 0.5331028375004665, + "learning_rate": 1.0640756534206412e-05, + "loss": 0.1837, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09721846878528595, + "step": 2960, + "valid_targets_mean": 3602.8, + "valid_targets_min": 987 + }, + { + "epoch": 4.8375510204081635, + "grad_norm": 0.5645277289402424, + "learning_rate": 1.0568925931235095e-05, + "loss": 0.1898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11298727989196777, + "step": 2965, + "valid_targets_mean": 4520.5, + "valid_targets_min": 2114 + }, + { + "epoch": 4.845714285714286, + "grad_norm": 0.4145507098470101, + "learning_rate": 1.0497251427474602e-05, + "loss": 0.1896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09271155297756195, + "step": 2970, + "valid_targets_mean": 6519.9, + "valid_targets_min": 2358 + }, + { + "epoch": 4.853877551020409, + "grad_norm": 0.5313111617754027, + "learning_rate": 1.0425734209251486e-05, + "loss": 0.2019, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09830297529697418, + "step": 2975, + "valid_targets_mean": 4779.5, + "valid_targets_min": 522 + }, + { + "epoch": 4.86204081632653, + "grad_norm": 0.5361791729354184, + "learning_rate": 1.035437546028896e-05, + "loss": 0.1796, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10205239057540894, + "step": 2980, + "valid_targets_mean": 4452.4, + "valid_targets_min": 2395 + }, + { + "epoch": 4.870204081632653, + "grad_norm": 0.5333208117311148, + "learning_rate": 1.028317636168732e-05, + "loss": 0.1872, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10787051916122437, + "step": 2985, + "valid_targets_mean": 5309.5, + "valid_targets_min": 1707 + }, + { + "epoch": 4.878367346938775, + "grad_norm": 0.4620967725507521, + "learning_rate": 1.0212138091904409e-05, + "loss": 0.1945, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10534008592367172, + "step": 2990, + "valid_targets_mean": 5099.0, + "valid_targets_min": 1974 + }, + { + "epoch": 4.886530612244898, + "grad_norm": 0.6802090469740967, + "learning_rate": 1.0141261826736067e-05, + "loss": 0.1889, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07617118954658508, + "step": 2995, + "valid_targets_mean": 2980.5, + "valid_targets_min": 557 + }, + { + "epoch": 4.8946938775510205, + "grad_norm": 0.5014624004024499, + "learning_rate": 1.0070548739296746e-05, + "loss": 0.1766, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06977512687444687, + "step": 3000, + "valid_targets_mean": 3862.0, + "valid_targets_min": 1983 + }, + { + "epoch": 4.902857142857143, + "grad_norm": 0.5733710902901722, + "learning_rate": 1.0000000000000006e-05, + "loss": 0.1967, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0930173248052597, + "step": 3005, + "valid_targets_mean": 4976.5, + "valid_targets_min": 2294 + }, + { + "epoch": 4.911020408163266, + "grad_norm": 0.5198355148927837, + "learning_rate": 9.929616776539199e-06, + "loss": 0.1966, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09638778120279312, + "step": 3010, + "valid_targets_mean": 5231.0, + "valid_targets_min": 1368 + }, + { + "epoch": 4.919183673469388, + "grad_norm": 0.49759517932274444, + "learning_rate": 9.859400233868159e-06, + "loss": 0.1829, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11821071803569794, + "step": 3015, + "valid_targets_mean": 5549.5, + "valid_targets_min": 3325 + }, + { + "epoch": 4.927346938775511, + "grad_norm": 0.5400270338184223, + "learning_rate": 9.789351534181837e-06, + "loss": 0.187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09970055520534515, + "step": 3020, + "valid_targets_mean": 4339.2, + "valid_targets_min": 2592 + }, + { + "epoch": 4.935510204081632, + "grad_norm": 0.518057347117569, + "learning_rate": 9.719471836897156e-06, + "loss": 0.1794, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10430602729320526, + "step": 3025, + "valid_targets_mean": 4877.8, + "valid_targets_min": 779 + }, + { + "epoch": 4.943673469387755, + "grad_norm": 0.5336142059133698, + "learning_rate": 9.649762298633747e-06, + "loss": 0.2045, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1243668645620346, + "step": 3030, + "valid_targets_mean": 5056.6, + "valid_targets_min": 2234 + }, + { + "epoch": 4.951836734693877, + "grad_norm": 0.5272384020096476, + "learning_rate": 9.58022407319488e-06, + "loss": 0.192, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09622962027788162, + "step": 3035, + "valid_targets_mean": 4204.1, + "valid_targets_min": 172 + }, + { + "epoch": 4.96, + "grad_norm": 0.4502669212181341, + "learning_rate": 9.510858311548306e-06, + "loss": 0.1919, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09366258978843689, + "step": 3040, + "valid_targets_mean": 6401.0, + "valid_targets_min": 1612 + }, + { + "epoch": 4.9681632653061225, + "grad_norm": 0.5272654404132141, + "learning_rate": 9.441666161807206e-06, + "loss": 0.1943, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10728336870670319, + "step": 3045, + "valid_targets_mean": 4781.5, + "valid_targets_min": 786 + }, + { + "epoch": 4.976326530612245, + "grad_norm": 0.40424101442932664, + "learning_rate": 9.372648769211259e-06, + "loss": 0.1918, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07456552982330322, + "step": 3050, + "valid_targets_mean": 4632.0, + "valid_targets_min": 2145 + }, + { + "epoch": 4.984489795918368, + "grad_norm": 0.5600282228675467, + "learning_rate": 9.30380727610759e-06, + "loss": 0.1999, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10491231083869934, + "step": 3055, + "valid_targets_mean": 5399.2, + "valid_targets_min": 2148 + }, + { + "epoch": 4.99265306122449, + "grad_norm": 0.3787674156466776, + "learning_rate": 9.235142821931928e-06, + "loss": 0.1861, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08586746454238892, + "step": 3060, + "valid_targets_mean": 8370.4, + "valid_targets_min": 4348 + }, + { + "epoch": 5.0, + "grad_norm": 0.7067710949929604, + "learning_rate": 9.166656543189733e-06, + "loss": 0.184, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18785783648490906, + "step": 3065, + "valid_targets_mean": 5040.1, + "valid_targets_min": 2867 + }, + { + "epoch": 5.0081632653061225, + "grad_norm": 0.4788393552290572, + "learning_rate": 9.098349573437368e-06, + "loss": 0.1672, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09013266861438751, + "step": 3070, + "valid_targets_mean": 4971.5, + "valid_targets_min": 2198 + }, + { + "epoch": 5.016326530612245, + "grad_norm": 0.5452058782222013, + "learning_rate": 9.03022304326336e-06, + "loss": 0.1836, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12489248812198639, + "step": 3075, + "valid_targets_mean": 6247.5, + "valid_targets_min": 3417 + }, + { + "epoch": 5.024489795918368, + "grad_norm": 0.4958317510599947, + "learning_rate": 8.962278080269649e-06, + "loss": 0.1868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09371688961982727, + "step": 3080, + "valid_targets_mean": 5708.5, + "valid_targets_min": 2568 + }, + { + "epoch": 5.03265306122449, + "grad_norm": 0.4874809774022175, + "learning_rate": 8.89451580905299e-06, + "loss": 0.1822, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09922803193330765, + "step": 3085, + "valid_targets_mean": 5210.1, + "valid_targets_min": 2680 + }, + { + "epoch": 5.040816326530612, + "grad_norm": 0.43812239820329635, + "learning_rate": 8.82693735118628e-06, + "loss": 0.1642, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0819573774933815, + "step": 3090, + "valid_targets_mean": 4633.8, + "valid_targets_min": 1840 + }, + { + "epoch": 5.048979591836734, + "grad_norm": 0.5062954532674303, + "learning_rate": 8.759543825200001e-06, + "loss": 0.1879, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12129976600408554, + "step": 3095, + "valid_targets_mean": 5984.4, + "valid_targets_min": 2152 + }, + { + "epoch": 5.057142857142857, + "grad_norm": 0.4048917822614195, + "learning_rate": 8.692336346563735e-06, + "loss": 0.1739, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.05262184143066406, + "step": 3100, + "valid_targets_mean": 4425.0, + "valid_targets_min": 1267 + }, + { + "epoch": 5.0653061224489795, + "grad_norm": 0.5806920910174197, + "learning_rate": 8.625316027667685e-06, + "loss": 0.1794, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09791389107704163, + "step": 3105, + "valid_targets_mean": 4411.6, + "valid_targets_min": 1893 + }, + { + "epoch": 5.073469387755102, + "grad_norm": 0.5005518459398053, + "learning_rate": 8.558483977804265e-06, + "loss": 0.1792, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07806237041950226, + "step": 3110, + "valid_targets_mean": 4310.2, + "valid_targets_min": 1229 + }, + { + "epoch": 5.081632653061225, + "grad_norm": 0.4362967329293203, + "learning_rate": 8.491841303149728e-06, + "loss": 0.1831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0794868916273117, + "step": 3115, + "valid_targets_mean": 6788.8, + "valid_targets_min": 1302 + }, + { + "epoch": 5.089795918367347, + "grad_norm": 0.5707602601433491, + "learning_rate": 8.425389106745881e-06, + "loss": 0.1739, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09465576708316803, + "step": 3120, + "valid_targets_mean": 4276.5, + "valid_targets_min": 1630 + }, + { + "epoch": 5.09795918367347, + "grad_norm": 0.5475463253467584, + "learning_rate": 8.35912848848181e-06, + "loss": 0.1748, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07283581793308258, + "step": 3125, + "valid_targets_mean": 3742.4, + "valid_targets_min": 2123 + }, + { + "epoch": 5.106122448979592, + "grad_norm": 0.5650276786893733, + "learning_rate": 8.293060545075667e-06, + "loss": 0.1844, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07957779616117477, + "step": 3130, + "valid_targets_mean": 3538.2, + "valid_targets_min": 209 + }, + { + "epoch": 5.114285714285714, + "grad_norm": 0.4742456395275569, + "learning_rate": 8.227186370056546e-06, + "loss": 0.1803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08860338479280472, + "step": 3135, + "valid_targets_mean": 6125.4, + "valid_targets_min": 1943 + }, + { + "epoch": 5.122448979591836, + "grad_norm": 0.4533762535853237, + "learning_rate": 8.16150705374636e-06, + "loss": 0.1826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06944446265697479, + "step": 3140, + "valid_targets_mean": 5134.6, + "valid_targets_min": 1513 + }, + { + "epoch": 5.130612244897959, + "grad_norm": 0.5879056841548346, + "learning_rate": 8.096023683241805e-06, + "loss": 0.1794, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08228239417076111, + "step": 3145, + "valid_targets_mean": 3871.9, + "valid_targets_min": 2275 + }, + { + "epoch": 5.1387755102040815, + "grad_norm": 0.45175621395683185, + "learning_rate": 8.030737342396362e-06, + "loss": 0.1816, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08341191709041595, + "step": 3150, + "valid_targets_mean": 6288.5, + "valid_targets_min": 1892 + }, + { + "epoch": 5.146938775510204, + "grad_norm": 0.5069533232855673, + "learning_rate": 7.965649111802358e-06, + "loss": 0.178, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08151441812515259, + "step": 3155, + "valid_targets_mean": 4190.8, + "valid_targets_min": 2441 + }, + { + "epoch": 5.155102040816327, + "grad_norm": 0.38454817482765574, + "learning_rate": 7.900760068773083e-06, + "loss": 0.1818, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08950209617614746, + "step": 3160, + "valid_targets_mean": 6927.8, + "valid_targets_min": 3518 + }, + { + "epoch": 5.163265306122449, + "grad_norm": 0.5503677993367798, + "learning_rate": 7.836071287324965e-06, + "loss": 0.1878, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08563916385173798, + "step": 3165, + "valid_targets_mean": 3916.5, + "valid_targets_min": 2768 + }, + { + "epoch": 5.171428571428572, + "grad_norm": 0.4633224599111477, + "learning_rate": 7.771583838159756e-06, + "loss": 0.171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0768350288271904, + "step": 3170, + "valid_targets_mean": 5713.6, + "valid_targets_min": 2432 + }, + { + "epoch": 5.179591836734694, + "grad_norm": 0.547107579940367, + "learning_rate": 7.70729878864689e-06, + "loss": 0.1774, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08860382437705994, + "step": 3175, + "valid_targets_mean": 5152.6, + "valid_targets_min": 2714 + }, + { + "epoch": 5.187755102040816, + "grad_norm": 0.5499484727345023, + "learning_rate": 7.643217202805715e-06, + "loss": 0.175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10342984646558762, + "step": 3180, + "valid_targets_mean": 5412.6, + "valid_targets_min": 2343 + }, + { + "epoch": 5.1959183673469385, + "grad_norm": 0.4611447181076451, + "learning_rate": 7.579340141287965e-06, + "loss": 0.1806, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0803723856806755, + "step": 3185, + "valid_targets_mean": 5043.5, + "valid_targets_min": 2553 + }, + { + "epoch": 5.204081632653061, + "grad_norm": 0.4987990372400629, + "learning_rate": 7.515668661360162e-06, + "loss": 0.1885, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08560259640216827, + "step": 3190, + "valid_targets_mean": 4151.0, + "valid_targets_min": 818 + }, + { + "epoch": 5.2122448979591836, + "grad_norm": 0.6385599519928179, + "learning_rate": 7.452203816886132e-06, + "loss": 0.182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09238561242818832, + "step": 3195, + "valid_targets_mean": 2983.1, + "valid_targets_min": 1985 + }, + { + "epoch": 5.220408163265306, + "grad_norm": 0.49972296239749786, + "learning_rate": 7.388946658309557e-06, + "loss": 0.1725, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07128877937793732, + "step": 3200, + "valid_targets_mean": 4395.1, + "valid_targets_min": 838 + }, + { + "epoch": 5.228571428571429, + "grad_norm": 0.4251048049138713, + "learning_rate": 7.3258982326365656e-06, + "loss": 0.1797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08632293343544006, + "step": 3205, + "valid_targets_mean": 7603.4, + "valid_targets_min": 3155 + }, + { + "epoch": 5.236734693877551, + "grad_norm": 0.5029326144715, + "learning_rate": 7.263059583418473e-06, + "loss": 0.1807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07719817012548447, + "step": 3210, + "valid_targets_mean": 4976.4, + "valid_targets_min": 3287 + }, + { + "epoch": 5.244897959183674, + "grad_norm": 0.61537978028631, + "learning_rate": 7.200431750734414e-06, + "loss": 0.1887, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07119373232126236, + "step": 3215, + "valid_targets_mean": 3600.5, + "valid_targets_min": 845 + }, + { + "epoch": 5.253061224489796, + "grad_norm": 0.6170700381181095, + "learning_rate": 7.138015771174196e-06, + "loss": 0.1709, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08307771384716034, + "step": 3220, + "valid_targets_mean": 4144.8, + "valid_targets_min": 2515 + }, + { + "epoch": 5.261224489795918, + "grad_norm": 0.56626652835952, + "learning_rate": 7.075812677821145e-06, + "loss": 0.1807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10066160559654236, + "step": 3225, + "valid_targets_mean": 3968.5, + "valid_targets_min": 927 + }, + { + "epoch": 5.2693877551020405, + "grad_norm": 0.48288948191263076, + "learning_rate": 7.013823500234938e-06, + "loss": 0.1757, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08535900712013245, + "step": 3230, + "valid_targets_mean": 5515.0, + "valid_targets_min": 2714 + }, + { + "epoch": 5.277551020408163, + "grad_norm": 0.4865175316764564, + "learning_rate": 6.952049264434644e-06, + "loss": 0.1711, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0826706737279892, + "step": 3235, + "valid_targets_mean": 5179.6, + "valid_targets_min": 1055 + }, + { + "epoch": 5.285714285714286, + "grad_norm": 0.5086830041890779, + "learning_rate": 6.890490992881671e-06, + "loss": 0.1656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08522078394889832, + "step": 3240, + "valid_targets_mean": 7090.1, + "valid_targets_min": 3606 + }, + { + "epoch": 5.293877551020408, + "grad_norm": 0.5137504324453931, + "learning_rate": 6.829149704462918e-06, + "loss": 0.1784, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08339082449674606, + "step": 3245, + "valid_targets_mean": 5650.1, + "valid_targets_min": 2900 + }, + { + "epoch": 5.302040816326531, + "grad_norm": 0.5367587892690145, + "learning_rate": 6.768026414473845e-06, + "loss": 0.1806, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09846776723861694, + "step": 3250, + "valid_targets_mean": 5025.4, + "valid_targets_min": 2477 + }, + { + "epoch": 5.310204081632653, + "grad_norm": 0.4109130891833122, + "learning_rate": 6.707122134601676e-06, + "loss": 0.1732, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09323020279407501, + "step": 3255, + "valid_targets_mean": 6455.5, + "valid_targets_min": 2791 + }, + { + "epoch": 5.318367346938776, + "grad_norm": 0.4363547754110573, + "learning_rate": 6.6464378729087196e-06, + "loss": 0.1698, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0628901794552803, + "step": 3260, + "valid_targets_mean": 3648.1, + "valid_targets_min": 1569 + }, + { + "epoch": 5.326530612244898, + "grad_norm": 0.6083738609525996, + "learning_rate": 6.585974633815584e-06, + "loss": 0.1833, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12459546327590942, + "step": 3265, + "valid_targets_mean": 5369.5, + "valid_targets_min": 2892 + }, + { + "epoch": 5.33469387755102, + "grad_norm": 0.5271929880336159, + "learning_rate": 6.525733418084635e-06, + "loss": 0.1849, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09798870235681534, + "step": 3270, + "valid_targets_mean": 5105.4, + "valid_targets_min": 1418 + }, + { + "epoch": 5.3428571428571425, + "grad_norm": 0.5526213337564067, + "learning_rate": 6.465715222803386e-06, + "loss": 0.178, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08597704768180847, + "step": 3275, + "valid_targets_mean": 4135.8, + "valid_targets_min": 2845 + }, + { + "epoch": 5.351020408163265, + "grad_norm": 0.5433902287654617, + "learning_rate": 6.4059210413680175e-06, + "loss": 0.1827, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09743403643369675, + "step": 3280, + "valid_targets_mean": 4350.8, + "valid_targets_min": 1555 + }, + { + "epoch": 5.359183673469388, + "grad_norm": 0.4767498259019695, + "learning_rate": 6.346351863466926e-06, + "loss": 0.1723, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07725220918655396, + "step": 3285, + "valid_targets_mean": 4635.6, + "valid_targets_min": 3068 + }, + { + "epoch": 5.36734693877551, + "grad_norm": 0.47409096876292317, + "learning_rate": 6.2870086750643215e-06, + "loss": 0.1831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07922723889350891, + "step": 3290, + "valid_targets_mean": 4892.5, + "valid_targets_min": 1427 + }, + { + "epoch": 5.375510204081633, + "grad_norm": 0.5084499811899263, + "learning_rate": 6.2278924583839665e-06, + "loss": 0.1777, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06283068656921387, + "step": 3295, + "valid_targets_mean": 3258.6, + "valid_targets_min": 1718 + }, + { + "epoch": 5.383673469387755, + "grad_norm": 0.6078381929182485, + "learning_rate": 6.16900419189286e-06, + "loss": 0.1859, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10925310105085373, + "step": 3300, + "valid_targets_mean": 4435.9, + "valid_targets_min": 1413 + }, + { + "epoch": 5.391836734693878, + "grad_norm": 0.6378926459368445, + "learning_rate": 6.110344850285059e-06, + "loss": 0.1834, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10677415132522583, + "step": 3305, + "valid_targets_mean": 3772.9, + "valid_targets_min": 1762 + }, + { + "epoch": 5.4, + "grad_norm": 0.4589984510633392, + "learning_rate": 6.0519154044655674e-06, + "loss": 0.1813, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09519022703170776, + "step": 3310, + "valid_targets_mean": 6158.0, + "valid_targets_min": 1700 + }, + { + "epoch": 5.408163265306122, + "grad_norm": 0.6567603574653257, + "learning_rate": 5.9937168215342345e-06, + "loss": 0.182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10160709172487259, + "step": 3315, + "valid_targets_mean": 4658.2, + "valid_targets_min": 655 + }, + { + "epoch": 5.416326530612245, + "grad_norm": 0.5338399798016586, + "learning_rate": 5.9357500647697786e-06, + "loss": 0.1711, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11896622180938721, + "step": 3320, + "valid_targets_mean": 5882.9, + "valid_targets_min": 3456 + }, + { + "epoch": 5.424489795918367, + "grad_norm": 0.5616038365610015, + "learning_rate": 5.878016093613816e-06, + "loss": 0.1936, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07236294448375702, + "step": 3325, + "valid_targets_mean": 3982.4, + "valid_targets_min": 1704 + }, + { + "epoch": 5.43265306122449, + "grad_norm": 0.515377184603261, + "learning_rate": 5.820515863654999e-06, + "loss": 0.1745, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07034678012132645, + "step": 3330, + "valid_targets_mean": 3606.8, + "valid_targets_min": 1744 + }, + { + "epoch": 5.440816326530612, + "grad_norm": 0.5377914529022676, + "learning_rate": 5.7632503266131925e-06, + "loss": 0.1734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09335826337337494, + "step": 3335, + "valid_targets_mean": 4594.0, + "valid_targets_min": 2669 + }, + { + "epoch": 5.448979591836735, + "grad_norm": 0.5255517607613142, + "learning_rate": 5.70622043032371e-06, + "loss": 0.1784, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0681927353143692, + "step": 3340, + "valid_targets_mean": 5194.6, + "valid_targets_min": 1358 + }, + { + "epoch": 5.457142857142857, + "grad_norm": 0.5117051396865993, + "learning_rate": 5.649427118721655e-06, + "loss": 0.1809, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08759818226099014, + "step": 3345, + "valid_targets_mean": 4859.9, + "valid_targets_min": 2107 + }, + { + "epoch": 5.46530612244898, + "grad_norm": 0.5608179829600851, + "learning_rate": 5.592871331826271e-06, + "loss": 0.1803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.092364102602005, + "step": 3350, + "valid_targets_mean": 4509.1, + "valid_targets_min": 182 + }, + { + "epoch": 5.473469387755102, + "grad_norm": 0.523015667386693, + "learning_rate": 5.536554005725392e-06, + "loss": 0.1687, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07398107647895813, + "step": 3355, + "valid_targets_mean": 3645.1, + "valid_targets_min": 802 + }, + { + "epoch": 5.481632653061224, + "grad_norm": 0.5440603818142956, + "learning_rate": 5.480476072559955e-06, + "loss": 0.1772, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10536229610443115, + "step": 3360, + "valid_targets_mean": 5585.2, + "valid_targets_min": 3009 + }, + { + "epoch": 5.489795918367347, + "grad_norm": 0.5106202139769874, + "learning_rate": 5.42463846050854e-06, + "loss": 0.1792, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08815483748912811, + "step": 3365, + "valid_targets_mean": 4938.4, + "valid_targets_min": 436 + }, + { + "epoch": 5.497959183673469, + "grad_norm": 0.5792561830469595, + "learning_rate": 5.369042093772068e-06, + "loss": 0.1864, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08561515808105469, + "step": 3370, + "valid_targets_mean": 3781.2, + "valid_targets_min": 2797 + }, + { + "epoch": 5.506122448979592, + "grad_norm": 0.5716983079904695, + "learning_rate": 5.313687892558455e-06, + "loss": 0.1797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07345669716596603, + "step": 3375, + "valid_targets_mean": 3362.6, + "valid_targets_min": 385 + }, + { + "epoch": 5.514285714285714, + "grad_norm": 0.5214998517548255, + "learning_rate": 5.258576773067366e-06, + "loss": 0.1731, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0814686045050621, + "step": 3380, + "valid_targets_mean": 4806.2, + "valid_targets_min": 2774 + }, + { + "epoch": 5.522448979591837, + "grad_norm": 0.5036864461547301, + "learning_rate": 5.203709647475128e-06, + "loss": 0.177, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09993688762187958, + "step": 3385, + "valid_targets_mean": 5561.8, + "valid_targets_min": 988 + }, + { + "epoch": 5.530612244897959, + "grad_norm": 0.45864417905162363, + "learning_rate": 5.149087423919541e-06, + "loss": 0.1638, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07489781081676483, + "step": 3390, + "valid_targets_mean": 5958.8, + "valid_targets_min": 1967 + }, + { + "epoch": 5.538775510204082, + "grad_norm": 0.4694299423214727, + "learning_rate": 5.094711006484907e-06, + "loss": 0.175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08611179888248444, + "step": 3395, + "valid_targets_mean": 5369.8, + "valid_targets_min": 1875 + }, + { + "epoch": 5.546938775510204, + "grad_norm": 0.5643467092019754, + "learning_rate": 5.040581295187048e-06, + "loss": 0.1736, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09875790774822235, + "step": 3400, + "valid_targets_mean": 4500.5, + "valid_targets_min": 2867 + }, + { + "epoch": 5.555102040816326, + "grad_norm": 0.4390252283312783, + "learning_rate": 4.986699185958403e-06, + "loss": 0.1745, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07030566036701202, + "step": 3405, + "valid_targets_mean": 4411.1, + "valid_targets_min": 2515 + }, + { + "epoch": 5.563265306122449, + "grad_norm": 0.4638144447712849, + "learning_rate": 4.933065570633215e-06, + "loss": 0.1768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07658712565898895, + "step": 3410, + "valid_targets_mean": 4903.1, + "valid_targets_min": 2572 + }, + { + "epoch": 5.571428571428571, + "grad_norm": 0.5537043507081186, + "learning_rate": 4.879681336932733e-06, + "loss": 0.1804, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11388731002807617, + "step": 3415, + "valid_targets_mean": 4796.4, + "valid_targets_min": 1771 + }, + { + "epoch": 5.579591836734694, + "grad_norm": 0.4731889808421747, + "learning_rate": 4.8265473684505844e-06, + "loss": 0.1799, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09846645593643188, + "step": 3420, + "valid_targets_mean": 6823.2, + "valid_targets_min": 3185 + }, + { + "epoch": 5.587755102040816, + "grad_norm": 0.4240237470679026, + "learning_rate": 4.773664544638079e-06, + "loss": 0.1751, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06629286706447601, + "step": 3425, + "valid_targets_mean": 5260.6, + "valid_targets_min": 755 + }, + { + "epoch": 5.595918367346939, + "grad_norm": 0.48722456054449315, + "learning_rate": 4.721033740789696e-06, + "loss": 0.1845, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12045013904571533, + "step": 3430, + "valid_targets_mean": 6874.5, + "valid_targets_min": 3698 + }, + { + "epoch": 5.604081632653061, + "grad_norm": 0.561411868538588, + "learning_rate": 4.6686558280285875e-06, + "loss": 0.1838, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09412881731987, + "step": 3435, + "valid_targets_mean": 4171.6, + "valid_targets_min": 1389 + }, + { + "epoch": 5.612244897959184, + "grad_norm": 0.5100313882410629, + "learning_rate": 4.6165316732921525e-06, + "loss": 0.1797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09020973742008209, + "step": 3440, + "valid_targets_mean": 4960.4, + "valid_targets_min": 3477 + }, + { + "epoch": 5.6204081632653065, + "grad_norm": 0.48037905657238017, + "learning_rate": 4.5646621393177e-06, + "loss": 0.1865, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10528279840946198, + "step": 3445, + "valid_targets_mean": 5070.9, + "valid_targets_min": 2828 + }, + { + "epoch": 5.628571428571428, + "grad_norm": 0.6129409856060618, + "learning_rate": 4.513048084628153e-06, + "loss": 0.1765, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06684743613004684, + "step": 3450, + "valid_targets_mean": 3636.2, + "valid_targets_min": 2132 + }, + { + "epoch": 5.636734693877551, + "grad_norm": 0.5188406585305368, + "learning_rate": 4.461690363517856e-06, + "loss": 0.1793, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0819583535194397, + "step": 3455, + "valid_targets_mean": 4631.9, + "valid_targets_min": 1520 + }, + { + "epoch": 5.644897959183673, + "grad_norm": 0.5261488357902977, + "learning_rate": 4.410589826038427e-06, + "loss": 0.1835, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09630433470010757, + "step": 3460, + "valid_targets_mean": 4737.4, + "valid_targets_min": 2158 + }, + { + "epoch": 5.653061224489796, + "grad_norm": 0.5564718916899687, + "learning_rate": 4.35974731798467e-06, + "loss": 0.17, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09089779853820801, + "step": 3465, + "valid_targets_mean": 4074.0, + "valid_targets_min": 2613 + }, + { + "epoch": 5.661224489795918, + "grad_norm": 0.49359667139456637, + "learning_rate": 4.309163680880619e-06, + "loss": 0.1857, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09821240603923798, + "step": 3470, + "valid_targets_mean": 5946.6, + "valid_targets_min": 1546 + }, + { + "epoch": 5.669387755102041, + "grad_norm": 0.5695633164379076, + "learning_rate": 4.258839751965566e-06, + "loss": 0.1803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08557415753602982, + "step": 3475, + "valid_targets_mean": 4009.6, + "valid_targets_min": 2417 + }, + { + "epoch": 5.677551020408163, + "grad_norm": 0.4790980762705507, + "learning_rate": 4.208776364180231e-06, + "loss": 0.1775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09636128693819046, + "step": 3480, + "valid_targets_mean": 6290.5, + "valid_targets_min": 1357 + }, + { + "epoch": 5.685714285714286, + "grad_norm": 0.49444289079244524, + "learning_rate": 4.15897434615296e-06, + "loss": 0.1768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07259830832481384, + "step": 3485, + "valid_targets_mean": 4316.1, + "valid_targets_min": 264 + }, + { + "epoch": 5.6938775510204085, + "grad_norm": 0.5166295899116976, + "learning_rate": 4.10943452218602e-06, + "loss": 0.1922, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1207120418548584, + "step": 3490, + "valid_targets_mean": 5905.9, + "valid_targets_min": 3679 + }, + { + "epoch": 5.70204081632653, + "grad_norm": 0.5250771519188173, + "learning_rate": 4.06015771224195e-06, + "loss": 0.1863, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09034144878387451, + "step": 3495, + "valid_targets_mean": 4832.6, + "valid_targets_min": 2040 + }, + { + "epoch": 5.710204081632653, + "grad_norm": 0.5483295619114166, + "learning_rate": 4.011144731929981e-06, + "loss": 0.1909, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11903704702854156, + "step": 3500, + "valid_targets_mean": 5562.9, + "valid_targets_min": 2074 + }, + { + "epoch": 5.718367346938775, + "grad_norm": 0.5227938259641866, + "learning_rate": 3.962396392492554e-06, + "loss": 0.1835, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09354373812675476, + "step": 3505, + "valid_targets_mean": 4609.5, + "valid_targets_min": 827 + }, + { + "epoch": 5.726530612244898, + "grad_norm": 0.46671920374295456, + "learning_rate": 3.9139135007919e-06, + "loss": 0.1815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10033905506134033, + "step": 3510, + "valid_targets_mean": 6064.2, + "valid_targets_min": 1319 + }, + { + "epoch": 5.73469387755102, + "grad_norm": 0.4910469291437556, + "learning_rate": 3.865696859296635e-06, + "loss": 0.1815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10732349753379822, + "step": 3515, + "valid_targets_mean": 5148.2, + "valid_targets_min": 2218 + }, + { + "epoch": 5.742857142857143, + "grad_norm": 0.43353585019350355, + "learning_rate": 3.817747266068541e-06, + "loss": 0.1756, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08565760403871536, + "step": 3520, + "valid_targets_mean": 4565.4, + "valid_targets_min": 1967 + }, + { + "epoch": 5.751020408163265, + "grad_norm": 0.5310832117874204, + "learning_rate": 3.770065514749317e-06, + "loss": 0.1731, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1077844426035881, + "step": 3525, + "valid_targets_mean": 5440.5, + "valid_targets_min": 2018 + }, + { + "epoch": 5.759183673469388, + "grad_norm": 0.5169834534688904, + "learning_rate": 3.7226523945474567e-06, + "loss": 0.1839, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10681848227977753, + "step": 3530, + "valid_targets_mean": 5627.4, + "valid_targets_min": 2586 + }, + { + "epoch": 5.7673469387755105, + "grad_norm": 0.5484383627618905, + "learning_rate": 3.6755086902251847e-06, + "loss": 0.1838, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0877349004149437, + "step": 3535, + "valid_targets_mean": 5774.1, + "valid_targets_min": 2666 + }, + { + "epoch": 5.775510204081632, + "grad_norm": 0.591238582407821, + "learning_rate": 3.6286351820854514e-06, + "loss": 0.1856, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09708157181739807, + "step": 3540, + "valid_targets_mean": 3561.5, + "valid_targets_min": 1409 + }, + { + "epoch": 5.783673469387755, + "grad_norm": 0.4832269943744314, + "learning_rate": 3.582032645959066e-06, + "loss": 0.1729, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08354311436414719, + "step": 3545, + "valid_targets_mean": 5434.9, + "valid_targets_min": 3216 + }, + { + "epoch": 5.791836734693877, + "grad_norm": 0.533477273302161, + "learning_rate": 3.535701853191784e-06, + "loss": 0.1854, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10452941060066223, + "step": 3550, + "valid_targets_mean": 4881.2, + "valid_targets_min": 2425 + }, + { + "epoch": 5.8, + "grad_norm": 0.4388726457273361, + "learning_rate": 3.4896435706316e-06, + "loss": 0.1783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0693935751914978, + "step": 3555, + "valid_targets_mean": 5472.5, + "valid_targets_min": 677 + }, + { + "epoch": 5.808163265306122, + "grad_norm": 0.6045179245376615, + "learning_rate": 3.443858560616029e-06, + "loss": 0.1716, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07769778370857239, + "step": 3560, + "valid_targets_mean": 2920.1, + "valid_targets_min": 1135 + }, + { + "epoch": 5.816326530612245, + "grad_norm": 0.4748517425016773, + "learning_rate": 3.3983475809594892e-06, + "loss": 0.1762, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07783222198486328, + "step": 3565, + "valid_targets_mean": 5231.5, + "valid_targets_min": 2791 + }, + { + "epoch": 5.8244897959183675, + "grad_norm": 0.5094001344648212, + "learning_rate": 3.353111384940775e-06, + "loss": 0.1751, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12287256121635437, + "step": 3570, + "valid_targets_mean": 6160.6, + "valid_targets_min": 3278 + }, + { + "epoch": 5.83265306122449, + "grad_norm": 0.5476080239908376, + "learning_rate": 3.308150721290546e-06, + "loss": 0.1829, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06965255737304688, + "step": 3575, + "valid_targets_mean": 3738.8, + "valid_targets_min": 2244 + }, + { + "epoch": 5.840816326530613, + "grad_norm": 0.5134167564842722, + "learning_rate": 3.263466334179002e-06, + "loss": 0.1783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09871196001768112, + "step": 3580, + "valid_targets_mean": 4696.0, + "valid_targets_min": 845 + }, + { + "epoch": 5.848979591836734, + "grad_norm": 0.5552614954950037, + "learning_rate": 3.2190589632035117e-06, + "loss": 0.181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08760549128055573, + "step": 3585, + "valid_targets_mean": 4645.8, + "valid_targets_min": 1498 + }, + { + "epoch": 5.857142857142857, + "grad_norm": 0.6017051275561736, + "learning_rate": 3.174929343376374e-06, + "loss": 0.1877, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09327945113182068, + "step": 3590, + "valid_targets_mean": 3505.9, + "valid_targets_min": 2137 + }, + { + "epoch": 5.865306122448979, + "grad_norm": 0.5419286866691388, + "learning_rate": 3.1310782051126986e-06, + "loss": 0.1859, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10793288052082062, + "step": 3595, + "valid_targets_mean": 5848.5, + "valid_targets_min": 2733 + }, + { + "epoch": 5.873469387755102, + "grad_norm": 0.8665054567909115, + "learning_rate": 3.0875062742182526e-06, + "loss": 0.1735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08918233215808868, + "step": 3600, + "valid_targets_mean": 5910.8, + "valid_targets_min": 1124 + }, + { + "epoch": 5.881632653061224, + "grad_norm": 0.4675391830481981, + "learning_rate": 3.0442142718774992e-06, + "loss": 0.1804, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09749846905469894, + "step": 3605, + "valid_targets_mean": 6391.9, + "valid_targets_min": 4286 + }, + { + "epoch": 5.889795918367347, + "grad_norm": 0.5536901285759013, + "learning_rate": 3.001202914641628e-06, + "loss": 0.1786, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08519423753023148, + "step": 3610, + "valid_targets_mean": 4084.0, + "valid_targets_min": 2777 + }, + { + "epoch": 5.8979591836734695, + "grad_norm": 0.5507724458934895, + "learning_rate": 2.958472914416721e-06, + "loss": 0.1814, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09706813097000122, + "step": 3615, + "valid_targets_mean": 3863.4, + "valid_targets_min": 1727 + }, + { + "epoch": 5.906122448979592, + "grad_norm": 0.49819909262377643, + "learning_rate": 2.916024978451948e-06, + "loss": 0.1914, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0789080485701561, + "step": 3620, + "valid_targets_mean": 5023.1, + "valid_targets_min": 1140 + }, + { + "epoch": 5.914285714285715, + "grad_norm": 0.5883950716908385, + "learning_rate": 2.8738598093278593e-06, + "loss": 0.1772, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09466518461704254, + "step": 3625, + "valid_targets_mean": 3618.1, + "valid_targets_min": 170 + }, + { + "epoch": 5.922448979591836, + "grad_norm": 0.5746947171019235, + "learning_rate": 2.8319781049447923e-06, + "loss": 0.1755, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09892357885837555, + "step": 3630, + "valid_targets_mean": 3903.4, + "valid_targets_min": 1718 + }, + { + "epoch": 5.930612244897959, + "grad_norm": 0.5179995327668454, + "learning_rate": 2.7903805585112674e-06, + "loss": 0.1751, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0768948346376419, + "step": 3635, + "valid_targets_mean": 4118.9, + "valid_targets_min": 2274 + }, + { + "epoch": 5.938775510204081, + "grad_norm": 0.4513464329838285, + "learning_rate": 2.7490678585325526e-06, + "loss": 0.1699, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06318273395299911, + "step": 3640, + "valid_targets_mean": 3500.8, + "valid_targets_min": 1673 + }, + { + "epoch": 5.946938775510204, + "grad_norm": 0.5186761444260091, + "learning_rate": 2.708040688799258e-06, + "loss": 0.1806, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09379004687070847, + "step": 3645, + "valid_targets_mean": 4657.5, + "valid_targets_min": 4182 + }, + { + "epoch": 5.955102040816326, + "grad_norm": 0.606778731280153, + "learning_rate": 2.6672997283760137e-06, + "loss": 0.1825, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11004135757684708, + "step": 3650, + "valid_targets_mean": 4382.9, + "valid_targets_min": 2676 + }, + { + "epoch": 5.963265306122449, + "grad_norm": 0.5229929454307417, + "learning_rate": 2.6268456515902283e-06, + "loss": 0.1801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09963811188936234, + "step": 3655, + "valid_targets_mean": 5354.8, + "valid_targets_min": 2892 + }, + { + "epoch": 5.9714285714285715, + "grad_norm": 0.5001733152158864, + "learning_rate": 2.5866791280209368e-06, + "loss": 0.1735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07187163084745407, + "step": 3660, + "valid_targets_mean": 3210.0, + "valid_targets_min": 684 + }, + { + "epoch": 5.979591836734694, + "grad_norm": 0.6159746098198164, + "learning_rate": 2.546800822487714e-06, + "loss": 0.1768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.052930839359760284, + "step": 3665, + "valid_targets_mean": 3052.1, + "valid_targets_min": 1593 + }, + { + "epoch": 5.987755102040817, + "grad_norm": 0.4193135572158273, + "learning_rate": 2.507211395039668e-06, + "loss": 0.1783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0880434513092041, + "step": 3670, + "valid_targets_mean": 8106.5, + "valid_targets_min": 177 + }, + { + "epoch": 5.995918367346938, + "grad_norm": 0.5144999909128704, + "learning_rate": 2.4679115009445132e-06, + "loss": 0.1636, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08727358281612396, + "step": 3675, + "valid_targets_mean": 4467.9, + "valid_targets_min": 3024 + }, + { + "epoch": 6.003265306122449, + "grad_norm": 0.47252041481974333, + "learning_rate": 2.428901790677731e-06, + "loss": 0.1667, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10447318851947784, + "step": 3680, + "valid_targets_mean": 6420.9, + "valid_targets_min": 2059 + }, + { + "epoch": 6.011428571428572, + "grad_norm": 0.4986949893337232, + "learning_rate": 2.3901829099118044e-06, + "loss": 0.1717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08693604916334152, + "step": 3685, + "valid_targets_mean": 4444.5, + "valid_targets_min": 2091 + }, + { + "epoch": 6.019591836734694, + "grad_norm": 0.5681658717936597, + "learning_rate": 2.351755499505526e-06, + "loss": 0.1726, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1184157133102417, + "step": 3690, + "valid_targets_mean": 4972.2, + "valid_targets_min": 2219 + }, + { + "epoch": 6.027755102040817, + "grad_norm": 0.4921539159180312, + "learning_rate": 2.313620195493389e-06, + "loss": 0.1731, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08808198571205139, + "step": 3695, + "valid_targets_mean": 4892.5, + "valid_targets_min": 2057 + }, + { + "epoch": 6.035918367346939, + "grad_norm": 0.49040858813536775, + "learning_rate": 2.2757776290750687e-06, + "loss": 0.1757, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1032937616109848, + "step": 3700, + "valid_targets_mean": 5997.8, + "valid_targets_min": 2168 + }, + { + "epoch": 6.044081632653061, + "grad_norm": 0.5624209287229496, + "learning_rate": 2.238228426604967e-06, + "loss": 0.1753, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0846734493970871, + "step": 3705, + "valid_targets_mean": 3722.6, + "valid_targets_min": 2394 + }, + { + "epoch": 6.052244897959183, + "grad_norm": 0.5530054487383088, + "learning_rate": 2.2009732095818425e-06, + "loss": 0.176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10292991250753403, + "step": 3710, + "valid_targets_mean": 4475.9, + "valid_targets_min": 1713 + }, + { + "epoch": 6.060408163265306, + "grad_norm": 0.49926699110953343, + "learning_rate": 2.1640125946385317e-06, + "loss": 0.1603, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08036244660615921, + "step": 3715, + "valid_targets_mean": 4533.5, + "valid_targets_min": 2572 + }, + { + "epoch": 6.0685714285714285, + "grad_norm": 0.5923091072775398, + "learning_rate": 2.127347193531757e-06, + "loss": 0.1647, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.053380295634269714, + "step": 3720, + "valid_targets_mean": 2151.5, + "valid_targets_min": 1027 + }, + { + "epoch": 6.076734693877551, + "grad_norm": 0.4861946492399618, + "learning_rate": 2.090977613131955e-06, + "loss": 0.1756, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08133894950151443, + "step": 3725, + "valid_targets_mean": 5612.1, + "valid_targets_min": 230 + }, + { + "epoch": 6.084897959183674, + "grad_norm": 0.47935348621159685, + "learning_rate": 2.0549044554132824e-06, + "loss": 0.169, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09912076592445374, + "step": 3730, + "valid_targets_mean": 6868.9, + "valid_targets_min": 2386 + }, + { + "epoch": 6.093061224489796, + "grad_norm": 0.47252848078690995, + "learning_rate": 2.019128317443626e-06, + "loss": 0.1807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06253606826066971, + "step": 3735, + "valid_targets_mean": 4328.8, + "valid_targets_min": 2804 + }, + { + "epoch": 6.101224489795919, + "grad_norm": 0.4222132937119648, + "learning_rate": 1.98364979137472e-06, + "loss": 0.1757, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07729291915893555, + "step": 3740, + "valid_targets_mean": 4720.0, + "valid_targets_min": 2142 + }, + { + "epoch": 6.109387755102041, + "grad_norm": 0.40428888724347384, + "learning_rate": 1.948469464432359e-06, + "loss": 0.176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.05827842652797699, + "step": 3745, + "valid_targets_mean": 5187.4, + "valid_targets_min": 1394 + }, + { + "epoch": 6.117551020408163, + "grad_norm": 0.5559197845200394, + "learning_rate": 1.9135879189066562e-06, + "loss": 0.1708, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10092753916978836, + "step": 3750, + "valid_targets_mean": 5105.6, + "valid_targets_min": 2940 + }, + { + "epoch": 6.1257142857142854, + "grad_norm": 0.6036578135073778, + "learning_rate": 1.8790057321424427e-06, + "loss": 0.1768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07620693743228912, + "step": 3755, + "valid_targets_mean": 3435.4, + "valid_targets_min": 751 + }, + { + "epoch": 6.133877551020408, + "grad_norm": 0.581844618201763, + "learning_rate": 1.8447234765296617e-06, + "loss": 0.1755, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08816036581993103, + "step": 3760, + "valid_targets_mean": 4790.1, + "valid_targets_min": 581 + }, + { + "epoch": 6.1420408163265305, + "grad_norm": 0.48917153992312773, + "learning_rate": 1.810741719493936e-06, + "loss": 0.1696, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07217949628829956, + "step": 3765, + "valid_targets_mean": 4503.4, + "valid_targets_min": 2449 + }, + { + "epoch": 6.150204081632653, + "grad_norm": 0.44242782826133753, + "learning_rate": 1.7770610234871588e-06, + "loss": 0.1654, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07949961721897125, + "step": 3770, + "valid_targets_mean": 6213.2, + "valid_targets_min": 2654 + }, + { + "epoch": 6.158367346938776, + "grad_norm": 0.5507329562626103, + "learning_rate": 1.743681945978184e-06, + "loss": 0.1786, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11011023819446564, + "step": 3775, + "valid_targets_mean": 5885.9, + "valid_targets_min": 2652 + }, + { + "epoch": 6.166530612244898, + "grad_norm": 0.4550015614312, + "learning_rate": 1.710605039443609e-06, + "loss": 0.1759, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07752910256385803, + "step": 3780, + "valid_targets_mean": 6354.8, + "valid_targets_min": 710 + }, + { + "epoch": 6.174693877551021, + "grad_norm": 0.5207522547176406, + "learning_rate": 1.6778308513586084e-06, + "loss": 0.1647, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08761303126811981, + "step": 3785, + "valid_targets_mean": 4992.5, + "valid_targets_min": 2047 + }, + { + "epoch": 6.182857142857143, + "grad_norm": 0.5273537447410652, + "learning_rate": 1.6453599241879014e-06, + "loss": 0.1634, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07733944058418274, + "step": 3790, + "valid_targets_mean": 4691.1, + "valid_targets_min": 961 + }, + { + "epoch": 6.191020408163265, + "grad_norm": 0.4161651854234828, + "learning_rate": 1.6131927953767567e-06, + "loss": 0.1674, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0671846866607666, + "step": 3795, + "valid_targets_mean": 6539.9, + "valid_targets_min": 1133 + }, + { + "epoch": 6.1991836734693875, + "grad_norm": 0.5463330466733015, + "learning_rate": 1.581329997342087e-06, + "loss": 0.1682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09839221835136414, + "step": 3800, + "valid_targets_mean": 4389.6, + "valid_targets_min": 1840 + }, + { + "epoch": 6.20734693877551, + "grad_norm": 0.583054445182557, + "learning_rate": 1.5497720574636632e-06, + "loss": 0.1778, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09077779203653336, + "step": 3805, + "valid_targets_mean": 4176.5, + "valid_targets_min": 2000 + }, + { + "epoch": 6.215510204081633, + "grad_norm": 0.47287015564436347, + "learning_rate": 1.518519498075366e-06, + "loss": 0.1696, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08704468607902527, + "step": 3810, + "valid_targets_mean": 5991.4, + "valid_targets_min": 3677 + }, + { + "epoch": 6.223673469387755, + "grad_norm": 0.48414217836633555, + "learning_rate": 1.4875728364565411e-06, + "loss": 0.1652, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08780664950609207, + "step": 3815, + "valid_targets_mean": 5148.4, + "valid_targets_min": 2605 + }, + { + "epoch": 6.231836734693878, + "grad_norm": 0.48754025209367896, + "learning_rate": 1.4569325848234473e-06, + "loss": 0.1755, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07883431017398834, + "step": 3820, + "valid_targets_mean": 5642.8, + "valid_targets_min": 2511 + }, + { + "epoch": 6.24, + "grad_norm": 0.5239266949028567, + "learning_rate": 1.4265992503207704e-06, + "loss": 0.1671, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09598593413829803, + "step": 3825, + "valid_targets_mean": 5202.4, + "valid_targets_min": 2352 + }, + { + "epoch": 6.248163265306123, + "grad_norm": 0.48189216098707605, + "learning_rate": 1.396573335013236e-06, + "loss": 0.1831, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07325664907693863, + "step": 3830, + "valid_targets_mean": 4080.9, + "valid_targets_min": 1525 + }, + { + "epoch": 6.256326530612245, + "grad_norm": 0.3815761140634215, + "learning_rate": 1.3668553358772863e-06, + "loss": 0.163, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07809938490390778, + "step": 3835, + "valid_targets_mean": 6897.1, + "valid_targets_min": 4401 + }, + { + "epoch": 6.264489795918367, + "grad_norm": 0.5540990052785123, + "learning_rate": 1.337445744792867e-06, + "loss": 0.1776, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09904327988624573, + "step": 3840, + "valid_targets_mean": 4890.2, + "valid_targets_min": 1419 + }, + { + "epoch": 6.2726530612244895, + "grad_norm": 0.566758209601489, + "learning_rate": 1.3083450485352867e-06, + "loss": 0.168, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12004449963569641, + "step": 3845, + "valid_targets_mean": 4876.9, + "valid_targets_min": 2036 + }, + { + "epoch": 6.280816326530612, + "grad_norm": 0.5280459490954896, + "learning_rate": 1.279553728767151e-06, + "loss": 0.1706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09008440375328064, + "step": 3850, + "valid_targets_mean": 4055.4, + "valid_targets_min": 2205 + }, + { + "epoch": 6.288979591836735, + "grad_norm": 0.4787362534748566, + "learning_rate": 1.2510722620303972e-06, + "loss": 0.1733, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11373571306467056, + "step": 3855, + "valid_targets_mean": 7750.4, + "valid_targets_min": 3349 + }, + { + "epoch": 6.297142857142857, + "grad_norm": 0.5015600982806857, + "learning_rate": 1.2229011197384021e-06, + "loss": 0.1695, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07090511918067932, + "step": 3860, + "valid_targets_mean": 5515.6, + "valid_targets_min": 3275 + }, + { + "epoch": 6.30530612244898, + "grad_norm": 0.5866542509141897, + "learning_rate": 1.1950407681681831e-06, + "loss": 0.1707, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07953464984893799, + "step": 3865, + "valid_targets_mean": 3868.6, + "valid_targets_min": 1761 + }, + { + "epoch": 6.313469387755102, + "grad_norm": 0.517511096201679, + "learning_rate": 1.1674916684526782e-06, + "loss": 0.1785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08121711760759354, + "step": 3870, + "valid_targets_mean": 4257.8, + "valid_targets_min": 2048 + }, + { + "epoch": 6.321632653061225, + "grad_norm": 0.4774974122619391, + "learning_rate": 1.1402542765731072e-06, + "loss": 0.1683, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0860278308391571, + "step": 3875, + "valid_targets_mean": 5665.1, + "valid_targets_min": 1144 + }, + { + "epoch": 6.329795918367347, + "grad_norm": 0.4311198664284439, + "learning_rate": 1.1133290433514522e-06, + "loss": 0.1691, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07120402157306671, + "step": 3880, + "valid_targets_mean": 4677.0, + "valid_targets_min": 2435 + }, + { + "epoch": 6.337959183673469, + "grad_norm": 0.5839912382789675, + "learning_rate": 1.086716414442952e-06, + "loss": 0.1736, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.075423963367939, + "step": 3885, + "valid_targets_mean": 4535.6, + "valid_targets_min": 1917 + }, + { + "epoch": 6.3461224489795915, + "grad_norm": 0.5308773482359681, + "learning_rate": 1.0604168303287588e-06, + "loss": 0.1711, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07817616313695908, + "step": 3890, + "valid_targets_mean": 3928.2, + "valid_targets_min": 3006 + }, + { + "epoch": 6.354285714285714, + "grad_norm": 0.4985657759227005, + "learning_rate": 1.0344307263086417e-06, + "loss": 0.1744, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08368688821792603, + "step": 3895, + "valid_targets_mean": 5520.8, + "valid_targets_min": 2725 + }, + { + "epoch": 6.362448979591837, + "grad_norm": 0.47797322901288064, + "learning_rate": 1.008758532493772e-06, + "loss": 0.1734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06277412176132202, + "step": 3900, + "valid_targets_mean": 4010.4, + "valid_targets_min": 844 + }, + { + "epoch": 6.370612244897959, + "grad_norm": 0.4862421427522955, + "learning_rate": 9.834006737996126e-07, + "loss": 0.1625, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0716070681810379, + "step": 3905, + "valid_targets_mean": 4795.8, + "valid_targets_min": 2504 + }, + { + "epoch": 6.378775510204082, + "grad_norm": 0.5535941599303137, + "learning_rate": 9.583575699388748e-07, + "loss": 0.1777, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09365393966436386, + "step": 3910, + "valid_targets_mean": 4736.4, + "valid_targets_min": 1733 + }, + { + "epoch": 6.386938775510204, + "grad_norm": 0.42700673020514873, + "learning_rate": 9.336296354145969e-07, + "loss": 0.17, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06854044646024704, + "step": 3915, + "valid_targets_mean": 4657.2, + "valid_targets_min": 2264 + }, + { + "epoch": 6.395102040816327, + "grad_norm": 0.6328165837439812, + "learning_rate": 9.092172795132459e-07, + "loss": 0.1727, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0929318219423294, + "step": 3920, + "valid_targets_mean": 3779.5, + "valid_targets_min": 1418 + }, + { + "epoch": 6.403265306122449, + "grad_norm": 0.5416202029620326, + "learning_rate": 8.851209062979715e-07, + "loss": 0.1805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10915129631757736, + "step": 3925, + "valid_targets_mean": 5157.9, + "valid_targets_min": 2472 + }, + { + "epoch": 6.411428571428571, + "grad_norm": 0.5240492479008688, + "learning_rate": 8.613409146019158e-07, + "loss": 0.1786, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09709012508392334, + "step": 3930, + "valid_targets_mean": 4688.8, + "valid_targets_min": 2118 + }, + { + "epoch": 6.419591836734694, + "grad_norm": 0.42698292167145513, + "learning_rate": 8.378776980215919e-07, + "loss": 0.1779, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07900860905647278, + "step": 3935, + "valid_targets_mean": 5823.5, + "valid_targets_min": 3881 + }, + { + "epoch": 6.427755102040816, + "grad_norm": 0.5744560900858467, + "learning_rate": 8.147316449103959e-07, + "loss": 0.1741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09831557422876358, + "step": 3940, + "valid_targets_mean": 4894.8, + "valid_targets_min": 469 + }, + { + "epoch": 6.435918367346939, + "grad_norm": 0.5299921373805195, + "learning_rate": 7.919031383721565e-07, + "loss": 0.1793, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08028759062290192, + "step": 3945, + "valid_targets_mean": 4523.4, + "valid_targets_min": 2392 + }, + { + "epoch": 6.444081632653061, + "grad_norm": 0.44582836009607935, + "learning_rate": 7.69392556254811e-07, + "loss": 0.166, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09207530319690704, + "step": 3950, + "valid_targets_mean": 6267.1, + "valid_targets_min": 3488 + }, + { + "epoch": 6.452244897959184, + "grad_norm": 0.4376049853971939, + "learning_rate": 7.472002711441418e-07, + "loss": 0.1782, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06368829309940338, + "step": 3955, + "valid_targets_mean": 4596.1, + "valid_targets_min": 2451 + }, + { + "epoch": 6.460408163265306, + "grad_norm": 0.5848626971055868, + "learning_rate": 7.253266503576051e-07, + "loss": 0.1787, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08905494213104248, + "step": 3960, + "valid_targets_mean": 3593.2, + "valid_targets_min": 2068 + }, + { + "epoch": 6.468571428571429, + "grad_norm": 0.555314669268285, + "learning_rate": 7.037720559382699e-07, + "loss": 0.1743, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08827593922615051, + "step": 3965, + "valid_targets_mean": 3865.5, + "valid_targets_min": 1420 + }, + { + "epoch": 6.476734693877551, + "grad_norm": 0.5514660209987079, + "learning_rate": 6.825368446488068e-07, + "loss": 0.1759, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10137411952018738, + "step": 3970, + "valid_targets_mean": 5435.1, + "valid_targets_min": 2282 + }, + { + "epoch": 6.484897959183673, + "grad_norm": 0.5054757359761405, + "learning_rate": 6.616213679655881e-07, + "loss": 0.1727, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0968596339225769, + "step": 3975, + "valid_targets_mean": 5499.6, + "valid_targets_min": 2797 + }, + { + "epoch": 6.493061224489796, + "grad_norm": 0.5159849008385041, + "learning_rate": 6.410259720728751e-07, + "loss": 0.17, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08823905885219574, + "step": 3980, + "valid_targets_mean": 4579.8, + "valid_targets_min": 2799 + }, + { + "epoch": 6.501224489795918, + "grad_norm": 0.44717355523296626, + "learning_rate": 6.207509978570848e-07, + "loss": 0.1648, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07926648110151291, + "step": 3985, + "valid_targets_mean": 5532.9, + "valid_targets_min": 2673 + }, + { + "epoch": 6.509387755102041, + "grad_norm": 0.5259133348625326, + "learning_rate": 6.007967809011517e-07, + "loss": 0.1694, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06820449978113174, + "step": 3990, + "valid_targets_mean": 3630.4, + "valid_targets_min": 1967 + }, + { + "epoch": 6.517551020408163, + "grad_norm": 0.4803407867104553, + "learning_rate": 5.811636514789598e-07, + "loss": 0.181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09790891408920288, + "step": 3995, + "valid_targets_mean": 5995.4, + "valid_targets_min": 230 + }, + { + "epoch": 6.525714285714286, + "grad_norm": 0.5829612841676888, + "learning_rate": 5.618519345498952e-07, + "loss": 0.1637, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07763680815696716, + "step": 4000, + "valid_targets_mean": 3797.0, + "valid_targets_min": 1159 + }, + { + "epoch": 6.533877551020408, + "grad_norm": 0.5543971635143561, + "learning_rate": 5.428619497534548e-07, + "loss": 0.1779, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0884358286857605, + "step": 4005, + "valid_targets_mean": 4588.5, + "valid_targets_min": 1806 + }, + { + "epoch": 6.542040816326531, + "grad_norm": 0.504734374948056, + "learning_rate": 5.2419401140396e-07, + "loss": 0.17, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07078321278095245, + "step": 4010, + "valid_targets_mean": 4664.2, + "valid_targets_min": 1845 + }, + { + "epoch": 6.550204081632653, + "grad_norm": 0.4242043874771602, + "learning_rate": 5.058484284853494e-07, + "loss": 0.1625, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07394272089004517, + "step": 4015, + "valid_targets_mean": 5321.9, + "valid_targets_min": 2081 + }, + { + "epoch": 6.558367346938775, + "grad_norm": 0.6707218641944701, + "learning_rate": 4.878255046460755e-07, + "loss": 0.179, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07270123064517975, + "step": 4020, + "valid_targets_mean": 2379.8, + "valid_targets_min": 1694 + }, + { + "epoch": 6.566530612244898, + "grad_norm": 0.54487029299083, + "learning_rate": 4.701255381940684e-07, + "loss": 0.1809, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11994723975658417, + "step": 4025, + "valid_targets_mean": 6316.6, + "valid_targets_min": 4205 + }, + { + "epoch": 6.57469387755102, + "grad_norm": 0.5040060467257438, + "learning_rate": 4.5274882209180195e-07, + "loss": 0.1795, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08845013380050659, + "step": 4030, + "valid_targets_mean": 5180.4, + "valid_targets_min": 2284 + }, + { + "epoch": 6.582857142857143, + "grad_norm": 0.5126754142819413, + "learning_rate": 4.3569564395144595e-07, + "loss": 0.1757, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07828289270401001, + "step": 4035, + "valid_targets_mean": 4312.9, + "valid_targets_min": 2007 + }, + { + "epoch": 6.591020408163265, + "grad_norm": 0.5011495581492582, + "learning_rate": 4.189662860301069e-07, + "loss": 0.1832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09557807445526123, + "step": 4040, + "valid_targets_mean": 6264.6, + "valid_targets_min": 3218 + }, + { + "epoch": 6.599183673469388, + "grad_norm": 0.4387049581215748, + "learning_rate": 4.025610252251433e-07, + "loss": 0.1747, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06180455535650253, + "step": 4045, + "valid_targets_mean": 5429.1, + "valid_targets_min": 802 + }, + { + "epoch": 6.60734693877551, + "grad_norm": 0.5220540147993991, + "learning_rate": 3.8648013306960664e-07, + "loss": 0.1721, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08135996013879776, + "step": 4050, + "valid_targets_mean": 4716.4, + "valid_targets_min": 2988 + }, + { + "epoch": 6.615510204081633, + "grad_norm": 0.4869619477437, + "learning_rate": 3.7072387572772317e-07, + "loss": 0.1798, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08093626797199249, + "step": 4055, + "valid_targets_mean": 5446.0, + "valid_targets_min": 201 + }, + { + "epoch": 6.6236734693877555, + "grad_norm": 0.5111229087974583, + "learning_rate": 3.552925139905083e-07, + "loss": 0.1815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11379435658454895, + "step": 4060, + "valid_targets_mean": 6852.5, + "valid_targets_min": 3094 + }, + { + "epoch": 6.631836734693877, + "grad_norm": 0.6008745315447422, + "learning_rate": 3.4018630327143474e-07, + "loss": 0.1734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10556414723396301, + "step": 4065, + "valid_targets_mean": 4508.4, + "valid_targets_min": 1910 + }, + { + "epoch": 6.64, + "grad_norm": 0.5057466157957375, + "learning_rate": 3.2540549360221105e-07, + "loss": 0.176, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09385024011135101, + "step": 4070, + "valid_targets_mean": 5681.0, + "valid_targets_min": 2830 + }, + { + "epoch": 6.648163265306122, + "grad_norm": 0.4568487337498165, + "learning_rate": 3.109503296286431e-07, + "loss": 0.1665, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0856882780790329, + "step": 4075, + "valid_targets_mean": 5635.8, + "valid_targets_min": 3360 + }, + { + "epoch": 6.656326530612245, + "grad_norm": 0.5328631009846894, + "learning_rate": 2.968210506065883e-07, + "loss": 0.1687, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0812063068151474, + "step": 4080, + "valid_targets_mean": 5037.9, + "valid_targets_min": 1807 + }, + { + "epoch": 6.664489795918367, + "grad_norm": 0.5606746811313319, + "learning_rate": 2.8301789039798523e-07, + "loss": 0.1722, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09690988063812256, + "step": 4085, + "valid_targets_mean": 4049.2, + "valid_targets_min": 1713 + }, + { + "epoch": 6.67265306122449, + "grad_norm": 0.477663793530669, + "learning_rate": 2.695410774669993e-07, + "loss": 0.1724, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07050501555204391, + "step": 4090, + "valid_targets_mean": 5059.9, + "valid_targets_min": 2563 + }, + { + "epoch": 6.680816326530612, + "grad_norm": 0.4907864488143317, + "learning_rate": 2.5639083487622116e-07, + "loss": 0.1777, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09305992722511292, + "step": 4095, + "valid_targets_mean": 4990.8, + "valid_targets_min": 1928 + }, + { + "epoch": 6.688979591836735, + "grad_norm": 0.5661942419711758, + "learning_rate": 2.435673802829941e-07, + "loss": 0.1822, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10230745375156403, + "step": 4100, + "valid_targets_mean": 4283.2, + "valid_targets_min": 1520 + }, + { + "epoch": 6.6971428571428575, + "grad_norm": 0.4826553611509298, + "learning_rate": 2.3107092593579905e-07, + "loss": 0.1735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08301982283592224, + "step": 4105, + "valid_targets_mean": 4510.2, + "valid_targets_min": 1516 + }, + { + "epoch": 6.705306122448979, + "grad_norm": 0.49400269132368113, + "learning_rate": 2.1890167867074208e-07, + "loss": 0.1664, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07874350249767303, + "step": 4110, + "valid_targets_mean": 4519.6, + "valid_targets_min": 2366 + }, + { + "epoch": 6.713469387755102, + "grad_norm": 0.5404819175512501, + "learning_rate": 2.0705983990813915e-07, + "loss": 0.187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09719543904066086, + "step": 4115, + "valid_targets_mean": 5836.2, + "valid_targets_min": 3722 + }, + { + "epoch": 6.721632653061224, + "grad_norm": 0.46187951904073626, + "learning_rate": 1.9554560564917001e-07, + "loss": 0.1713, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0860443189740181, + "step": 4120, + "valid_targets_mean": 6422.9, + "valid_targets_min": 1910 + }, + { + "epoch": 6.729795918367347, + "grad_norm": 0.5749101600772076, + "learning_rate": 1.843591664726496e-07, + "loss": 0.1804, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08156899362802505, + "step": 4125, + "valid_targets_mean": 3767.6, + "valid_targets_min": 1886 + }, + { + "epoch": 6.737959183673469, + "grad_norm": 0.5184829896214048, + "learning_rate": 1.7350070753186176e-07, + "loss": 0.175, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0800006240606308, + "step": 4130, + "valid_targets_mean": 4578.6, + "valid_targets_min": 2523 + }, + { + "epoch": 6.746122448979592, + "grad_norm": 0.5493945003783083, + "learning_rate": 1.6297040855149493e-07, + "loss": 0.1785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11112290620803833, + "step": 4135, + "valid_targets_mean": 4718.0, + "valid_targets_min": 1843 + }, + { + "epoch": 6.7542857142857144, + "grad_norm": 0.5579444536692725, + "learning_rate": 1.5276844382467125e-07, + "loss": 0.1704, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08533066511154175, + "step": 4140, + "valid_targets_mean": 3846.4, + "valid_targets_min": 1548 + }, + { + "epoch": 6.762448979591837, + "grad_norm": 0.5697066680123272, + "learning_rate": 1.4289498221006893e-07, + "loss": 0.1775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07441797852516174, + "step": 4145, + "valid_targets_mean": 3496.2, + "valid_targets_min": 1210 + }, + { + "epoch": 6.7706122448979595, + "grad_norm": 0.5354391536115609, + "learning_rate": 1.333501871291065e-07, + "loss": 0.1598, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08926458656787872, + "step": 4150, + "valid_targets_mean": 5302.8, + "valid_targets_min": 2814 + }, + { + "epoch": 6.778775510204081, + "grad_norm": 0.4695044365380142, + "learning_rate": 1.2413421656326307e-07, + "loss": 0.1788, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10014846175909042, + "step": 4155, + "valid_targets_mean": 6973.4, + "valid_targets_min": 1842 + }, + { + "epoch": 6.786938775510204, + "grad_norm": 0.5838126357143397, + "learning_rate": 1.1524722305144231e-07, + "loss": 0.1791, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07630135864019394, + "step": 4160, + "valid_targets_mean": 4727.8, + "valid_targets_min": 1813 + }, + { + "epoch": 6.795102040816326, + "grad_norm": 0.5070279879159841, + "learning_rate": 1.0668935368746136e-07, + "loss": 0.1801, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12365451455116272, + "step": 4165, + "valid_targets_mean": 6618.4, + "valid_targets_min": 4383 + }, + { + "epoch": 6.803265306122449, + "grad_norm": 0.5953179410730541, + "learning_rate": 9.846075011761048e-08, + "loss": 0.1712, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09822401404380798, + "step": 4170, + "valid_targets_mean": 3929.1, + "valid_targets_min": 1845 + }, + { + "epoch": 6.811428571428571, + "grad_norm": 0.697967990542345, + "learning_rate": 9.056154853830823e-08, + "loss": 0.1786, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08507151901721954, + "step": 4175, + "valid_targets_mean": 3861.1, + "valid_targets_min": 2219 + }, + { + "epoch": 6.819591836734694, + "grad_norm": 0.48169611662455075, + "learning_rate": 8.299187969385226e-08, + "loss": 0.1685, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07968790829181671, + "step": 4180, + "valid_targets_mean": 4887.4, + "valid_targets_min": 2593 + }, + { + "epoch": 6.8277551020408165, + "grad_norm": 0.4578362029303609, + "learning_rate": 7.575186887424757e-08, + "loss": 0.1696, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08985200524330139, + "step": 4185, + "valid_targets_mean": 7675.6, + "valid_targets_min": 2966 + }, + { + "epoch": 6.835918367346939, + "grad_norm": 0.5139405442593589, + "learning_rate": 6.884163591313941e-08, + "loss": 0.162, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08233918249607086, + "step": 4190, + "valid_targets_mean": 4719.0, + "valid_targets_min": 1192 + }, + { + "epoch": 6.844081632653062, + "grad_norm": 0.5600721868850346, + "learning_rate": 6.226129518582369e-08, + "loss": 0.1722, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08688675612211227, + "step": 4195, + "valid_targets_mean": 4590.5, + "valid_targets_min": 2579 + }, + { + "epoch": 6.852244897959183, + "grad_norm": 0.4247434382218982, + "learning_rate": 5.6010955607366245e-08, + "loss": 0.1681, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.05935890972614288, + "step": 4200, + "valid_targets_mean": 3941.9, + "valid_targets_min": 2516 + }, + { + "epoch": 6.860408163265306, + "grad_norm": 0.6070505109956861, + "learning_rate": 5.009072063077769e-08, + "loss": 0.185, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06743767112493515, + "step": 4205, + "valid_targets_mean": 3737.9, + "valid_targets_min": 758 + }, + { + "epoch": 6.868571428571428, + "grad_norm": 0.5891898407341091, + "learning_rate": 4.450068824532583e-08, + "loss": 0.1779, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08612845838069916, + "step": 4210, + "valid_targets_mean": 4001.5, + "valid_targets_min": 191 + }, + { + "epoch": 6.876734693877551, + "grad_norm": 0.49097204196746064, + "learning_rate": 3.924095097489922e-08, + "loss": 0.173, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09206587821245193, + "step": 4215, + "valid_targets_mean": 5461.0, + "valid_targets_min": 3681 + }, + { + "epoch": 6.884897959183673, + "grad_norm": 0.5234098272363491, + "learning_rate": 3.4311595876479475e-08, + "loss": 0.1761, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07597896456718445, + "step": 4220, + "valid_targets_mean": 4421.2, + "valid_targets_min": 1266 + }, + { + "epoch": 6.893061224489796, + "grad_norm": 0.6033382656401394, + "learning_rate": 2.971270453869579e-08, + "loss": 0.1733, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08271235227584839, + "step": 4225, + "valid_targets_mean": 3503.1, + "valid_targets_min": 1923 + }, + { + "epoch": 6.9012244897959185, + "grad_norm": 0.6792301772472162, + "learning_rate": 2.5444353080485983e-08, + "loss": 0.1784, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09355944395065308, + "step": 4230, + "valid_targets_mean": 6051.4, + "valid_targets_min": 2813 + }, + { + "epoch": 6.909387755102041, + "grad_norm": 0.48141128673204825, + "learning_rate": 2.150661214982641e-08, + "loss": 0.1679, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09260822832584381, + "step": 4235, + "valid_targets_mean": 5953.1, + "valid_targets_min": 3984 + }, + { + "epoch": 6.917551020408164, + "grad_norm": 0.5441935587073681, + "learning_rate": 1.7899546922564016e-08, + "loss": 0.182, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0769985094666481, + "step": 4240, + "valid_targets_mean": 3709.2, + "valid_targets_min": 1469 + }, + { + "epoch": 6.925714285714285, + "grad_norm": 0.5536238988411614, + "learning_rate": 1.4623217101339405e-08, + "loss": 0.1655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07819455862045288, + "step": 4245, + "valid_targets_mean": 3355.0, + "valid_targets_min": 160 + }, + { + "epoch": 6.933877551020408, + "grad_norm": 0.5080659448282431, + "learning_rate": 1.1677676914603197e-08, + "loss": 0.1645, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07104445993900299, + "step": 4250, + "valid_targets_mean": 4236.8, + "valid_targets_min": 2233 + }, + { + "epoch": 6.94204081632653, + "grad_norm": 0.5853986089854679, + "learning_rate": 9.06297511570564e-09, + "loss": 0.178, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1054919958114624, + "step": 4255, + "valid_targets_mean": 4447.1, + "valid_targets_min": 1903 + }, + { + "epoch": 6.950204081632653, + "grad_norm": 0.5012639100276955, + "learning_rate": 6.779154982103908e-09, + "loss": 0.1617, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07082746922969818, + "step": 4260, + "valid_targets_mean": 4263.4, + "valid_targets_min": 1060 + }, + { + "epoch": 6.9583673469387755, + "grad_norm": 0.5178754273320104, + "learning_rate": 4.8262543146382345e-09, + "loss": 0.1806, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07891260087490082, + "step": 4265, + "valid_targets_mean": 4791.5, + "valid_targets_min": 2335 + }, + { + "epoch": 6.966530612244898, + "grad_norm": 0.4976765498227958, + "learning_rate": 3.2043054369057523e-09, + "loss": 0.1783, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10171708464622498, + "step": 4270, + "valid_targets_mean": 5939.1, + "valid_targets_min": 4245 + }, + { + "epoch": 6.9746938775510205, + "grad_norm": 0.511741522266631, + "learning_rate": 1.9133351947320246e-09, + "loss": 0.1772, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08228804171085358, + "step": 4275, + "valid_targets_mean": 3950.4, + "valid_targets_min": 3062 + }, + { + "epoch": 6.982857142857143, + "grad_norm": 0.580838054903444, + "learning_rate": 9.53364955715852e-10, + "loss": 0.1724, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08037175238132477, + "step": 4280, + "valid_targets_mean": 3290.5, + "valid_targets_min": 1396 + }, + { + "epoch": 6.991020408163266, + "grad_norm": 0.5444763889021368, + "learning_rate": 3.244106088873267e-10, + "loss": 0.1788, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08252573758363724, + "step": 4285, + "valid_targets_mean": 3990.9, + "valid_targets_min": 1991 + }, + { + "epoch": 6.999183673469387, + "grad_norm": 0.5481609381293677, + "learning_rate": 2.648256443471553e-11, + "loss": 0.1759, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0896102786064148, + "step": 4290, + "valid_targets_mean": 4104.1, + "valid_targets_min": 2246 + }, + { + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15556100010871887, + "step": 4291, + "total_flos": 1.820182243006546e+18, + "train_loss": 0.2264533686425352, + "train_runtime": 56587.5923, + "train_samples_per_second": 1.212, + "train_steps_per_second": 0.076, + "valid_targets_mean": 4983.8, + "valid_targets_min": 1127 + } + ], + "logging_steps": 5, + "max_steps": 4291, + "num_input_tokens_seen": 0, + "num_train_epochs": 7, + "save_steps": 200, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 1.820182243006546e+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..cc9900e --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61a37cd53e41d43e718c68efb84f5579f2575062e7d1e9b5b8310a2b5afde102 +size 8657 diff --git a/training_loss.png b/training_loss.png new file mode 100644 index 0000000..36bb57a 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