commit ca4d22c13224d3904f3d4063c2250afbdc3db593 Author: ModelHub XC Date: Tue Aug 11 23:03:12 2026 +0800 初始化项目,由ModelHub XC社区提供模型 Model: laion/GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..024b75d --- /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 + +training_args.bin filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text +vocab.json filter=lfs diff=lfs merge=lfs -text +model-00004-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +model-00003-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +merges.txt filter=lfs diff=lfs merge=lfs -text +model-00001-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text +model-00002-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..063e00b --- /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: GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink + results: [] +--- + + + +# GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink + +This model is a fine-tuned version of [Qwen/Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) on the /data/cat/ws/befe330h-befe330h-otagent/huggingface/hub/datasets--DCAgent2--GLM-4.7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k/snapshots/e209a88db18950c3ce4e72a45a6088561d99d1bf_thinking_preprocessed 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 4.57.6 +- Pytorch 2.9.0+cu128 +- Datasets 4.4.1 +- Tokenizers 0.22.2 diff --git a/added_tokens.json b/added_tokens.json new file mode 100644 index 0000000..b54f913 --- /dev/null +++ b/added_tokens.json @@ -0,0 +1,28 @@ +{ + "": 151668, + "": 151658, + "": 151666, + "": 151667, + "": 151657, + "": 151665, + "<|box_end|>": 151649, + "<|box_start|>": 151648, + "<|endoftext|>": 151643, + "<|file_sep|>": 151664, + "<|fim_middle|>": 151660, + "<|fim_pad|>": 151662, + "<|fim_prefix|>": 151659, + "<|fim_suffix|>": 151661, + "<|im_end|>": 151645, + "<|im_start|>": 151644, + "<|image_pad|>": 151655, + "<|object_ref_end|>": 151647, + "<|object_ref_start|>": 151646, + "<|quad_end|>": 151651, + "<|quad_start|>": 151650, + "<|repo_name|>": 151663, + "<|video_pad|>": 151656, + "<|vision_end|>": 151653, + "<|vision_pad|>": 151654, + "<|vision_start|>": 151652 +} diff --git a/all_results.json b/all_results.json new file mode 100644 index 0000000..d93ac21 --- /dev/null +++ b/all_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 5.12016799029668, + "achieved_tflops_per_gpu_theoretical": 216.82589202068274, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07740560173988342, + "mfu_percent": 0.5177116269258523, + "mfu_percent_theoretical": 21.923750457096332, + "total_flos": 2.6502668329167094e+18, + "train_loss": 0.3043324818911333, + "train_runtime": 64701.6572, + "train_samples_per_second": 1.021, + "train_steps_per_second": 0.064, + "valid_targets_mean": 6206.0, + "valid_targets_min": 3223 +} \ 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..59b4193 --- /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": "4.57.6", + "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..9adbb28 --- /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": "4.57.6" +} 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..44b38ea --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ecef380ff336f223c6aac4756c6d6fd6d21e9d319ee46b8ebcd4cb69a29271a +size 4902257696 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..91a5933 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4fd971690d2521796d3975f5f0559681e59ee4d0a63e1d95cbb1a77cf98595f6 +size 4915960368 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..4dd6c24 --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1397a6fad7e9c237f3b76e4686aaa2c19d12832293a6cdcf726258a2dc0fffea +size 4983068496 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..253db6d --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fb69283d1ca34768ff64e3ae89f233bbc5522913aa832d6a7e53bc98dbad5ffe +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..332e7ac --- /dev/null +++ b/run_summary.json @@ -0,0 +1,12 @@ +{ + "agent_name": "e209a88db18950c3ce4e72a45a6088561d99d1bf_thinking_preprocessed", + "training_start": null, + "training_end": null, + "created_by": "DCAgent", + "base_model_name": "Qwen/Qwen3-8B", + "dataset_name": "/data/cat/ws/befe330h-befe330h-otagent/huggingface/hub/datasets--DCAgent2--GLM-4.7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k/snapshots/e209a88db18950c3ce4e72a45a6088561d99d1bf_thinking_preprocessed", + "training_type": "SFT", + "training_parameters": "https://huggingface.co/laion/GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k-fixthink/blob/main/config.json", + "wandb_link": "https://wandb.ai/dogml/OpenThoughts-Agent/runs/sft_GLM-4-7-swesmith-sandboxes-with_tests-oracle_verified_120s-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..d93ac21 --- /dev/null +++ b/train_results.json @@ -0,0 +1,16 @@ +{ + "achieved_tflops_per_gpu": 5.12016799029668, + "achieved_tflops_per_gpu_theoretical": 216.82589202068274, + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07740560173988342, + "mfu_percent": 0.5177116269258523, + "mfu_percent_theoretical": 21.923750457096332, + "total_flos": 2.6502668329167094e+18, + "train_loss": 0.3043324818911333, + "train_runtime": 64701.6572, + "train_samples_per_second": 1.021, + "train_steps_per_second": 0.064, + "valid_targets_mean": 6206.0, + "valid_targets_min": 3223 +} \ No newline at end of file diff --git a/trainer_log.jsonl b/trainer_log.jsonl new file mode 100644 index 0000000..9296441 --- /dev/null +++ b/trainer_log.jsonl @@ -0,0 +1,827 @@ +{"current_steps": 5, "total_steps": 4130, "loss": 0.8805, "lr": 3.8740920096852305e-07, "epoch": 0.00847457627118644, "percentage": 0.12, "elapsed_time": "0:01:22", "remaining_time": "18:55:17"} +{"current_steps": 10, "total_steps": 4130, "loss": 0.8479, "lr": 8.716707021791768e-07, "epoch": 0.01694915254237288, "percentage": 0.24, "elapsed_time": "0:02:44", "remaining_time": "18:51:57"} +{"current_steps": 15, "total_steps": 4130, "loss": 0.7925, "lr": 1.3559322033898307e-06, "epoch": 0.025423728813559324, "percentage": 0.36, "elapsed_time": "0:04:05", "remaining_time": "18:41:22"} +{"current_steps": 20, "total_steps": 4130, "loss": 0.7771, "lr": 1.8401937046004845e-06, "epoch": 0.03389830508474576, "percentage": 0.48, "elapsed_time": "0:05:27", "remaining_time": "18:42:01"} +{"current_steps": 25, "total_steps": 4130, "loss": 0.6944, "lr": 2.324455205811138e-06, "epoch": 0.0423728813559322, "percentage": 0.61, "elapsed_time": "0:06:46", "remaining_time": "18:33:48"} +{"current_steps": 30, "total_steps": 4130, "loss": 0.681, "lr": 2.808716707021792e-06, "epoch": 0.05084745762711865, "percentage": 0.73, "elapsed_time": "0:08:08", "remaining_time": "18:32:42"} +{"current_steps": 35, "total_steps": 4130, "loss": 0.6805, "lr": 3.2929782082324455e-06, "epoch": 0.059322033898305086, "percentage": 0.85, "elapsed_time": "0:09:30", "remaining_time": "18:32:59"} +{"current_steps": 40, "total_steps": 4130, "loss": 0.5617, "lr": 3.7772397094430994e-06, "epoch": 0.06779661016949153, "percentage": 0.97, "elapsed_time": "0:10:49", "remaining_time": "18:27:13"} +{"current_steps": 45, "total_steps": 4130, "loss": 0.6141, "lr": 4.261501210653753e-06, "epoch": 0.07627118644067797, "percentage": 1.09, "elapsed_time": "0:12:09", "remaining_time": "18:23:55"} +{"current_steps": 50, "total_steps": 4130, "loss": 0.5575, "lr": 4.745762711864408e-06, "epoch": 0.0847457627118644, "percentage": 1.21, "elapsed_time": "0:13:29", "remaining_time": "18:21:35"} +{"current_steps": 55, "total_steps": 4130, "loss": 0.5185, "lr": 5.230024213075061e-06, "epoch": 0.09322033898305085, "percentage": 1.33, "elapsed_time": "0:14:46", "remaining_time": "18:14:13"} +{"current_steps": 60, "total_steps": 4130, "loss": 0.5227, "lr": 5.7142857142857145e-06, "epoch": 0.1016949152542373, "percentage": 1.45, "elapsed_time": "0:16:07", "remaining_time": "18:14:01"} +{"current_steps": 65, "total_steps": 4130, "loss": 0.5268, "lr": 6.198547215496369e-06, "epoch": 0.11016949152542373, "percentage": 1.57, "elapsed_time": "0:17:25", "remaining_time": "18:09:55"} +{"current_steps": 70, "total_steps": 4130, "loss": 0.5186, "lr": 6.682808716707022e-06, "epoch": 0.11864406779661017, "percentage": 1.69, "elapsed_time": "0:18:44", "remaining_time": "18:07:13"} +{"current_steps": 75, "total_steps": 4130, "loss": 0.5181, "lr": 7.1670702179176766e-06, "epoch": 0.1271186440677966, "percentage": 1.82, "elapsed_time": "0:20:02", "remaining_time": "18:03:40"} +{"current_steps": 80, "total_steps": 4130, "loss": 0.4839, "lr": 7.65133171912833e-06, "epoch": 0.13559322033898305, "percentage": 1.94, "elapsed_time": "0:21:18", "remaining_time": "17:58:30"} +{"current_steps": 85, "total_steps": 4130, "loss": 0.5197, "lr": 8.135593220338983e-06, "epoch": 0.1440677966101695, "percentage": 2.06, "elapsed_time": "0:22:40", "remaining_time": "17:59:19"} +{"current_steps": 90, "total_steps": 4130, "loss": 0.517, "lr": 8.619854721549637e-06, "epoch": 0.15254237288135594, "percentage": 2.18, "elapsed_time": "0:24:01", "remaining_time": "17:58:24"} +{"current_steps": 95, "total_steps": 4130, "loss": 0.4704, "lr": 9.10411622276029e-06, "epoch": 0.16101694915254236, "percentage": 2.3, "elapsed_time": "0:25:24", "remaining_time": "17:59:02"} +{"current_steps": 100, "total_steps": 4130, "loss": 0.4631, "lr": 9.588377723970946e-06, "epoch": 0.1694915254237288, "percentage": 2.42, "elapsed_time": "0:26:42", "remaining_time": "17:56:31"} +{"current_steps": 105, "total_steps": 4130, "loss": 0.4495, "lr": 1.0072639225181599e-05, "epoch": 0.17796610169491525, "percentage": 2.54, "elapsed_time": "0:28:04", "remaining_time": "17:56:09"} +{"current_steps": 110, "total_steps": 4130, "loss": 0.4323, "lr": 1.0556900726392252e-05, "epoch": 0.1864406779661017, "percentage": 2.66, "elapsed_time": "0:29:27", "remaining_time": "17:56:22"} +{"current_steps": 115, "total_steps": 4130, "loss": 0.4595, "lr": 1.1041162227602906e-05, "epoch": 0.19491525423728814, "percentage": 2.78, "elapsed_time": "0:30:48", "remaining_time": "17:55:43"} +{"current_steps": 120, "total_steps": 4130, "loss": 0.4257, "lr": 1.1525423728813561e-05, "epoch": 0.2033898305084746, "percentage": 2.91, "elapsed_time": "0:32:05", "remaining_time": "17:52:08"} +{"current_steps": 125, "total_steps": 4130, "loss": 0.4552, "lr": 1.2009685230024215e-05, "epoch": 0.211864406779661, "percentage": 3.03, "elapsed_time": "0:33:24", "remaining_time": "17:50:20"} +{"current_steps": 130, "total_steps": 4130, "loss": 0.4537, "lr": 1.2493946731234868e-05, "epoch": 0.22033898305084745, "percentage": 3.15, "elapsed_time": "0:34:42", "remaining_time": "17:47:48"} +{"current_steps": 135, "total_steps": 4130, "loss": 0.4045, "lr": 1.2978208232445521e-05, "epoch": 0.2288135593220339, "percentage": 3.27, "elapsed_time": "0:36:01", "remaining_time": "17:46:06"} +{"current_steps": 140, "total_steps": 4130, "loss": 0.4456, "lr": 1.3462469733656177e-05, "epoch": 0.23728813559322035, "percentage": 3.39, "elapsed_time": "0:37:22", "remaining_time": "17:45:07"} +{"current_steps": 145, "total_steps": 4130, "loss": 0.4839, "lr": 1.3946731234866828e-05, "epoch": 0.2457627118644068, "percentage": 3.51, "elapsed_time": "0:38:44", "remaining_time": "17:44:51"} +{"current_steps": 150, "total_steps": 4130, "loss": 0.3795, "lr": 1.4430992736077482e-05, "epoch": 0.2542372881355932, "percentage": 3.63, "elapsed_time": "0:40:07", "remaining_time": "17:44:35"} +{"current_steps": 155, "total_steps": 4130, "loss": 0.422, "lr": 1.4915254237288137e-05, "epoch": 0.2627118644067797, "percentage": 3.75, "elapsed_time": "0:41:28", "remaining_time": "17:43:32"} +{"current_steps": 160, "total_steps": 4130, "loss": 0.3944, "lr": 1.5399515738498792e-05, "epoch": 0.2711864406779661, "percentage": 3.87, "elapsed_time": "0:42:46", "remaining_time": "17:41:32"} +{"current_steps": 165, "total_steps": 4130, "loss": 0.3938, "lr": 1.5883777239709442e-05, "epoch": 0.2796610169491525, "percentage": 4.0, "elapsed_time": "0:44:08", "remaining_time": "17:40:38"} +{"current_steps": 170, "total_steps": 4130, "loss": 0.4304, "lr": 1.63680387409201e-05, "epoch": 0.288135593220339, "percentage": 4.12, "elapsed_time": "0:45:30", "remaining_time": "17:40:03"} +{"current_steps": 175, "total_steps": 4130, "loss": 0.3997, "lr": 1.6852300242130752e-05, "epoch": 0.2966101694915254, "percentage": 4.24, "elapsed_time": "0:46:48", "remaining_time": "17:37:56"} +{"current_steps": 180, "total_steps": 4130, "loss": 0.4092, "lr": 1.7336561743341406e-05, "epoch": 0.3050847457627119, "percentage": 4.36, "elapsed_time": "0:48:10", "remaining_time": "17:37:17"} +{"current_steps": 185, "total_steps": 4130, "loss": 0.4097, "lr": 1.782082324455206e-05, "epoch": 0.3135593220338983, "percentage": 4.48, "elapsed_time": "0:49:27", "remaining_time": "17:34:48"} +{"current_steps": 190, "total_steps": 4130, "loss": 0.4365, "lr": 1.8305084745762713e-05, "epoch": 0.3220338983050847, "percentage": 4.6, "elapsed_time": "0:50:50", "remaining_time": "17:34:08"} +{"current_steps": 195, "total_steps": 4130, "loss": 0.3837, "lr": 1.8789346246973366e-05, "epoch": 0.3305084745762712, "percentage": 4.72, "elapsed_time": "0:52:10", "remaining_time": "17:32:50"} +{"current_steps": 200, "total_steps": 4130, "loss": 0.4048, "lr": 1.927360774818402e-05, "epoch": 0.3389830508474576, "percentage": 4.84, "elapsed_time": "0:53:31", "remaining_time": "17:31:55"} +{"current_steps": 205, "total_steps": 4130, "loss": 0.4291, "lr": 1.9757869249394673e-05, "epoch": 0.3474576271186441, "percentage": 4.96, "elapsed_time": "0:54:54", "remaining_time": "17:31:14"} +{"current_steps": 210, "total_steps": 4130, "loss": 0.3686, "lr": 2.0242130750605327e-05, "epoch": 0.3559322033898305, "percentage": 5.08, "elapsed_time": "0:56:15", "remaining_time": "17:30:12"} +{"current_steps": 215, "total_steps": 4130, "loss": 0.3826, "lr": 2.072639225181598e-05, "epoch": 0.3644067796610169, "percentage": 5.21, "elapsed_time": "0:57:29", "remaining_time": "17:26:54"} +{"current_steps": 220, "total_steps": 4130, "loss": 0.3656, "lr": 2.1210653753026637e-05, "epoch": 0.3728813559322034, "percentage": 5.33, "elapsed_time": "0:58:47", "remaining_time": "17:24:59"} +{"current_steps": 225, "total_steps": 4130, "loss": 0.3937, "lr": 2.169491525423729e-05, "epoch": 0.3813559322033898, "percentage": 5.45, "elapsed_time": "1:00:02", "remaining_time": "17:22:03"} +{"current_steps": 230, "total_steps": 4130, "loss": 0.3614, "lr": 2.2179176755447944e-05, "epoch": 0.3898305084745763, "percentage": 5.57, "elapsed_time": "1:01:20", "remaining_time": "17:20:11"} +{"current_steps": 235, "total_steps": 4130, "loss": 0.4322, "lr": 2.2663438256658597e-05, "epoch": 0.3983050847457627, "percentage": 5.69, "elapsed_time": "1:02:41", "remaining_time": "17:19:11"} +{"current_steps": 240, "total_steps": 4130, "loss": 0.3513, "lr": 2.3147699757869254e-05, "epoch": 0.4067796610169492, "percentage": 5.81, "elapsed_time": "1:04:02", "remaining_time": "17:17:54"} +{"current_steps": 245, "total_steps": 4130, "loss": 0.3925, "lr": 2.3631961259079904e-05, "epoch": 0.4152542372881356, "percentage": 5.93, "elapsed_time": "1:05:22", "remaining_time": "17:16:37"} +{"current_steps": 250, "total_steps": 4130, "loss": 0.3952, "lr": 2.4116222760290558e-05, "epoch": 0.423728813559322, "percentage": 6.05, "elapsed_time": "1:06:44", "remaining_time": "17:15:42"} +{"current_steps": 255, "total_steps": 4130, "loss": 0.3749, "lr": 2.460048426150121e-05, "epoch": 0.4322033898305085, "percentage": 6.17, "elapsed_time": "1:08:05", "remaining_time": "17:14:39"} +{"current_steps": 260, "total_steps": 4130, "loss": 0.3864, "lr": 2.5084745762711865e-05, "epoch": 0.4406779661016949, "percentage": 6.3, "elapsed_time": "1:09:21", "remaining_time": "17:12:27"} +{"current_steps": 265, "total_steps": 4130, "loss": 0.3826, "lr": 2.556900726392252e-05, "epoch": 0.4491525423728814, "percentage": 6.42, "elapsed_time": "1:10:39", "remaining_time": "17:10:25"} +{"current_steps": 270, "total_steps": 4130, "loss": 0.3712, "lr": 2.6053268765133175e-05, "epoch": 0.4576271186440678, "percentage": 6.54, "elapsed_time": "1:11:57", "remaining_time": "17:08:48"} +{"current_steps": 275, "total_steps": 4130, "loss": 0.4284, "lr": 2.653753026634383e-05, "epoch": 0.4661016949152542, "percentage": 6.66, "elapsed_time": "1:13:20", "remaining_time": "17:08:00"} +{"current_steps": 280, "total_steps": 4130, "loss": 0.4028, "lr": 2.702179176755448e-05, "epoch": 0.4745762711864407, "percentage": 6.78, "elapsed_time": "1:14:39", "remaining_time": "17:06:38"} +{"current_steps": 285, "total_steps": 4130, "loss": 0.3862, "lr": 2.7506053268765135e-05, "epoch": 0.4830508474576271, "percentage": 6.9, "elapsed_time": "1:16:00", "remaining_time": "17:05:32"} +{"current_steps": 290, "total_steps": 4130, "loss": 0.3911, "lr": 2.799031476997579e-05, "epoch": 0.4915254237288136, "percentage": 7.02, "elapsed_time": "1:17:20", "remaining_time": "17:04:12"} +{"current_steps": 295, "total_steps": 4130, "loss": 0.3677, "lr": 2.8474576271186442e-05, "epoch": 0.5, "percentage": 7.14, "elapsed_time": "1:18:39", "remaining_time": "17:02:38"} +{"current_steps": 300, "total_steps": 4130, "loss": 0.3671, "lr": 2.8958837772397096e-05, "epoch": 0.5084745762711864, "percentage": 7.26, "elapsed_time": "1:19:58", "remaining_time": "17:01:04"} +{"current_steps": 305, "total_steps": 4130, "loss": 0.3601, "lr": 2.9443099273607753e-05, "epoch": 0.5169491525423728, "percentage": 7.38, "elapsed_time": "1:21:17", "remaining_time": "16:59:33"} +{"current_steps": 310, "total_steps": 4130, "loss": 0.3657, "lr": 2.9927360774818406e-05, "epoch": 0.5254237288135594, "percentage": 7.51, "elapsed_time": "1:22:36", "remaining_time": "16:57:51"} +{"current_steps": 315, "total_steps": 4130, "loss": 0.3619, "lr": 3.041162227602906e-05, "epoch": 0.5338983050847458, "percentage": 7.63, "elapsed_time": "1:23:55", "remaining_time": "16:56:28"} +{"current_steps": 320, "total_steps": 4130, "loss": 0.3822, "lr": 3.089588377723971e-05, "epoch": 0.5423728813559322, "percentage": 7.75, "elapsed_time": "1:25:15", "remaining_time": "16:55:04"} +{"current_steps": 325, "total_steps": 4130, "loss": 0.3587, "lr": 3.138014527845036e-05, "epoch": 0.5508474576271186, "percentage": 7.87, "elapsed_time": "1:26:36", "remaining_time": "16:54:00"} +{"current_steps": 330, "total_steps": 4130, "loss": 0.357, "lr": 3.186440677966102e-05, "epoch": 0.559322033898305, "percentage": 7.99, "elapsed_time": "1:27:52", "remaining_time": "16:51:49"} +{"current_steps": 335, "total_steps": 4130, "loss": 0.3714, "lr": 3.234866828087168e-05, "epoch": 0.5677966101694916, "percentage": 8.11, "elapsed_time": "1:29:13", "remaining_time": "16:50:41"} +{"current_steps": 340, "total_steps": 4130, "loss": 0.3821, "lr": 3.283292978208233e-05, "epoch": 0.576271186440678, "percentage": 8.23, "elapsed_time": "1:30:33", "remaining_time": "16:49:27"} +{"current_steps": 345, "total_steps": 4130, "loss": 0.388, "lr": 3.3317191283292984e-05, "epoch": 0.5847457627118644, "percentage": 8.35, "elapsed_time": "1:31:50", "remaining_time": "16:47:37"} +{"current_steps": 350, "total_steps": 4130, "loss": 0.3823, "lr": 3.3801452784503634e-05, "epoch": 0.5932203389830508, "percentage": 8.47, "elapsed_time": "1:33:08", "remaining_time": "16:45:55"} +{"current_steps": 355, "total_steps": 4130, "loss": 0.3749, "lr": 3.4285714285714284e-05, "epoch": 0.6016949152542372, "percentage": 8.6, "elapsed_time": "1:34:29", "remaining_time": "16:44:44"} +{"current_steps": 360, "total_steps": 4130, "loss": 0.3682, "lr": 3.476997578692494e-05, "epoch": 0.6101694915254238, "percentage": 8.72, "elapsed_time": "1:35:50", "remaining_time": "16:43:42"} +{"current_steps": 365, "total_steps": 4130, "loss": 0.3903, "lr": 3.52542372881356e-05, "epoch": 0.6186440677966102, "percentage": 8.84, "elapsed_time": "1:37:11", "remaining_time": "16:42:37"} +{"current_steps": 370, "total_steps": 4130, "loss": 0.3743, "lr": 3.573849878934625e-05, "epoch": 0.6271186440677966, "percentage": 8.96, "elapsed_time": "1:38:32", "remaining_time": "16:41:26"} +{"current_steps": 375, "total_steps": 4130, "loss": 0.3868, "lr": 3.6222760290556904e-05, "epoch": 0.635593220338983, "percentage": 9.08, "elapsed_time": "1:39:51", "remaining_time": "16:39:53"} +{"current_steps": 380, "total_steps": 4130, "loss": 0.3817, "lr": 3.670702179176756e-05, "epoch": 0.6440677966101694, "percentage": 9.2, "elapsed_time": "1:41:10", "remaining_time": "16:38:28"} +{"current_steps": 385, "total_steps": 4130, "loss": 0.4001, "lr": 3.719128329297821e-05, "epoch": 0.652542372881356, "percentage": 9.32, "elapsed_time": "1:42:30", "remaining_time": "16:37:11"} +{"current_steps": 390, "total_steps": 4130, "loss": 0.3746, "lr": 3.767554479418886e-05, "epoch": 0.6610169491525424, "percentage": 9.44, "elapsed_time": "1:43:48", "remaining_time": "16:35:29"} +{"current_steps": 395, "total_steps": 4130, "loss": 0.3509, "lr": 3.815980629539952e-05, "epoch": 0.6694915254237288, "percentage": 9.56, "elapsed_time": "1:45:06", "remaining_time": "16:33:52"} +{"current_steps": 400, "total_steps": 4130, "loss": 0.3717, "lr": 3.8644067796610175e-05, "epoch": 0.6779661016949152, "percentage": 9.69, "elapsed_time": "1:46:27", "remaining_time": "16:32:46"} +{"current_steps": 405, "total_steps": 4130, "loss": 0.3594, "lr": 3.9128329297820825e-05, "epoch": 0.6864406779661016, "percentage": 9.81, "elapsed_time": "1:47:40", "remaining_time": "16:30:19"} +{"current_steps": 410, "total_steps": 4130, "loss": 0.3868, "lr": 3.961259079903148e-05, "epoch": 0.6949152542372882, "percentage": 9.93, "elapsed_time": "1:48:57", "remaining_time": "16:28:38"} +{"current_steps": 415, "total_steps": 4130, "loss": 0.351, "lr": 3.9999992856441635e-05, "epoch": 0.7033898305084746, "percentage": 10.05, "elapsed_time": "1:50:17", "remaining_time": "16:27:14"} +{"current_steps": 420, "total_steps": 4130, "loss": 0.366, "lr": 3.999974283243472e-05, "epoch": 0.711864406779661, "percentage": 10.17, "elapsed_time": "1:51:38", "remaining_time": "16:26:06"} +{"current_steps": 425, "total_steps": 4130, "loss": 0.4029, "lr": 3.9999135635612606e-05, "epoch": 0.7203389830508474, "percentage": 10.29, "elapsed_time": "1:52:59", "remaining_time": "16:25:00"} +{"current_steps": 430, "total_steps": 4130, "loss": 0.3795, "lr": 3.999817127681917e-05, "epoch": 0.7288135593220338, "percentage": 10.41, "elapsed_time": "1:54:19", "remaining_time": "16:23:44"} +{"current_steps": 435, "total_steps": 4130, "loss": 0.3473, "lr": 3.9996849773276764e-05, "epoch": 0.7372881355932204, "percentage": 10.53, "elapsed_time": "1:55:36", "remaining_time": "16:21:59"} +{"current_steps": 440, "total_steps": 4130, "loss": 0.3192, "lr": 3.9995171148585944e-05, "epoch": 0.7457627118644068, "percentage": 10.65, "elapsed_time": "1:56:53", "remaining_time": "16:20:17"} +{"current_steps": 445, "total_steps": 4130, "loss": 0.3584, "lr": 3.999313543272505e-05, "epoch": 0.7542372881355932, "percentage": 10.77, "elapsed_time": "1:58:10", "remaining_time": "16:18:37"} +{"current_steps": 450, "total_steps": 4130, "loss": 0.3279, "lr": 3.999074266204967e-05, "epoch": 0.7627118644067796, "percentage": 10.9, "elapsed_time": "1:59:26", "remaining_time": "16:16:48"} +{"current_steps": 455, "total_steps": 4130, "loss": 0.3391, "lr": 3.998799287929199e-05, "epoch": 0.7711864406779662, "percentage": 11.02, "elapsed_time": "2:00:47", "remaining_time": "16:15:34"} +{"current_steps": 460, "total_steps": 4130, "loss": 0.3482, "lr": 3.998488613356002e-05, "epoch": 0.7796610169491526, "percentage": 11.14, "elapsed_time": "2:02:05", "remaining_time": "16:14:07"} +{"current_steps": 465, "total_steps": 4130, "loss": 0.3809, "lr": 3.998142248033673e-05, "epoch": 0.788135593220339, "percentage": 11.26, "elapsed_time": "2:03:26", "remaining_time": "16:12:55"} +{"current_steps": 470, "total_steps": 4130, "loss": 0.3509, "lr": 3.997760198147904e-05, "epoch": 0.7966101694915254, "percentage": 11.38, "elapsed_time": "2:04:43", "remaining_time": "16:11:13"} +{"current_steps": 475, "total_steps": 4130, "loss": 0.3805, "lr": 3.9973424705216774e-05, "epoch": 0.8050847457627118, "percentage": 11.5, "elapsed_time": "2:05:59", "remaining_time": "16:09:28"} +{"current_steps": 480, "total_steps": 4130, "loss": 0.3276, "lr": 3.996889072615134e-05, "epoch": 0.8135593220338984, "percentage": 11.62, "elapsed_time": "2:07:19", "remaining_time": "16:08:14"} +{"current_steps": 485, "total_steps": 4130, "loss": 0.3765, "lr": 3.99640001252545e-05, "epoch": 0.8220338983050848, "percentage": 11.74, "elapsed_time": "2:08:40", "remaining_time": "16:07:02"} +{"current_steps": 490, "total_steps": 4130, "loss": 0.3689, "lr": 3.9958752989866856e-05, "epoch": 0.8305084745762712, "percentage": 11.86, "elapsed_time": "2:09:58", "remaining_time": "16:05:35"} +{"current_steps": 495, "total_steps": 4130, "loss": 0.3401, "lr": 3.995314941369631e-05, "epoch": 0.8389830508474576, "percentage": 11.99, "elapsed_time": "2:11:09", "remaining_time": "16:03:09"} +{"current_steps": 500, "total_steps": 4130, "loss": 0.3257, "lr": 3.994718949681642e-05, "epoch": 0.847457627118644, "percentage": 12.11, "elapsed_time": "2:12:26", "remaining_time": "16:01:27"} +{"current_steps": 505, "total_steps": 4130, "loss": 0.3487, "lr": 3.994087334566455e-05, "epoch": 0.8559322033898306, "percentage": 12.23, "elapsed_time": "2:13:43", "remaining_time": "15:59:52"} +{"current_steps": 510, "total_steps": 4130, "loss": 0.3608, "lr": 3.9934201073040035e-05, "epoch": 0.864406779661017, "percentage": 12.35, "elapsed_time": "2:15:02", "remaining_time": "15:58:33"} +{"current_steps": 515, "total_steps": 4130, "loss": 0.3558, "lr": 3.992717279810213e-05, "epoch": 0.8728813559322034, "percentage": 12.47, "elapsed_time": "2:16:21", "remaining_time": "15:57:11"} +{"current_steps": 520, "total_steps": 4130, "loss": 0.344, "lr": 3.991978864636788e-05, "epoch": 0.8813559322033898, "percentage": 12.59, "elapsed_time": "2:17:39", "remaining_time": "15:55:42"} +{"current_steps": 525, "total_steps": 4130, "loss": 0.3512, "lr": 3.9912048749709896e-05, "epoch": 0.8898305084745762, "percentage": 12.71, "elapsed_time": "2:18:58", "remaining_time": "15:54:17"} +{"current_steps": 530, "total_steps": 4130, "loss": 0.3311, "lr": 3.990395324635399e-05, "epoch": 0.8983050847457628, "percentage": 12.83, "elapsed_time": "2:20:18", "remaining_time": "15:53:03"} +{"current_steps": 535, "total_steps": 4130, "loss": 0.3513, "lr": 3.989550228087671e-05, "epoch": 0.9067796610169492, "percentage": 12.95, "elapsed_time": "2:21:38", "remaining_time": "15:51:43"} +{"current_steps": 540, "total_steps": 4130, "loss": 0.3215, "lr": 3.988669600420275e-05, "epoch": 0.9152542372881356, "percentage": 13.08, "elapsed_time": "2:22:57", "remaining_time": "15:50:24"} +{"current_steps": 545, "total_steps": 4130, "loss": 0.3634, "lr": 3.9877534573602254e-05, "epoch": 0.923728813559322, "percentage": 13.2, "elapsed_time": "2:24:14", "remaining_time": "15:48:49"} +{"current_steps": 550, "total_steps": 4130, "loss": 0.3573, "lr": 3.9868018152688044e-05, "epoch": 0.9322033898305084, "percentage": 13.32, "elapsed_time": "2:25:34", "remaining_time": "15:47:31"} +{"current_steps": 555, "total_steps": 4130, "loss": 0.3615, "lr": 3.985814691141263e-05, "epoch": 0.940677966101695, "percentage": 13.44, "elapsed_time": "2:26:51", "remaining_time": "15:45:57"} +{"current_steps": 560, "total_steps": 4130, "loss": 0.3839, "lr": 3.984792102606524e-05, "epoch": 0.9491525423728814, "percentage": 13.56, "elapsed_time": "2:28:06", "remaining_time": "15:44:12"} +{"current_steps": 565, "total_steps": 4130, "loss": 0.3165, "lr": 3.9837340679268634e-05, "epoch": 0.9576271186440678, "percentage": 13.68, "elapsed_time": "2:29:26", "remaining_time": "15:42:58"} +{"current_steps": 570, "total_steps": 4130, "loss": 0.3308, "lr": 3.982640605997585e-05, "epoch": 0.9661016949152542, "percentage": 13.8, "elapsed_time": "2:30:47", "remaining_time": "15:41:48"} +{"current_steps": 575, "total_steps": 4130, "loss": 0.353, "lr": 3.981511736346684e-05, "epoch": 0.9745762711864406, "percentage": 13.92, "elapsed_time": "2:32:02", "remaining_time": "15:40:03"} +{"current_steps": 580, "total_steps": 4130, "loss": 0.3341, "lr": 3.980347479134498e-05, "epoch": 0.9830508474576272, "percentage": 14.04, "elapsed_time": "2:33:23", "remaining_time": "15:38:52"} +{"current_steps": 585, "total_steps": 4130, "loss": 0.3257, "lr": 3.9791478551533435e-05, "epoch": 0.9915254237288136, "percentage": 14.16, "elapsed_time": "2:34:40", "remaining_time": "15:37:20"} +{"current_steps": 590, "total_steps": 4130, "loss": 0.3438, "lr": 3.977912885827151e-05, "epoch": 1.0, "percentage": 14.29, "elapsed_time": "2:35:56", "remaining_time": "15:35:39"} +{"current_steps": 595, "total_steps": 4130, "loss": 0.3632, "lr": 3.976642593211079e-05, "epoch": 1.0084745762711864, "percentage": 14.41, "elapsed_time": "2:37:16", "remaining_time": "15:34:23"} +{"current_steps": 600, "total_steps": 4130, "loss": 0.3324, "lr": 3.9753369999911166e-05, "epoch": 1.0169491525423728, "percentage": 14.53, "elapsed_time": "2:38:34", "remaining_time": "15:32:59"} +{"current_steps": 605, "total_steps": 4130, "loss": 0.3188, "lr": 3.973996129483684e-05, "epoch": 1.0254237288135593, "percentage": 14.65, "elapsed_time": "2:39:54", "remaining_time": "15:31:42"} +{"current_steps": 610, "total_steps": 4130, "loss": 0.2951, "lr": 3.972620005635216e-05, "epoch": 1.0338983050847457, "percentage": 14.77, "elapsed_time": "2:41:06", "remaining_time": "15:29:38"} +{"current_steps": 615, "total_steps": 4130, "loss": 0.3193, "lr": 3.971208653021727e-05, "epoch": 1.042372881355932, "percentage": 14.89, "elapsed_time": "2:42:25", "remaining_time": "15:28:19"} +{"current_steps": 620, "total_steps": 4130, "loss": 0.3693, "lr": 3.969762096848383e-05, "epoch": 1.0508474576271187, "percentage": 15.01, "elapsed_time": "2:43:46", "remaining_time": "15:27:09"} +{"current_steps": 625, "total_steps": 4130, "loss": 0.3308, "lr": 3.968280362949042e-05, "epoch": 1.0593220338983051, "percentage": 15.13, "elapsed_time": "2:44:59", "remaining_time": "15:25:17"} +{"current_steps": 630, "total_steps": 4130, "loss": 0.3605, "lr": 3.9667634777857975e-05, "epoch": 1.0677966101694916, "percentage": 15.25, "elapsed_time": "2:46:19", "remaining_time": "15:24:00"} +{"current_steps": 635, "total_steps": 4130, "loss": 0.3408, "lr": 3.965211468448505e-05, "epoch": 1.076271186440678, "percentage": 15.38, "elapsed_time": "2:47:37", "remaining_time": "15:22:34"} +{"current_steps": 640, "total_steps": 4130, "loss": 0.3339, "lr": 3.963624362654299e-05, "epoch": 1.0847457627118644, "percentage": 15.5, "elapsed_time": "2:48:57", "remaining_time": "15:21:21"} +{"current_steps": 645, "total_steps": 4130, "loss": 0.3253, "lr": 3.962002188747096e-05, "epoch": 1.0932203389830508, "percentage": 15.62, "elapsed_time": "2:50:12", "remaining_time": "15:19:40"} +{"current_steps": 650, "total_steps": 4130, "loss": 0.3232, "lr": 3.9603449756970877e-05, "epoch": 1.1016949152542372, "percentage": 15.74, "elapsed_time": "2:51:33", "remaining_time": "15:18:30"} +{"current_steps": 655, "total_steps": 4130, "loss": 0.3291, "lr": 3.95865275310023e-05, "epoch": 1.1101694915254237, "percentage": 15.86, "elapsed_time": "2:52:51", "remaining_time": "15:17:05"} +{"current_steps": 660, "total_steps": 4130, "loss": 0.3269, "lr": 3.9569255511777054e-05, "epoch": 1.11864406779661, "percentage": 15.98, "elapsed_time": "2:54:10", "remaining_time": "15:15:41"} +{"current_steps": 665, "total_steps": 4130, "loss": 0.3623, "lr": 3.955163400775389e-05, "epoch": 1.1271186440677967, "percentage": 16.1, "elapsed_time": "2:55:28", "remaining_time": "15:14:19"} +{"current_steps": 670, "total_steps": 4130, "loss": 0.324, "lr": 3.953366333363297e-05, "epoch": 1.1355932203389831, "percentage": 16.22, "elapsed_time": "2:56:42", "remaining_time": "15:12:35"} +{"current_steps": 675, "total_steps": 4130, "loss": 0.3109, "lr": 3.9515343810350236e-05, "epoch": 1.1440677966101696, "percentage": 16.34, "elapsed_time": "2:57:58", "remaining_time": "15:11:00"} +{"current_steps": 680, "total_steps": 4130, "loss": 0.3356, "lr": 3.949667576507168e-05, "epoch": 1.152542372881356, "percentage": 16.46, "elapsed_time": "2:59:13", "remaining_time": "15:09:20"} +{"current_steps": 685, "total_steps": 4130, "loss": 0.3321, "lr": 3.9477659531187484e-05, "epoch": 1.1610169491525424, "percentage": 16.59, "elapsed_time": "3:00:31", "remaining_time": "15:07:52"} +{"current_steps": 690, "total_steps": 4130, "loss": 0.329, "lr": 3.9458295448306134e-05, "epoch": 1.1694915254237288, "percentage": 16.71, "elapsed_time": "3:01:46", "remaining_time": "15:06:14"} +{"current_steps": 695, "total_steps": 4130, "loss": 0.3561, "lr": 3.943858386224825e-05, "epoch": 1.1779661016949152, "percentage": 16.83, "elapsed_time": "3:03:05", "remaining_time": "15:04:56"} +{"current_steps": 700, "total_steps": 4130, "loss": 0.3091, "lr": 3.941852512504052e-05, "epoch": 1.1864406779661016, "percentage": 16.95, "elapsed_time": "3:04:22", "remaining_time": "15:03:24"} +{"current_steps": 705, "total_steps": 4130, "loss": 0.3206, "lr": 3.939811959490931e-05, "epoch": 1.194915254237288, "percentage": 17.07, "elapsed_time": "3:05:38", "remaining_time": "15:01:50"} +{"current_steps": 710, "total_steps": 4130, "loss": 0.3178, "lr": 3.937736763627435e-05, "epoch": 1.2033898305084745, "percentage": 17.19, "elapsed_time": "3:06:55", "remaining_time": "15:00:23"} +{"current_steps": 715, "total_steps": 4130, "loss": 0.3624, "lr": 3.935626961974217e-05, "epoch": 1.211864406779661, "percentage": 17.31, "elapsed_time": "3:08:15", "remaining_time": "14:59:10"} +{"current_steps": 720, "total_steps": 4130, "loss": 0.3332, "lr": 3.933482592209951e-05, "epoch": 1.2203389830508475, "percentage": 17.43, "elapsed_time": "3:09:31", "remaining_time": "14:57:36"} +{"current_steps": 725, "total_steps": 4130, "loss": 0.3116, "lr": 3.931303692630661e-05, "epoch": 1.228813559322034, "percentage": 17.55, "elapsed_time": "3:10:51", "remaining_time": "14:56:23"} +{"current_steps": 730, "total_steps": 4130, "loss": 0.3284, "lr": 3.92909030214903e-05, "epoch": 1.2372881355932204, "percentage": 17.68, "elapsed_time": "3:12:10", "remaining_time": "14:55:04"} +{"current_steps": 735, "total_steps": 4130, "loss": 0.3215, "lr": 3.9268424602937124e-05, "epoch": 1.2457627118644068, "percentage": 17.8, "elapsed_time": "3:13:29", "remaining_time": "14:53:43"} +{"current_steps": 740, "total_steps": 4130, "loss": 0.3755, "lr": 3.9245602072086246e-05, "epoch": 1.2542372881355932, "percentage": 17.92, "elapsed_time": "3:14:44", "remaining_time": "14:52:05"} +{"current_steps": 745, "total_steps": 4130, "loss": 0.3391, "lr": 3.9222435836522286e-05, "epoch": 1.2627118644067796, "percentage": 18.04, "elapsed_time": "3:16:04", "remaining_time": "14:50:53"} +{"current_steps": 750, "total_steps": 4130, "loss": 0.3372, "lr": 3.919892630996804e-05, "epoch": 1.271186440677966, "percentage": 18.16, "elapsed_time": "3:17:20", "remaining_time": "14:49:22"} +{"current_steps": 755, "total_steps": 4130, "loss": 0.307, "lr": 3.9175073912277107e-05, "epoch": 1.2796610169491525, "percentage": 18.28, "elapsed_time": "3:18:39", "remaining_time": "14:48:01"} +{"current_steps": 760, "total_steps": 4130, "loss": 0.3215, "lr": 3.915087906942635e-05, "epoch": 1.288135593220339, "percentage": 18.4, "elapsed_time": "3:19:52", "remaining_time": "14:46:19"} +{"current_steps": 765, "total_steps": 4130, "loss": 0.3504, "lr": 3.912634221350834e-05, "epoch": 1.2966101694915255, "percentage": 18.52, "elapsed_time": "3:21:10", "remaining_time": "14:44:56"} +{"current_steps": 770, "total_steps": 4130, "loss": 0.3183, "lr": 3.910146378272361e-05, "epoch": 1.305084745762712, "percentage": 18.64, "elapsed_time": "3:22:27", "remaining_time": "14:43:27"} +{"current_steps": 775, "total_steps": 4130, "loss": 0.323, "lr": 3.9076244221372824e-05, "epoch": 1.3135593220338984, "percentage": 18.77, "elapsed_time": "3:23:44", "remaining_time": "14:41:58"} +{"current_steps": 780, "total_steps": 4130, "loss": 0.3181, "lr": 3.905068397984886e-05, "epoch": 1.3220338983050848, "percentage": 18.89, "elapsed_time": "3:25:03", "remaining_time": "14:40:43"} +{"current_steps": 785, "total_steps": 4130, "loss": 0.3586, "lr": 3.9024783514628765e-05, "epoch": 1.3305084745762712, "percentage": 19.01, "elapsed_time": "3:26:22", "remaining_time": "14:39:25"} +{"current_steps": 790, "total_steps": 4130, "loss": 0.3441, "lr": 3.899854328826559e-05, "epoch": 1.3389830508474576, "percentage": 19.13, "elapsed_time": "3:27:38", "remaining_time": "14:37:51"} +{"current_steps": 795, "total_steps": 4130, "loss": 0.3201, "lr": 3.897196376938013e-05, "epoch": 1.347457627118644, "percentage": 19.25, "elapsed_time": "3:28:55", "remaining_time": "14:36:24"} +{"current_steps": 800, "total_steps": 4130, "loss": 0.3419, "lr": 3.8945045432652575e-05, "epoch": 1.3559322033898304, "percentage": 19.37, "elapsed_time": "3:30:09", "remaining_time": "14:34:46"} +{"current_steps": 805, "total_steps": 4130, "loss": 0.3191, "lr": 3.8917788758814e-05, "epoch": 1.3644067796610169, "percentage": 19.49, "elapsed_time": "3:31:27", "remaining_time": "14:33:23"} +{"current_steps": 810, "total_steps": 4130, "loss": 0.3171, "lr": 3.889019423463783e-05, "epoch": 1.3728813559322033, "percentage": 19.61, "elapsed_time": "3:32:46", "remaining_time": "14:32:05"} +{"current_steps": 815, "total_steps": 4130, "loss": 0.3312, "lr": 3.8862262352931075e-05, "epoch": 1.3813559322033897, "percentage": 19.73, "elapsed_time": "3:34:03", "remaining_time": "14:30:42"} +{"current_steps": 820, "total_steps": 4130, "loss": 0.3178, "lr": 3.88339936125256e-05, "epoch": 1.3898305084745763, "percentage": 19.85, "elapsed_time": "3:35:24", "remaining_time": "14:29:31"} +{"current_steps": 825, "total_steps": 4130, "loss": 0.2978, "lr": 3.8805388518269184e-05, "epoch": 1.3983050847457628, "percentage": 19.98, "elapsed_time": "3:36:42", "remaining_time": "14:28:09"} +{"current_steps": 830, "total_steps": 4130, "loss": 0.3173, "lr": 3.877644758101648e-05, "epoch": 1.4067796610169492, "percentage": 20.1, "elapsed_time": "3:38:03", "remaining_time": "14:26:57"} +{"current_steps": 835, "total_steps": 4130, "loss": 0.3159, "lr": 3.8747171317619955e-05, "epoch": 1.4152542372881356, "percentage": 20.22, "elapsed_time": "3:39:14", "remaining_time": "14:25:10"} +{"current_steps": 840, "total_steps": 4130, "loss": 0.3505, "lr": 3.871756025092059e-05, "epoch": 1.423728813559322, "percentage": 20.34, "elapsed_time": "3:40:34", "remaining_time": "14:23:53"} +{"current_steps": 845, "total_steps": 4130, "loss": 0.3087, "lr": 3.868761490973859e-05, "epoch": 1.4322033898305084, "percentage": 20.46, "elapsed_time": "3:41:53", "remaining_time": "14:22:37"} +{"current_steps": 850, "total_steps": 4130, "loss": 0.3337, "lr": 3.8657335828863924e-05, "epoch": 1.4406779661016949, "percentage": 20.58, "elapsed_time": "3:43:13", "remaining_time": "14:21:24"} +{"current_steps": 855, "total_steps": 4130, "loss": 0.3289, "lr": 3.8626723549046774e-05, "epoch": 1.4491525423728815, "percentage": 20.7, "elapsed_time": "3:44:30", "remaining_time": "14:19:56"} +{"current_steps": 860, "total_steps": 4130, "loss": 0.2938, "lr": 3.859577861698787e-05, "epoch": 1.457627118644068, "percentage": 20.82, "elapsed_time": "3:45:38", "remaining_time": "14:17:59"} +{"current_steps": 865, "total_steps": 4130, "loss": 0.3247, "lr": 3.856450158532875e-05, "epoch": 1.4661016949152543, "percentage": 20.94, "elapsed_time": "3:46:57", "remaining_time": "14:16:40"} +{"current_steps": 870, "total_steps": 4130, "loss": 0.3374, "lr": 3.853289301264187e-05, "epoch": 1.4745762711864407, "percentage": 21.07, "elapsed_time": "3:48:12", "remaining_time": "14:15:06"} +{"current_steps": 875, "total_steps": 4130, "loss": 0.3295, "lr": 3.850095346342064e-05, "epoch": 1.4830508474576272, "percentage": 21.19, "elapsed_time": "3:49:29", "remaining_time": "14:13:42"} +{"current_steps": 880, "total_steps": 4130, "loss": 0.3117, "lr": 3.84686835080693e-05, "epoch": 1.4915254237288136, "percentage": 21.31, "elapsed_time": "3:50:45", "remaining_time": "14:12:15"} +{"current_steps": 885, "total_steps": 4130, "loss": 0.329, "lr": 3.843608372289283e-05, "epoch": 1.5, "percentage": 21.43, "elapsed_time": "3:52:04", "remaining_time": "14:10:56"} +{"current_steps": 890, "total_steps": 4130, "loss": 0.3383, "lr": 3.8403154690086564e-05, "epoch": 1.5084745762711864, "percentage": 21.55, "elapsed_time": "3:53:22", "remaining_time": "14:09:33"} +{"current_steps": 895, "total_steps": 4130, "loss": 0.3245, "lr": 3.836989699772582e-05, "epoch": 1.5169491525423728, "percentage": 21.67, "elapsed_time": "3:54:38", "remaining_time": "14:08:08"} +{"current_steps": 900, "total_steps": 4130, "loss": 0.3638, "lr": 3.8336311239755424e-05, "epoch": 1.5254237288135593, "percentage": 21.79, "elapsed_time": "3:55:59", "remaining_time": "14:06:55"} +{"current_steps": 905, "total_steps": 4130, "loss": 0.3142, "lr": 3.830239801597907e-05, "epoch": 1.5338983050847457, "percentage": 21.91, "elapsed_time": "3:57:15", "remaining_time": "14:05:28"} +{"current_steps": 910, "total_steps": 4130, "loss": 0.3029, "lr": 3.826815793204863e-05, "epoch": 1.542372881355932, "percentage": 22.03, "elapsed_time": "3:58:34", "remaining_time": "14:04:11"} +{"current_steps": 915, "total_steps": 4130, "loss": 0.3208, "lr": 3.823359159945332e-05, "epoch": 1.5508474576271185, "percentage": 22.15, "elapsed_time": "3:59:53", "remaining_time": "14:02:53"} +{"current_steps": 920, "total_steps": 4130, "loss": 0.2998, "lr": 3.8198699635508805e-05, "epoch": 1.559322033898305, "percentage": 22.28, "elapsed_time": "4:01:10", "remaining_time": "14:01:29"} +{"current_steps": 925, "total_steps": 4130, "loss": 0.3202, "lr": 3.816348266334614e-05, "epoch": 1.5677966101694916, "percentage": 22.4, "elapsed_time": "4:02:30", "remaining_time": "14:00:15"} +{"current_steps": 930, "total_steps": 4130, "loss": 0.3326, "lr": 3.812794131190066e-05, "epoch": 1.576271186440678, "percentage": 22.52, "elapsed_time": "4:03:50", "remaining_time": "13:59:01"} +{"current_steps": 935, "total_steps": 4130, "loss": 0.3042, "lr": 3.8092076215900765e-05, "epoch": 1.5847457627118644, "percentage": 22.64, "elapsed_time": "4:05:06", "remaining_time": "13:57:32"} +{"current_steps": 940, "total_steps": 4130, "loss": 0.3207, "lr": 3.805588801585654e-05, "epoch": 1.5932203389830508, "percentage": 22.76, "elapsed_time": "4:06:26", "remaining_time": "13:56:19"} +{"current_steps": 945, "total_steps": 4130, "loss": 0.3457, "lr": 3.801937735804838e-05, "epoch": 1.6016949152542372, "percentage": 22.88, "elapsed_time": "4:07:46", "remaining_time": "13:55:05"} +{"current_steps": 950, "total_steps": 4130, "loss": 0.3241, "lr": 3.798254489451539e-05, "epoch": 1.6101694915254239, "percentage": 23.0, "elapsed_time": "4:09:05", "remaining_time": "13:53:49"} +{"current_steps": 955, "total_steps": 4130, "loss": 0.3342, "lr": 3.794539128304374e-05, "epoch": 1.6186440677966103, "percentage": 23.12, "elapsed_time": "4:10:24", "remaining_time": "13:52:31"} +{"current_steps": 960, "total_steps": 4130, "loss": 0.3383, "lr": 3.790791718715498e-05, "epoch": 1.6271186440677967, "percentage": 23.24, "elapsed_time": "4:11:45", "remaining_time": "13:51:19"} +{"current_steps": 965, "total_steps": 4130, "loss": 0.308, "lr": 3.7870123276094134e-05, "epoch": 1.6355932203389831, "percentage": 23.37, "elapsed_time": "4:13:03", "remaining_time": "13:49:58"} +{"current_steps": 970, "total_steps": 4130, "loss": 0.3088, "lr": 3.783201022481775e-05, "epoch": 1.6440677966101696, "percentage": 23.49, "elapsed_time": "4:14:16", "remaining_time": "13:48:20"} +{"current_steps": 975, "total_steps": 4130, "loss": 0.2726, "lr": 3.7793578713981876e-05, "epoch": 1.652542372881356, "percentage": 23.61, "elapsed_time": "4:15:28", "remaining_time": "13:46:43"} +{"current_steps": 980, "total_steps": 4130, "loss": 0.2978, "lr": 3.775482942992989e-05, "epoch": 1.6610169491525424, "percentage": 23.73, "elapsed_time": "4:16:39", "remaining_time": "13:44:59"} +{"current_steps": 985, "total_steps": 4130, "loss": 0.3333, "lr": 3.771576306468022e-05, "epoch": 1.6694915254237288, "percentage": 23.85, "elapsed_time": "4:17:58", "remaining_time": "13:43:41"} +{"current_steps": 990, "total_steps": 4130, "loss": 0.3329, "lr": 3.7676380315914025e-05, "epoch": 1.6779661016949152, "percentage": 23.97, "elapsed_time": "4:19:16", "remaining_time": "13:42:19"} +{"current_steps": 995, "total_steps": 4130, "loss": 0.3682, "lr": 3.7636681886962724e-05, "epoch": 1.6864406779661016, "percentage": 24.09, "elapsed_time": "4:20:36", "remaining_time": "13:41:07"} +{"current_steps": 1000, "total_steps": 4130, "loss": 0.3052, "lr": 3.759666848679539e-05, "epoch": 1.694915254237288, "percentage": 24.21, "elapsed_time": "4:21:53", "remaining_time": "13:39:44"} +{"current_steps": 1005, "total_steps": 4130, "loss": 0.331, "lr": 3.755634083000617e-05, "epoch": 1.7033898305084745, "percentage": 24.33, "elapsed_time": "4:23:13", "remaining_time": "13:38:29"} +{"current_steps": 1010, "total_steps": 4130, "loss": 0.2965, "lr": 3.751569963680146e-05, "epoch": 1.711864406779661, "percentage": 24.46, "elapsed_time": "4:24:31", "remaining_time": "13:37:09"} +{"current_steps": 1015, "total_steps": 4130, "loss": 0.3306, "lr": 3.747474563298705e-05, "epoch": 1.7203389830508473, "percentage": 24.58, "elapsed_time": "4:25:50", "remaining_time": "13:35:51"} +{"current_steps": 1020, "total_steps": 4130, "loss": 0.3287, "lr": 3.743347954995519e-05, "epoch": 1.7288135593220337, "percentage": 24.7, "elapsed_time": "4:27:07", "remaining_time": "13:34:28"} +{"current_steps": 1025, "total_steps": 4130, "loss": 0.3264, "lr": 3.7391902124671516e-05, "epoch": 1.7372881355932204, "percentage": 24.82, "elapsed_time": "4:28:27", "remaining_time": "13:33:14"} +{"current_steps": 1030, "total_steps": 4130, "loss": 0.3084, "lr": 3.7350014099661874e-05, "epoch": 1.7457627118644068, "percentage": 24.94, "elapsed_time": "4:29:46", "remaining_time": "13:31:56"} +{"current_steps": 1035, "total_steps": 4130, "loss": 0.3232, "lr": 3.730781622299907e-05, "epoch": 1.7542372881355932, "percentage": 25.06, "elapsed_time": "4:31:02", "remaining_time": "13:30:31"} +{"current_steps": 1040, "total_steps": 4130, "loss": 0.3472, "lr": 3.7265309248289516e-05, "epoch": 1.7627118644067796, "percentage": 25.18, "elapsed_time": "4:32:17", "remaining_time": "13:29:02"} +{"current_steps": 1045, "total_steps": 4130, "loss": 0.3486, "lr": 3.7222493934659765e-05, "epoch": 1.7711864406779663, "percentage": 25.3, "elapsed_time": "4:33:38", "remaining_time": "13:27:49"} +{"current_steps": 1050, "total_steps": 4130, "loss": 0.2993, "lr": 3.717937104674296e-05, "epoch": 1.7796610169491527, "percentage": 25.42, "elapsed_time": "4:34:55", "remaining_time": "13:26:27"} +{"current_steps": 1055, "total_steps": 4130, "loss": 0.3245, "lr": 3.713594135466515e-05, "epoch": 1.788135593220339, "percentage": 25.54, "elapsed_time": "4:36:11", "remaining_time": "13:25:01"} +{"current_steps": 1060, "total_steps": 4130, "loss": 0.3495, "lr": 3.7092205634031595e-05, "epoch": 1.7966101694915255, "percentage": 25.67, "elapsed_time": "4:37:26", "remaining_time": "13:23:31"} +{"current_steps": 1065, "total_steps": 4130, "loss": 0.3181, "lr": 3.704816466591286e-05, "epoch": 1.805084745762712, "percentage": 25.79, "elapsed_time": "4:38:43", "remaining_time": "13:22:07"} +{"current_steps": 1070, "total_steps": 4130, "loss": 0.3383, "lr": 3.700381923683087e-05, "epoch": 1.8135593220338984, "percentage": 25.91, "elapsed_time": "4:40:03", "remaining_time": "13:20:54"} +{"current_steps": 1075, "total_steps": 4130, "loss": 0.2844, "lr": 3.695917013874491e-05, "epoch": 1.8220338983050848, "percentage": 26.03, "elapsed_time": "4:41:23", "remaining_time": "13:19:39"} +{"current_steps": 1080, "total_steps": 4130, "loss": 0.3396, "lr": 3.691421816903744e-05, "epoch": 1.8305084745762712, "percentage": 26.15, "elapsed_time": "4:42:42", "remaining_time": "13:18:22"} +{"current_steps": 1085, "total_steps": 4130, "loss": 0.2903, "lr": 3.686896413049985e-05, "epoch": 1.8389830508474576, "percentage": 26.27, "elapsed_time": "4:44:01", "remaining_time": "13:17:07"} +{"current_steps": 1090, "total_steps": 4130, "loss": 0.3118, "lr": 3.6823408831318154e-05, "epoch": 1.847457627118644, "percentage": 26.39, "elapsed_time": "4:45:22", "remaining_time": "13:15:53"} +{"current_steps": 1095, "total_steps": 4130, "loss": 0.344, "lr": 3.677755308505854e-05, "epoch": 1.8559322033898304, "percentage": 26.51, "elapsed_time": "4:46:39", "remaining_time": "13:14:32"} +{"current_steps": 1100, "total_steps": 4130, "loss": 0.3249, "lr": 3.673139771065282e-05, "epoch": 1.8644067796610169, "percentage": 26.63, "elapsed_time": "4:47:57", "remaining_time": "13:13:12"} +{"current_steps": 1105, "total_steps": 4130, "loss": 0.3469, "lr": 3.6684943532383866e-05, "epoch": 1.8728813559322033, "percentage": 26.76, "elapsed_time": "4:49:16", "remaining_time": "13:11:55"} +{"current_steps": 1110, "total_steps": 4130, "loss": 0.3365, "lr": 3.663819137987081e-05, "epoch": 1.8813559322033897, "percentage": 26.88, "elapsed_time": "4:50:37", "remaining_time": "13:10:42"} +{"current_steps": 1115, "total_steps": 4130, "loss": 0.3122, "lr": 3.65911420880543e-05, "epoch": 1.8898305084745761, "percentage": 27.0, "elapsed_time": "4:51:55", "remaining_time": "13:09:22"} +{"current_steps": 1120, "total_steps": 4130, "loss": 0.3077, "lr": 3.654379649718153e-05, "epoch": 1.8983050847457628, "percentage": 27.12, "elapsed_time": "4:53:10", "remaining_time": "13:07:54"} +{"current_steps": 1125, "total_steps": 4130, "loss": 0.3467, "lr": 3.649615545279127e-05, "epoch": 1.9067796610169492, "percentage": 27.24, "elapsed_time": "4:54:30", "remaining_time": "13:06:40"} +{"current_steps": 1130, "total_steps": 4130, "loss": 0.3308, "lr": 3.644821980569876e-05, "epoch": 1.9152542372881356, "percentage": 27.36, "elapsed_time": "4:55:48", "remaining_time": "13:05:21"} +{"current_steps": 1135, "total_steps": 4130, "loss": 0.312, "lr": 3.639999041198051e-05, "epoch": 1.923728813559322, "percentage": 27.48, "elapsed_time": "4:57:03", "remaining_time": "13:03:52"} +{"current_steps": 1140, "total_steps": 4130, "loss": 0.3366, "lr": 3.635146813295902e-05, "epoch": 1.9322033898305084, "percentage": 27.6, "elapsed_time": "4:58:20", "remaining_time": "13:02:28"} +{"current_steps": 1145, "total_steps": 4130, "loss": 0.3165, "lr": 3.6302653835187366e-05, "epoch": 1.940677966101695, "percentage": 27.72, "elapsed_time": "4:59:34", "remaining_time": "13:00:59"} +{"current_steps": 1150, "total_steps": 4130, "loss": 0.2881, "lr": 3.625354839043377e-05, "epoch": 1.9491525423728815, "percentage": 27.85, "elapsed_time": "5:00:53", "remaining_time": "12:59:43"} +{"current_steps": 1155, "total_steps": 4130, "loss": 0.3325, "lr": 3.620415267566601e-05, "epoch": 1.957627118644068, "percentage": 27.97, "elapsed_time": "5:02:14", "remaining_time": "12:58:29"} +{"current_steps": 1160, "total_steps": 4130, "loss": 0.324, "lr": 3.615446757303575e-05, "epoch": 1.9661016949152543, "percentage": 28.09, "elapsed_time": "5:03:34", "remaining_time": "12:57:15"} +{"current_steps": 1165, "total_steps": 4130, "loss": 0.3102, "lr": 3.610449396986281e-05, "epoch": 1.9745762711864407, "percentage": 28.21, "elapsed_time": "5:04:50", "remaining_time": "12:55:50"} +{"current_steps": 1170, "total_steps": 4130, "loss": 0.3178, "lr": 3.6054232758619274e-05, "epoch": 1.9830508474576272, "percentage": 28.33, "elapsed_time": "5:06:05", "remaining_time": "12:54:24"} +{"current_steps": 1175, "total_steps": 4130, "loss": 0.3299, "lr": 3.600368483691361e-05, "epoch": 1.9915254237288136, "percentage": 28.45, "elapsed_time": "5:07:26", "remaining_time": "12:53:10"} +{"current_steps": 1180, "total_steps": 4130, "loss": 0.3012, "lr": 3.59528511074746e-05, "epoch": 2.0, "percentage": 28.57, "elapsed_time": "5:08:43", "remaining_time": "12:51:49"} +{"current_steps": 1185, "total_steps": 4130, "loss": 0.3156, "lr": 3.5901732478135235e-05, "epoch": 2.0084745762711864, "percentage": 28.69, "elapsed_time": "5:09:59", "remaining_time": "12:50:25"} +{"current_steps": 1190, "total_steps": 4130, "loss": 0.2836, "lr": 3.5850329861816475e-05, "epoch": 2.016949152542373, "percentage": 28.81, "elapsed_time": "5:11:14", "remaining_time": "12:48:56"} +{"current_steps": 1195, "total_steps": 4130, "loss": 0.2993, "lr": 3.5798644176511e-05, "epoch": 2.0254237288135593, "percentage": 28.93, "elapsed_time": "5:12:30", "remaining_time": "12:47:31"} +{"current_steps": 1200, "total_steps": 4130, "loss": 0.3142, "lr": 3.574667634526676e-05, "epoch": 2.0338983050847457, "percentage": 29.06, "elapsed_time": "5:13:45", "remaining_time": "12:46:05"} +{"current_steps": 1205, "total_steps": 4130, "loss": 0.3068, "lr": 3.5694427296170514e-05, "epoch": 2.042372881355932, "percentage": 29.18, "elapsed_time": "5:15:01", "remaining_time": "12:44:42"} +{"current_steps": 1210, "total_steps": 4130, "loss": 0.316, "lr": 3.564189796233125e-05, "epoch": 2.0508474576271185, "percentage": 29.3, "elapsed_time": "5:16:21", "remaining_time": "12:43:27"} +{"current_steps": 1215, "total_steps": 4130, "loss": 0.2988, "lr": 3.5589089281863547e-05, "epoch": 2.059322033898305, "percentage": 29.42, "elapsed_time": "5:17:37", "remaining_time": "12:42:02"} +{"current_steps": 1220, "total_steps": 4130, "loss": 0.2926, "lr": 3.553600219787079e-05, "epoch": 2.0677966101694913, "percentage": 29.54, "elapsed_time": "5:18:56", "remaining_time": "12:40:44"} +{"current_steps": 1225, "total_steps": 4130, "loss": 0.2965, "lr": 3.5482637658428315e-05, "epoch": 2.0762711864406778, "percentage": 29.66, "elapsed_time": "5:20:13", "remaining_time": "12:39:22"} +{"current_steps": 1230, "total_steps": 4130, "loss": 0.3234, "lr": 3.542899661656653e-05, "epoch": 2.084745762711864, "percentage": 29.78, "elapsed_time": "5:21:33", "remaining_time": "12:38:09"} +{"current_steps": 1235, "total_steps": 4130, "loss": 0.3451, "lr": 3.5375080030253826e-05, "epoch": 2.093220338983051, "percentage": 29.9, "elapsed_time": "5:22:54", "remaining_time": "12:36:55"} +{"current_steps": 1240, "total_steps": 4130, "loss": 0.304, "lr": 3.532088886237956e-05, "epoch": 2.1016949152542375, "percentage": 30.02, "elapsed_time": "5:24:11", "remaining_time": "12:35:35"} +{"current_steps": 1245, "total_steps": 4130, "loss": 0.3092, "lr": 3.5266424080736765e-05, "epoch": 2.110169491525424, "percentage": 30.15, "elapsed_time": "5:25:30", "remaining_time": "12:34:17"} +{"current_steps": 1250, "total_steps": 4130, "loss": 0.3236, "lr": 3.521168665800491e-05, "epoch": 2.1186440677966103, "percentage": 30.27, "elapsed_time": "5:26:50", "remaining_time": "12:33:03"} +{"current_steps": 1255, "total_steps": 4130, "loss": 0.2971, "lr": 3.515667757173254e-05, "epoch": 2.1271186440677967, "percentage": 30.39, "elapsed_time": "5:28:09", "remaining_time": "12:31:44"} +{"current_steps": 1260, "total_steps": 4130, "loss": 0.2878, "lr": 3.5101397804319794e-05, "epoch": 2.135593220338983, "percentage": 30.51, "elapsed_time": "5:29:25", "remaining_time": "12:30:20"} +{"current_steps": 1265, "total_steps": 4130, "loss": 0.2871, "lr": 3.5045848343000876e-05, "epoch": 2.1440677966101696, "percentage": 30.63, "elapsed_time": "5:30:41", "remaining_time": "12:28:56"} +{"current_steps": 1270, "total_steps": 4130, "loss": 0.3271, "lr": 3.4990030179826414e-05, "epoch": 2.152542372881356, "percentage": 30.75, "elapsed_time": "5:31:57", "remaining_time": "12:27:32"} +{"current_steps": 1275, "total_steps": 4130, "loss": 0.2948, "lr": 3.493394431164576e-05, "epoch": 2.1610169491525424, "percentage": 30.87, "elapsed_time": "5:33:16", "remaining_time": "12:26:17"} +{"current_steps": 1280, "total_steps": 4130, "loss": 0.3051, "lr": 3.4877591740089144e-05, "epoch": 2.169491525423729, "percentage": 30.99, "elapsed_time": "5:34:37", "remaining_time": "12:25:03"} +{"current_steps": 1285, "total_steps": 4130, "loss": 0.2968, "lr": 3.482097347154986e-05, "epoch": 2.1779661016949152, "percentage": 31.11, "elapsed_time": "5:35:52", "remaining_time": "12:23:38"} +{"current_steps": 1290, "total_steps": 4130, "loss": 0.3091, "lr": 3.476409051716621e-05, "epoch": 2.1864406779661016, "percentage": 31.23, "elapsed_time": "5:37:11", "remaining_time": "12:22:20"} +{"current_steps": 1295, "total_steps": 4130, "loss": 0.2911, "lr": 3.470694389280352e-05, "epoch": 2.194915254237288, "percentage": 31.36, "elapsed_time": "5:38:25", "remaining_time": "12:20:51"} +{"current_steps": 1300, "total_steps": 4130, "loss": 0.283, "lr": 3.464953461903593e-05, "epoch": 2.2033898305084745, "percentage": 31.48, "elapsed_time": "5:39:39", "remaining_time": "12:19:24"} +{"current_steps": 1305, "total_steps": 4130, "loss": 0.308, "lr": 3.459186372112824e-05, "epoch": 2.211864406779661, "percentage": 31.6, "elapsed_time": "5:40:54", "remaining_time": "12:17:58"} +{"current_steps": 1310, "total_steps": 4130, "loss": 0.3253, "lr": 3.453393222901753e-05, "epoch": 2.2203389830508473, "percentage": 31.72, "elapsed_time": "5:42:12", "remaining_time": "12:16:39"} +{"current_steps": 1315, "total_steps": 4130, "loss": 0.3152, "lr": 3.44757411772948e-05, "epoch": 2.2288135593220337, "percentage": 31.84, "elapsed_time": "5:43:31", "remaining_time": "12:15:22"} +{"current_steps": 1320, "total_steps": 4130, "loss": 0.3282, "lr": 3.441729160518652e-05, "epoch": 2.23728813559322, "percentage": 31.96, "elapsed_time": "5:44:50", "remaining_time": "12:14:05"} +{"current_steps": 1325, "total_steps": 4130, "loss": 0.3084, "lr": 3.435858455653601e-05, "epoch": 2.2457627118644066, "percentage": 32.08, "elapsed_time": "5:46:10", "remaining_time": "12:12:51"} +{"current_steps": 1330, "total_steps": 4130, "loss": 0.3119, "lr": 3.429962107978485e-05, "epoch": 2.2542372881355934, "percentage": 32.2, "elapsed_time": "5:47:29", "remaining_time": "12:11:34"} +{"current_steps": 1335, "total_steps": 4130, "loss": 0.303, "lr": 3.424040222795412e-05, "epoch": 2.26271186440678, "percentage": 32.32, "elapsed_time": "5:48:47", "remaining_time": "12:10:14"} +{"current_steps": 1340, "total_steps": 4130, "loss": 0.294, "lr": 3.418092905862563e-05, "epoch": 2.2711864406779663, "percentage": 32.45, "elapsed_time": "5:50:06", "remaining_time": "12:08:56"} +{"current_steps": 1345, "total_steps": 4130, "loss": 0.2951, "lr": 3.412120263392301e-05, "epoch": 2.2796610169491527, "percentage": 32.57, "elapsed_time": "5:51:25", "remaining_time": "12:07:41"} +{"current_steps": 1350, "total_steps": 4130, "loss": 0.3212, "lr": 3.406122402049272e-05, "epoch": 2.288135593220339, "percentage": 32.69, "elapsed_time": "5:52:44", "remaining_time": "12:06:23"} +{"current_steps": 1355, "total_steps": 4130, "loss": 0.2984, "lr": 3.400099428948505e-05, "epoch": 2.2966101694915255, "percentage": 32.81, "elapsed_time": "5:54:04", "remaining_time": "12:05:08"} +{"current_steps": 1360, "total_steps": 4130, "loss": 0.3003, "lr": 3.394051451653496e-05, "epoch": 2.305084745762712, "percentage": 32.93, "elapsed_time": "5:55:23", "remaining_time": "12:03:50"} +{"current_steps": 1365, "total_steps": 4130, "loss": 0.3187, "lr": 3.3879785781742875e-05, "epoch": 2.3135593220338984, "percentage": 33.05, "elapsed_time": "5:56:41", "remaining_time": "12:02:31"} +{"current_steps": 1370, "total_steps": 4130, "loss": 0.2955, "lr": 3.381880916965541e-05, "epoch": 2.3220338983050848, "percentage": 33.17, "elapsed_time": "5:57:55", "remaining_time": "12:01:04"} +{"current_steps": 1375, "total_steps": 4130, "loss": 0.3139, "lr": 3.3757585769245964e-05, "epoch": 2.330508474576271, "percentage": 33.29, "elapsed_time": "5:59:15", "remaining_time": "11:59:49"} +{"current_steps": 1380, "total_steps": 4130, "loss": 0.3236, "lr": 3.3696116673895314e-05, "epoch": 2.3389830508474576, "percentage": 33.41, "elapsed_time": "6:00:36", "remaining_time": "11:58:35"} +{"current_steps": 1385, "total_steps": 4130, "loss": 0.3091, "lr": 3.363440298137207e-05, "epoch": 2.347457627118644, "percentage": 33.54, "elapsed_time": "6:01:52", "remaining_time": "11:57:13"} +{"current_steps": 1390, "total_steps": 4130, "loss": 0.2928, "lr": 3.357244579381306e-05, "epoch": 2.3559322033898304, "percentage": 33.66, "elapsed_time": "6:03:11", "remaining_time": "11:55:55"} +{"current_steps": 1395, "total_steps": 4130, "loss": 0.3004, "lr": 3.351024621770369e-05, "epoch": 2.364406779661017, "percentage": 33.78, "elapsed_time": "6:04:25", "remaining_time": "11:54:28"} +{"current_steps": 1400, "total_steps": 4130, "loss": 0.2874, "lr": 3.34478053638581e-05, "epoch": 2.3728813559322033, "percentage": 33.9, "elapsed_time": "6:05:44", "remaining_time": "11:53:11"} +{"current_steps": 1405, "total_steps": 4130, "loss": 0.2771, "lr": 3.3385124347399415e-05, "epoch": 2.3813559322033897, "percentage": 34.02, "elapsed_time": "6:07:01", "remaining_time": "11:51:50"} +{"current_steps": 1410, "total_steps": 4130, "loss": 0.3024, "lr": 3.33222042877398e-05, "epoch": 2.389830508474576, "percentage": 34.14, "elapsed_time": "6:08:20", "remaining_time": "11:50:32"} +{"current_steps": 1415, "total_steps": 4130, "loss": 0.2977, "lr": 3.325904630856042e-05, "epoch": 2.3983050847457625, "percentage": 34.26, "elapsed_time": "6:09:40", "remaining_time": "11:49:17"} +{"current_steps": 1420, "total_steps": 4130, "loss": 0.3493, "lr": 3.319565153779146e-05, "epoch": 2.406779661016949, "percentage": 34.38, "elapsed_time": "6:10:55", "remaining_time": "11:47:53"} +{"current_steps": 1425, "total_steps": 4130, "loss": 0.3226, "lr": 3.31320211075919e-05, "epoch": 2.415254237288136, "percentage": 34.5, "elapsed_time": "6:12:11", "remaining_time": "11:46:31"} +{"current_steps": 1430, "total_steps": 4130, "loss": 0.3404, "lr": 3.306815615432936e-05, "epoch": 2.423728813559322, "percentage": 34.62, "elapsed_time": "6:13:32", "remaining_time": "11:45:17"} +{"current_steps": 1435, "total_steps": 4130, "loss": 0.2958, "lr": 3.300405781855975e-05, "epoch": 2.4322033898305087, "percentage": 34.75, "elapsed_time": "6:14:48", "remaining_time": "11:43:54"} +{"current_steps": 1440, "total_steps": 4130, "loss": 0.3017, "lr": 3.2939727245006956e-05, "epoch": 2.440677966101695, "percentage": 34.87, "elapsed_time": "6:16:05", "remaining_time": "11:42:33"} +{"current_steps": 1445, "total_steps": 4130, "loss": 0.3066, "lr": 3.2875165582542347e-05, "epoch": 2.4491525423728815, "percentage": 34.99, "elapsed_time": "6:17:23", "remaining_time": "11:41:14"} +{"current_steps": 1450, "total_steps": 4130, "loss": 0.2713, "lr": 3.281037398416427e-05, "epoch": 2.457627118644068, "percentage": 35.11, "elapsed_time": "6:18:41", "remaining_time": "11:39:54"} +{"current_steps": 1455, "total_steps": 4130, "loss": 0.3181, "lr": 3.2745353606977506e-05, "epoch": 2.4661016949152543, "percentage": 35.23, "elapsed_time": "6:19:58", "remaining_time": "11:38:34"} +{"current_steps": 1460, "total_steps": 4130, "loss": 0.3097, "lr": 3.268010561217253e-05, "epoch": 2.4745762711864407, "percentage": 35.35, "elapsed_time": "6:21:14", "remaining_time": "11:37:12"} +{"current_steps": 1465, "total_steps": 4130, "loss": 0.2961, "lr": 3.261463116500483e-05, "epoch": 2.483050847457627, "percentage": 35.47, "elapsed_time": "6:22:32", "remaining_time": "11:35:53"} +{"current_steps": 1470, "total_steps": 4130, "loss": 0.2933, "lr": 3.254893143477408e-05, "epoch": 2.4915254237288136, "percentage": 35.59, "elapsed_time": "6:23:49", "remaining_time": "11:34:32"} +{"current_steps": 1475, "total_steps": 4130, "loss": 0.3205, "lr": 3.248300759480323e-05, "epoch": 2.5, "percentage": 35.71, "elapsed_time": "6:25:09", "remaining_time": "11:33:17"} +{"current_steps": 1480, "total_steps": 4130, "loss": 0.2873, "lr": 3.241686082241761e-05, "epoch": 2.5084745762711864, "percentage": 35.84, "elapsed_time": "6:26:30", "remaining_time": "11:32:02"} +{"current_steps": 1485, "total_steps": 4130, "loss": 0.2667, "lr": 3.235049229892384e-05, "epoch": 2.516949152542373, "percentage": 35.96, "elapsed_time": "6:27:48", "remaining_time": "11:30:44"} +{"current_steps": 1490, "total_steps": 4130, "loss": 0.3194, "lr": 3.2283903209588795e-05, "epoch": 2.5254237288135593, "percentage": 36.08, "elapsed_time": "6:29:07", "remaining_time": "11:29:28"} +{"current_steps": 1495, "total_steps": 4130, "loss": 0.3028, "lr": 3.221709474361838e-05, "epoch": 2.5338983050847457, "percentage": 36.2, "elapsed_time": "6:30:23", "remaining_time": "11:28:04"} +{"current_steps": 1500, "total_steps": 4130, "loss": 0.3209, "lr": 3.215006809413633e-05, "epoch": 2.542372881355932, "percentage": 36.32, "elapsed_time": "6:31:43", "remaining_time": "11:26:49"} +{"current_steps": 1505, "total_steps": 4130, "loss": 0.2896, "lr": 3.20828244581629e-05, "epoch": 2.5508474576271185, "percentage": 36.44, "elapsed_time": "6:33:40", "remaining_time": "11:26:38"} +{"current_steps": 1510, "total_steps": 4130, "loss": 0.3325, "lr": 3.2015365036593466e-05, "epoch": 2.559322033898305, "percentage": 36.56, "elapsed_time": "6:35:01", "remaining_time": "11:25:23"} +{"current_steps": 1515, "total_steps": 4130, "loss": 0.2919, "lr": 3.194769103417709e-05, "epoch": 2.5677966101694913, "percentage": 36.68, "elapsed_time": "6:36:19", "remaining_time": "11:24:05"} +{"current_steps": 1520, "total_steps": 4130, "loss": 0.2994, "lr": 3.1879803659495015e-05, "epoch": 2.576271186440678, "percentage": 36.8, "elapsed_time": "6:37:39", "remaining_time": "11:22:49"} +{"current_steps": 1525, "total_steps": 4130, "loss": 0.2972, "lr": 3.181170412493906e-05, "epoch": 2.584745762711864, "percentage": 36.92, "elapsed_time": "6:38:56", "remaining_time": "11:21:28"} +{"current_steps": 1530, "total_steps": 4130, "loss": 0.2709, "lr": 3.1743393646689996e-05, "epoch": 2.593220338983051, "percentage": 37.05, "elapsed_time": "6:40:10", "remaining_time": "11:20:02"} +{"current_steps": 1535, "total_steps": 4130, "loss": 0.2859, "lr": 3.1674873444695804e-05, "epoch": 2.601694915254237, "percentage": 37.17, "elapsed_time": "6:41:30", "remaining_time": "11:18:46"} +{"current_steps": 1540, "total_steps": 4130, "loss": 0.2991, "lr": 3.1606144742649875e-05, "epoch": 2.610169491525424, "percentage": 37.29, "elapsed_time": "6:42:50", "remaining_time": "11:17:30"} +{"current_steps": 1545, "total_steps": 4130, "loss": 0.271, "lr": 3.15372087679692e-05, "epoch": 2.6186440677966103, "percentage": 37.41, "elapsed_time": "6:44:06", "remaining_time": "11:16:07"} +{"current_steps": 1550, "total_steps": 4130, "loss": 0.3008, "lr": 3.1468066751772415e-05, "epoch": 2.6271186440677967, "percentage": 37.53, "elapsed_time": "6:45:23", "remaining_time": "11:14:47"} +{"current_steps": 1555, "total_steps": 4130, "loss": 0.2868, "lr": 3.1398719928857834e-05, "epoch": 2.635593220338983, "percentage": 37.65, "elapsed_time": "6:46:37", "remaining_time": "11:13:20"} +{"current_steps": 1560, "total_steps": 4130, "loss": 0.2862, "lr": 3.132916953768137e-05, "epoch": 2.6440677966101696, "percentage": 37.77, "elapsed_time": "6:47:54", "remaining_time": "11:12:00"} +{"current_steps": 1565, "total_steps": 4130, "loss": 0.2933, "lr": 3.1259416820334446e-05, "epoch": 2.652542372881356, "percentage": 37.89, "elapsed_time": "6:49:11", "remaining_time": "11:10:39"} +{"current_steps": 1570, "total_steps": 4130, "loss": 0.2766, "lr": 3.118946302252181e-05, "epoch": 2.6610169491525424, "percentage": 38.01, "elapsed_time": "6:50:26", "remaining_time": "11:09:14"} +{"current_steps": 1575, "total_steps": 4130, "loss": 0.278, "lr": 3.1119309393539256e-05, "epoch": 2.669491525423729, "percentage": 38.14, "elapsed_time": "6:51:42", "remaining_time": "11:07:53"} +{"current_steps": 1580, "total_steps": 4130, "loss": 0.2893, "lr": 3.104895718625136e-05, "epoch": 2.6779661016949152, "percentage": 38.26, "elapsed_time": "6:53:02", "remaining_time": "11:06:37"} +{"current_steps": 1585, "total_steps": 4130, "loss": 0.3102, "lr": 3.097840765706908e-05, "epoch": 2.6864406779661016, "percentage": 38.38, "elapsed_time": "6:54:21", "remaining_time": "11:05:20"} +{"current_steps": 1590, "total_steps": 4130, "loss": 0.3017, "lr": 3.0907662065927297e-05, "epoch": 2.694915254237288, "percentage": 38.5, "elapsed_time": "6:55:40", "remaining_time": "11:04:02"} +{"current_steps": 1595, "total_steps": 4130, "loss": 0.308, "lr": 3.083672167626237e-05, "epoch": 2.7033898305084745, "percentage": 38.62, "elapsed_time": "6:57:01", "remaining_time": "11:02:47"} +{"current_steps": 1600, "total_steps": 4130, "loss": 0.2901, "lr": 3.0765587754989516e-05, "epoch": 2.711864406779661, "percentage": 38.74, "elapsed_time": "6:58:21", "remaining_time": "11:01:31"} +{"current_steps": 1605, "total_steps": 4130, "loss": 0.2903, "lr": 3.069426157248022e-05, "epoch": 2.7203389830508473, "percentage": 38.86, "elapsed_time": "6:59:34", "remaining_time": "11:00:05"} +{"current_steps": 1610, "total_steps": 4130, "loss": 0.3259, "lr": 3.062274440253953e-05, "epoch": 2.7288135593220337, "percentage": 38.98, "elapsed_time": "7:00:54", "remaining_time": "10:58:49"} +{"current_steps": 1615, "total_steps": 4130, "loss": 0.3096, "lr": 3.0551037522383325e-05, "epoch": 2.7372881355932206, "percentage": 39.1, "elapsed_time": "7:02:13", "remaining_time": "10:57:30"} +{"current_steps": 1620, "total_steps": 4130, "loss": 0.3053, "lr": 3.0479142212615457e-05, "epoch": 2.7457627118644066, "percentage": 39.23, "elapsed_time": "7:03:33", "remaining_time": "10:56:14"} +{"current_steps": 1625, "total_steps": 4130, "loss": 0.3126, "lr": 3.0407059757204968e-05, "epoch": 2.7542372881355934, "percentage": 39.35, "elapsed_time": "7:04:54", "remaining_time": "10:55:00"} +{"current_steps": 1630, "total_steps": 4130, "loss": 0.3394, "lr": 3.033479144346308e-05, "epoch": 2.7627118644067794, "percentage": 39.47, "elapsed_time": "7:06:14", "remaining_time": "10:53:44"} +{"current_steps": 1635, "total_steps": 4130, "loss": 0.3046, "lr": 3.0262338562020246e-05, "epoch": 2.7711864406779663, "percentage": 39.59, "elapsed_time": "7:07:34", "remaining_time": "10:52:29"} +{"current_steps": 1640, "total_steps": 4130, "loss": 0.3088, "lr": 3.018970240680308e-05, "epoch": 2.7796610169491527, "percentage": 39.71, "elapsed_time": "7:08:54", "remaining_time": "10:51:12"} +{"current_steps": 1645, "total_steps": 4130, "loss": 0.3076, "lr": 3.011688427501126e-05, "epoch": 2.788135593220339, "percentage": 39.83, "elapsed_time": "7:10:11", "remaining_time": "10:49:52"} +{"current_steps": 1650, "total_steps": 4130, "loss": 0.314, "lr": 3.0043885467094382e-05, "epoch": 2.7966101694915255, "percentage": 39.95, "elapsed_time": "7:11:31", "remaining_time": "10:48:36"} +{"current_steps": 1655, "total_steps": 4130, "loss": 0.3015, "lr": 2.9970707286728676e-05, "epoch": 2.805084745762712, "percentage": 40.07, "elapsed_time": "7:12:50", "remaining_time": "10:47:17"} +{"current_steps": 1660, "total_steps": 4130, "loss": 0.2792, "lr": 2.9897351040793786e-05, "epoch": 2.8135593220338984, "percentage": 40.19, "elapsed_time": "7:14:05", "remaining_time": "10:45:54"} +{"current_steps": 1665, "total_steps": 4130, "loss": 0.2803, "lr": 2.9823818039349407e-05, "epoch": 2.8220338983050848, "percentage": 40.31, "elapsed_time": "7:15:23", "remaining_time": "10:44:35"} +{"current_steps": 1670, "total_steps": 4130, "loss": 0.29, "lr": 2.9750109595611884e-05, "epoch": 2.830508474576271, "percentage": 40.44, "elapsed_time": "7:16:41", "remaining_time": "10:43:16"} +{"current_steps": 1675, "total_steps": 4130, "loss": 0.287, "lr": 2.967622702593074e-05, "epoch": 2.8389830508474576, "percentage": 40.56, "elapsed_time": "7:18:02", "remaining_time": "10:42:01"} +{"current_steps": 1680, "total_steps": 4130, "loss": 0.3084, "lr": 2.9602171649765235e-05, "epoch": 2.847457627118644, "percentage": 40.68, "elapsed_time": "7:19:20", "remaining_time": "10:40:42"} +{"current_steps": 1685, "total_steps": 4130, "loss": 0.2907, "lr": 2.9527944789660732e-05, "epoch": 2.8559322033898304, "percentage": 40.8, "elapsed_time": "7:20:40", "remaining_time": "10:39:25"} +{"current_steps": 1690, "total_steps": 4130, "loss": 0.2887, "lr": 2.9453547771225088e-05, "epoch": 2.864406779661017, "percentage": 40.92, "elapsed_time": "7:21:58", "remaining_time": "10:38:07"} +{"current_steps": 1695, "total_steps": 4130, "loss": 0.2823, "lr": 2.9378981923105026e-05, "epoch": 2.8728813559322033, "percentage": 41.04, "elapsed_time": "7:23:17", "remaining_time": "10:36:49"} +{"current_steps": 1700, "total_steps": 4130, "loss": 0.3211, "lr": 2.930424857696236e-05, "epoch": 2.8813559322033897, "percentage": 41.16, "elapsed_time": "7:24:33", "remaining_time": "10:35:27"} +{"current_steps": 1705, "total_steps": 4130, "loss": 0.3209, "lr": 2.922934906745024e-05, "epoch": 2.889830508474576, "percentage": 41.28, "elapsed_time": "7:25:52", "remaining_time": "10:34:10"} +{"current_steps": 1710, "total_steps": 4130, "loss": 0.3164, "lr": 2.9154284732189285e-05, "epoch": 2.898305084745763, "percentage": 41.4, "elapsed_time": "7:27:11", "remaining_time": "10:32:52"} +{"current_steps": 1715, "total_steps": 4130, "loss": 0.2904, "lr": 2.907905691174374e-05, "epoch": 2.906779661016949, "percentage": 41.53, "elapsed_time": "7:28:32", "remaining_time": "10:31:36"} +{"current_steps": 1720, "total_steps": 4130, "loss": 0.3213, "lr": 2.90036669495975e-05, "epoch": 2.915254237288136, "percentage": 41.65, "elapsed_time": "7:29:50", "remaining_time": "10:30:18"} +{"current_steps": 1725, "total_steps": 4130, "loss": 0.3215, "lr": 2.8928116192130127e-05, "epoch": 2.923728813559322, "percentage": 41.77, "elapsed_time": "7:31:10", "remaining_time": "10:29:01"} +{"current_steps": 1730, "total_steps": 4130, "loss": 0.2815, "lr": 2.8852405988592802e-05, "epoch": 2.9322033898305087, "percentage": 41.89, "elapsed_time": "7:32:28", "remaining_time": "10:27:42"} +{"current_steps": 1735, "total_steps": 4130, "loss": 0.2986, "lr": 2.8776537691084247e-05, "epoch": 2.940677966101695, "percentage": 42.01, "elapsed_time": "7:33:41", "remaining_time": "10:26:16"} +{"current_steps": 1740, "total_steps": 4130, "loss": 0.3046, "lr": 2.8700512654526543e-05, "epoch": 2.9491525423728815, "percentage": 42.13, "elapsed_time": "7:35:00", "remaining_time": "10:24:58"} +{"current_steps": 1745, "total_steps": 4130, "loss": 0.2869, "lr": 2.8624332236640975e-05, "epoch": 2.957627118644068, "percentage": 42.25, "elapsed_time": "7:36:14", "remaining_time": "10:23:34"} +{"current_steps": 1750, "total_steps": 4130, "loss": 0.2825, "lr": 2.854799779792375e-05, "epoch": 2.9661016949152543, "percentage": 42.37, "elapsed_time": "7:37:35", "remaining_time": "10:22:19"} +{"current_steps": 1755, "total_steps": 4130, "loss": 0.2853, "lr": 2.8471510701621708e-05, "epoch": 2.9745762711864407, "percentage": 42.49, "elapsed_time": "7:38:56", "remaining_time": "10:21:04"} +{"current_steps": 1760, "total_steps": 4130, "loss": 0.3144, "lr": 2.8394872313707998e-05, "epoch": 2.983050847457627, "percentage": 42.62, "elapsed_time": "7:40:16", "remaining_time": "10:19:48"} +{"current_steps": 1765, "total_steps": 4130, "loss": 0.3017, "lr": 2.8318084002857654e-05, "epoch": 2.9915254237288136, "percentage": 42.74, "elapsed_time": "7:41:35", "remaining_time": "10:18:30"} +{"current_steps": 1770, "total_steps": 4130, "loss": 0.2819, "lr": 2.824114714042317e-05, "epoch": 3.0, "percentage": 42.86, "elapsed_time": "7:42:52", "remaining_time": "10:17:10"} +{"current_steps": 1775, "total_steps": 4130, "loss": 0.3075, "lr": 2.8164063100409992e-05, "epoch": 3.0084745762711864, "percentage": 42.98, "elapsed_time": "7:44:13", "remaining_time": "10:15:54"} +{"current_steps": 1780, "total_steps": 4130, "loss": 0.2576, "lr": 2.8086833259452006e-05, "epoch": 3.016949152542373, "percentage": 43.1, "elapsed_time": "7:45:28", "remaining_time": "10:14:31"} +{"current_steps": 1785, "total_steps": 4130, "loss": 0.2891, "lr": 2.8009458996786934e-05, "epoch": 3.0254237288135593, "percentage": 43.22, "elapsed_time": "7:46:47", "remaining_time": "10:13:14"} +{"current_steps": 1790, "total_steps": 4130, "loss": 0.3207, "lr": 2.7931941694231705e-05, "epoch": 3.0338983050847457, "percentage": 43.34, "elapsed_time": "7:48:08", "remaining_time": "10:11:58"} +{"current_steps": 1795, "total_steps": 4130, "loss": 0.2785, "lr": 2.7854282736157777e-05, "epoch": 3.042372881355932, "percentage": 43.46, "elapsed_time": "7:49:25", "remaining_time": "10:10:38"} +{"current_steps": 1800, "total_steps": 4130, "loss": 0.3067, "lr": 2.777648350946642e-05, "epoch": 3.0508474576271185, "percentage": 43.58, "elapsed_time": "7:50:43", "remaining_time": "10:09:20"} +{"current_steps": 1805, "total_steps": 4130, "loss": 0.2858, "lr": 2.7698545403563934e-05, "epoch": 3.059322033898305, "percentage": 43.7, "elapsed_time": "7:52:01", "remaining_time": "10:08:00"} +{"current_steps": 1810, "total_steps": 4130, "loss": 0.2608, "lr": 2.7620469810336854e-05, "epoch": 3.0677966101694913, "percentage": 43.83, "elapsed_time": "7:53:17", "remaining_time": "10:06:38"} +{"current_steps": 1815, "total_steps": 4130, "loss": 0.2882, "lr": 2.754225812412708e-05, "epoch": 3.0762711864406778, "percentage": 43.95, "elapsed_time": "7:54:29", "remaining_time": "10:05:12"} +{"current_steps": 1820, "total_steps": 4130, "loss": 0.3186, "lr": 2.7463911741706982e-05, "epoch": 3.084745762711864, "percentage": 44.07, "elapsed_time": "7:55:50", "remaining_time": "10:03:56"} +{"current_steps": 1825, "total_steps": 4130, "loss": 0.3186, "lr": 2.738543206225445e-05, "epoch": 3.093220338983051, "percentage": 44.19, "elapsed_time": "7:57:06", "remaining_time": "10:02:35"} +{"current_steps": 1830, "total_steps": 4130, "loss": 0.2562, "lr": 2.7306820487327906e-05, "epoch": 3.1016949152542375, "percentage": 44.31, "elapsed_time": "7:58:22", "remaining_time": "10:01:14"} +{"current_steps": 1835, "total_steps": 4130, "loss": 0.2873, "lr": 2.7228078420841277e-05, "epoch": 3.110169491525424, "percentage": 44.43, "elapsed_time": "7:59:41", "remaining_time": "9:59:56"} +{"current_steps": 1840, "total_steps": 4130, "loss": 0.2612, "lr": 2.714920726903892e-05, "epoch": 3.1186440677966103, "percentage": 44.55, "elapsed_time": "8:01:01", "remaining_time": "9:58:39"} +{"current_steps": 1845, "total_steps": 4130, "loss": 0.3008, "lr": 2.7070208440470525e-05, "epoch": 3.1271186440677967, "percentage": 44.67, "elapsed_time": "8:02:19", "remaining_time": "9:57:21"} +{"current_steps": 1850, "total_steps": 4130, "loss": 0.2956, "lr": 2.699108334596592e-05, "epoch": 3.135593220338983, "percentage": 44.79, "elapsed_time": "8:03:35", "remaining_time": "9:56:00"} +{"current_steps": 1855, "total_steps": 4130, "loss": 0.2518, "lr": 2.6911833398609923e-05, "epoch": 3.1440677966101696, "percentage": 44.92, "elapsed_time": "8:04:47", "remaining_time": "9:54:33"} +{"current_steps": 1860, "total_steps": 4130, "loss": 0.2862, "lr": 2.683246001371706e-05, "epoch": 3.152542372881356, "percentage": 45.04, "elapsed_time": "8:06:08", "remaining_time": "9:53:17"} +{"current_steps": 1865, "total_steps": 4130, "loss": 0.2854, "lr": 2.6752964608806338e-05, "epoch": 3.1610169491525424, "percentage": 45.16, "elapsed_time": "8:07:27", "remaining_time": "9:52:00"} +{"current_steps": 1870, "total_steps": 4130, "loss": 0.2775, "lr": 2.6673348603575884e-05, "epoch": 3.169491525423729, "percentage": 45.28, "elapsed_time": "8:08:46", "remaining_time": "9:50:42"} +{"current_steps": 1875, "total_steps": 4130, "loss": 0.2927, "lr": 2.6593613419877615e-05, "epoch": 3.1779661016949152, "percentage": 45.4, "elapsed_time": "8:10:02", "remaining_time": "9:49:20"} +{"current_steps": 1880, "total_steps": 4130, "loss": 0.3094, "lr": 2.6513760481691842e-05, "epoch": 3.1864406779661016, "percentage": 45.52, "elapsed_time": "8:11:23", "remaining_time": "9:48:05"} +{"current_steps": 1885, "total_steps": 4130, "loss": 0.2881, "lr": 2.6433791215101848e-05, "epoch": 3.194915254237288, "percentage": 45.64, "elapsed_time": "8:12:37", "remaining_time": "9:46:42"} +{"current_steps": 1890, "total_steps": 4130, "loss": 0.3341, "lr": 2.6353707048268397e-05, "epoch": 3.2033898305084745, "percentage": 45.76, "elapsed_time": "8:13:57", "remaining_time": "9:45:26"} +{"current_steps": 1895, "total_steps": 4130, "loss": 0.2995, "lr": 2.6273509411404234e-05, "epoch": 3.211864406779661, "percentage": 45.88, "elapsed_time": "8:15:17", "remaining_time": "9:44:09"} +{"current_steps": 1900, "total_steps": 4130, "loss": 0.2862, "lr": 2.6193199736748572e-05, "epoch": 3.2203389830508473, "percentage": 46.0, "elapsed_time": "8:16:35", "remaining_time": "9:42:50"} +{"current_steps": 1905, "total_steps": 4130, "loss": 0.3022, "lr": 2.611277945854148e-05, "epoch": 3.2288135593220337, "percentage": 46.13, "elapsed_time": "8:17:52", "remaining_time": "9:41:30"} +{"current_steps": 1910, "total_steps": 4130, "loss": 0.3021, "lr": 2.603225001299829e-05, "epoch": 3.23728813559322, "percentage": 46.25, "elapsed_time": "8:19:10", "remaining_time": "9:40:11"} +{"current_steps": 1915, "total_steps": 4130, "loss": 0.3117, "lr": 2.595161283828392e-05, "epoch": 3.2457627118644066, "percentage": 46.37, "elapsed_time": "8:20:28", "remaining_time": "9:38:52"} +{"current_steps": 1920, "total_steps": 4130, "loss": 0.2706, "lr": 2.5870869374487227e-05, "epoch": 3.2542372881355934, "percentage": 46.49, "elapsed_time": "8:21:45", "remaining_time": "9:37:32"} +{"current_steps": 1925, "total_steps": 4130, "loss": 0.2969, "lr": 2.579002106359527e-05, "epoch": 3.26271186440678, "percentage": 46.61, "elapsed_time": "8:23:04", "remaining_time": "9:36:15"} +{"current_steps": 1930, "total_steps": 4130, "loss": 0.3404, "lr": 2.5709069349467562e-05, "epoch": 3.2711864406779663, "percentage": 46.73, "elapsed_time": "8:24:25", "remaining_time": "9:34:59"} +{"current_steps": 1935, "total_steps": 4130, "loss": 0.3009, "lr": 2.562801567781026e-05, "epoch": 3.2796610169491527, "percentage": 46.85, "elapsed_time": "8:25:45", "remaining_time": "9:33:43"} +{"current_steps": 1940, "total_steps": 4130, "loss": 0.2584, "lr": 2.554686149615038e-05, "epoch": 3.288135593220339, "percentage": 46.97, "elapsed_time": "8:27:00", "remaining_time": "9:32:20"} +{"current_steps": 1945, "total_steps": 4130, "loss": 0.2659, "lr": 2.5465608253809944e-05, "epoch": 3.2966101694915255, "percentage": 47.09, "elapsed_time": "8:28:17", "remaining_time": "9:31:01"} +{"current_steps": 1950, "total_steps": 4130, "loss": 0.2707, "lr": 2.5384257401880077e-05, "epoch": 3.305084745762712, "percentage": 47.22, "elapsed_time": "8:29:36", "remaining_time": "9:29:42"} +{"current_steps": 1955, "total_steps": 4130, "loss": 0.2552, "lr": 2.530281039319509e-05, "epoch": 3.3135593220338984, "percentage": 47.34, "elapsed_time": "8:30:51", "remaining_time": "9:28:20"} +{"current_steps": 1960, "total_steps": 4130, "loss": 0.2629, "lr": 2.5221268682306565e-05, "epoch": 3.3220338983050848, "percentage": 47.46, "elapsed_time": "8:32:08", "remaining_time": "9:27:01"} +{"current_steps": 1965, "total_steps": 4130, "loss": 0.3055, "lr": 2.5139633725457345e-05, "epoch": 3.330508474576271, "percentage": 47.58, "elapsed_time": "8:33:27", "remaining_time": "9:25:42"} +{"current_steps": 1970, "total_steps": 4130, "loss": 0.2804, "lr": 2.505790698055554e-05, "epoch": 3.3389830508474576, "percentage": 47.7, "elapsed_time": "8:34:41", "remaining_time": "9:24:20"} +{"current_steps": 1975, "total_steps": 4130, "loss": 0.2928, "lr": 2.49760899071485e-05, "epoch": 3.347457627118644, "percentage": 47.82, "elapsed_time": "8:36:01", "remaining_time": "9:23:03"} +{"current_steps": 1980, "total_steps": 4130, "loss": 0.2705, "lr": 2.4894183966396735e-05, "epoch": 3.3559322033898304, "percentage": 47.94, "elapsed_time": "8:37:17", "remaining_time": "9:21:42"} +{"current_steps": 1985, "total_steps": 4130, "loss": 0.2963, "lr": 2.4812190621047822e-05, "epoch": 3.364406779661017, "percentage": 48.06, "elapsed_time": "8:38:38", "remaining_time": "9:20:26"} +{"current_steps": 1990, "total_steps": 4130, "loss": 0.3058, "lr": 2.4730111335410298e-05, "epoch": 3.3728813559322033, "percentage": 48.18, "elapsed_time": "8:39:58", "remaining_time": "9:19:10"} +{"current_steps": 1995, "total_steps": 4130, "loss": 0.2798, "lr": 2.4647947575327468e-05, "epoch": 3.3813559322033897, "percentage": 48.31, "elapsed_time": "8:41:14", "remaining_time": "9:17:49"} +{"current_steps": 2000, "total_steps": 4130, "loss": 0.2525, "lr": 2.4565700808151297e-05, "epoch": 3.389830508474576, "percentage": 48.43, "elapsed_time": "8:42:29", "remaining_time": "9:16:27"} +{"current_steps": 2005, "total_steps": 4130, "loss": 0.282, "lr": 2.4483372502716124e-05, "epoch": 3.3983050847457625, "percentage": 48.55, "elapsed_time": "8:43:48", "remaining_time": "9:15:09"} +{"current_steps": 2010, "total_steps": 4130, "loss": 0.29, "lr": 2.44009641293125e-05, "epoch": 3.406779661016949, "percentage": 48.67, "elapsed_time": "8:45:07", "remaining_time": "9:13:52"} +{"current_steps": 2015, "total_steps": 4130, "loss": 0.2811, "lr": 2.431847715966087e-05, "epoch": 3.415254237288136, "percentage": 48.79, "elapsed_time": "8:46:27", "remaining_time": "9:12:35"} +{"current_steps": 2020, "total_steps": 4130, "loss": 0.2722, "lr": 2.423591306688534e-05, "epoch": 3.423728813559322, "percentage": 48.91, "elapsed_time": "8:47:42", "remaining_time": "9:11:12"} +{"current_steps": 2025, "total_steps": 4130, "loss": 0.2813, "lr": 2.4153273325487346e-05, "epoch": 3.4322033898305087, "percentage": 49.03, "elapsed_time": "8:48:58", "remaining_time": "9:09:52"} +{"current_steps": 2030, "total_steps": 4130, "loss": 0.2675, "lr": 2.407055941131932e-05, "epoch": 3.440677966101695, "percentage": 49.15, "elapsed_time": "8:50:18", "remaining_time": "9:08:36"} +{"current_steps": 2035, "total_steps": 4130, "loss": 0.2873, "lr": 2.3987772801558328e-05, "epoch": 3.4491525423728815, "percentage": 49.27, "elapsed_time": "8:51:36", "remaining_time": "9:07:16"} +{"current_steps": 2040, "total_steps": 4130, "loss": 0.2845, "lr": 2.3904914974679705e-05, "epoch": 3.457627118644068, "percentage": 49.39, "elapsed_time": "8:52:56", "remaining_time": "9:05:59"} +{"current_steps": 2045, "total_steps": 4130, "loss": 0.2671, "lr": 2.3821987410430646e-05, "epoch": 3.4661016949152543, "percentage": 49.52, "elapsed_time": "8:54:12", "remaining_time": "9:04:39"} +{"current_steps": 2050, "total_steps": 4130, "loss": 0.2978, "lr": 2.3738991589803783e-05, "epoch": 3.4745762711864407, "percentage": 49.64, "elapsed_time": "8:55:30", "remaining_time": "9:03:20"} +{"current_steps": 2055, "total_steps": 4130, "loss": 0.281, "lr": 2.36559289950107e-05, "epoch": 3.483050847457627, "percentage": 49.76, "elapsed_time": "8:56:49", "remaining_time": "9:02:03"} +{"current_steps": 2060, "total_steps": 4130, "loss": 0.2985, "lr": 2.357280110945552e-05, "epoch": 3.4915254237288136, "percentage": 49.88, "elapsed_time": "8:58:03", "remaining_time": "9:00:40"} +{"current_steps": 2065, "total_steps": 4130, "loss": 0.3108, "lr": 2.3489609417708383e-05, "epoch": 3.5, "percentage": 50.0, "elapsed_time": "8:59:21", "remaining_time": "8:59:21"} +{"current_steps": 2070, "total_steps": 4130, "loss": 0.2611, "lr": 2.340635540547892e-05, "epoch": 3.5084745762711864, "percentage": 50.12, "elapsed_time": "9:00:37", "remaining_time": "8:58:00"} +{"current_steps": 2075, "total_steps": 4130, "loss": 0.2853, "lr": 2.3323040559589755e-05, "epoch": 3.516949152542373, "percentage": 50.24, "elapsed_time": "9:01:57", "remaining_time": "8:56:44"} +{"current_steps": 2080, "total_steps": 4130, "loss": 0.2843, "lr": 2.3239666367949917e-05, "epoch": 3.5254237288135593, "percentage": 50.36, "elapsed_time": "9:03:15", "remaining_time": "8:55:24"} +{"current_steps": 2085, "total_steps": 4130, "loss": 0.2665, "lr": 2.3156234319528295e-05, "epoch": 3.5338983050847457, "percentage": 50.48, "elapsed_time": "9:04:34", "remaining_time": "8:54:07"} +{"current_steps": 2090, "total_steps": 4130, "loss": 0.2958, "lr": 2.3072745904327015e-05, "epoch": 3.542372881355932, "percentage": 50.61, "elapsed_time": "9:05:50", "remaining_time": "8:52:47"} +{"current_steps": 2095, "total_steps": 4130, "loss": 0.2588, "lr": 2.298920261335488e-05, "epoch": 3.5508474576271185, "percentage": 50.73, "elapsed_time": "9:07:11", "remaining_time": "8:51:30"} +{"current_steps": 2100, "total_steps": 4130, "loss": 0.2911, "lr": 2.2905605938600685e-05, "epoch": 3.559322033898305, "percentage": 50.85, "elapsed_time": "9:08:26", "remaining_time": "8:50:09"} +{"current_steps": 2105, "total_steps": 4130, "loss": 0.2634, "lr": 2.2821957373006617e-05, "epoch": 3.5677966101694913, "percentage": 50.97, "elapsed_time": "9:09:46", "remaining_time": "8:48:52"} +{"current_steps": 2110, "total_steps": 4130, "loss": 0.2958, "lr": 2.273825841044156e-05, "epoch": 3.576271186440678, "percentage": 51.09, "elapsed_time": "9:11:04", "remaining_time": "8:47:34"} +{"current_steps": 2115, "total_steps": 4130, "loss": 0.2788, "lr": 2.2654510545674442e-05, "epoch": 3.584745762711864, "percentage": 51.21, "elapsed_time": "9:12:20", "remaining_time": "8:46:13"} +{"current_steps": 2120, "total_steps": 4130, "loss": 0.2984, "lr": 2.257071527434753e-05, "epoch": 3.593220338983051, "percentage": 51.33, "elapsed_time": "9:13:33", "remaining_time": "8:44:50"} +{"current_steps": 2125, "total_steps": 4130, "loss": 0.2826, "lr": 2.2486874092949708e-05, "epoch": 3.601694915254237, "percentage": 51.45, "elapsed_time": "9:14:54", "remaining_time": "8:43:33"} +{"current_steps": 2130, "total_steps": 4130, "loss": 0.2928, "lr": 2.2402988498789767e-05, "epoch": 3.610169491525424, "percentage": 51.57, "elapsed_time": "9:16:14", "remaining_time": "8:42:17"} +{"current_steps": 2135, "total_steps": 4130, "loss": 0.2946, "lr": 2.2319059989969668e-05, "epoch": 3.6186440677966103, "percentage": 51.69, "elapsed_time": "9:17:34", "remaining_time": "8:41:00"} +{"current_steps": 2140, "total_steps": 4130, "loss": 0.296, "lr": 2.2235090065357773e-05, "epoch": 3.6271186440677967, "percentage": 51.82, "elapsed_time": "9:18:52", "remaining_time": "8:39:41"} +{"current_steps": 2145, "total_steps": 4130, "loss": 0.3039, "lr": 2.215108022456208e-05, "epoch": 3.635593220338983, "percentage": 51.94, "elapsed_time": "9:20:10", "remaining_time": "8:38:23"} +{"current_steps": 2150, "total_steps": 4130, "loss": 0.2561, "lr": 2.2067031967903443e-05, "epoch": 3.6440677966101696, "percentage": 52.06, "elapsed_time": "9:21:30", "remaining_time": "8:37:06"} +{"current_steps": 2155, "total_steps": 4130, "loss": 0.3076, "lr": 2.1982946796388788e-05, "epoch": 3.652542372881356, "percentage": 52.18, "elapsed_time": "9:22:43", "remaining_time": "8:35:43"} +{"current_steps": 2160, "total_steps": 4130, "loss": 0.2756, "lr": 2.1898826211684297e-05, "epoch": 3.6610169491525424, "percentage": 52.3, "elapsed_time": "9:23:58", "remaining_time": "8:34:21"} +{"current_steps": 2165, "total_steps": 4130, "loss": 0.3082, "lr": 2.1814671716088593e-05, "epoch": 3.669491525423729, "percentage": 52.42, "elapsed_time": "9:25:14", "remaining_time": "8:33:01"} +{"current_steps": 2170, "total_steps": 4130, "loss": 0.294, "lr": 2.17304848125059e-05, "epoch": 3.6779661016949152, "percentage": 52.54, "elapsed_time": "9:26:31", "remaining_time": "8:31:41"} +{"current_steps": 2175, "total_steps": 4130, "loss": 0.2758, "lr": 2.1646267004419218e-05, "epoch": 3.6864406779661016, "percentage": 52.66, "elapsed_time": "9:27:51", "remaining_time": "8:30:25"} +{"current_steps": 2180, "total_steps": 4130, "loss": 0.2974, "lr": 2.1562019795863463e-05, "epoch": 3.694915254237288, "percentage": 52.78, "elapsed_time": "9:29:10", "remaining_time": "8:29:07"} +{"current_steps": 2185, "total_steps": 4130, "loss": 0.2534, "lr": 2.147774469139863e-05, "epoch": 3.7033898305084745, "percentage": 52.91, "elapsed_time": "9:30:30", "remaining_time": "8:27:50"} +{"current_steps": 2190, "total_steps": 4130, "loss": 0.2851, "lr": 2.1393443196082867e-05, "epoch": 3.711864406779661, "percentage": 53.03, "elapsed_time": "9:31:46", "remaining_time": "8:26:29"} +{"current_steps": 2195, "total_steps": 4130, "loss": 0.2772, "lr": 2.1309116815445665e-05, "epoch": 3.7203389830508473, "percentage": 53.15, "elapsed_time": "9:32:59", "remaining_time": "8:25:07"} +{"current_steps": 2200, "total_steps": 4130, "loss": 0.2984, "lr": 2.1224767055460922e-05, "epoch": 3.7288135593220337, "percentage": 53.27, "elapsed_time": "9:34:20", "remaining_time": "8:23:51"} +{"current_steps": 2205, "total_steps": 4130, "loss": 0.3004, "lr": 2.114039542252008e-05, "epoch": 3.7372881355932206, "percentage": 53.39, "elapsed_time": "9:35:40", "remaining_time": "8:22:34"} +{"current_steps": 2210, "total_steps": 4130, "loss": 0.3037, "lr": 2.1056003423405178e-05, "epoch": 3.7457627118644066, "percentage": 53.51, "elapsed_time": "9:37:00", "remaining_time": "8:21:17"} +{"current_steps": 2215, "total_steps": 4130, "loss": 0.2938, "lr": 2.0971592565262005e-05, "epoch": 3.7542372881355934, "percentage": 53.63, "elapsed_time": "9:38:17", "remaining_time": "8:19:58"} +{"current_steps": 2220, "total_steps": 4130, "loss": 0.2809, "lr": 2.088716435557313e-05, "epoch": 3.7627118644067794, "percentage": 53.75, "elapsed_time": "9:39:36", "remaining_time": "8:18:40"} +{"current_steps": 2225, "total_steps": 4130, "loss": 0.3082, "lr": 2.0802720302131016e-05, "epoch": 3.7711864406779663, "percentage": 53.87, "elapsed_time": "9:40:57", "remaining_time": "8:17:23"} +{"current_steps": 2230, "total_steps": 4130, "loss": 0.2891, "lr": 2.0718261913011055e-05, "epoch": 3.7796610169491527, "percentage": 54.0, "elapsed_time": "9:42:17", "remaining_time": "8:16:07"} +{"current_steps": 2235, "total_steps": 4130, "loss": 0.2434, "lr": 2.063379069654468e-05, "epoch": 3.788135593220339, "percentage": 54.12, "elapsed_time": "9:43:32", "remaining_time": "8:14:46"} +{"current_steps": 2240, "total_steps": 4130, "loss": 0.3023, "lr": 2.0549308161292405e-05, "epoch": 3.7966101694915255, "percentage": 54.24, "elapsed_time": "9:44:47", "remaining_time": "8:13:24"} +{"current_steps": 2245, "total_steps": 4130, "loss": 0.2698, "lr": 2.0464815816016866e-05, "epoch": 3.805084745762712, "percentage": 54.36, "elapsed_time": "9:46:04", "remaining_time": "8:12:05"} +{"current_steps": 2250, "total_steps": 4130, "loss": 0.3099, "lr": 2.0380315169655905e-05, "epoch": 3.8135593220338984, "percentage": 54.48, "elapsed_time": "9:47:23", "remaining_time": "8:10:48"} +{"current_steps": 2255, "total_steps": 4130, "loss": 0.2861, "lr": 2.0295807731295628e-05, "epoch": 3.8220338983050848, "percentage": 54.6, "elapsed_time": "9:48:40", "remaining_time": "8:09:28"} +{"current_steps": 2260, "total_steps": 4130, "loss": 0.2637, "lr": 2.021129501014342e-05, "epoch": 3.830508474576271, "percentage": 54.72, "elapsed_time": "9:50:00", "remaining_time": "8:08:11"} +{"current_steps": 2265, "total_steps": 4130, "loss": 0.2859, "lr": 2.0126778515501e-05, "epoch": 3.8389830508474576, "percentage": 54.84, "elapsed_time": "9:51:19", "remaining_time": "8:06:53"} +{"current_steps": 2270, "total_steps": 4130, "loss": 0.3067, "lr": 2.0042259756737508e-05, "epoch": 3.847457627118644, "percentage": 54.96, "elapsed_time": "9:52:36", "remaining_time": "8:05:34"} +{"current_steps": 2275, "total_steps": 4130, "loss": 0.2925, "lr": 1.99577402432625e-05, "epoch": 3.8559322033898304, "percentage": 55.08, "elapsed_time": "9:53:54", "remaining_time": "8:04:16"} +{"current_steps": 2280, "total_steps": 4130, "loss": 0.2674, "lr": 1.9873221484499006e-05, "epoch": 3.864406779661017, "percentage": 55.21, "elapsed_time": "9:55:12", "remaining_time": "8:02:57"} +{"current_steps": 2285, "total_steps": 4130, "loss": 0.2946, "lr": 1.978870498985659e-05, "epoch": 3.8728813559322033, "percentage": 55.33, "elapsed_time": "9:56:29", "remaining_time": "8:01:37"} +{"current_steps": 2290, "total_steps": 4130, "loss": 0.2703, "lr": 1.9704192268704372e-05, "epoch": 3.8813559322033897, "percentage": 55.45, "elapsed_time": "9:57:48", "remaining_time": "8:00:20"} +{"current_steps": 2295, "total_steps": 4130, "loss": 0.2871, "lr": 1.96196848303441e-05, "epoch": 3.889830508474576, "percentage": 55.57, "elapsed_time": "9:59:03", "remaining_time": "7:58:59"} +{"current_steps": 2300, "total_steps": 4130, "loss": 0.2751, "lr": 1.953518418398314e-05, "epoch": 3.898305084745763, "percentage": 55.69, "elapsed_time": "10:00:23", "remaining_time": "7:57:42"} +{"current_steps": 2305, "total_steps": 4130, "loss": 0.284, "lr": 1.94506918387076e-05, "epoch": 3.906779661016949, "percentage": 55.81, "elapsed_time": "10:01:43", "remaining_time": "7:56:25"} +{"current_steps": 2310, "total_steps": 4130, "loss": 0.2901, "lr": 1.9366209303455322e-05, "epoch": 3.915254237288136, "percentage": 55.93, "elapsed_time": "10:03:01", "remaining_time": "7:55:06"} +{"current_steps": 2315, "total_steps": 4130, "loss": 0.2884, "lr": 1.928173808698895e-05, "epoch": 3.923728813559322, "percentage": 56.05, "elapsed_time": "10:04:17", "remaining_time": "7:53:46"} +{"current_steps": 2320, "total_steps": 4130, "loss": 0.2848, "lr": 1.919727969786899e-05, "epoch": 3.9322033898305087, "percentage": 56.17, "elapsed_time": "10:05:38", "remaining_time": "7:52:30"} +{"current_steps": 2325, "total_steps": 4130, "loss": 0.2735, "lr": 1.911283564442687e-05, "epoch": 3.940677966101695, "percentage": 56.3, "elapsed_time": "10:06:56", "remaining_time": "7:51:11"} +{"current_steps": 2330, "total_steps": 4130, "loss": 0.2789, "lr": 1.9028407434737998e-05, "epoch": 3.9491525423728815, "percentage": 56.42, "elapsed_time": "10:08:14", "remaining_time": "7:49:53"} +{"current_steps": 2335, "total_steps": 4130, "loss": 0.2846, "lr": 1.894399657659483e-05, "epoch": 3.957627118644068, "percentage": 56.54, "elapsed_time": "10:09:32", "remaining_time": "7:48:34"} +{"current_steps": 2340, "total_steps": 4130, "loss": 0.2699, "lr": 1.8859604577479927e-05, "epoch": 3.9661016949152543, "percentage": 56.66, "elapsed_time": "10:10:50", "remaining_time": "7:47:16"} +{"current_steps": 2345, "total_steps": 4130, "loss": 0.2759, "lr": 1.877523294453908e-05, "epoch": 3.9745762711864407, "percentage": 56.78, "elapsed_time": "10:12:07", "remaining_time": "7:45:56"} +{"current_steps": 2350, "total_steps": 4130, "loss": 0.284, "lr": 1.869088318455434e-05, "epoch": 3.983050847457627, "percentage": 56.9, "elapsed_time": "10:13:27", "remaining_time": "7:44:39"} +{"current_steps": 2355, "total_steps": 4130, "loss": 0.2346, "lr": 1.8606556803917136e-05, "epoch": 3.9915254237288136, "percentage": 57.02, "elapsed_time": "10:14:43", "remaining_time": "7:43:19"} +{"current_steps": 2360, "total_steps": 4130, "loss": 0.2733, "lr": 1.8522255308601378e-05, "epoch": 4.0, "percentage": 57.14, "elapsed_time": "10:15:59", "remaining_time": "7:41:59"} +{"current_steps": 2365, "total_steps": 4130, "loss": 0.2855, "lr": 1.8437980204136537e-05, "epoch": 4.008474576271187, "percentage": 57.26, "elapsed_time": "10:17:15", "remaining_time": "7:40:39"} +{"current_steps": 2370, "total_steps": 4130, "loss": 0.248, "lr": 1.8353732995580786e-05, "epoch": 4.016949152542373, "percentage": 57.38, "elapsed_time": "10:18:35", "remaining_time": "7:39:22"} +{"current_steps": 2375, "total_steps": 4130, "loss": 0.2807, "lr": 1.826951518749411e-05, "epoch": 4.02542372881356, "percentage": 57.51, "elapsed_time": "10:19:53", "remaining_time": "7:38:03"} +{"current_steps": 2380, "total_steps": 4130, "loss": 0.2649, "lr": 1.8185328283911417e-05, "epoch": 4.033898305084746, "percentage": 57.63, "elapsed_time": "10:21:13", "remaining_time": "7:36:46"} +{"current_steps": 2385, "total_steps": 4130, "loss": 0.3013, "lr": 1.8101173788315707e-05, "epoch": 4.0423728813559325, "percentage": 57.75, "elapsed_time": "10:22:30", "remaining_time": "7:35:27"} +{"current_steps": 2390, "total_steps": 4130, "loss": 0.2583, "lr": 1.8017053203611215e-05, "epoch": 4.0508474576271185, "percentage": 57.87, "elapsed_time": "10:23:49", "remaining_time": "7:34:09"} +{"current_steps": 2395, "total_steps": 4130, "loss": 0.2779, "lr": 1.7932968032096564e-05, "epoch": 4.059322033898305, "percentage": 57.99, "elapsed_time": "10:25:06", "remaining_time": "7:32:50"} +{"current_steps": 2400, "total_steps": 4130, "loss": 0.2679, "lr": 1.7848919775437924e-05, "epoch": 4.067796610169491, "percentage": 58.11, "elapsed_time": "10:26:17", "remaining_time": "7:31:27"} +{"current_steps": 2405, "total_steps": 4130, "loss": 0.2886, "lr": 1.776490993464223e-05, "epoch": 4.076271186440678, "percentage": 58.23, "elapsed_time": "10:27:38", "remaining_time": "7:30:10"} +{"current_steps": 2410, "total_steps": 4130, "loss": 0.2592, "lr": 1.768094001003033e-05, "epoch": 4.084745762711864, "percentage": 58.35, "elapsed_time": "10:28:56", "remaining_time": "7:28:52"} +{"current_steps": 2415, "total_steps": 4130, "loss": 0.2474, "lr": 1.759701150121024e-05, "epoch": 4.093220338983051, "percentage": 58.47, "elapsed_time": "10:30:12", "remaining_time": "7:27:32"} +{"current_steps": 2420, "total_steps": 4130, "loss": 0.2916, "lr": 1.7513125907050302e-05, "epoch": 4.101694915254237, "percentage": 58.6, "elapsed_time": "10:31:28", "remaining_time": "7:26:12"} +{"current_steps": 2425, "total_steps": 4130, "loss": 0.2718, "lr": 1.742928472565248e-05, "epoch": 4.110169491525424, "percentage": 58.72, "elapsed_time": "10:32:43", "remaining_time": "7:24:51"} +{"current_steps": 2430, "total_steps": 4130, "loss": 0.2717, "lr": 1.7345489454325564e-05, "epoch": 4.11864406779661, "percentage": 58.84, "elapsed_time": "10:34:00", "remaining_time": "7:23:32"} +{"current_steps": 2435, "total_steps": 4130, "loss": 0.2892, "lr": 1.7261741589558448e-05, "epoch": 4.127118644067797, "percentage": 58.96, "elapsed_time": "10:35:19", "remaining_time": "7:22:14"} +{"current_steps": 2440, "total_steps": 4130, "loss": 0.273, "lr": 1.717804262699339e-05, "epoch": 4.135593220338983, "percentage": 59.08, "elapsed_time": "10:36:39", "remaining_time": "7:20:58"} +{"current_steps": 2445, "total_steps": 4130, "loss": 0.2589, "lr": 1.7094394061399318e-05, "epoch": 4.1440677966101696, "percentage": 59.2, "elapsed_time": "10:37:57", "remaining_time": "7:19:39"} +{"current_steps": 2450, "total_steps": 4130, "loss": 0.2622, "lr": 1.701079738664512e-05, "epoch": 4.1525423728813555, "percentage": 59.32, "elapsed_time": "10:39:13", "remaining_time": "7:18:19"} +{"current_steps": 2455, "total_steps": 4130, "loss": 0.266, "lr": 1.6927254095672992e-05, "epoch": 4.161016949152542, "percentage": 59.44, "elapsed_time": "10:40:31", "remaining_time": "7:17:01"} +{"current_steps": 2460, "total_steps": 4130, "loss": 0.2761, "lr": 1.6843765680471715e-05, "epoch": 4.169491525423728, "percentage": 59.56, "elapsed_time": "10:41:48", "remaining_time": "7:15:41"} +{"current_steps": 2465, "total_steps": 4130, "loss": 0.2851, "lr": 1.6760333632050086e-05, "epoch": 4.177966101694915, "percentage": 59.69, "elapsed_time": "10:43:06", "remaining_time": "7:14:23"} +{"current_steps": 2470, "total_steps": 4130, "loss": 0.2744, "lr": 1.667695944041025e-05, "epoch": 4.186440677966102, "percentage": 59.81, "elapsed_time": "10:44:26", "remaining_time": "7:13:06"} +{"current_steps": 2475, "total_steps": 4130, "loss": 0.301, "lr": 1.6593644594521082e-05, "epoch": 4.194915254237288, "percentage": 59.93, "elapsed_time": "10:45:44", "remaining_time": "7:11:47"} +{"current_steps": 2480, "total_steps": 4130, "loss": 0.2627, "lr": 1.6510390582291624e-05, "epoch": 4.203389830508475, "percentage": 60.05, "elapsed_time": "10:47:03", "remaining_time": "7:10:30"} +{"current_steps": 2485, "total_steps": 4130, "loss": 0.27, "lr": 1.6427198890544483e-05, "epoch": 4.211864406779661, "percentage": 60.17, "elapsed_time": "10:48:23", "remaining_time": "7:09:13"} +{"current_steps": 2490, "total_steps": 4130, "loss": 0.2656, "lr": 1.634407100498931e-05, "epoch": 4.220338983050848, "percentage": 60.29, "elapsed_time": "10:49:37", "remaining_time": "7:07:51"} +{"current_steps": 2495, "total_steps": 4130, "loss": 0.2649, "lr": 1.626100841019623e-05, "epoch": 4.228813559322034, "percentage": 60.41, "elapsed_time": "10:50:56", "remaining_time": "7:06:34"} +{"current_steps": 2500, "total_steps": 4130, "loss": 0.29, "lr": 1.6178012589569358e-05, "epoch": 4.237288135593221, "percentage": 60.53, "elapsed_time": "10:52:14", "remaining_time": "7:05:15"} +{"current_steps": 2505, "total_steps": 4130, "loss": 0.2792, "lr": 1.60950850253203e-05, "epoch": 4.245762711864407, "percentage": 60.65, "elapsed_time": "10:53:32", "remaining_time": "7:03:57"} +{"current_steps": 2510, "total_steps": 4130, "loss": 0.2614, "lr": 1.601222719844168e-05, "epoch": 4.254237288135593, "percentage": 60.77, "elapsed_time": "10:54:50", "remaining_time": "7:02:39"} +{"current_steps": 2515, "total_steps": 4130, "loss": 0.2405, "lr": 1.5929440588680688e-05, "epoch": 4.262711864406779, "percentage": 60.9, "elapsed_time": "10:56:08", "remaining_time": "7:01:20"} +{"current_steps": 2520, "total_steps": 4130, "loss": 0.2487, "lr": 1.5846726674512654e-05, "epoch": 4.271186440677966, "percentage": 61.02, "elapsed_time": "10:57:23", "remaining_time": "7:00:00"} +{"current_steps": 2525, "total_steps": 4130, "loss": 0.269, "lr": 1.576408693311466e-05, "epoch": 4.279661016949152, "percentage": 61.14, "elapsed_time": "10:58:42", "remaining_time": "6:58:42"} +{"current_steps": 2530, "total_steps": 4130, "loss": 0.2741, "lr": 1.5681522840339143e-05, "epoch": 4.288135593220339, "percentage": 61.26, "elapsed_time": "11:00:00", "remaining_time": "6:57:23"} +{"current_steps": 2535, "total_steps": 4130, "loss": 0.2872, "lr": 1.5599035870687515e-05, "epoch": 4.296610169491525, "percentage": 61.38, "elapsed_time": "11:01:18", "remaining_time": "6:56:05"} +{"current_steps": 2540, "total_steps": 4130, "loss": 0.2769, "lr": 1.5516627497283882e-05, "epoch": 4.305084745762712, "percentage": 61.5, "elapsed_time": "11:02:38", "remaining_time": "6:54:48"} +{"current_steps": 2545, "total_steps": 4130, "loss": 0.2819, "lr": 1.5434299191848713e-05, "epoch": 4.313559322033898, "percentage": 61.62, "elapsed_time": "11:03:58", "remaining_time": "6:53:31"} +{"current_steps": 2550, "total_steps": 4130, "loss": 0.2662, "lr": 1.5352052424672535e-05, "epoch": 4.322033898305085, "percentage": 61.74, "elapsed_time": "11:05:15", "remaining_time": "6:52:11"} +{"current_steps": 2555, "total_steps": 4130, "loss": 0.2604, "lr": 1.5269888664589712e-05, "epoch": 4.330508474576272, "percentage": 61.86, "elapsed_time": "11:06:32", "remaining_time": "6:50:52"} +{"current_steps": 2560, "total_steps": 4130, "loss": 0.2572, "lr": 1.5187809378952182e-05, "epoch": 4.338983050847458, "percentage": 61.99, "elapsed_time": "11:07:43", "remaining_time": "6:49:30"} +{"current_steps": 2565, "total_steps": 4130, "loss": 0.2446, "lr": 1.5105816033603266e-05, "epoch": 4.3474576271186445, "percentage": 62.11, "elapsed_time": "11:08:59", "remaining_time": "6:48:10"} +{"current_steps": 2570, "total_steps": 4130, "loss": 0.2713, "lr": 1.5023910092851509e-05, "epoch": 4.3559322033898304, "percentage": 62.23, "elapsed_time": "11:10:17", "remaining_time": "6:46:52"} +{"current_steps": 2575, "total_steps": 4130, "loss": 0.2729, "lr": 1.4942093019444468e-05, "epoch": 4.364406779661017, "percentage": 62.35, "elapsed_time": "11:11:37", "remaining_time": "6:45:35"} +{"current_steps": 2580, "total_steps": 4130, "loss": 0.2594, "lr": 1.4860366274542663e-05, "epoch": 4.372881355932203, "percentage": 62.47, "elapsed_time": "11:12:54", "remaining_time": "6:44:15"} +{"current_steps": 2585, "total_steps": 4130, "loss": 0.249, "lr": 1.4778731317693442e-05, "epoch": 4.38135593220339, "percentage": 62.59, "elapsed_time": "11:14:13", "remaining_time": "6:42:58"} +{"current_steps": 2590, "total_steps": 4130, "loss": 0.2504, "lr": 1.4697189606804914e-05, "epoch": 4.389830508474576, "percentage": 62.71, "elapsed_time": "11:15:33", "remaining_time": "6:41:41"} +{"current_steps": 2595, "total_steps": 4130, "loss": 0.2638, "lr": 1.4615742598119927e-05, "epoch": 4.398305084745763, "percentage": 62.83, "elapsed_time": "11:16:52", "remaining_time": "6:40:22"} +{"current_steps": 2600, "total_steps": 4130, "loss": 0.2781, "lr": 1.4534391746190056e-05, "epoch": 4.406779661016949, "percentage": 62.95, "elapsed_time": "11:18:09", "remaining_time": "6:39:04"} +{"current_steps": 2605, "total_steps": 4130, "loss": 0.3089, "lr": 1.4453138503849622e-05, "epoch": 4.415254237288136, "percentage": 63.08, "elapsed_time": "11:19:26", "remaining_time": "6:37:45"} +{"current_steps": 2610, "total_steps": 4130, "loss": 0.266, "lr": 1.4371984322189754e-05, "epoch": 4.423728813559322, "percentage": 63.2, "elapsed_time": "11:20:44", "remaining_time": "6:36:26"} +{"current_steps": 2615, "total_steps": 4130, "loss": 0.2693, "lr": 1.4290930650532448e-05, "epoch": 4.432203389830509, "percentage": 63.32, "elapsed_time": "11:22:03", "remaining_time": "6:35:08"} +{"current_steps": 2620, "total_steps": 4130, "loss": 0.2896, "lr": 1.4209978936404731e-05, "epoch": 4.440677966101695, "percentage": 63.44, "elapsed_time": "11:23:23", "remaining_time": "6:33:51"} +{"current_steps": 2625, "total_steps": 4130, "loss": 0.2989, "lr": 1.4129130625512776e-05, "epoch": 4.4491525423728815, "percentage": 63.56, "elapsed_time": "11:24:44", "remaining_time": "6:32:35"} +{"current_steps": 2630, "total_steps": 4130, "loss": 0.2752, "lr": 1.4048387161716086e-05, "epoch": 4.4576271186440675, "percentage": 63.68, "elapsed_time": "11:26:05", "remaining_time": "6:31:18"} +{"current_steps": 2635, "total_steps": 4130, "loss": 0.2661, "lr": 1.3967749987001717e-05, "epoch": 4.466101694915254, "percentage": 63.8, "elapsed_time": "11:27:25", "remaining_time": "6:30:01"} +{"current_steps": 2640, "total_steps": 4130, "loss": 0.2741, "lr": 1.3887220541458521e-05, "epoch": 4.47457627118644, "percentage": 63.92, "elapsed_time": "11:28:45", "remaining_time": "6:28:43"} +{"current_steps": 2645, "total_steps": 4130, "loss": 0.2373, "lr": 1.380680026325143e-05, "epoch": 4.483050847457627, "percentage": 64.04, "elapsed_time": "11:29:59", "remaining_time": "6:27:23"} +{"current_steps": 2650, "total_steps": 4130, "loss": 0.2693, "lr": 1.3726490588595776e-05, "epoch": 4.491525423728813, "percentage": 64.16, "elapsed_time": "11:31:16", "remaining_time": "6:26:04"} +{"current_steps": 2655, "total_steps": 4130, "loss": 0.2637, "lr": 1.3646292951731615e-05, "epoch": 4.5, "percentage": 64.29, "elapsed_time": "11:32:36", "remaining_time": "6:24:46"} +{"current_steps": 2660, "total_steps": 4130, "loss": 0.2708, "lr": 1.3566208784898157e-05, "epoch": 4.508474576271187, "percentage": 64.41, "elapsed_time": "11:33:56", "remaining_time": "6:23:29"} +{"current_steps": 2665, "total_steps": 4130, "loss": 0.2614, "lr": 1.348623951830816e-05, "epoch": 4.516949152542373, "percentage": 64.53, "elapsed_time": "11:35:16", "remaining_time": "6:22:12"} +{"current_steps": 2670, "total_steps": 4130, "loss": 0.2893, "lr": 1.3406386580122389e-05, "epoch": 4.52542372881356, "percentage": 64.65, "elapsed_time": "11:36:34", "remaining_time": "6:20:54"} +{"current_steps": 2675, "total_steps": 4130, "loss": 0.2752, "lr": 1.332665139642412e-05, "epoch": 4.533898305084746, "percentage": 64.77, "elapsed_time": "11:37:55", "remaining_time": "6:19:37"} +{"current_steps": 2680, "total_steps": 4130, "loss": 0.269, "lr": 1.3247035391193664e-05, "epoch": 4.5423728813559325, "percentage": 64.89, "elapsed_time": "11:39:12", "remaining_time": "6:18:18"} +{"current_steps": 2685, "total_steps": 4130, "loss": 0.2426, "lr": 1.3167539986282938e-05, "epoch": 4.5508474576271185, "percentage": 65.01, "elapsed_time": "11:40:28", "remaining_time": "6:16:58"} +{"current_steps": 2690, "total_steps": 4130, "loss": 0.2769, "lr": 1.3088166601390087e-05, "epoch": 4.559322033898305, "percentage": 65.13, "elapsed_time": "11:41:47", "remaining_time": "6:15:40"} +{"current_steps": 2695, "total_steps": 4130, "loss": 0.3127, "lr": 1.3008916654034085e-05, "epoch": 4.567796610169491, "percentage": 65.25, "elapsed_time": "11:43:04", "remaining_time": "6:14:21"} +{"current_steps": 2700, "total_steps": 4130, "loss": 0.2956, "lr": 1.2929791559529484e-05, "epoch": 4.576271186440678, "percentage": 65.38, "elapsed_time": "11:44:25", "remaining_time": "6:13:04"} +{"current_steps": 2705, "total_steps": 4130, "loss": 0.2884, "lr": 1.2850792730961082e-05, "epoch": 4.584745762711864, "percentage": 65.5, "elapsed_time": "11:45:45", "remaining_time": "6:11:47"} +{"current_steps": 2710, "total_steps": 4130, "loss": 0.2869, "lr": 1.2771921579158728e-05, "epoch": 4.593220338983051, "percentage": 65.62, "elapsed_time": "11:46:57", "remaining_time": "6:10:25"} +{"current_steps": 2715, "total_steps": 4130, "loss": 0.2668, "lr": 1.26931795126721e-05, "epoch": 4.601694915254237, "percentage": 65.74, "elapsed_time": "11:48:13", "remaining_time": "6:09:06"} +{"current_steps": 2720, "total_steps": 4130, "loss": 0.276, "lr": 1.2614567937745554e-05, "epoch": 4.610169491525424, "percentage": 65.86, "elapsed_time": "11:49:33", "remaining_time": "6:07:49"} +{"current_steps": 2725, "total_steps": 4130, "loss": 0.2745, "lr": 1.2536088258293018e-05, "epoch": 4.61864406779661, "percentage": 65.98, "elapsed_time": "11:50:51", "remaining_time": "6:06:31"} +{"current_steps": 2730, "total_steps": 4130, "loss": 0.2664, "lr": 1.2457741875872928e-05, "epoch": 4.627118644067797, "percentage": 66.1, "elapsed_time": "11:52:12", "remaining_time": "6:05:13"} +{"current_steps": 2735, "total_steps": 4130, "loss": 0.2709, "lr": 1.2379530189663156e-05, "epoch": 4.635593220338983, "percentage": 66.22, "elapsed_time": "11:53:31", "remaining_time": "6:03:56"} +{"current_steps": 2740, "total_steps": 4130, "loss": 0.2594, "lr": 1.2301454596436076e-05, "epoch": 4.6440677966101696, "percentage": 66.34, "elapsed_time": "11:54:50", "remaining_time": "6:02:38"} +{"current_steps": 2745, "total_steps": 4130, "loss": 0.2476, "lr": 1.222351649053359e-05, "epoch": 4.652542372881356, "percentage": 66.46, "elapsed_time": "11:56:10", "remaining_time": "6:01:21"} +{"current_steps": 2750, "total_steps": 4130, "loss": 0.2767, "lr": 1.2145717263842228e-05, "epoch": 4.661016949152542, "percentage": 66.59, "elapsed_time": "11:57:30", "remaining_time": "6:00:03"} +{"current_steps": 2755, "total_steps": 4130, "loss": 0.2791, "lr": 1.2068058305768298e-05, "epoch": 4.669491525423728, "percentage": 66.71, "elapsed_time": "11:58:50", "remaining_time": "5:58:45"} +{"current_steps": 2760, "total_steps": 4130, "loss": 0.28, "lr": 1.1990541003213072e-05, "epoch": 4.677966101694915, "percentage": 66.83, "elapsed_time": "12:00:06", "remaining_time": "5:57:26"} +{"current_steps": 2765, "total_steps": 4130, "loss": 0.2634, "lr": 1.1913166740547999e-05, "epoch": 4.686440677966102, "percentage": 66.95, "elapsed_time": "12:01:26", "remaining_time": "5:56:09"} +{"current_steps": 2770, "total_steps": 4130, "loss": 0.2779, "lr": 1.1835936899590017e-05, "epoch": 4.694915254237288, "percentage": 67.07, "elapsed_time": "12:02:45", "remaining_time": "5:54:51"} +{"current_steps": 2775, "total_steps": 4130, "loss": 0.2816, "lr": 1.1758852859576842e-05, "epoch": 4.703389830508475, "percentage": 67.19, "elapsed_time": "12:04:06", "remaining_time": "5:53:34"} +{"current_steps": 2780, "total_steps": 4130, "loss": 0.2716, "lr": 1.1681915997142354e-05, "epoch": 4.711864406779661, "percentage": 67.31, "elapsed_time": "12:05:20", "remaining_time": "5:52:14"} +{"current_steps": 2785, "total_steps": 4130, "loss": 0.2706, "lr": 1.1605127686292009e-05, "epoch": 4.720338983050848, "percentage": 67.43, "elapsed_time": "12:06:40", "remaining_time": "5:50:56"} +{"current_steps": 2790, "total_steps": 4130, "loss": 0.2803, "lr": 1.15284892983783e-05, "epoch": 4.728813559322034, "percentage": 67.55, "elapsed_time": "12:07:59", "remaining_time": "5:49:38"} +{"current_steps": 2795, "total_steps": 4130, "loss": 0.2907, "lr": 1.1452002202076256e-05, "epoch": 4.737288135593221, "percentage": 67.68, "elapsed_time": "12:09:19", "remaining_time": "5:48:21"} +{"current_steps": 2800, "total_steps": 4130, "loss": 0.2662, "lr": 1.1375667763359031e-05, "epoch": 4.745762711864407, "percentage": 67.8, "elapsed_time": "12:10:35", "remaining_time": "5:47:02"} +{"current_steps": 2805, "total_steps": 4130, "loss": 0.2793, "lr": 1.1299487345473457e-05, "epoch": 4.754237288135593, "percentage": 67.92, "elapsed_time": "12:11:56", "remaining_time": "5:45:44"} +{"current_steps": 2810, "total_steps": 4130, "loss": 0.2785, "lr": 1.1223462308915767e-05, "epoch": 4.762711864406779, "percentage": 68.04, "elapsed_time": "12:13:15", "remaining_time": "5:44:26"} +{"current_steps": 2815, "total_steps": 4130, "loss": 0.2935, "lr": 1.1147594011407203e-05, "epoch": 4.771186440677966, "percentage": 68.16, "elapsed_time": "12:14:35", "remaining_time": "5:43:09"} +{"current_steps": 2820, "total_steps": 4130, "loss": 0.2785, "lr": 1.1071883807869886e-05, "epoch": 4.779661016949152, "percentage": 68.28, "elapsed_time": "12:15:54", "remaining_time": "5:41:51"} +{"current_steps": 2825, "total_steps": 4130, "loss": 0.291, "lr": 1.0996333050402503e-05, "epoch": 4.788135593220339, "percentage": 68.4, "elapsed_time": "12:17:14", "remaining_time": "5:40:34"} +{"current_steps": 2830, "total_steps": 4130, "loss": 0.2941, "lr": 1.0920943088256266e-05, "epoch": 4.796610169491525, "percentage": 68.52, "elapsed_time": "12:18:33", "remaining_time": "5:39:15"} +{"current_steps": 2835, "total_steps": 4130, "loss": 0.2702, "lr": 1.0845715267810716e-05, "epoch": 4.805084745762712, "percentage": 68.64, "elapsed_time": "12:19:49", "remaining_time": "5:37:56"} +{"current_steps": 2840, "total_steps": 4130, "loss": 0.2785, "lr": 1.0770650932549768e-05, "epoch": 4.813559322033898, "percentage": 68.77, "elapsed_time": "12:21:07", "remaining_time": "5:36:38"} +{"current_steps": 2845, "total_steps": 4130, "loss": 0.2612, "lr": 1.0695751423037639e-05, "epoch": 4.822033898305085, "percentage": 68.89, "elapsed_time": "12:22:24", "remaining_time": "5:35:19"} +{"current_steps": 2850, "total_steps": 4130, "loss": 0.271, "lr": 1.0621018076894984e-05, "epoch": 4.830508474576272, "percentage": 69.01, "elapsed_time": "12:23:39", "remaining_time": "5:33:59"} +{"current_steps": 2855, "total_steps": 4130, "loss": 0.2955, "lr": 1.0546452228774919e-05, "epoch": 4.838983050847458, "percentage": 69.13, "elapsed_time": "12:24:59", "remaining_time": "5:32:42"} +{"current_steps": 2860, "total_steps": 4130, "loss": 0.2676, "lr": 1.047205521033928e-05, "epoch": 4.847457627118644, "percentage": 69.25, "elapsed_time": "12:26:14", "remaining_time": "5:31:22"} +{"current_steps": 2865, "total_steps": 4130, "loss": 0.2733, "lr": 1.0397828350234765e-05, "epoch": 4.8559322033898304, "percentage": 69.37, "elapsed_time": "12:27:31", "remaining_time": "5:30:03"} +{"current_steps": 2870, "total_steps": 4130, "loss": 0.2686, "lr": 1.032377297406926e-05, "epoch": 4.864406779661017, "percentage": 69.49, "elapsed_time": "12:28:52", "remaining_time": "5:28:46"} +{"current_steps": 2875, "total_steps": 4130, "loss": 0.2786, "lr": 1.0249890404388121e-05, "epoch": 4.872881355932203, "percentage": 69.61, "elapsed_time": "12:30:07", "remaining_time": "5:27:26"} +{"current_steps": 2880, "total_steps": 4130, "loss": 0.2617, "lr": 1.0176181960650595e-05, "epoch": 4.88135593220339, "percentage": 69.73, "elapsed_time": "12:31:27", "remaining_time": "5:26:09"} +{"current_steps": 2885, "total_steps": 4130, "loss": 0.2485, "lr": 1.0102648959206212e-05, "epoch": 4.889830508474576, "percentage": 69.85, "elapsed_time": "12:32:47", "remaining_time": "5:24:51"} +{"current_steps": 2890, "total_steps": 4130, "loss": 0.2661, "lr": 1.0029292713271334e-05, "epoch": 4.898305084745763, "percentage": 69.98, "elapsed_time": "12:34:04", "remaining_time": "5:23:32"} +{"current_steps": 2895, "total_steps": 4130, "loss": 0.3071, "lr": 9.956114532905625e-06, "epoch": 4.906779661016949, "percentage": 70.1, "elapsed_time": "12:35:21", "remaining_time": "5:22:13"} +{"current_steps": 2900, "total_steps": 4130, "loss": 0.2764, "lr": 9.883115724988743e-06, "epoch": 4.915254237288136, "percentage": 70.22, "elapsed_time": "12:36:38", "remaining_time": "5:20:55"} +{"current_steps": 2905, "total_steps": 4130, "loss": 0.2858, "lr": 9.810297593196922e-06, "epoch": 4.923728813559322, "percentage": 70.34, "elapsed_time": "12:37:55", "remaining_time": "5:19:36"} +{"current_steps": 2910, "total_steps": 4130, "loss": 0.2991, "lr": 9.73766143797976e-06, "epoch": 4.932203389830509, "percentage": 70.46, "elapsed_time": "12:39:13", "remaining_time": "5:18:17"} +{"current_steps": 2915, "total_steps": 4130, "loss": 0.2572, "lr": 9.665208556536918e-06, "epoch": 4.940677966101695, "percentage": 70.58, "elapsed_time": "12:40:31", "remaining_time": "5:16:59"} +{"current_steps": 2920, "total_steps": 4130, "loss": 0.2705, "lr": 9.592940242795035e-06, "epoch": 4.9491525423728815, "percentage": 70.7, "elapsed_time": "12:41:50", "remaining_time": "5:15:41"} +{"current_steps": 2925, "total_steps": 4130, "loss": 0.2931, "lr": 9.520857787384548e-06, "epoch": 4.9576271186440675, "percentage": 70.82, "elapsed_time": "12:43:10", "remaining_time": "5:14:24"} +{"current_steps": 2930, "total_steps": 4130, "loss": 0.288, "lr": 9.44896247761669e-06, "epoch": 4.966101694915254, "percentage": 70.94, "elapsed_time": "12:44:29", "remaining_time": "5:13:06"} +{"current_steps": 2935, "total_steps": 4130, "loss": 0.2711, "lr": 9.377255597460469e-06, "epoch": 4.97457627118644, "percentage": 71.07, "elapsed_time": "12:45:47", "remaining_time": "5:11:47"} +{"current_steps": 2940, "total_steps": 4130, "loss": 0.2963, "lr": 9.305738427519782e-06, "epoch": 4.983050847457627, "percentage": 71.19, "elapsed_time": "12:47:06", "remaining_time": "5:10:29"} +{"current_steps": 2945, "total_steps": 4130, "loss": 0.2432, "lr": 9.234412245010482e-06, "epoch": 4.991525423728813, "percentage": 71.31, "elapsed_time": "12:48:25", "remaining_time": "5:09:11"} +{"current_steps": 2950, "total_steps": 4130, "loss": 0.2877, "lr": 9.163278323737635e-06, "epoch": 5.0, "percentage": 71.43, "elapsed_time": "12:49:46", "remaining_time": "5:07:54"} +{"current_steps": 2955, "total_steps": 4130, "loss": 0.2718, "lr": 9.092337934072703e-06, "epoch": 5.008474576271187, "percentage": 71.55, "elapsed_time": "12:51:04", "remaining_time": "5:06:36"} +{"current_steps": 2960, "total_steps": 4130, "loss": 0.2325, "lr": 9.021592342930928e-06, "epoch": 5.016949152542373, "percentage": 71.67, "elapsed_time": "12:52:19", "remaining_time": "5:05:16"} +{"current_steps": 2965, "total_steps": 4130, "loss": 0.2686, "lr": 8.951042813748645e-06, "epoch": 5.02542372881356, "percentage": 71.79, "elapsed_time": "12:53:37", "remaining_time": "5:03:58"} +{"current_steps": 2970, "total_steps": 4130, "loss": 0.2938, "lr": 8.880690606460754e-06, "epoch": 5.033898305084746, "percentage": 71.91, "elapsed_time": "12:54:57", "remaining_time": "5:02:40"} +{"current_steps": 2975, "total_steps": 4130, "loss": 0.2736, "lr": 8.8105369774782e-06, "epoch": 5.0423728813559325, "percentage": 72.03, "elapsed_time": "12:56:17", "remaining_time": "5:01:23"} +{"current_steps": 2980, "total_steps": 4130, "loss": 0.2502, "lr": 8.74058317966556e-06, "epoch": 5.0508474576271185, "percentage": 72.15, "elapsed_time": "12:57:35", "remaining_time": "5:00:04"} +{"current_steps": 2985, "total_steps": 4130, "loss": 0.2536, "lr": 8.670830462318633e-06, "epoch": 5.059322033898305, "percentage": 72.28, "elapsed_time": "12:58:54", "remaining_time": "4:58:46"} +{"current_steps": 2990, "total_steps": 4130, "loss": 0.2721, "lr": 8.60128007114217e-06, "epoch": 5.067796610169491, "percentage": 72.4, "elapsed_time": "13:00:14", "remaining_time": "4:57:28"} +{"current_steps": 2995, "total_steps": 4130, "loss": 0.2402, "lr": 8.531933248227582e-06, "epoch": 5.076271186440678, "percentage": 72.52, "elapsed_time": "13:01:31", "remaining_time": "4:56:10"} +{"current_steps": 3000, "total_steps": 4130, "loss": 0.2775, "lr": 8.462791232030803e-06, "epoch": 5.084745762711864, "percentage": 72.64, "elapsed_time": "13:02:50", "remaining_time": "4:54:52"} +{"current_steps": 3005, "total_steps": 4130, "loss": 0.253, "lr": 8.393855257350132e-06, "epoch": 5.093220338983051, "percentage": 72.76, "elapsed_time": "13:04:49", "remaining_time": "4:53:49"} +{"current_steps": 3010, "total_steps": 4130, "loss": 0.2628, "lr": 8.325126555304208e-06, "epoch": 5.101694915254237, "percentage": 72.88, "elapsed_time": "13:06:07", "remaining_time": "4:52:30"} +{"current_steps": 3015, "total_steps": 4130, "loss": 0.2487, "lr": 8.256606353310003e-06, "epoch": 5.110169491525424, "percentage": 73.0, "elapsed_time": "13:07:26", "remaining_time": "4:51:12"} +{"current_steps": 3020, "total_steps": 4130, "loss": 0.2767, "lr": 8.188295875060943e-06, "epoch": 5.11864406779661, "percentage": 73.12, "elapsed_time": "13:08:46", "remaining_time": "4:49:54"} +{"current_steps": 3025, "total_steps": 4130, "loss": 0.2284, "lr": 8.12019634050499e-06, "epoch": 5.127118644067797, "percentage": 73.24, "elapsed_time": "13:10:00", "remaining_time": "4:48:34"} +{"current_steps": 3030, "total_steps": 4130, "loss": 0.2484, "lr": 8.052308965822916e-06, "epoch": 5.135593220338983, "percentage": 73.37, "elapsed_time": "13:11:20", "remaining_time": "4:47:17"} +{"current_steps": 3035, "total_steps": 4130, "loss": 0.2513, "lr": 7.984634963406537e-06, "epoch": 5.1440677966101696, "percentage": 73.49, "elapsed_time": "13:12:34", "remaining_time": "4:45:57"} +{"current_steps": 3040, "total_steps": 4130, "loss": 0.2862, "lr": 7.917175541837101e-06, "epoch": 5.1525423728813555, "percentage": 73.61, "elapsed_time": "13:13:55", "remaining_time": "4:44:39"} +{"current_steps": 3045, "total_steps": 4130, "loss": 0.29, "lr": 7.849931905863672e-06, "epoch": 5.161016949152542, "percentage": 73.73, "elapsed_time": "13:15:12", "remaining_time": "4:43:20"} +{"current_steps": 3050, "total_steps": 4130, "loss": 0.259, "lr": 7.78290525638163e-06, "epoch": 5.169491525423728, "percentage": 73.85, "elapsed_time": "13:16:29", "remaining_time": "4:42:02"} +{"current_steps": 3055, "total_steps": 4130, "loss": 0.252, "lr": 7.71609679041121e-06, "epoch": 5.177966101694915, "percentage": 73.97, "elapsed_time": "13:17:49", "remaining_time": "4:40:44"} +{"current_steps": 3060, "total_steps": 4130, "loss": 0.264, "lr": 7.649507701076164e-06, "epoch": 5.186440677966102, "percentage": 74.09, "elapsed_time": "13:19:07", "remaining_time": "4:39:26"} +{"current_steps": 3065, "total_steps": 4130, "loss": 0.2651, "lr": 7.583139177582395e-06, "epoch": 5.194915254237288, "percentage": 74.21, "elapsed_time": "13:20:27", "remaining_time": "4:38:08"} +{"current_steps": 3070, "total_steps": 4130, "loss": 0.2671, "lr": 7.516992405196772e-06, "epoch": 5.203389830508475, "percentage": 74.33, "elapsed_time": "13:21:47", "remaining_time": "4:36:50"} +{"current_steps": 3075, "total_steps": 4130, "loss": 0.2698, "lr": 7.45106856522593e-06, "epoch": 5.211864406779661, "percentage": 74.46, "elapsed_time": "13:23:07", "remaining_time": "4:35:32"} +{"current_steps": 3080, "total_steps": 4130, "loss": 0.2312, "lr": 7.385368834995168e-06, "epoch": 5.220338983050848, "percentage": 74.58, "elapsed_time": "13:24:25", "remaining_time": "4:34:14"} +{"current_steps": 3085, "total_steps": 4130, "loss": 0.2531, "lr": 7.319894387827473e-06, "epoch": 5.228813559322034, "percentage": 74.7, "elapsed_time": "13:25:40", "remaining_time": "4:32:54"} +{"current_steps": 3090, "total_steps": 4130, "loss": 0.2826, "lr": 7.254646393022502e-06, "epoch": 5.237288135593221, "percentage": 74.82, "elapsed_time": "13:26:58", "remaining_time": "4:31:36"} +{"current_steps": 3095, "total_steps": 4130, "loss": 0.2851, "lr": 7.189626015835733e-06, "epoch": 5.245762711864407, "percentage": 74.94, "elapsed_time": "13:28:18", "remaining_time": "4:30:18"} +{"current_steps": 3100, "total_steps": 4130, "loss": 0.2511, "lr": 7.124834417457664e-06, "epoch": 5.254237288135593, "percentage": 75.06, "elapsed_time": "13:29:29", "remaining_time": "4:28:57"} +{"current_steps": 3105, "total_steps": 4130, "loss": 0.2538, "lr": 7.060272754993051e-06, "epoch": 5.262711864406779, "percentage": 75.18, "elapsed_time": "13:30:45", "remaining_time": "4:27:38"} +{"current_steps": 3110, "total_steps": 4130, "loss": 0.2703, "lr": 6.9959421814402494e-06, "epoch": 5.271186440677966, "percentage": 75.3, "elapsed_time": "13:32:04", "remaining_time": "4:26:20"} +{"current_steps": 3115, "total_steps": 4130, "loss": 0.2802, "lr": 6.931843845670647e-06, "epoch": 5.279661016949152, "percentage": 75.42, "elapsed_time": "13:33:19", "remaining_time": "4:25:01"} +{"current_steps": 3120, "total_steps": 4130, "loss": 0.2827, "lr": 6.867978892408101e-06, "epoch": 5.288135593220339, "percentage": 75.54, "elapsed_time": "13:34:38", "remaining_time": "4:23:42"} +{"current_steps": 3125, "total_steps": 4130, "loss": 0.2517, "lr": 6.804348462208548e-06, "epoch": 5.296610169491525, "percentage": 75.67, "elapsed_time": "13:35:55", "remaining_time": "4:22:24"} +{"current_steps": 3130, "total_steps": 4130, "loss": 0.2565, "lr": 6.7409536914395866e-06, "epoch": 5.305084745762712, "percentage": 75.79, "elapsed_time": "13:37:13", "remaining_time": "4:21:05"} +{"current_steps": 3135, "total_steps": 4130, "loss": 0.2706, "lr": 6.677795712260206e-06, "epoch": 5.313559322033898, "percentage": 75.91, "elapsed_time": "13:38:31", "remaining_time": "4:19:47"} +{"current_steps": 3140, "total_steps": 4130, "loss": 0.2646, "lr": 6.6148756526005855e-06, "epoch": 5.322033898305085, "percentage": 76.03, "elapsed_time": "13:39:49", "remaining_time": "4:18:28"} +{"current_steps": 3145, "total_steps": 4130, "loss": 0.2458, "lr": 6.552194636141909e-06, "epoch": 5.330508474576272, "percentage": 76.15, "elapsed_time": "13:41:02", "remaining_time": "4:17:08"} +{"current_steps": 3150, "total_steps": 4130, "loss": 0.2616, "lr": 6.489753782296315e-06, "epoch": 5.338983050847458, "percentage": 76.27, "elapsed_time": "13:42:16", "remaining_time": "4:15:49"} +{"current_steps": 3155, "total_steps": 4130, "loss": 0.2611, "lr": 6.427554206186939e-06, "epoch": 5.3474576271186445, "percentage": 76.39, "elapsed_time": "13:43:36", "remaining_time": "4:14:31"} +{"current_steps": 3160, "total_steps": 4130, "loss": 0.2804, "lr": 6.3655970186279314e-06, "epoch": 5.3559322033898304, "percentage": 76.51, "elapsed_time": "13:44:55", "remaining_time": "4:13:13"} +{"current_steps": 3165, "total_steps": 4130, "loss": 0.2494, "lr": 6.30388332610469e-06, "epoch": 5.364406779661017, "percentage": 76.63, "elapsed_time": "13:46:10", "remaining_time": "4:11:54"} +{"current_steps": 3170, "total_steps": 4130, "loss": 0.2668, "lr": 6.242414230754044e-06, "epoch": 5.372881355932203, "percentage": 76.76, "elapsed_time": "13:47:30", "remaining_time": "4:10:35"} +{"current_steps": 3175, "total_steps": 4130, "loss": 0.2883, "lr": 6.181190830344601e-06, "epoch": 5.38135593220339, "percentage": 76.88, "elapsed_time": "13:48:50", "remaining_time": "4:09:18"} +{"current_steps": 3180, "total_steps": 4130, "loss": 0.2572, "lr": 6.120214218257128e-06, "epoch": 5.389830508474576, "percentage": 77.0, "elapsed_time": "13:50:10", "remaining_time": "4:08:00"} +{"current_steps": 3185, "total_steps": 4130, "loss": 0.2501, "lr": 6.059485483465048e-06, "epoch": 5.398305084745763, "percentage": 77.12, "elapsed_time": "13:51:28", "remaining_time": "4:06:42"} +{"current_steps": 3190, "total_steps": 4130, "loss": 0.2464, "lr": 5.999005710514956e-06, "epoch": 5.406779661016949, "percentage": 77.24, "elapsed_time": "13:52:41", "remaining_time": "4:05:22"} +{"current_steps": 3195, "total_steps": 4130, "loss": 0.2645, "lr": 5.938775979507287e-06, "epoch": 5.415254237288136, "percentage": 77.36, "elapsed_time": "13:53:59", "remaining_time": "4:04:03"} +{"current_steps": 3200, "total_steps": 4130, "loss": 0.2437, "lr": 5.878797366076994e-06, "epoch": 5.423728813559322, "percentage": 77.48, "elapsed_time": "13:55:19", "remaining_time": "4:02:45"} +{"current_steps": 3205, "total_steps": 4130, "loss": 0.2493, "lr": 5.819070941374368e-06, "epoch": 5.432203389830509, "percentage": 77.6, "elapsed_time": "13:56:35", "remaining_time": "4:01:27"} +{"current_steps": 3210, "total_steps": 4130, "loss": 0.2822, "lr": 5.759597772045882e-06, "epoch": 5.440677966101695, "percentage": 77.72, "elapsed_time": "13:57:50", "remaining_time": "4:00:07"} +{"current_steps": 3215, "total_steps": 4130, "loss": 0.2749, "lr": 5.700378920215159e-06, "epoch": 5.4491525423728815, "percentage": 77.85, "elapsed_time": "13:59:06", "remaining_time": "3:58:48"} +{"current_steps": 3220, "total_steps": 4130, "loss": 0.263, "lr": 5.641415443463994e-06, "epoch": 5.4576271186440675, "percentage": 77.97, "elapsed_time": "14:00:26", "remaining_time": "3:57:31"} +{"current_steps": 3225, "total_steps": 4130, "loss": 0.2835, "lr": 5.5827083948134876e-06, "epoch": 5.466101694915254, "percentage": 78.09, "elapsed_time": "14:01:47", "remaining_time": "3:56:13"} +{"current_steps": 3230, "total_steps": 4130, "loss": 0.231, "lr": 5.524258822705202e-06, "epoch": 5.47457627118644, "percentage": 78.21, "elapsed_time": "14:02:59", "remaining_time": "3:54:53"} +{"current_steps": 3235, "total_steps": 4130, "loss": 0.2954, "lr": 5.466067770982482e-06, "epoch": 5.483050847457627, "percentage": 78.33, "elapsed_time": "14:04:15", "remaining_time": "3:53:34"} +{"current_steps": 3240, "total_steps": 4130, "loss": 0.2926, "lr": 5.4081362788717625e-06, "epoch": 5.491525423728813, "percentage": 78.45, "elapsed_time": "14:05:34", "remaining_time": "3:52:16"} +{"current_steps": 3245, "total_steps": 4130, "loss": 0.2708, "lr": 5.350465380964076e-06, "epoch": 5.5, "percentage": 78.57, "elapsed_time": "14:06:52", "remaining_time": "3:50:57"} +{"current_steps": 3250, "total_steps": 4130, "loss": 0.2832, "lr": 5.293056107196488e-06, "epoch": 5.508474576271187, "percentage": 78.69, "elapsed_time": "14:08:11", "remaining_time": "3:49:39"} +{"current_steps": 3255, "total_steps": 4130, "loss": 0.2734, "lr": 5.235909482833797e-06, "epoch": 5.516949152542373, "percentage": 78.81, "elapsed_time": "14:09:28", "remaining_time": "3:48:21"} +{"current_steps": 3260, "total_steps": 4130, "loss": 0.2656, "lr": 5.179026528450146e-06, "epoch": 5.52542372881356, "percentage": 78.93, "elapsed_time": "14:10:43", "remaining_time": "3:47:01"} +{"current_steps": 3265, "total_steps": 4130, "loss": 0.2453, "lr": 5.1224082599108584e-06, "epoch": 5.533898305084746, "percentage": 79.06, "elapsed_time": "14:12:02", "remaining_time": "3:45:43"} +{"current_steps": 3270, "total_steps": 4130, "loss": 0.2778, "lr": 5.066055688354246e-06, "epoch": 5.5423728813559325, "percentage": 79.18, "elapsed_time": "14:13:22", "remaining_time": "3:44:26"} +{"current_steps": 3275, "total_steps": 4130, "loss": 0.2627, "lr": 5.0099698201735855e-06, "epoch": 5.5508474576271185, "percentage": 79.3, "elapsed_time": "14:14:41", "remaining_time": "3:43:07"} +{"current_steps": 3280, "total_steps": 4130, "loss": 0.2614, "lr": 4.954151656999122e-06, "epoch": 5.559322033898305, "percentage": 79.42, "elapsed_time": "14:15:56", "remaining_time": "3:41:48"} +{"current_steps": 3285, "total_steps": 4130, "loss": 0.2586, "lr": 4.898602195680214e-06, "epoch": 5.567796610169491, "percentage": 79.54, "elapsed_time": "14:17:10", "remaining_time": "3:40:29"} +{"current_steps": 3290, "total_steps": 4130, "loss": 0.2595, "lr": 4.843322428267465e-06, "epoch": 5.576271186440678, "percentage": 79.66, "elapsed_time": "14:18:27", "remaining_time": "3:39:10"} +{"current_steps": 3295, "total_steps": 4130, "loss": 0.3024, "lr": 4.788313341995096e-06, "epoch": 5.584745762711864, "percentage": 79.78, "elapsed_time": "14:19:47", "remaining_time": "3:37:53"} +{"current_steps": 3300, "total_steps": 4130, "loss": 0.2738, "lr": 4.733575919263238e-06, "epoch": 5.593220338983051, "percentage": 79.9, "elapsed_time": "14:21:05", "remaining_time": "3:36:34"} +{"current_steps": 3305, "total_steps": 4130, "loss": 0.2778, "lr": 4.679111137620442e-06, "epoch": 5.601694915254237, "percentage": 80.02, "elapsed_time": "14:22:20", "remaining_time": "3:35:15"} +{"current_steps": 3310, "total_steps": 4130, "loss": 0.2771, "lr": 4.624919969746171e-06, "epoch": 5.610169491525424, "percentage": 80.15, "elapsed_time": "14:23:36", "remaining_time": "3:33:56"} +{"current_steps": 3315, "total_steps": 4130, "loss": 0.2915, "lr": 4.571003383433481e-06, "epoch": 5.61864406779661, "percentage": 80.27, "elapsed_time": "14:24:56", "remaining_time": "3:32:38"} +{"current_steps": 3320, "total_steps": 4130, "loss": 0.2775, "lr": 4.517362341571687e-06, "epoch": 5.627118644067797, "percentage": 80.39, "elapsed_time": "14:26:17", "remaining_time": "3:31:21"} +{"current_steps": 3325, "total_steps": 4130, "loss": 0.3019, "lr": 4.463997802129221e-06, "epoch": 5.635593220338983, "percentage": 80.51, "elapsed_time": "14:27:35", "remaining_time": "3:30:02"} +{"current_steps": 3330, "total_steps": 4130, "loss": 0.2666, "lr": 4.410910718136454e-06, "epoch": 5.6440677966101696, "percentage": 80.63, "elapsed_time": "14:28:53", "remaining_time": "3:28:44"} +{"current_steps": 3335, "total_steps": 4130, "loss": 0.2634, "lr": 4.3581020376687566e-06, "epoch": 5.652542372881356, "percentage": 80.75, "elapsed_time": "14:30:09", "remaining_time": "3:27:25"} +{"current_steps": 3340, "total_steps": 4130, "loss": 0.2525, "lr": 4.305572703829495e-06, "epoch": 5.661016949152542, "percentage": 80.87, "elapsed_time": "14:31:27", "remaining_time": "3:26:07"} +{"current_steps": 3345, "total_steps": 4130, "loss": 0.268, "lr": 4.253323654733248e-06, "epoch": 5.669491525423728, "percentage": 80.99, "elapsed_time": "14:32:42", "remaining_time": "3:24:48"} +{"current_steps": 3350, "total_steps": 4130, "loss": 0.2693, "lr": 4.201355823488999e-06, "epoch": 5.677966101694915, "percentage": 81.11, "elapsed_time": "14:34:02", "remaining_time": "3:23:30"} +{"current_steps": 3355, "total_steps": 4130, "loss": 0.252, "lr": 4.149670138183526e-06, "epoch": 5.686440677966102, "percentage": 81.23, "elapsed_time": "14:35:19", "remaining_time": "3:22:12"} +{"current_steps": 3360, "total_steps": 4130, "loss": 0.2404, "lr": 4.098267521864772e-06, "epoch": 5.694915254237288, "percentage": 81.36, "elapsed_time": "14:36:30", "remaining_time": "3:20:52"} +{"current_steps": 3365, "total_steps": 4130, "loss": 0.2303, "lr": 4.047148892525403e-06, "epoch": 5.703389830508475, "percentage": 81.48, "elapsed_time": "14:37:47", "remaining_time": "3:19:33"} +{"current_steps": 3370, "total_steps": 4130, "loss": 0.2597, "lr": 3.996315163086391e-06, "epoch": 5.711864406779661, "percentage": 81.6, "elapsed_time": "14:39:05", "remaining_time": "3:18:15"} +{"current_steps": 3375, "total_steps": 4130, "loss": 0.2911, "lr": 3.945767241380732e-06, "epoch": 5.720338983050848, "percentage": 81.72, "elapsed_time": "14:40:25", "remaining_time": "3:16:57"} +{"current_steps": 3380, "total_steps": 4130, "loss": 0.2489, "lr": 3.895506030137195e-06, "epoch": 5.728813559322034, "percentage": 81.84, "elapsed_time": "14:41:40", "remaining_time": "3:15:38"} +{"current_steps": 3385, "total_steps": 4130, "loss": 0.2523, "lr": 3.845532426964252e-06, "epoch": 5.737288135593221, "percentage": 81.96, "elapsed_time": "14:43:00", "remaining_time": "3:14:20"} +{"current_steps": 3390, "total_steps": 4130, "loss": 0.2838, "lr": 3.79584732433399e-06, "epoch": 5.745762711864407, "percentage": 82.08, "elapsed_time": "14:44:18", "remaining_time": "3:13:02"} +{"current_steps": 3395, "total_steps": 4130, "loss": 0.2682, "lr": 3.746451609566233e-06, "epoch": 5.754237288135593, "percentage": 82.2, "elapsed_time": "14:45:33", "remaining_time": "3:11:43"} +{"current_steps": 3400, "total_steps": 4130, "loss": 0.2436, "lr": 3.697346164812643e-06, "epoch": 5.762711864406779, "percentage": 82.32, "elapsed_time": "14:46:48", "remaining_time": "3:10:24"} +{"current_steps": 3405, "total_steps": 4130, "loss": 0.263, "lr": 3.6485318670409896e-06, "epoch": 5.771186440677966, "percentage": 82.45, "elapsed_time": "14:48:04", "remaining_time": "3:09:05"} +{"current_steps": 3410, "total_steps": 4130, "loss": 0.2997, "lr": 3.6000095880194905e-06, "epoch": 5.779661016949152, "percentage": 82.57, "elapsed_time": "14:49:24", "remaining_time": "3:07:47"} +{"current_steps": 3415, "total_steps": 4130, "loss": 0.2655, "lr": 3.5517801943012443e-06, "epoch": 5.788135593220339, "percentage": 82.69, "elapsed_time": "14:50:44", "remaining_time": "3:06:29"} +{"current_steps": 3420, "total_steps": 4130, "loss": 0.2769, "lr": 3.5038445472087324e-06, "epoch": 5.796610169491525, "percentage": 82.81, "elapsed_time": "14:52:05", "remaining_time": "3:05:11"} +{"current_steps": 3425, "total_steps": 4130, "loss": 0.2603, "lr": 3.456203502818476e-06, "epoch": 5.805084745762712, "percentage": 82.93, "elapsed_time": "14:53:24", "remaining_time": "3:03:53"} +{"current_steps": 3430, "total_steps": 4130, "loss": 0.292, "lr": 3.408857911945702e-06, "epoch": 5.813559322033898, "percentage": 83.05, "elapsed_time": "14:54:44", "remaining_time": "3:02:36"} +{"current_steps": 3435, "total_steps": 4130, "loss": 0.2665, "lr": 3.36180862012919e-06, "epoch": 5.822033898305085, "percentage": 83.17, "elapsed_time": "14:56:04", "remaining_time": "3:01:18"} +{"current_steps": 3440, "total_steps": 4130, "loss": 0.2371, "lr": 3.3150564676161402e-06, "epoch": 5.830508474576272, "percentage": 83.29, "elapsed_time": "14:57:23", "remaining_time": "2:59:59"} +{"current_steps": 3445, "total_steps": 4130, "loss": 0.2585, "lr": 3.268602289347185e-06, "epoch": 5.838983050847458, "percentage": 83.41, "elapsed_time": "14:58:40", "remaining_time": "2:58:41"} +{"current_steps": 3450, "total_steps": 4130, "loss": 0.2765, "lr": 3.222446914941468e-06, "epoch": 5.847457627118644, "percentage": 83.54, "elapsed_time": "15:00:00", "remaining_time": "2:57:23"} +{"current_steps": 3455, "total_steps": 4130, "loss": 0.2596, "lr": 3.176591168681851e-06, "epoch": 5.8559322033898304, "percentage": 83.66, "elapsed_time": "15:01:18", "remaining_time": "2:56:05"} +{"current_steps": 3460, "total_steps": 4130, "loss": 0.2477, "lr": 3.131035869500152e-06, "epoch": 5.864406779661017, "percentage": 83.78, "elapsed_time": "15:02:30", "remaining_time": "2:54:45"} +{"current_steps": 3465, "total_steps": 4130, "loss": 0.2613, "lr": 3.085781830962564e-06, "epoch": 5.872881355932203, "percentage": 83.9, "elapsed_time": "15:03:48", "remaining_time": "2:53:27"} +{"current_steps": 3470, "total_steps": 4130, "loss": 0.2359, "lr": 3.0408298612550878e-06, "epoch": 5.88135593220339, "percentage": 84.02, "elapsed_time": "15:05:06", "remaining_time": "2:52:09"} +{"current_steps": 3475, "total_steps": 4130, "loss": 0.2515, "lr": 2.996180763169132e-06, "epoch": 5.889830508474576, "percentage": 84.14, "elapsed_time": "15:06:21", "remaining_time": "2:50:50"} +{"current_steps": 3480, "total_steps": 4130, "loss": 0.2575, "lr": 2.95183533408715e-06, "epoch": 5.898305084745763, "percentage": 84.26, "elapsed_time": "15:07:39", "remaining_time": "2:49:32"} +{"current_steps": 3485, "total_steps": 4130, "loss": 0.2625, "lr": 2.907794365968413e-06, "epoch": 5.906779661016949, "percentage": 84.38, "elapsed_time": "15:08:58", "remaining_time": "2:48:13"} +{"current_steps": 3490, "total_steps": 4130, "loss": 0.2456, "lr": 2.864058645334853e-06, "epoch": 5.915254237288136, "percentage": 84.5, "elapsed_time": "15:10:17", "remaining_time": "2:46:55"} +{"current_steps": 3495, "total_steps": 4130, "loss": 0.2739, "lr": 2.820628953257052e-06, "epoch": 5.923728813559322, "percentage": 84.62, "elapsed_time": "15:11:37", "remaining_time": "2:45:37"} +{"current_steps": 3500, "total_steps": 4130, "loss": 0.2747, "lr": 2.7775060653402387e-06, "epoch": 5.932203389830509, "percentage": 84.75, "elapsed_time": "15:12:53", "remaining_time": "2:44:19"} +{"current_steps": 3505, "total_steps": 4130, "loss": 0.2957, "lr": 2.7346907517104894e-06, "epoch": 5.940677966101695, "percentage": 84.87, "elapsed_time": "15:14:13", "remaining_time": "2:43:01"} +{"current_steps": 3510, "total_steps": 4130, "loss": 0.2679, "lr": 2.692183777000932e-06, "epoch": 5.9491525423728815, "percentage": 84.99, "elapsed_time": "15:15:34", "remaining_time": "2:41:43"} +{"current_steps": 3515, "total_steps": 4130, "loss": 0.27, "lr": 2.64998590033813e-06, "epoch": 5.9576271186440675, "percentage": 85.11, "elapsed_time": "15:16:52", "remaining_time": "2:40:25"} +{"current_steps": 3520, "total_steps": 4130, "loss": 0.2487, "lr": 2.608097875328488e-06, "epoch": 5.966101694915254, "percentage": 85.23, "elapsed_time": "15:18:12", "remaining_time": "2:39:07"} +{"current_steps": 3525, "total_steps": 4130, "loss": 0.2494, "lr": 2.5665204500448137e-06, "epoch": 5.97457627118644, "percentage": 85.35, "elapsed_time": "15:19:27", "remaining_time": "2:37:48"} +{"current_steps": 3530, "total_steps": 4130, "loss": 0.2695, "lr": 2.525254367012955e-06, "epoch": 5.983050847457627, "percentage": 85.47, "elapsed_time": "15:20:45", "remaining_time": "2:36:30"} +{"current_steps": 3535, "total_steps": 4130, "loss": 0.2968, "lr": 2.4843003631985486e-06, "epoch": 5.991525423728813, "percentage": 85.59, "elapsed_time": "15:22:02", "remaining_time": "2:35:11"} +{"current_steps": 3540, "total_steps": 4130, "loss": 0.2477, "lr": 2.4436591699938305e-06, "epoch": 6.0, "percentage": 85.71, "elapsed_time": "15:23:20", "remaining_time": "2:33:53"} +{"current_steps": 3545, "total_steps": 4130, "loss": 0.2467, "lr": 2.403331513204612e-06, "epoch": 6.008474576271187, "percentage": 85.84, "elapsed_time": "15:24:40", "remaining_time": "2:32:35"} +{"current_steps": 3550, "total_steps": 4130, "loss": 0.2437, "lr": 2.363318113037283e-06, "epoch": 6.016949152542373, "percentage": 85.96, "elapsed_time": "15:25:57", "remaining_time": "2:31:17"} +{"current_steps": 3555, "total_steps": 4130, "loss": 0.2427, "lr": 2.323619684085976e-06, "epoch": 6.02542372881356, "percentage": 86.08, "elapsed_time": "15:27:17", "remaining_time": "2:29:58"} +{"current_steps": 3560, "total_steps": 4130, "loss": 0.2412, "lr": 2.2842369353197858e-06, "epoch": 6.033898305084746, "percentage": 86.2, "elapsed_time": "15:28:34", "remaining_time": "2:28:40"} +{"current_steps": 3565, "total_steps": 4130, "loss": 0.2457, "lr": 2.2451705700701185e-06, "epoch": 6.0423728813559325, "percentage": 86.32, "elapsed_time": "15:29:51", "remaining_time": "2:27:22"} +{"current_steps": 3570, "total_steps": 4130, "loss": 0.2364, "lr": 2.2064212860181258e-06, "epoch": 6.0508474576271185, "percentage": 86.44, "elapsed_time": "15:31:07", "remaining_time": "2:26:03"} +{"current_steps": 3575, "total_steps": 4130, "loss": 0.2462, "lr": 2.1679897751822532e-06, "epoch": 6.059322033898305, "percentage": 86.56, "elapsed_time": "15:32:27", "remaining_time": "2:24:45"} +{"current_steps": 3580, "total_steps": 4130, "loss": 0.2797, "lr": 2.1298767239058684e-06, "epoch": 6.067796610169491, "percentage": 86.68, "elapsed_time": "15:33:45", "remaining_time": "2:23:27"} +{"current_steps": 3585, "total_steps": 4130, "loss": 0.2426, "lr": 2.0920828128450197e-06, "epoch": 6.076271186440678, "percentage": 86.8, "elapsed_time": "15:35:03", "remaining_time": "2:22:09"} +{"current_steps": 3590, "total_steps": 4130, "loss": 0.2509, "lr": 2.054608716956259e-06, "epoch": 6.084745762711864, "percentage": 86.92, "elapsed_time": "15:36:23", "remaining_time": "2:20:50"} +{"current_steps": 3595, "total_steps": 4130, "loss": 0.2423, "lr": 2.0174551054846158e-06, "epoch": 6.093220338983051, "percentage": 87.05, "elapsed_time": "15:37:38", "remaining_time": "2:19:32"} +{"current_steps": 3600, "total_steps": 4130, "loss": 0.286, "lr": 1.9806226419516195e-06, "epoch": 6.101694915254237, "percentage": 87.17, "elapsed_time": "15:38:58", "remaining_time": "2:18:14"} +{"current_steps": 3605, "total_steps": 4130, "loss": 0.2674, "lr": 1.94411198414346e-06, "epoch": 6.110169491525424, "percentage": 87.29, "elapsed_time": "15:40:17", "remaining_time": "2:16:56"} +{"current_steps": 3610, "total_steps": 4130, "loss": 0.2341, "lr": 1.9079237840992416e-06, "epoch": 6.11864406779661, "percentage": 87.41, "elapsed_time": "15:41:36", "remaining_time": "2:15:37"} +{"current_steps": 3615, "total_steps": 4130, "loss": 0.2768, "lr": 1.8720586880993452e-06, "epoch": 6.127118644067797, "percentage": 87.53, "elapsed_time": "15:42:56", "remaining_time": "2:14:20"} +{"current_steps": 3620, "total_steps": 4130, "loss": 0.3004, "lr": 1.8365173366538647e-06, "epoch": 6.135593220338983, "percentage": 87.65, "elapsed_time": "15:44:13", "remaining_time": "2:13:01"} +{"current_steps": 3625, "total_steps": 4130, "loss": 0.2891, "lr": 1.8013003644911987e-06, "epoch": 6.1440677966101696, "percentage": 87.77, "elapsed_time": "15:45:33", "remaining_time": "2:11:43"} +{"current_steps": 3630, "total_steps": 4130, "loss": 0.2599, "lr": 1.7664084005466796e-06, "epoch": 6.1525423728813555, "percentage": 87.89, "elapsed_time": "15:46:52", "remaining_time": "2:10:25"} +{"current_steps": 3635, "total_steps": 4130, "loss": 0.2685, "lr": 1.7318420679513725e-06, "epoch": 6.161016949152542, "percentage": 88.01, "elapsed_time": "15:48:13", "remaining_time": "2:09:07"} +{"current_steps": 3640, "total_steps": 4130, "loss": 0.2503, "lr": 1.6976019840209334e-06, "epoch": 6.169491525423728, "percentage": 88.14, "elapsed_time": "15:49:33", "remaining_time": "2:07:49"} +{"current_steps": 3645, "total_steps": 4130, "loss": 0.2503, "lr": 1.6636887602445839e-06, "epoch": 6.177966101694915, "percentage": 88.26, "elapsed_time": "15:50:50", "remaining_time": "2:06:31"} +{"current_steps": 3650, "total_steps": 4130, "loss": 0.2667, "lr": 1.6301030022741837e-06, "epoch": 6.186440677966102, "percentage": 88.38, "elapsed_time": "15:52:10", "remaining_time": "2:05:13"} +{"current_steps": 3655, "total_steps": 4130, "loss": 0.2781, "lr": 1.596845309913444e-06, "epoch": 6.194915254237288, "percentage": 88.5, "elapsed_time": "15:53:30", "remaining_time": "2:03:55"} +{"current_steps": 3660, "total_steps": 4130, "loss": 0.2598, "lr": 1.5639162771071736e-06, "epoch": 6.203389830508475, "percentage": 88.62, "elapsed_time": "15:54:47", "remaining_time": "2:02:36"} +{"current_steps": 3665, "total_steps": 4130, "loss": 0.2644, "lr": 1.5313164919307033e-06, "epoch": 6.211864406779661, "percentage": 88.74, "elapsed_time": "15:56:04", "remaining_time": "2:01:18"} +{"current_steps": 3670, "total_steps": 4130, "loss": 0.2837, "lr": 1.499046536579374e-06, "epoch": 6.220338983050848, "percentage": 88.86, "elapsed_time": "15:57:25", "remaining_time": "2:00:00"} +{"current_steps": 3675, "total_steps": 4130, "loss": 0.2868, "lr": 1.4671069873581312e-06, "epoch": 6.228813559322034, "percentage": 88.98, "elapsed_time": "15:58:42", "remaining_time": "1:58:41"} +{"current_steps": 3680, "total_steps": 4130, "loss": 0.2643, "lr": 1.4354984146712503e-06, "epoch": 6.237288135593221, "percentage": 89.1, "elapsed_time": "16:00:02", "remaining_time": "1:57:23"} +{"current_steps": 3685, "total_steps": 4130, "loss": 0.2519, "lr": 1.4042213830121344e-06, "epoch": 6.245762711864407, "percentage": 89.23, "elapsed_time": "16:01:19", "remaining_time": "1:56:05"} +{"current_steps": 3690, "total_steps": 4130, "loss": 0.2585, "lr": 1.3732764509532316e-06, "epoch": 6.254237288135593, "percentage": 89.35, "elapsed_time": "16:02:40", "remaining_time": "1:54:47"} +{"current_steps": 3695, "total_steps": 4130, "loss": 0.2557, "lr": 1.3426641711360788e-06, "epoch": 6.262711864406779, "percentage": 89.47, "elapsed_time": "16:04:00", "remaining_time": "1:53:29"} +{"current_steps": 3700, "total_steps": 4130, "loss": 0.2447, "lr": 1.3123850902614143e-06, "epoch": 6.271186440677966, "percentage": 89.59, "elapsed_time": "16:05:17", "remaining_time": "1:52:10"} +{"current_steps": 3705, "total_steps": 4130, "loss": 0.2811, "lr": 1.2824397490794115e-06, "epoch": 6.279661016949152, "percentage": 89.71, "elapsed_time": "16:06:37", "remaining_time": "1:50:52"} +{"current_steps": 3710, "total_steps": 4130, "loss": 0.2507, "lr": 1.2528286823800495e-06, "epoch": 6.288135593220339, "percentage": 89.83, "elapsed_time": "16:07:52", "remaining_time": "1:49:34"} +{"current_steps": 3715, "total_steps": 4130, "loss": 0.2665, "lr": 1.223552418983518e-06, "epoch": 6.296610169491525, "percentage": 89.95, "elapsed_time": "16:09:11", "remaining_time": "1:48:16"} +{"current_steps": 3720, "total_steps": 4130, "loss": 0.2466, "lr": 1.1946114817308207e-06, "epoch": 6.305084745762712, "percentage": 90.07, "elapsed_time": "16:10:31", "remaining_time": "1:46:57"} +{"current_steps": 3725, "total_steps": 4130, "loss": 0.2446, "lr": 1.166006387474401e-06, "epoch": 6.313559322033898, "percentage": 90.19, "elapsed_time": "16:11:49", "remaining_time": "1:45:39"} +{"current_steps": 3730, "total_steps": 4130, "loss": 0.2735, "lr": 1.1377376470689328e-06, "epoch": 6.322033898305085, "percentage": 90.31, "elapsed_time": "16:13:09", "remaining_time": "1:44:21"} +{"current_steps": 3735, "total_steps": 4130, "loss": 0.2554, "lr": 1.1098057653621775e-06, "epoch": 6.330508474576272, "percentage": 90.44, "elapsed_time": "16:14:29", "remaining_time": "1:43:03"} +{"current_steps": 3740, "total_steps": 4130, "loss": 0.2669, "lr": 1.0822112411860042e-06, "epoch": 6.338983050847458, "percentage": 90.56, "elapsed_time": "16:15:48", "remaining_time": "1:41:45"} +{"current_steps": 3745, "total_steps": 4130, "loss": 0.2473, "lr": 1.0549545673474304e-06, "epoch": 6.3474576271186445, "percentage": 90.68, "elapsed_time": "16:17:07", "remaining_time": "1:40:27"} +{"current_steps": 3750, "total_steps": 4130, "loss": 0.2618, "lr": 1.0280362306198732e-06, "epoch": 6.3559322033898304, "percentage": 90.8, "elapsed_time": "16:18:24", "remaining_time": "1:39:08"} +{"current_steps": 3755, "total_steps": 4130, "loss": 0.2702, "lr": 1.0014567117344121e-06, "epoch": 6.364406779661017, "percentage": 90.92, "elapsed_time": "16:19:44", "remaining_time": "1:37:50"} +{"current_steps": 3760, "total_steps": 4130, "loss": 0.2445, "lr": 9.752164853712355e-07, "epoch": 6.372881355932203, "percentage": 91.04, "elapsed_time": "16:21:04", "remaining_time": "1:36:32"} +{"current_steps": 3765, "total_steps": 4130, "loss": 0.2669, "lr": 9.493160201511409e-07, "epoch": 6.38135593220339, "percentage": 91.16, "elapsed_time": "16:22:24", "remaining_time": "1:35:14"} +{"current_steps": 3770, "total_steps": 4130, "loss": 0.2653, "lr": 9.2375577862718e-07, "epoch": 6.389830508474576, "percentage": 91.28, "elapsed_time": "16:23:38", "remaining_time": "1:33:55"} +{"current_steps": 3775, "total_steps": 4130, "loss": 0.2463, "lr": 8.985362172763933e-07, "epoch": 6.398305084745763, "percentage": 91.4, "elapsed_time": "16:24:57", "remaining_time": "1:32:37"} +{"current_steps": 3780, "total_steps": 4130, "loss": 0.2513, "lr": 8.736577864916617e-07, "epoch": 6.406779661016949, "percentage": 91.53, "elapsed_time": "16:26:17", "remaining_time": "1:31:19"} +{"current_steps": 3785, "total_steps": 4130, "loss": 0.2546, "lr": 8.491209305736525e-07, "epoch": 6.415254237288136, "percentage": 91.65, "elapsed_time": "16:27:36", "remaining_time": "1:30:01"} +{"current_steps": 3790, "total_steps": 4130, "loss": 0.2876, "lr": 8.249260877228993e-07, "epoch": 6.423728813559322, "percentage": 91.77, "elapsed_time": "16:28:56", "remaining_time": "1:28:43"} +{"current_steps": 3795, "total_steps": 4130, "loss": 0.274, "lr": 8.010736900319615e-07, "epoch": 6.432203389830509, "percentage": 91.89, "elapsed_time": "16:30:15", "remaining_time": "1:27:24"} +{"current_steps": 3800, "total_steps": 4130, "loss": 0.2621, "lr": 7.775641634777198e-07, "epoch": 6.440677966101695, "percentage": 92.01, "elapsed_time": "16:31:36", "remaining_time": "1:26:06"} +{"current_steps": 3805, "total_steps": 4130, "loss": 0.2596, "lr": 7.543979279137592e-07, "epoch": 6.4491525423728815, "percentage": 92.13, "elapsed_time": "16:32:56", "remaining_time": "1:24:48"} +{"current_steps": 3810, "total_steps": 4130, "loss": 0.2499, "lr": 7.315753970628825e-07, "epoch": 6.4576271186440675, "percentage": 92.25, "elapsed_time": "16:34:12", "remaining_time": "1:23:30"} +{"current_steps": 3815, "total_steps": 4130, "loss": 0.2659, "lr": 7.090969785097046e-07, "epoch": 6.466101694915254, "percentage": 92.37, "elapsed_time": "16:35:33", "remaining_time": "1:22:12"} +{"current_steps": 3820, "total_steps": 4130, "loss": 0.2437, "lr": 6.869630736933963e-07, "epoch": 6.47457627118644, "percentage": 92.49, "elapsed_time": "16:36:49", "remaining_time": "1:20:53"} +{"current_steps": 3825, "total_steps": 4130, "loss": 0.2665, "lr": 6.651740779004878e-07, "epoch": 6.483050847457627, "percentage": 92.62, "elapsed_time": "16:38:02", "remaining_time": "1:19:34"} +{"current_steps": 3830, "total_steps": 4130, "loss": 0.2636, "lr": 6.437303802578365e-07, "epoch": 6.491525423728813, "percentage": 92.74, "elapsed_time": "16:39:21", "remaining_time": "1:18:16"} +{"current_steps": 3835, "total_steps": 4130, "loss": 0.2584, "lr": 6.226323637256592e-07, "epoch": 6.5, "percentage": 92.86, "elapsed_time": "16:40:38", "remaining_time": "1:16:58"} +{"current_steps": 3840, "total_steps": 4130, "loss": 0.2607, "lr": 6.018804050906957e-07, "epoch": 6.508474576271187, "percentage": 92.98, "elapsed_time": "16:41:55", "remaining_time": "1:15:40"} +{"current_steps": 3845, "total_steps": 4130, "loss": 0.2934, "lr": 5.814748749594845e-07, "epoch": 6.516949152542373, "percentage": 93.1, "elapsed_time": "16:43:16", "remaining_time": "1:14:21"} +{"current_steps": 3850, "total_steps": 4130, "loss": 0.2407, "lr": 5.614161377517491e-07, "epoch": 6.52542372881356, "percentage": 93.22, "elapsed_time": "16:44:36", "remaining_time": "1:13:03"} +{"current_steps": 3855, "total_steps": 4130, "loss": 0.262, "lr": 5.417045516938712e-07, "epoch": 6.533898305084746, "percentage": 93.34, "elapsed_time": "16:45:56", "remaining_time": "1:11:45"} +{"current_steps": 3860, "total_steps": 4130, "loss": 0.266, "lr": 5.223404688125189e-07, "epoch": 6.5423728813559325, "percentage": 93.46, "elapsed_time": "16:47:12", "remaining_time": "1:10:27"} +{"current_steps": 3865, "total_steps": 4130, "loss": 0.2479, "lr": 5.033242349283307e-07, "epoch": 6.5508474576271185, "percentage": 93.58, "elapsed_time": "16:48:31", "remaining_time": "1:09:08"} +{"current_steps": 3870, "total_steps": 4130, "loss": 0.2898, "lr": 4.846561896497682e-07, "epoch": 6.559322033898305, "percentage": 93.7, "elapsed_time": "16:49:51", "remaining_time": "1:07:50"} +{"current_steps": 3875, "total_steps": 4130, "loss": 0.2574, "lr": 4.6633666636703325e-07, "epoch": 6.567796610169491, "percentage": 93.83, "elapsed_time": "16:51:10", "remaining_time": "1:06:32"} +{"current_steps": 3880, "total_steps": 4130, "loss": 0.2534, "lr": 4.4836599224611586e-07, "epoch": 6.576271186440678, "percentage": 93.95, "elapsed_time": "16:52:28", "remaining_time": "1:05:14"} +{"current_steps": 3885, "total_steps": 4130, "loss": 0.2677, "lr": 4.3074448822295165e-07, "epoch": 6.584745762711864, "percentage": 94.07, "elapsed_time": "16:53:47", "remaining_time": "1:03:55"} +{"current_steps": 3890, "total_steps": 4130, "loss": 0.2531, "lr": 4.1347246899770435e-07, "epoch": 6.593220338983051, "percentage": 94.19, "elapsed_time": "16:55:00", "remaining_time": "1:02:37"} +{"current_steps": 3895, "total_steps": 4130, "loss": 0.2603, "lr": 3.965502430291235e-07, "epoch": 6.601694915254237, "percentage": 94.31, "elapsed_time": "16:56:21", "remaining_time": "1:01:19"} +{"current_steps": 3900, "total_steps": 4130, "loss": 0.2469, "lr": 3.7997811252905135e-07, "epoch": 6.610169491525424, "percentage": 94.43, "elapsed_time": "16:57:39", "remaining_time": "1:00:00"} +{"current_steps": 3905, "total_steps": 4130, "loss": 0.2932, "lr": 3.6375637345701376e-07, "epoch": 6.61864406779661, "percentage": 94.55, "elapsed_time": "16:59:00", "remaining_time": "0:58:42"} +{"current_steps": 3910, "total_steps": 4130, "loss": 0.2353, "lr": 3.4788531551495307e-07, "epoch": 6.627118644067797, "percentage": 94.67, "elapsed_time": "17:00:15", "remaining_time": "0:57:24"} +{"current_steps": 3915, "total_steps": 4130, "loss": 0.2673, "lr": 3.323652221420326e-07, "epoch": 6.635593220338983, "percentage": 94.79, "elapsed_time": "17:01:35", "remaining_time": "0:56:06"} +{"current_steps": 3920, "total_steps": 4130, "loss": 0.2459, "lr": 3.1719637050958706e-07, "epoch": 6.6440677966101696, "percentage": 94.92, "elapsed_time": "17:02:53", "remaining_time": "0:54:47"} +{"current_steps": 3925, "total_steps": 4130, "loss": 0.2355, "lr": 3.023790315161734e-07, "epoch": 6.652542372881356, "percentage": 95.04, "elapsed_time": "17:04:12", "remaining_time": "0:53:29"} +{"current_steps": 3930, "total_steps": 4130, "loss": 0.2474, "lr": 2.8791346978273015e-07, "epoch": 6.661016949152542, "percentage": 95.16, "elapsed_time": "17:05:29", "remaining_time": "0:52:11"} +{"current_steps": 3935, "total_steps": 4130, "loss": 0.2761, "lr": 2.7379994364784556e-07, "epoch": 6.669491525423728, "percentage": 95.28, "elapsed_time": "17:06:47", "remaining_time": "0:50:52"} +{"current_steps": 3940, "total_steps": 4130, "loss": 0.247, "lr": 2.60038705163157e-07, "epoch": 6.677966101694915, "percentage": 95.4, "elapsed_time": "17:08:06", "remaining_time": "0:49:34"} +{"current_steps": 3945, "total_steps": 4130, "loss": 0.2969, "lr": 2.4663000008883664e-07, "epoch": 6.686440677966102, "percentage": 95.52, "elapsed_time": "17:09:26", "remaining_time": "0:48:16"} +{"current_steps": 3950, "total_steps": 4130, "loss": 0.2584, "lr": 2.335740678892129e-07, "epoch": 6.694915254237288, "percentage": 95.64, "elapsed_time": "17:10:40", "remaining_time": "0:46:58"} +{"current_steps": 3955, "total_steps": 4130, "loss": 0.2768, "lr": 2.20871141728487e-07, "epoch": 6.703389830508475, "percentage": 95.76, "elapsed_time": "17:11:57", "remaining_time": "0:45:39"} +{"current_steps": 3960, "total_steps": 4130, "loss": 0.2552, "lr": 2.0852144846656764e-07, "epoch": 6.711864406779661, "percentage": 95.88, "elapsed_time": "17:13:15", "remaining_time": "0:44:21"} +{"current_steps": 3965, "total_steps": 4130, "loss": 0.268, "lr": 1.9652520865502734e-07, "epoch": 6.720338983050848, "percentage": 96.0, "elapsed_time": "17:14:33", "remaining_time": "0:43:03"} +{"current_steps": 3970, "total_steps": 4130, "loss": 0.2741, "lr": 1.848826365331613e-07, "epoch": 6.728813559322034, "percentage": 96.13, "elapsed_time": "17:15:50", "remaining_time": "0:41:44"} +{"current_steps": 3975, "total_steps": 4130, "loss": 0.255, "lr": 1.7359394002415265e-07, "epoch": 6.737288135593221, "percentage": 96.25, "elapsed_time": "17:17:10", "remaining_time": "0:40:26"} +{"current_steps": 3980, "total_steps": 4130, "loss": 0.2504, "lr": 1.626593207313709e-07, "epoch": 6.745762711864407, "percentage": 96.37, "elapsed_time": "17:18:29", "remaining_time": "0:39:08"} +{"current_steps": 3985, "total_steps": 4130, "loss": 0.2742, "lr": 1.520789739347617e-07, "epoch": 6.754237288135593, "percentage": 96.49, "elapsed_time": "17:19:49", "remaining_time": "0:37:50"} +{"current_steps": 3990, "total_steps": 4130, "loss": 0.2872, "lr": 1.4185308858737367e-07, "epoch": 6.762711864406779, "percentage": 96.61, "elapsed_time": "17:21:09", "remaining_time": "0:36:31"} +{"current_steps": 3995, "total_steps": 4130, "loss": 0.2565, "lr": 1.3198184731196162e-07, "epoch": 6.771186440677966, "percentage": 96.73, "elapsed_time": "17:22:24", "remaining_time": "0:35:13"} +{"current_steps": 4000, "total_steps": 4130, "loss": 0.2911, "lr": 1.2246542639774871e-07, "epoch": 6.779661016949152, "percentage": 96.85, "elapsed_time": "17:23:43", "remaining_time": "0:33:55"} +{"current_steps": 4005, "total_steps": 4130, "loss": 0.2516, "lr": 1.1330399579725814e-07, "epoch": 6.788135593220339, "percentage": 96.97, "elapsed_time": "17:25:01", "remaining_time": "0:32:36"} +{"current_steps": 4010, "total_steps": 4130, "loss": 0.2692, "lr": 1.0449771912329543e-07, "epoch": 6.796610169491525, "percentage": 97.09, "elapsed_time": "17:26:20", "remaining_time": "0:31:18"} +{"current_steps": 4015, "total_steps": 4130, "loss": 0.2814, "lr": 9.604675364601079e-08, "epoch": 6.805084745762712, "percentage": 97.22, "elapsed_time": "17:27:41", "remaining_time": "0:30:00"} +{"current_steps": 4020, "total_steps": 4130, "loss": 0.2358, "lr": 8.795125029010587e-08, "epoch": 6.813559322033898, "percentage": 97.34, "elapsed_time": "17:28:58", "remaining_time": "0:28:42"} +{"current_steps": 4025, "total_steps": 4130, "loss": 0.2683, "lr": 8.021135363212252e-08, "epoch": 6.822033898305085, "percentage": 97.46, "elapsed_time": "17:30:16", "remaining_time": "0:27:23"} +{"current_steps": 4030, "total_steps": 4130, "loss": 0.2589, "lr": 7.282720189787374e-08, "epoch": 6.830508474576272, "percentage": 97.58, "elapsed_time": "17:31:36", "remaining_time": "0:26:05"} +{"current_steps": 4035, "total_steps": 4130, "loss": 0.2547, "lr": 6.579892695996571e-08, "epoch": 6.838983050847458, "percentage": 97.7, "elapsed_time": "17:32:53", "remaining_time": "0:24:47"} +{"current_steps": 4040, "total_steps": 4130, "loss": 0.2601, "lr": 5.912665433545517e-08, "epoch": 6.847457627118644, "percentage": 97.82, "elapsed_time": "17:34:13", "remaining_time": "0:23:29"} +{"current_steps": 4045, "total_steps": 4130, "loss": 0.2419, "lr": 5.281050318358683e-08, "epoch": 6.8559322033898304, "percentage": 97.94, "elapsed_time": "17:35:32", "remaining_time": "0:22:10"} +{"current_steps": 4050, "total_steps": 4130, "loss": 0.2775, "lr": 4.685058630369055e-08, "epoch": 6.864406779661017, "percentage": 98.06, "elapsed_time": "17:36:52", "remaining_time": "0:20:52"} +{"current_steps": 4055, "total_steps": 4130, "loss": 0.2836, "lr": 4.124701013314969e-08, "epoch": 6.872881355932203, "percentage": 98.18, "elapsed_time": "17:38:12", "remaining_time": "0:19:34"} +{"current_steps": 4060, "total_steps": 4130, "loss": 0.2677, "lr": 3.599987474550259e-08, "epoch": 6.88135593220339, "percentage": 98.31, "elapsed_time": "17:39:31", "remaining_time": "0:18:16"} +{"current_steps": 4065, "total_steps": 4130, "loss": 0.2785, "lr": 3.110927384865958e-08, "epoch": 6.889830508474576, "percentage": 98.43, "elapsed_time": "17:40:49", "remaining_time": "0:16:57"} +{"current_steps": 4070, "total_steps": 4130, "loss": 0.2329, "lr": 2.6575294783230954e-08, "epoch": 6.898305084745763, "percentage": 98.55, "elapsed_time": "17:42:01", "remaining_time": "0:15:39"} +{"current_steps": 4075, "total_steps": 4130, "loss": 0.2457, "lr": 2.239801852095935e-08, "epoch": 6.906779661016949, "percentage": 98.67, "elapsed_time": "17:43:11", "remaining_time": "0:14:20"} +{"current_steps": 4080, "total_steps": 4130, "loss": 0.2409, "lr": 1.8577519663278697e-08, "epoch": 6.915254237288136, "percentage": 98.79, "elapsed_time": "17:44:24", "remaining_time": "0:13:02"} +{"current_steps": 4085, "total_steps": 4130, "loss": 0.285, "lr": 1.5113866439986357e-08, "epoch": 6.923728813559322, "percentage": 98.91, "elapsed_time": "17:45:40", "remaining_time": "0:11:44"} +{"current_steps": 4090, "total_steps": 4130, "loss": 0.2731, "lr": 1.2007120708015241e-08, "epoch": 6.932203389830509, "percentage": 99.03, "elapsed_time": "17:46:59", "remaining_time": "0:10:26"} +{"current_steps": 4095, "total_steps": 4130, "loss": 0.2669, "lr": 9.257337950332456e-09, "epoch": 6.940677966101695, "percentage": 99.15, "elapsed_time": "17:48:19", "remaining_time": "0:09:07"} +{"current_steps": 4100, "total_steps": 4130, "loss": 0.2706, "lr": 6.864567274955658e-09, "epoch": 6.9491525423728815, "percentage": 99.27, "elapsed_time": "17:49:39", "remaining_time": "0:07:49"} +{"current_steps": 4105, "total_steps": 4130, "loss": 0.2787, "lr": 4.828851414062641e-09, "epoch": 6.9576271186440675, "percentage": 99.39, "elapsed_time": "17:50:59", "remaining_time": "0:06:31"} +{"current_steps": 4110, "total_steps": 4130, "loss": 0.2701, "lr": 3.150226723243055e-09, "epoch": 6.966101694915254, "percentage": 99.52, "elapsed_time": "17:52:20", "remaining_time": "0:05:13"} +{"current_steps": 4115, "total_steps": 4130, "loss": 0.2611, "lr": 1.8287231808322702e-09, "epoch": 6.97457627118644, "percentage": 99.64, "elapsed_time": "17:53:38", "remaining_time": "0:03:54"} +{"current_steps": 4120, "total_steps": 4130, "loss": 0.246, "lr": 8.643643873962326e-10, "epoch": 6.983050847457627, "percentage": 99.76, "elapsed_time": "17:54:57", "remaining_time": "0:02:36"} +{"current_steps": 4125, "total_steps": 4130, "loss": 0.2559, "lr": 2.5716756528959553e-10, "epoch": 6.991525423728813, "percentage": 99.88, "elapsed_time": "17:56:16", "remaining_time": "0:01:18"} +{"current_steps": 4130, "total_steps": 4130, "loss": 0.2482, "lr": 7.143558364841597e-12, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "17:57:36", "remaining_time": "0:00:00"} +{"current_steps": 4130, "total_steps": 4130, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "17:58:18", "remaining_time": "0:00:00"} diff --git a/trainer_state.json b/trainer_state.json new file mode 100644 index 0000000..e23a90a --- /dev/null +++ b/trainer_state.json @@ -0,0 +1,9133 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 7.0, + "eval_steps": 500, + "global_step": 4130, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.00847457627118644, + "grad_norm": 22.647816413525792, + "learning_rate": 3.8740920096852305e-07, + "loss": 0.8805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.288921058177948, + "step": 5, + "valid_targets_mean": 4341.6, + "valid_targets_min": 2919 + }, + { + "epoch": 0.01694915254237288, + "grad_norm": 21.990651536413747, + "learning_rate": 8.716707021791768e-07, + "loss": 0.8479, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.5880306363105774, + "step": 10, + "valid_targets_mean": 9507.8, + "valid_targets_min": 4513 + }, + { + "epoch": 0.025423728813559324, + "grad_norm": 17.586196084839443, + "learning_rate": 1.3559322033898307e-06, + "loss": 0.7925, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.41249993443489075, + "step": 15, + "valid_targets_mean": 9530.0, + "valid_targets_min": 5082 + }, + { + "epoch": 0.03389830508474576, + "grad_norm": 12.681774144921047, + "learning_rate": 1.8401937046004845e-06, + "loss": 0.7771, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.42156755924224854, + "step": 20, + "valid_targets_mean": 8183.4, + "valid_targets_min": 3531 + }, + { + "epoch": 0.0423728813559322, + "grad_norm": 6.663565745843069, + "learning_rate": 2.324455205811138e-06, + "loss": 0.6944, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.38799262046813965, + "step": 25, + "valid_targets_mean": 10096.0, + "valid_targets_min": 4076 + }, + { + "epoch": 0.05084745762711865, + "grad_norm": 3.7100294856635565, + "learning_rate": 2.808716707021792e-06, + "loss": 0.681, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2823716700077057, + "step": 30, + "valid_targets_mean": 8515.1, + "valid_targets_min": 3821 + }, + { + "epoch": 0.059322033898305086, + "grad_norm": 2.2606726637163868, + "learning_rate": 3.2929782082324455e-06, + "loss": 0.6805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2876054644584656, + "step": 35, + "valid_targets_mean": 5941.0, + "valid_targets_min": 2971 + }, + { + "epoch": 0.06779661016949153, + "grad_norm": 1.456504970480383, + "learning_rate": 3.7772397094430994e-06, + "loss": 0.5617, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2882470190525055, + "step": 40, + "valid_targets_mean": 6889.1, + "valid_targets_min": 3245 + }, + { + "epoch": 0.07627118644067797, + "grad_norm": 1.172636032370015, + "learning_rate": 4.261501210653753e-06, + "loss": 0.6141, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3414521813392639, + "step": 45, + "valid_targets_mean": 9074.0, + "valid_targets_min": 2893 + }, + { + "epoch": 0.0847457627118644, + "grad_norm": 0.9481635217519315, + "learning_rate": 4.745762711864408e-06, + "loss": 0.5575, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3006157875061035, + "step": 50, + "valid_targets_mean": 9120.2, + "valid_targets_min": 4524 + }, + { + "epoch": 0.09322033898305085, + "grad_norm": 0.7620100721554565, + "learning_rate": 5.230024213075061e-06, + "loss": 0.5185, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26493698358535767, + "step": 55, + "valid_targets_mean": 7610.0, + "valid_targets_min": 2214 + }, + { + "epoch": 0.1016949152542373, + "grad_norm": 0.6750070679245781, + "learning_rate": 5.7142857142857145e-06, + "loss": 0.5227, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33802706003189087, + "step": 60, + "valid_targets_mean": 9920.9, + "valid_targets_min": 6570 + }, + { + "epoch": 0.11016949152542373, + "grad_norm": 0.5834293025203762, + "learning_rate": 6.198547215496369e-06, + "loss": 0.5268, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3178923726081848, + "step": 65, + "valid_targets_mean": 8564.9, + "valid_targets_min": 3453 + }, + { + "epoch": 0.11864406779661017, + "grad_norm": 0.6433832325871438, + "learning_rate": 6.682808716707022e-06, + "loss": 0.5186, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2924896478652954, + "step": 70, + "valid_targets_mean": 8542.8, + "valid_targets_min": 4081 + }, + { + "epoch": 0.1271186440677966, + "grad_norm": 0.5102951653191009, + "learning_rate": 7.1670702179176766e-06, + "loss": 0.5181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2909778356552124, + "step": 75, + "valid_targets_mean": 9478.9, + "valid_targets_min": 2078 + }, + { + "epoch": 0.13559322033898305, + "grad_norm": 0.514344781405155, + "learning_rate": 7.65133171912833e-06, + "loss": 0.4839, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19710806012153625, + "step": 80, + "valid_targets_mean": 6526.2, + "valid_targets_min": 2923 + }, + { + "epoch": 0.1440677966101695, + "grad_norm": 0.5918854407115377, + "learning_rate": 8.135593220338983e-06, + "loss": 0.5197, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.44993823766708374, + "step": 85, + "valid_targets_mean": 12081.6, + "valid_targets_min": 3606 + }, + { + "epoch": 0.15254237288135594, + "grad_norm": 0.48936835430375303, + "learning_rate": 8.619854721549637e-06, + "loss": 0.517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.28323233127593994, + "step": 90, + "valid_targets_mean": 8565.9, + "valid_targets_min": 3485 + }, + { + "epoch": 0.16101694915254236, + "grad_norm": 0.47441357303678267, + "learning_rate": 9.10411622276029e-06, + "loss": 0.4704, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1535816788673401, + "step": 95, + "valid_targets_mean": 6248.6, + "valid_targets_min": 3001 + }, + { + "epoch": 0.1694915254237288, + "grad_norm": 0.43781999486196965, + "learning_rate": 9.588377723970946e-06, + "loss": 0.4631, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20473748445510864, + "step": 100, + "valid_targets_mean": 7446.6, + "valid_targets_min": 3441 + }, + { + "epoch": 0.17796610169491525, + "grad_norm": 0.44793786503274047, + "learning_rate": 1.0072639225181599e-05, + "loss": 0.4495, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20548701286315918, + "step": 105, + "valid_targets_mean": 6873.8, + "valid_targets_min": 2833 + }, + { + "epoch": 0.1864406779661017, + "grad_norm": 0.49319289641225694, + "learning_rate": 1.0556900726392252e-05, + "loss": 0.4323, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2931867837905884, + "step": 110, + "valid_targets_mean": 8893.8, + "valid_targets_min": 1702 + }, + { + "epoch": 0.19491525423728814, + "grad_norm": 0.4386454650682578, + "learning_rate": 1.1041162227602906e-05, + "loss": 0.4595, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15317876636981964, + "step": 115, + "valid_targets_mean": 6561.0, + "valid_targets_min": 1806 + }, + { + "epoch": 0.2033898305084746, + "grad_norm": 0.49070460068044697, + "learning_rate": 1.1525423728813561e-05, + "loss": 0.4257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2104090005159378, + "step": 120, + "valid_targets_mean": 6806.9, + "valid_targets_min": 2801 + }, + { + "epoch": 0.211864406779661, + "grad_norm": 0.48702889996642884, + "learning_rate": 1.2009685230024215e-05, + "loss": 0.4552, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2909752130508423, + "step": 125, + "valid_targets_mean": 9399.8, + "valid_targets_min": 2333 + }, + { + "epoch": 0.22033898305084745, + "grad_norm": 0.5090713737402093, + "learning_rate": 1.2493946731234868e-05, + "loss": 0.4537, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.3254137933254242, + "step": 130, + "valid_targets_mean": 9565.8, + "valid_targets_min": 3205 + }, + { + "epoch": 0.2288135593220339, + "grad_norm": 0.48330297583304416, + "learning_rate": 1.2978208232445521e-05, + "loss": 0.4045, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22189545631408691, + "step": 135, + "valid_targets_mean": 7905.2, + "valid_targets_min": 1760 + }, + { + "epoch": 0.23728813559322035, + "grad_norm": 0.46052557354238716, + "learning_rate": 1.3462469733656177e-05, + "loss": 0.4456, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19440916180610657, + "step": 140, + "valid_targets_mean": 6993.1, + "valid_targets_min": 3707 + }, + { + "epoch": 0.2457627118644068, + "grad_norm": 0.4909819940196489, + "learning_rate": 1.3946731234866828e-05, + "loss": 0.4839, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2511296272277832, + "step": 145, + "valid_targets_mean": 7719.5, + "valid_targets_min": 3264 + }, + { + "epoch": 0.2542372881355932, + "grad_norm": 0.5594539670454729, + "learning_rate": 1.4430992736077482e-05, + "loss": 0.3795, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1961612105369568, + "step": 150, + "valid_targets_mean": 7736.5, + "valid_targets_min": 2212 + }, + { + "epoch": 0.2627118644067797, + "grad_norm": 0.4514188587237233, + "learning_rate": 1.4915254237288137e-05, + "loss": 0.422, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20295317471027374, + "step": 155, + "valid_targets_mean": 6204.5, + "valid_targets_min": 2781 + }, + { + "epoch": 0.2711864406779661, + "grad_norm": 0.4999364100558755, + "learning_rate": 1.5399515738498792e-05, + "loss": 0.3944, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19053910672664642, + "step": 160, + "valid_targets_mean": 7032.2, + "valid_targets_min": 3634 + }, + { + "epoch": 0.2796610169491525, + "grad_norm": 0.45346810997487935, + "learning_rate": 1.5883777239709442e-05, + "loss": 0.3938, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14634712040424347, + "step": 165, + "valid_targets_mean": 7056.1, + "valid_targets_min": 3444 + }, + { + "epoch": 0.288135593220339, + "grad_norm": 0.46398811821797475, + "learning_rate": 1.63680387409201e-05, + "loss": 0.4304, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1805558204650879, + "step": 170, + "valid_targets_mean": 8803.9, + "valid_targets_min": 3245 + }, + { + "epoch": 0.2966101694915254, + "grad_norm": 0.515402662275605, + "learning_rate": 1.6852300242130752e-05, + "loss": 0.3997, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23095428943634033, + "step": 175, + "valid_targets_mean": 8863.0, + "valid_targets_min": 3852 + }, + { + "epoch": 0.3050847457627119, + "grad_norm": 0.4283835795039386, + "learning_rate": 1.7336561743341406e-05, + "loss": 0.4092, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22036141157150269, + "step": 180, + "valid_targets_mean": 9179.9, + "valid_targets_min": 4131 + }, + { + "epoch": 0.3135593220338983, + "grad_norm": 0.4781421216849134, + "learning_rate": 1.782082324455206e-05, + "loss": 0.4097, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20116378366947174, + "step": 185, + "valid_targets_mean": 8578.9, + "valid_targets_min": 5608 + }, + { + "epoch": 0.3220338983050847, + "grad_norm": 0.5200480518561265, + "learning_rate": 1.8305084745762713e-05, + "loss": 0.4365, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15508520603179932, + "step": 190, + "valid_targets_mean": 5893.8, + "valid_targets_min": 2951 + }, + { + "epoch": 0.3305084745762712, + "grad_norm": 0.5134986617901804, + "learning_rate": 1.8789346246973366e-05, + "loss": 0.3837, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2245933711528778, + "step": 195, + "valid_targets_mean": 6491.8, + "valid_targets_min": 1908 + }, + { + "epoch": 0.3389830508474576, + "grad_norm": 0.39547736013974827, + "learning_rate": 1.927360774818402e-05, + "loss": 0.4048, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23196324706077576, + "step": 200, + "valid_targets_mean": 11168.2, + "valid_targets_min": 5671 + }, + { + "epoch": 0.3474576271186441, + "grad_norm": 0.6662190158998941, + "learning_rate": 1.9757869249394673e-05, + "loss": 0.4291, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.29568159580230713, + "step": 205, + "valid_targets_mean": 12355.8, + "valid_targets_min": 5576 + }, + { + "epoch": 0.3559322033898305, + "grad_norm": 0.481082014069215, + "learning_rate": 2.0242130750605327e-05, + "loss": 0.3686, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22347992658615112, + "step": 210, + "valid_targets_mean": 9118.5, + "valid_targets_min": 3127 + }, + { + "epoch": 0.3644067796610169, + "grad_norm": 0.50188354175535, + "learning_rate": 2.072639225181598e-05, + "loss": 0.3826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14265790581703186, + "step": 215, + "valid_targets_mean": 6121.6, + "valid_targets_min": 2475 + }, + { + "epoch": 0.3728813559322034, + "grad_norm": 0.45850382713351473, + "learning_rate": 2.1210653753026637e-05, + "loss": 0.3656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.30828917026519775, + "step": 220, + "valid_targets_mean": 10314.8, + "valid_targets_min": 4913 + }, + { + "epoch": 0.3813559322033898, + "grad_norm": 0.4704525379286944, + "learning_rate": 2.169491525423729e-05, + "loss": 0.3937, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18286801874637604, + "step": 225, + "valid_targets_mean": 8533.8, + "valid_targets_min": 6406 + }, + { + "epoch": 0.3898305084745763, + "grad_norm": 0.5390826850869097, + "learning_rate": 2.2179176755447944e-05, + "loss": 0.3614, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14486080408096313, + "step": 230, + "valid_targets_mean": 6356.0, + "valid_targets_min": 2784 + }, + { + "epoch": 0.3983050847457627, + "grad_norm": 0.45845359728031587, + "learning_rate": 2.2663438256658597e-05, + "loss": 0.4322, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12409990280866623, + "step": 235, + "valid_targets_mean": 6001.4, + "valid_targets_min": 2401 + }, + { + "epoch": 0.4067796610169492, + "grad_norm": 0.44770892119992406, + "learning_rate": 2.3147699757869254e-05, + "loss": 0.3513, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15416109561920166, + "step": 240, + "valid_targets_mean": 7657.6, + "valid_targets_min": 2306 + }, + { + "epoch": 0.4152542372881356, + "grad_norm": 0.48992403126988787, + "learning_rate": 2.3631961259079904e-05, + "loss": 0.3925, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2023354172706604, + "step": 245, + "valid_targets_mean": 6805.5, + "valid_targets_min": 2903 + }, + { + "epoch": 0.423728813559322, + "grad_norm": 0.49021429282553164, + "learning_rate": 2.4116222760290558e-05, + "loss": 0.3952, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21003684401512146, + "step": 250, + "valid_targets_mean": 8469.0, + "valid_targets_min": 2457 + }, + { + "epoch": 0.4322033898305085, + "grad_norm": 0.4640908570139275, + "learning_rate": 2.460048426150121e-05, + "loss": 0.3749, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18187469244003296, + "step": 255, + "valid_targets_mean": 7923.5, + "valid_targets_min": 3374 + }, + { + "epoch": 0.4406779661016949, + "grad_norm": 0.4725941251523362, + "learning_rate": 2.5084745762711865e-05, + "loss": 0.3864, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21986892819404602, + "step": 260, + "valid_targets_mean": 8163.0, + "valid_targets_min": 4396 + }, + { + "epoch": 0.4491525423728814, + "grad_norm": 0.504672773087851, + "learning_rate": 2.556900726392252e-05, + "loss": 0.3826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18904973566532135, + "step": 265, + "valid_targets_mean": 7939.8, + "valid_targets_min": 2622 + }, + { + "epoch": 0.4576271186440678, + "grad_norm": 0.5254794688428569, + "learning_rate": 2.6053268765133175e-05, + "loss": 0.3712, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14334720373153687, + "step": 270, + "valid_targets_mean": 6861.9, + "valid_targets_min": 2824 + }, + { + "epoch": 0.4661016949152542, + "grad_norm": 0.48544091835627723, + "learning_rate": 2.653753026634383e-05, + "loss": 0.4284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25785499811172485, + "step": 275, + "valid_targets_mean": 10133.1, + "valid_targets_min": 2843 + }, + { + "epoch": 0.4745762711864407, + "grad_norm": 0.45398873282760527, + "learning_rate": 2.702179176755448e-05, + "loss": 0.4028, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26890379190444946, + "step": 280, + "valid_targets_mean": 10401.2, + "valid_targets_min": 2623 + }, + { + "epoch": 0.4830508474576271, + "grad_norm": 0.4653503731278924, + "learning_rate": 2.7506053268765135e-05, + "loss": 0.3862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23974227905273438, + "step": 285, + "valid_targets_mean": 9559.5, + "valid_targets_min": 4730 + }, + { + "epoch": 0.4915254237288136, + "grad_norm": 0.5221118689000261, + "learning_rate": 2.799031476997579e-05, + "loss": 0.3911, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18821915984153748, + "step": 290, + "valid_targets_mean": 7395.4, + "valid_targets_min": 2127 + }, + { + "epoch": 0.5, + "grad_norm": 0.4816270445038903, + "learning_rate": 2.8474576271186442e-05, + "loss": 0.3677, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2023887038230896, + "step": 295, + "valid_targets_mean": 7899.9, + "valid_targets_min": 1955 + }, + { + "epoch": 0.5084745762711864, + "grad_norm": 0.4768576322729555, + "learning_rate": 2.8958837772397096e-05, + "loss": 0.3671, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1648670732975006, + "step": 300, + "valid_targets_mean": 7478.8, + "valid_targets_min": 1736 + }, + { + "epoch": 0.5169491525423728, + "grad_norm": 0.5293806390444394, + "learning_rate": 2.9443099273607753e-05, + "loss": 0.3601, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2311454862356186, + "step": 305, + "valid_targets_mean": 10215.6, + "valid_targets_min": 6512 + }, + { + "epoch": 0.5254237288135594, + "grad_norm": 0.5210685990067819, + "learning_rate": 2.9927360774818406e-05, + "loss": 0.3657, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15663181245326996, + "step": 310, + "valid_targets_mean": 6098.5, + "valid_targets_min": 2342 + }, + { + "epoch": 0.5338983050847458, + "grad_norm": 0.48005310800960843, + "learning_rate": 3.041162227602906e-05, + "loss": 0.3619, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14108559489250183, + "step": 315, + "valid_targets_mean": 7621.5, + "valid_targets_min": 5209 + }, + { + "epoch": 0.5423728813559322, + "grad_norm": 0.49482029985465964, + "learning_rate": 3.089588377723971e-05, + "loss": 0.3822, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19488084316253662, + "step": 320, + "valid_targets_mean": 7531.2, + "valid_targets_min": 1613 + }, + { + "epoch": 0.5508474576271186, + "grad_norm": 0.47675644302446346, + "learning_rate": 3.138014527845036e-05, + "loss": 0.3587, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12048106640577316, + "step": 325, + "valid_targets_mean": 5926.4, + "valid_targets_min": 4436 + }, + { + "epoch": 0.559322033898305, + "grad_norm": 0.49744425649436924, + "learning_rate": 3.186440677966102e-05, + "loss": 0.357, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16427424550056458, + "step": 330, + "valid_targets_mean": 7583.2, + "valid_targets_min": 2576 + }, + { + "epoch": 0.5677966101694916, + "grad_norm": 0.5088179460735696, + "learning_rate": 3.234866828087168e-05, + "loss": 0.3714, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1933363527059555, + "step": 335, + "valid_targets_mean": 8849.0, + "valid_targets_min": 2818 + }, + { + "epoch": 0.576271186440678, + "grad_norm": 0.5041082367216122, + "learning_rate": 3.283292978208233e-05, + "loss": 0.3821, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16098180413246155, + "step": 340, + "valid_targets_mean": 6717.6, + "valid_targets_min": 1670 + }, + { + "epoch": 0.5847457627118644, + "grad_norm": 0.48735048980977763, + "learning_rate": 3.3317191283292984e-05, + "loss": 0.388, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1445331871509552, + "step": 345, + "valid_targets_mean": 7253.1, + "valid_targets_min": 1976 + }, + { + "epoch": 0.5932203389830508, + "grad_norm": 0.5011045691225819, + "learning_rate": 3.3801452784503634e-05, + "loss": 0.3823, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23252391815185547, + "step": 350, + "valid_targets_mean": 9214.6, + "valid_targets_min": 2553 + }, + { + "epoch": 0.6016949152542372, + "grad_norm": 0.5129390690315153, + "learning_rate": 3.4285714285714284e-05, + "loss": 0.3749, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2209414392709732, + "step": 355, + "valid_targets_mean": 8149.8, + "valid_targets_min": 2851 + }, + { + "epoch": 0.6101694915254238, + "grad_norm": 0.4345827629814721, + "learning_rate": 3.476997578692494e-05, + "loss": 0.3682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1325065791606903, + "step": 360, + "valid_targets_mean": 8140.2, + "valid_targets_min": 4997 + }, + { + "epoch": 0.6186440677966102, + "grad_norm": 0.476244427890107, + "learning_rate": 3.52542372881356e-05, + "loss": 0.3903, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15084779262542725, + "step": 365, + "valid_targets_mean": 6385.0, + "valid_targets_min": 3888 + }, + { + "epoch": 0.6271186440677966, + "grad_norm": 0.47924393386440584, + "learning_rate": 3.573849878934625e-05, + "loss": 0.3743, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16552937030792236, + "step": 370, + "valid_targets_mean": 7364.0, + "valid_targets_min": 2941 + }, + { + "epoch": 0.635593220338983, + "grad_norm": 0.4954509125530307, + "learning_rate": 3.6222760290556904e-05, + "loss": 0.3868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22299477458000183, + "step": 375, + "valid_targets_mean": 7803.1, + "valid_targets_min": 2953 + }, + { + "epoch": 0.6440677966101694, + "grad_norm": 0.4979006616910799, + "learning_rate": 3.670702179176756e-05, + "loss": 0.3817, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2504671812057495, + "step": 380, + "valid_targets_mean": 8191.4, + "valid_targets_min": 5642 + }, + { + "epoch": 0.652542372881356, + "grad_norm": 0.46347755745616037, + "learning_rate": 3.719128329297821e-05, + "loss": 0.4001, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19989028573036194, + "step": 385, + "valid_targets_mean": 9660.4, + "valid_targets_min": 4356 + }, + { + "epoch": 0.6610169491525424, + "grad_norm": 0.4650135947081464, + "learning_rate": 3.767554479418886e-05, + "loss": 0.3746, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17294305562973022, + "step": 390, + "valid_targets_mean": 7292.4, + "valid_targets_min": 2899 + }, + { + "epoch": 0.6694915254237288, + "grad_norm": 0.43785966553865663, + "learning_rate": 3.815980629539952e-05, + "loss": 0.3509, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15276168286800385, + "step": 395, + "valid_targets_mean": 6736.5, + "valid_targets_min": 2329 + }, + { + "epoch": 0.6779661016949152, + "grad_norm": 0.4393502110188653, + "learning_rate": 3.8644067796610175e-05, + "loss": 0.3717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17951694130897522, + "step": 400, + "valid_targets_mean": 7865.9, + "valid_targets_min": 3219 + }, + { + "epoch": 0.6864406779661016, + "grad_norm": 0.5646370364355308, + "learning_rate": 3.9128329297820825e-05, + "loss": 0.3594, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13970831036567688, + "step": 405, + "valid_targets_mean": 5126.8, + "valid_targets_min": 3110 + }, + { + "epoch": 0.6949152542372882, + "grad_norm": 0.5059385994762761, + "learning_rate": 3.961259079903148e-05, + "loss": 0.3868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1319550722837448, + "step": 410, + "valid_targets_mean": 6735.0, + "valid_targets_min": 3663 + }, + { + "epoch": 0.7033898305084746, + "grad_norm": 0.4441536863390733, + "learning_rate": 3.9999992856441635e-05, + "loss": 0.351, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17535431683063507, + "step": 415, + "valid_targets_mean": 7174.1, + "valid_targets_min": 2377 + }, + { + "epoch": 0.711864406779661, + "grad_norm": 0.47803871319263513, + "learning_rate": 3.999974283243472e-05, + "loss": 0.366, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10861378908157349, + "step": 420, + "valid_targets_mean": 6504.9, + "valid_targets_min": 2418 + }, + { + "epoch": 0.7203389830508474, + "grad_norm": 0.6006114256852154, + "learning_rate": 3.9999135635612606e-05, + "loss": 0.4029, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23201918601989746, + "step": 425, + "valid_targets_mean": 8352.6, + "valid_targets_min": 3709 + }, + { + "epoch": 0.7288135593220338, + "grad_norm": 0.5075799125704737, + "learning_rate": 3.999817127681917e-05, + "loss": 0.3795, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.33688515424728394, + "step": 430, + "valid_targets_mean": 11800.5, + "valid_targets_min": 1920 + }, + { + "epoch": 0.7372881355932204, + "grad_norm": 0.5535367710374015, + "learning_rate": 3.9996849773276764e-05, + "loss": 0.3473, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12052987515926361, + "step": 435, + "valid_targets_mean": 4661.8, + "valid_targets_min": 3397 + }, + { + "epoch": 0.7457627118644068, + "grad_norm": 0.512294047106741, + "learning_rate": 3.9995171148585944e-05, + "loss": 0.3192, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16196998953819275, + "step": 440, + "valid_targets_mean": 6370.0, + "valid_targets_min": 3295 + }, + { + "epoch": 0.7542372881355932, + "grad_norm": 0.4672133373894365, + "learning_rate": 3.999313543272505e-05, + "loss": 0.3584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1513005495071411, + "step": 445, + "valid_targets_mean": 6997.0, + "valid_targets_min": 3496 + }, + { + "epoch": 0.7627118644067796, + "grad_norm": 0.4494037029390078, + "learning_rate": 3.999074266204967e-05, + "loss": 0.3279, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16777727007865906, + "step": 450, + "valid_targets_mean": 8542.5, + "valid_targets_min": 3298 + }, + { + "epoch": 0.7711864406779662, + "grad_norm": 0.53674461562687, + "learning_rate": 3.998799287929199e-05, + "loss": 0.3391, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20486463606357574, + "step": 455, + "valid_targets_mean": 7580.0, + "valid_targets_min": 1230 + }, + { + "epoch": 0.7796610169491526, + "grad_norm": 0.4598145218135828, + "learning_rate": 3.998488613356002e-05, + "loss": 0.3482, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2002931833267212, + "step": 460, + "valid_targets_mean": 9587.5, + "valid_targets_min": 4036 + }, + { + "epoch": 0.788135593220339, + "grad_norm": 0.4901495202397604, + "learning_rate": 3.998142248033673e-05, + "loss": 0.3809, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20463253557682037, + "step": 465, + "valid_targets_mean": 7521.1, + "valid_targets_min": 3494 + }, + { + "epoch": 0.7966101694915254, + "grad_norm": 0.46038899128304067, + "learning_rate": 3.997760198147904e-05, + "loss": 0.3509, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15787944197654724, + "step": 470, + "valid_targets_mean": 6793.0, + "valid_targets_min": 2710 + }, + { + "epoch": 0.8050847457627118, + "grad_norm": 0.48737085076940234, + "learning_rate": 3.9973424705216774e-05, + "loss": 0.3805, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17634077370166779, + "step": 475, + "valid_targets_mean": 7378.4, + "valid_targets_min": 1929 + }, + { + "epoch": 0.8135593220338984, + "grad_norm": 0.43829721409671807, + "learning_rate": 3.996889072615134e-05, + "loss": 0.3276, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21389906108379364, + "step": 480, + "valid_targets_mean": 9768.9, + "valid_targets_min": 2989 + }, + { + "epoch": 0.8220338983050848, + "grad_norm": 0.4887523283448908, + "learning_rate": 3.99640001252545e-05, + "loss": 0.3765, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13324296474456787, + "step": 485, + "valid_targets_mean": 5801.4, + "valid_targets_min": 3758 + }, + { + "epoch": 0.8305084745762712, + "grad_norm": 0.4837948992026841, + "learning_rate": 3.9958752989866856e-05, + "loss": 0.3689, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20884181559085846, + "step": 490, + "valid_targets_mean": 7791.8, + "valid_targets_min": 3525 + }, + { + "epoch": 0.8389830508474576, + "grad_norm": 0.5188813221623474, + "learning_rate": 3.995314941369631e-05, + "loss": 0.3401, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19880643486976624, + "step": 495, + "valid_targets_mean": 7705.0, + "valid_targets_min": 3223 + }, + { + "epoch": 0.847457627118644, + "grad_norm": 0.41769434994726196, + "learning_rate": 3.994718949681642e-05, + "loss": 0.3257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08604508638381958, + "step": 500, + "valid_targets_mean": 4836.0, + "valid_targets_min": 1361 + }, + { + "epoch": 0.8559322033898306, + "grad_norm": 0.7506187012955293, + "learning_rate": 3.994087334566455e-05, + "loss": 0.3487, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17222467064857483, + "step": 505, + "valid_targets_mean": 7686.1, + "valid_targets_min": 4200 + }, + { + "epoch": 0.864406779661017, + "grad_norm": 0.4967778065574364, + "learning_rate": 3.9934201073040035e-05, + "loss": 0.3608, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13858087360858917, + "step": 510, + "valid_targets_mean": 7109.0, + "valid_targets_min": 4815 + }, + { + "epoch": 0.8728813559322034, + "grad_norm": 0.4683226298397944, + "learning_rate": 3.992717279810213e-05, + "loss": 0.3558, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17592290043830872, + "step": 515, + "valid_targets_mean": 7741.9, + "valid_targets_min": 4364 + }, + { + "epoch": 0.8813559322033898, + "grad_norm": 0.5347159130011266, + "learning_rate": 3.991978864636788e-05, + "loss": 0.344, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12308762222528458, + "step": 520, + "valid_targets_mean": 6528.5, + "valid_targets_min": 2710 + }, + { + "epoch": 0.8898305084745762, + "grad_norm": 0.44304888546304205, + "learning_rate": 3.9912048749709896e-05, + "loss": 0.3512, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14396895468235016, + "step": 525, + "valid_targets_mean": 8291.0, + "valid_targets_min": 3067 + }, + { + "epoch": 0.8983050847457628, + "grad_norm": 0.4991466113617328, + "learning_rate": 3.990395324635399e-05, + "loss": 0.3311, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1871216595172882, + "step": 530, + "valid_targets_mean": 8318.8, + "valid_targets_min": 4083 + }, + { + "epoch": 0.9067796610169492, + "grad_norm": 0.453620077636563, + "learning_rate": 3.989550228087671e-05, + "loss": 0.3513, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1634388267993927, + "step": 535, + "valid_targets_mean": 7201.2, + "valid_targets_min": 2887 + }, + { + "epoch": 0.9152542372881356, + "grad_norm": 0.41349895631346173, + "learning_rate": 3.988669600420275e-05, + "loss": 0.3215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1667180359363556, + "step": 540, + "valid_targets_mean": 9424.4, + "valid_targets_min": 4268 + }, + { + "epoch": 0.923728813559322, + "grad_norm": 0.4601829959588257, + "learning_rate": 3.9877534573602254e-05, + "loss": 0.3634, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2731122672557831, + "step": 545, + "valid_targets_mean": 8933.6, + "valid_targets_min": 2811 + }, + { + "epoch": 0.9322033898305084, + "grad_norm": 0.4640913220277186, + "learning_rate": 3.9868018152688044e-05, + "loss": 0.3573, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17383530735969543, + "step": 550, + "valid_targets_mean": 7304.9, + "valid_targets_min": 1703 + }, + { + "epoch": 0.940677966101695, + "grad_norm": 0.5448982634240432, + "learning_rate": 3.985814691141263e-05, + "loss": 0.3615, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22727759182453156, + "step": 555, + "valid_targets_mean": 9085.8, + "valid_targets_min": 3315 + }, + { + "epoch": 0.9491525423728814, + "grad_norm": 0.45385983205548086, + "learning_rate": 3.984792102606524e-05, + "loss": 0.3839, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1131582260131836, + "step": 560, + "valid_targets_mean": 6142.0, + "valid_targets_min": 2981 + }, + { + "epoch": 0.9576271186440678, + "grad_norm": 0.4465361575153305, + "learning_rate": 3.9837340679268634e-05, + "loss": 0.3165, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11280742287635803, + "step": 565, + "valid_targets_mean": 6052.8, + "valid_targets_min": 3070 + }, + { + "epoch": 0.9661016949152542, + "grad_norm": 0.40935340628275685, + "learning_rate": 3.982640605997585e-05, + "loss": 0.3308, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14417824149131775, + "step": 570, + "valid_targets_mean": 7390.8, + "valid_targets_min": 3158 + }, + { + "epoch": 0.9745762711864406, + "grad_norm": 0.4743822005506336, + "learning_rate": 3.981511736346684e-05, + "loss": 0.353, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1906016767024994, + "step": 575, + "valid_targets_mean": 8296.5, + "valid_targets_min": 3963 + }, + { + "epoch": 0.9830508474576272, + "grad_norm": 0.48645691171935773, + "learning_rate": 3.980347479134498e-05, + "loss": 0.3341, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16270534694194794, + "step": 580, + "valid_targets_mean": 7576.6, + "valid_targets_min": 3410 + }, + { + "epoch": 0.9915254237288136, + "grad_norm": 0.4953969392901077, + "learning_rate": 3.9791478551533435e-05, + "loss": 0.3257, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17336884140968323, + "step": 585, + "valid_targets_mean": 7680.1, + "valid_targets_min": 3140 + }, + { + "epoch": 1.0, + "grad_norm": 0.6068448212487687, + "learning_rate": 3.977912885827151e-05, + "loss": 0.3438, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14036622643470764, + "step": 590, + "valid_targets_mean": 6497.4, + "valid_targets_min": 2977 + }, + { + "epoch": 1.0084745762711864, + "grad_norm": 0.46278763763653086, + "learning_rate": 3.976642593211079e-05, + "loss": 0.3632, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25034165382385254, + "step": 595, + "valid_targets_mean": 10331.9, + "valid_targets_min": 3875 + }, + { + "epoch": 1.0169491525423728, + "grad_norm": 0.46464694059275796, + "learning_rate": 3.9753369999911166e-05, + "loss": 0.3324, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15899217128753662, + "step": 600, + "valid_targets_mean": 7868.1, + "valid_targets_min": 3683 + }, + { + "epoch": 1.0254237288135593, + "grad_norm": 0.4186988069340215, + "learning_rate": 3.973996129483684e-05, + "loss": 0.3188, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10641027241945267, + "step": 605, + "valid_targets_mean": 7382.8, + "valid_targets_min": 2512 + }, + { + "epoch": 1.0338983050847457, + "grad_norm": 0.48458724521260527, + "learning_rate": 3.972620005635216e-05, + "loss": 0.2951, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17370054125785828, + "step": 610, + "valid_targets_mean": 7550.0, + "valid_targets_min": 2807 + }, + { + "epoch": 1.042372881355932, + "grad_norm": 0.4682943686309724, + "learning_rate": 3.971208653021727e-05, + "loss": 0.3193, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20658673346042633, + "step": 615, + "valid_targets_mean": 8164.0, + "valid_targets_min": 2490 + }, + { + "epoch": 1.0508474576271187, + "grad_norm": 0.4518264543716217, + "learning_rate": 3.969762096848383e-05, + "loss": 0.3693, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15463273227214813, + "step": 620, + "valid_targets_mean": 7257.2, + "valid_targets_min": 708 + }, + { + "epoch": 1.0593220338983051, + "grad_norm": 0.44674028472988253, + "learning_rate": 3.968280362949042e-05, + "loss": 0.3308, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20777608454227448, + "step": 625, + "valid_targets_mean": 9344.2, + "valid_targets_min": 4668 + }, + { + "epoch": 1.0677966101694916, + "grad_norm": 0.40848401717039223, + "learning_rate": 3.9667634777857975e-05, + "loss": 0.3605, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14285874366760254, + "step": 630, + "valid_targets_mean": 8423.5, + "valid_targets_min": 3299 + }, + { + "epoch": 1.076271186440678, + "grad_norm": 0.5064466280506356, + "learning_rate": 3.965211468448505e-05, + "loss": 0.3408, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14683064818382263, + "step": 635, + "valid_targets_mean": 7132.4, + "valid_targets_min": 3807 + }, + { + "epoch": 1.0847457627118644, + "grad_norm": 0.4438801074138089, + "learning_rate": 3.963624362654299e-05, + "loss": 0.3339, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13330812752246857, + "step": 640, + "valid_targets_mean": 7115.0, + "valid_targets_min": 1920 + }, + { + "epoch": 1.0932203389830508, + "grad_norm": 0.4567192090970232, + "learning_rate": 3.962002188747096e-05, + "loss": 0.3253, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2908390164375305, + "step": 645, + "valid_targets_mean": 10478.0, + "valid_targets_min": 2951 + }, + { + "epoch": 1.1016949152542372, + "grad_norm": 0.4981661086451052, + "learning_rate": 3.9603449756970877e-05, + "loss": 0.3232, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0967227965593338, + "step": 650, + "valid_targets_mean": 6367.0, + "valid_targets_min": 2070 + }, + { + "epoch": 1.1101694915254237, + "grad_norm": 0.4383118368331649, + "learning_rate": 3.95865275310023e-05, + "loss": 0.3291, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09382268786430359, + "step": 655, + "valid_targets_mean": 4957.8, + "valid_targets_min": 2135 + }, + { + "epoch": 1.11864406779661, + "grad_norm": 0.4665208990365514, + "learning_rate": 3.9569255511777054e-05, + "loss": 0.3269, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13936468958854675, + "step": 660, + "valid_targets_mean": 6068.1, + "valid_targets_min": 2937 + }, + { + "epoch": 1.1271186440677967, + "grad_norm": 0.41077826271905876, + "learning_rate": 3.955163400775389e-05, + "loss": 0.3623, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12347327917814255, + "step": 665, + "valid_targets_mean": 6312.8, + "valid_targets_min": 3508 + }, + { + "epoch": 1.1355932203389831, + "grad_norm": 0.49343240139841715, + "learning_rate": 3.953366333363297e-05, + "loss": 0.324, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12519244849681854, + "step": 670, + "valid_targets_mean": 6849.1, + "valid_targets_min": 1397 + }, + { + "epoch": 1.1440677966101696, + "grad_norm": 0.39545538040209194, + "learning_rate": 3.9515343810350236e-05, + "loss": 0.3109, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25974246859550476, + "step": 675, + "valid_targets_mean": 11069.4, + "valid_targets_min": 2809 + }, + { + "epoch": 1.152542372881356, + "grad_norm": 0.46689959668479314, + "learning_rate": 3.949667576507168e-05, + "loss": 0.3356, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2463110387325287, + "step": 680, + "valid_targets_mean": 7808.6, + "valid_targets_min": 2797 + }, + { + "epoch": 1.1610169491525424, + "grad_norm": 0.4984035434942114, + "learning_rate": 3.9477659531187484e-05, + "loss": 0.3321, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2779242992401123, + "step": 685, + "valid_targets_mean": 9492.8, + "valid_targets_min": 1889 + }, + { + "epoch": 1.1694915254237288, + "grad_norm": 0.431916021655085, + "learning_rate": 3.9458295448306134e-05, + "loss": 0.329, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17030349373817444, + "step": 690, + "valid_targets_mean": 8139.1, + "valid_targets_min": 3175 + }, + { + "epoch": 1.1779661016949152, + "grad_norm": 0.4843078612720861, + "learning_rate": 3.943858386224825e-05, + "loss": 0.3561, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20718806982040405, + "step": 695, + "valid_targets_mean": 7875.4, + "valid_targets_min": 2953 + }, + { + "epoch": 1.1864406779661016, + "grad_norm": 0.5669068651728717, + "learning_rate": 3.941852512504052e-05, + "loss": 0.3091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17090219259262085, + "step": 700, + "valid_targets_mean": 8395.6, + "valid_targets_min": 3439 + }, + { + "epoch": 1.194915254237288, + "grad_norm": 0.43839002881668476, + "learning_rate": 3.939811959490931e-05, + "loss": 0.3206, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22540715336799622, + "step": 705, + "valid_targets_mean": 11013.1, + "valid_targets_min": 4393 + }, + { + "epoch": 1.2033898305084745, + "grad_norm": 0.4738148722127186, + "learning_rate": 3.937736763627435e-05, + "loss": 0.3178, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19129815697669983, + "step": 710, + "valid_targets_mean": 8546.8, + "valid_targets_min": 3957 + }, + { + "epoch": 1.211864406779661, + "grad_norm": 0.4721947538647376, + "learning_rate": 3.935626961974217e-05, + "loss": 0.3624, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1431015133857727, + "step": 715, + "valid_targets_mean": 6937.4, + "valid_targets_min": 3007 + }, + { + "epoch": 1.2203389830508475, + "grad_norm": 0.4843590613492343, + "learning_rate": 3.933482592209951e-05, + "loss": 0.3332, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.27055826783180237, + "step": 720, + "valid_targets_mean": 9273.4, + "valid_targets_min": 3369 + }, + { + "epoch": 1.228813559322034, + "grad_norm": 0.46403463903278935, + "learning_rate": 3.931303692630661e-05, + "loss": 0.3116, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16579920053482056, + "step": 725, + "valid_targets_mean": 7017.1, + "valid_targets_min": 2382 + }, + { + "epoch": 1.2372881355932204, + "grad_norm": 0.4996210142837503, + "learning_rate": 3.92909030214903e-05, + "loss": 0.3284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11564907431602478, + "step": 730, + "valid_targets_mean": 6940.6, + "valid_targets_min": 3964 + }, + { + "epoch": 1.2457627118644068, + "grad_norm": 0.4831401711242658, + "learning_rate": 3.9268424602937124e-05, + "loss": 0.3215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17444834113121033, + "step": 735, + "valid_targets_mean": 8816.0, + "valid_targets_min": 2620 + }, + { + "epoch": 1.2542372881355932, + "grad_norm": 0.4791730738361785, + "learning_rate": 3.9245602072086246e-05, + "loss": 0.3755, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10753200203180313, + "step": 740, + "valid_targets_mean": 5726.8, + "valid_targets_min": 4618 + }, + { + "epoch": 1.2627118644067796, + "grad_norm": 0.43306794158718565, + "learning_rate": 3.9222435836522286e-05, + "loss": 0.3391, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16541627049446106, + "step": 745, + "valid_targets_mean": 9554.5, + "valid_targets_min": 1638 + }, + { + "epoch": 1.271186440677966, + "grad_norm": 0.5259721546030984, + "learning_rate": 3.919892630996804e-05, + "loss": 0.3372, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14351120591163635, + "step": 750, + "valid_targets_mean": 6029.6, + "valid_targets_min": 2423 + }, + { + "epoch": 1.2796610169491525, + "grad_norm": 0.4468402378838669, + "learning_rate": 3.9175073912277107e-05, + "loss": 0.307, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12015321850776672, + "step": 755, + "valid_targets_mean": 6729.4, + "valid_targets_min": 3491 + }, + { + "epoch": 1.288135593220339, + "grad_norm": 0.47941342524051483, + "learning_rate": 3.915087906942635e-05, + "loss": 0.3215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18960008025169373, + "step": 760, + "valid_targets_mean": 7721.4, + "valid_targets_min": 2945 + }, + { + "epoch": 1.2966101694915255, + "grad_norm": 0.4187927303370573, + "learning_rate": 3.912634221350834e-05, + "loss": 0.3504, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20863047242164612, + "step": 765, + "valid_targets_mean": 10637.6, + "valid_targets_min": 2466 + }, + { + "epoch": 1.305084745762712, + "grad_norm": 0.4663828688691197, + "learning_rate": 3.910146378272361e-05, + "loss": 0.3183, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10256993025541306, + "step": 770, + "valid_targets_mean": 5322.9, + "valid_targets_min": 2783 + }, + { + "epoch": 1.3135593220338984, + "grad_norm": 0.452140249596733, + "learning_rate": 3.9076244221372824e-05, + "loss": 0.323, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15320637822151184, + "step": 775, + "valid_targets_mean": 6534.4, + "valid_targets_min": 2414 + }, + { + "epoch": 1.3220338983050848, + "grad_norm": 0.4216191524447884, + "learning_rate": 3.905068397984886e-05, + "loss": 0.3181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16810233891010284, + "step": 780, + "valid_targets_mean": 8753.8, + "valid_targets_min": 4126 + }, + { + "epoch": 1.3305084745762712, + "grad_norm": 0.4171527484018184, + "learning_rate": 3.9024783514628765e-05, + "loss": 0.3586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1767878234386444, + "step": 785, + "valid_targets_mean": 10436.5, + "valid_targets_min": 3606 + }, + { + "epoch": 1.3389830508474576, + "grad_norm": 0.5381131810276069, + "learning_rate": 3.899854328826559e-05, + "loss": 0.3441, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09073320776224136, + "step": 790, + "valid_targets_mean": 4006.0, + "valid_targets_min": 1408 + }, + { + "epoch": 1.347457627118644, + "grad_norm": 0.4928947240171949, + "learning_rate": 3.897196376938013e-05, + "loss": 0.3201, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17408603429794312, + "step": 795, + "valid_targets_mean": 6354.2, + "valid_targets_min": 2403 + }, + { + "epoch": 1.3559322033898304, + "grad_norm": 0.4223258490595182, + "learning_rate": 3.8945045432652575e-05, + "loss": 0.3419, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08797114342451096, + "step": 800, + "valid_targets_mean": 6019.8, + "valid_targets_min": 3531 + }, + { + "epoch": 1.3644067796610169, + "grad_norm": 0.44759924805112894, + "learning_rate": 3.8917788758814e-05, + "loss": 0.3191, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1312120258808136, + "step": 805, + "valid_targets_mean": 7411.5, + "valid_targets_min": 3038 + }, + { + "epoch": 1.3728813559322033, + "grad_norm": 0.4123257465614314, + "learning_rate": 3.889019423463783e-05, + "loss": 0.3171, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09574757516384125, + "step": 810, + "valid_targets_mean": 6820.2, + "valid_targets_min": 2889 + }, + { + "epoch": 1.3813559322033897, + "grad_norm": 0.4699957340929783, + "learning_rate": 3.8862262352931075e-05, + "loss": 0.3312, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15508131682872772, + "step": 815, + "valid_targets_mean": 6407.5, + "valid_targets_min": 2094 + }, + { + "epoch": 1.3898305084745763, + "grad_norm": 0.44153035912518707, + "learning_rate": 3.88339936125256e-05, + "loss": 0.3178, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1751125454902649, + "step": 820, + "valid_targets_mean": 8232.1, + "valid_targets_min": 2725 + }, + { + "epoch": 1.3983050847457628, + "grad_norm": 0.42694559401651966, + "learning_rate": 3.8805388518269184e-05, + "loss": 0.2978, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1752178966999054, + "step": 825, + "valid_targets_mean": 8601.1, + "valid_targets_min": 4140 + }, + { + "epoch": 1.4067796610169492, + "grad_norm": 0.42969135374896245, + "learning_rate": 3.877644758101648e-05, + "loss": 0.3173, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18027272820472717, + "step": 830, + "valid_targets_mean": 8545.4, + "valid_targets_min": 1880 + }, + { + "epoch": 1.4152542372881356, + "grad_norm": 0.49654623461497327, + "learning_rate": 3.8747171317619955e-05, + "loss": 0.3159, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0949745625257492, + "step": 835, + "valid_targets_mean": 5100.9, + "valid_targets_min": 2239 + }, + { + "epoch": 1.423728813559322, + "grad_norm": 0.4720165036240917, + "learning_rate": 3.871756025092059e-05, + "loss": 0.3505, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1421668529510498, + "step": 840, + "valid_targets_mean": 6590.9, + "valid_targets_min": 2769 + }, + { + "epoch": 1.4322033898305084, + "grad_norm": 0.3994755374635355, + "learning_rate": 3.868761490973859e-05, + "loss": 0.3087, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13265396654605865, + "step": 845, + "valid_targets_mean": 6718.4, + "valid_targets_min": 3008 + }, + { + "epoch": 1.4406779661016949, + "grad_norm": 0.44229555376258, + "learning_rate": 3.8657335828863924e-05, + "loss": 0.3337, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16455960273742676, + "step": 850, + "valid_targets_mean": 7452.9, + "valid_targets_min": 3129 + }, + { + "epoch": 1.4491525423728815, + "grad_norm": 0.42144368020373396, + "learning_rate": 3.8626723549046774e-05, + "loss": 0.3289, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12426987290382385, + "step": 855, + "valid_targets_mean": 6588.8, + "valid_targets_min": 2820 + }, + { + "epoch": 1.457627118644068, + "grad_norm": 0.44582761725314035, + "learning_rate": 3.859577861698787e-05, + "loss": 0.2938, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1588899791240692, + "step": 860, + "valid_targets_mean": 7472.6, + "valid_targets_min": 3010 + }, + { + "epoch": 1.4661016949152543, + "grad_norm": 0.4474961099293207, + "learning_rate": 3.856450158532875e-05, + "loss": 0.3247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1527513563632965, + "step": 865, + "valid_targets_mean": 7205.9, + "valid_targets_min": 3821 + }, + { + "epoch": 1.4745762711864407, + "grad_norm": 0.6963468535745492, + "learning_rate": 3.853289301264187e-05, + "loss": 0.3374, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11420440673828125, + "step": 870, + "valid_targets_mean": 6204.8, + "valid_targets_min": 1600 + }, + { + "epoch": 1.4830508474576272, + "grad_norm": 0.4670711062662348, + "learning_rate": 3.850095346342064e-05, + "loss": 0.3295, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16958177089691162, + "step": 875, + "valid_targets_mean": 7305.1, + "valid_targets_min": 3831 + }, + { + "epoch": 1.4915254237288136, + "grad_norm": 0.43457657809243977, + "learning_rate": 3.84686835080693e-05, + "loss": 0.3117, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11202533543109894, + "step": 880, + "valid_targets_mean": 6075.1, + "valid_targets_min": 3763 + }, + { + "epoch": 1.5, + "grad_norm": 0.4312755699648669, + "learning_rate": 3.843608372289283e-05, + "loss": 0.329, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15226225554943085, + "step": 885, + "valid_targets_mean": 7362.9, + "valid_targets_min": 3271 + }, + { + "epoch": 1.5084745762711864, + "grad_norm": 0.4512007864570461, + "learning_rate": 3.8403154690086564e-05, + "loss": 0.3383, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13130944967269897, + "step": 890, + "valid_targets_mean": 7195.8, + "valid_targets_min": 3209 + }, + { + "epoch": 1.5169491525423728, + "grad_norm": 0.4552864375986253, + "learning_rate": 3.836989699772582e-05, + "loss": 0.3245, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18861161172389984, + "step": 895, + "valid_targets_mean": 6992.8, + "valid_targets_min": 2171 + }, + { + "epoch": 1.5254237288135593, + "grad_norm": 0.41688741758257297, + "learning_rate": 3.8336311239755424e-05, + "loss": 0.3638, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1524004489183426, + "step": 900, + "valid_targets_mean": 7088.0, + "valid_targets_min": 2186 + }, + { + "epoch": 1.5338983050847457, + "grad_norm": 0.4112741724687864, + "learning_rate": 3.830239801597907e-05, + "loss": 0.3142, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1504945456981659, + "step": 905, + "valid_targets_mean": 7589.8, + "valid_targets_min": 2334 + }, + { + "epoch": 1.542372881355932, + "grad_norm": 0.40116464525893, + "learning_rate": 3.826815793204863e-05, + "loss": 0.3029, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18146401643753052, + "step": 910, + "valid_targets_mean": 9366.8, + "valid_targets_min": 4155 + }, + { + "epoch": 1.5508474576271185, + "grad_norm": 0.4180686453785114, + "learning_rate": 3.823359159945332e-05, + "loss": 0.3208, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19999852776527405, + "step": 915, + "valid_targets_mean": 8296.5, + "valid_targets_min": 3215 + }, + { + "epoch": 1.559322033898305, + "grad_norm": 0.4225510654881023, + "learning_rate": 3.8198699635508805e-05, + "loss": 0.2998, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1913813203573227, + "step": 920, + "valid_targets_mean": 9003.4, + "valid_targets_min": 3408 + }, + { + "epoch": 1.5677966101694916, + "grad_norm": 0.38733909243601283, + "learning_rate": 3.816348266334614e-05, + "loss": 0.3202, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12754607200622559, + "step": 925, + "valid_targets_mean": 8739.4, + "valid_targets_min": 4950 + }, + { + "epoch": 1.576271186440678, + "grad_norm": 0.3831668563774191, + "learning_rate": 3.812794131190066e-05, + "loss": 0.3326, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13470211625099182, + "step": 930, + "valid_targets_mean": 7801.5, + "valid_targets_min": 3955 + }, + { + "epoch": 1.5847457627118644, + "grad_norm": 0.40233056550625984, + "learning_rate": 3.8092076215900765e-05, + "loss": 0.3042, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2096737027168274, + "step": 935, + "valid_targets_mean": 11177.6, + "valid_targets_min": 5314 + }, + { + "epoch": 1.5932203389830508, + "grad_norm": 0.4309360432918033, + "learning_rate": 3.805588801585654e-05, + "loss": 0.3207, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08947759121656418, + "step": 940, + "valid_targets_mean": 4647.1, + "valid_targets_min": 2823 + }, + { + "epoch": 1.6016949152542372, + "grad_norm": 0.4454706160268094, + "learning_rate": 3.801937735804838e-05, + "loss": 0.3457, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17006026208400726, + "step": 945, + "valid_targets_mean": 8202.2, + "valid_targets_min": 2278 + }, + { + "epoch": 1.6101694915254239, + "grad_norm": 0.4766527726184422, + "learning_rate": 3.798254489451539e-05, + "loss": 0.3241, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2619522213935852, + "step": 950, + "valid_targets_mean": 11633.5, + "valid_targets_min": 2082 + }, + { + "epoch": 1.6186440677966103, + "grad_norm": 0.4299809742585944, + "learning_rate": 3.794539128304374e-05, + "loss": 0.3342, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2215835154056549, + "step": 955, + "valid_targets_mean": 10074.8, + "valid_targets_min": 4879 + }, + { + "epoch": 1.6271186440677967, + "grad_norm": 0.6379205273438897, + "learning_rate": 3.790791718715498e-05, + "loss": 0.3383, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16873620450496674, + "step": 960, + "valid_targets_mean": 7615.6, + "valid_targets_min": 2981 + }, + { + "epoch": 1.6355932203389831, + "grad_norm": 0.444917907184848, + "learning_rate": 3.7870123276094134e-05, + "loss": 0.308, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14380893111228943, + "step": 965, + "valid_targets_mean": 6693.8, + "valid_targets_min": 2020 + }, + { + "epoch": 1.6440677966101696, + "grad_norm": 0.42501094874543216, + "learning_rate": 3.783201022481775e-05, + "loss": 0.3088, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1657855212688446, + "step": 970, + "valid_targets_mean": 7619.8, + "valid_targets_min": 3984 + }, + { + "epoch": 1.652542372881356, + "grad_norm": 0.4370649274939684, + "learning_rate": 3.7793578713981876e-05, + "loss": 0.2726, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18086126446723938, + "step": 975, + "valid_targets_mean": 7186.5, + "valid_targets_min": 2052 + }, + { + "epoch": 1.6610169491525424, + "grad_norm": 0.4709213922563684, + "learning_rate": 3.775482942992989e-05, + "loss": 0.2978, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10135913640260696, + "step": 980, + "valid_targets_mean": 5315.8, + "valid_targets_min": 3661 + }, + { + "epoch": 1.6694915254237288, + "grad_norm": 0.39829590022078026, + "learning_rate": 3.771576306468022e-05, + "loss": 0.3333, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13913916051387787, + "step": 985, + "valid_targets_mean": 8544.4, + "valid_targets_min": 2525 + }, + { + "epoch": 1.6779661016949152, + "grad_norm": 0.4958009972855573, + "learning_rate": 3.7676380315914025e-05, + "loss": 0.3329, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09229759126901627, + "step": 990, + "valid_targets_mean": 5690.1, + "valid_targets_min": 1946 + }, + { + "epoch": 1.6864406779661016, + "grad_norm": 0.46063794892512244, + "learning_rate": 3.7636681886962724e-05, + "loss": 0.3682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15866561233997345, + "step": 995, + "valid_targets_mean": 7674.0, + "valid_targets_min": 2544 + }, + { + "epoch": 1.694915254237288, + "grad_norm": 0.4095748837132922, + "learning_rate": 3.759666848679539e-05, + "loss": 0.3052, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15614467859268188, + "step": 1000, + "valid_targets_mean": 9040.6, + "valid_targets_min": 5002 + }, + { + "epoch": 1.7033898305084745, + "grad_norm": 0.40817498338779584, + "learning_rate": 3.755634083000617e-05, + "loss": 0.331, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1329311728477478, + "step": 1005, + "valid_targets_mean": 6859.1, + "valid_targets_min": 2941 + }, + { + "epoch": 1.711864406779661, + "grad_norm": 0.4186391165550277, + "learning_rate": 3.751569963680146e-05, + "loss": 0.2965, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20867162942886353, + "step": 1010, + "valid_targets_mean": 9555.4, + "valid_targets_min": 5043 + }, + { + "epoch": 1.7203389830508473, + "grad_norm": 0.41304904819896177, + "learning_rate": 3.747474563298705e-05, + "loss": 0.3306, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1333586573600769, + "step": 1015, + "valid_targets_mean": 6414.8, + "valid_targets_min": 1864 + }, + { + "epoch": 1.7288135593220337, + "grad_norm": 0.4814872970064052, + "learning_rate": 3.743347954995519e-05, + "loss": 0.3287, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13862814009189606, + "step": 1020, + "valid_targets_mean": 6174.1, + "valid_targets_min": 3819 + }, + { + "epoch": 1.7372881355932204, + "grad_norm": 0.4250075191298139, + "learning_rate": 3.7391902124671516e-05, + "loss": 0.3264, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19187797605991364, + "step": 1025, + "valid_targets_mean": 8820.8, + "valid_targets_min": 3660 + }, + { + "epoch": 1.7457627118644068, + "grad_norm": 0.38841628466641565, + "learning_rate": 3.7350014099661874e-05, + "loss": 0.3084, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20551234483718872, + "step": 1030, + "valid_targets_mean": 9656.0, + "valid_targets_min": 3528 + }, + { + "epoch": 1.7542372881355932, + "grad_norm": 0.4136870657166387, + "learning_rate": 3.730781622299907e-05, + "loss": 0.3232, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17536289989948273, + "step": 1035, + "valid_targets_mean": 8021.1, + "valid_targets_min": 3783 + }, + { + "epoch": 1.7627118644067796, + "grad_norm": 0.7311742841410761, + "learning_rate": 3.7265309248289516e-05, + "loss": 0.3472, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17020183801651, + "step": 1040, + "valid_targets_mean": 7756.2, + "valid_targets_min": 3890 + }, + { + "epoch": 1.7711864406779663, + "grad_norm": 0.4565783832214843, + "learning_rate": 3.7222493934659765e-05, + "loss": 0.3486, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1536499261856079, + "step": 1045, + "valid_targets_mean": 9314.8, + "valid_targets_min": 4946 + }, + { + "epoch": 1.7796610169491527, + "grad_norm": 0.4252286803029882, + "learning_rate": 3.717937104674296e-05, + "loss": 0.2993, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13162419199943542, + "step": 1050, + "valid_targets_mean": 7717.6, + "valid_targets_min": 4375 + }, + { + "epoch": 1.788135593220339, + "grad_norm": 0.42315018442974783, + "learning_rate": 3.713594135466515e-05, + "loss": 0.3245, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17134219408035278, + "step": 1055, + "valid_targets_mean": 8660.0, + "valid_targets_min": 2787 + }, + { + "epoch": 1.7966101694915255, + "grad_norm": 0.4460311450253956, + "learning_rate": 3.7092205634031595e-05, + "loss": 0.3495, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08716882765293121, + "step": 1060, + "valid_targets_mean": 4970.2, + "valid_targets_min": 2356 + }, + { + "epoch": 1.805084745762712, + "grad_norm": 0.407857276812608, + "learning_rate": 3.704816466591286e-05, + "loss": 0.3181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1297275722026825, + "step": 1065, + "valid_targets_mean": 7571.6, + "valid_targets_min": 4007 + }, + { + "epoch": 1.8135593220338984, + "grad_norm": 0.3977408996615863, + "learning_rate": 3.700381923683087e-05, + "loss": 0.3383, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11389312893152237, + "step": 1070, + "valid_targets_mean": 7162.1, + "valid_targets_min": 2781 + }, + { + "epoch": 1.8220338983050848, + "grad_norm": 0.43601706810282037, + "learning_rate": 3.695917013874491e-05, + "loss": 0.2844, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1395803689956665, + "step": 1075, + "valid_targets_mean": 7036.4, + "valid_targets_min": 3527 + }, + { + "epoch": 1.8305084745762712, + "grad_norm": 0.427769673869337, + "learning_rate": 3.691421816903744e-05, + "loss": 0.3396, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2047993391752243, + "step": 1080, + "valid_targets_mean": 8439.5, + "valid_targets_min": 1779 + }, + { + "epoch": 1.8389830508474576, + "grad_norm": 0.4696044107301725, + "learning_rate": 3.686896413049985e-05, + "loss": 0.2903, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12242710590362549, + "step": 1085, + "valid_targets_mean": 6735.5, + "valid_targets_min": 3348 + }, + { + "epoch": 1.847457627118644, + "grad_norm": 0.5013621044652345, + "learning_rate": 3.6823408831318154e-05, + "loss": 0.3118, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16277314722537994, + "step": 1090, + "valid_targets_mean": 7934.6, + "valid_targets_min": 2514 + }, + { + "epoch": 1.8559322033898304, + "grad_norm": 0.41708780569974935, + "learning_rate": 3.677755308505854e-05, + "loss": 0.344, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16447079181671143, + "step": 1095, + "valid_targets_mean": 7324.9, + "valid_targets_min": 3842 + }, + { + "epoch": 1.8644067796610169, + "grad_norm": 0.44156963603568017, + "learning_rate": 3.673139771065282e-05, + "loss": 0.3249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11587251722812653, + "step": 1100, + "valid_targets_mean": 6803.6, + "valid_targets_min": 3852 + }, + { + "epoch": 1.8728813559322033, + "grad_norm": 0.38630784106679994, + "learning_rate": 3.6684943532383866e-05, + "loss": 0.3469, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19013455510139465, + "step": 1105, + "valid_targets_mean": 10397.4, + "valid_targets_min": 3561 + }, + { + "epoch": 1.8813559322033897, + "grad_norm": 0.4085933479293492, + "learning_rate": 3.663819137987081e-05, + "loss": 0.3365, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16472648084163666, + "step": 1110, + "valid_targets_mean": 9081.0, + "valid_targets_min": 6186 + }, + { + "epoch": 1.8898305084745761, + "grad_norm": 0.43391283451388735, + "learning_rate": 3.65911420880543e-05, + "loss": 0.3122, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20973503589630127, + "step": 1115, + "valid_targets_mean": 8154.1, + "valid_targets_min": 3421 + }, + { + "epoch": 1.8983050847457628, + "grad_norm": 0.4175542637223881, + "learning_rate": 3.654379649718153e-05, + "loss": 0.3077, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16133342683315277, + "step": 1120, + "valid_targets_mean": 8112.1, + "valid_targets_min": 2321 + }, + { + "epoch": 1.9067796610169492, + "grad_norm": 0.4726007732664944, + "learning_rate": 3.649615545279127e-05, + "loss": 0.3467, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2352825105190277, + "step": 1125, + "valid_targets_mean": 8948.4, + "valid_targets_min": 1580 + }, + { + "epoch": 1.9152542372881356, + "grad_norm": 0.4238072738722841, + "learning_rate": 3.644821980569876e-05, + "loss": 0.3308, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1875908374786377, + "step": 1130, + "valid_targets_mean": 9306.8, + "valid_targets_min": 2877 + }, + { + "epoch": 1.923728813559322, + "grad_norm": 0.4391878996198538, + "learning_rate": 3.639999041198051e-05, + "loss": 0.312, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1598246842622757, + "step": 1135, + "valid_targets_mean": 7715.6, + "valid_targets_min": 2222 + }, + { + "epoch": 1.9322033898305084, + "grad_norm": 0.4313660549167382, + "learning_rate": 3.635146813295902e-05, + "loss": 0.3366, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12922486662864685, + "step": 1140, + "valid_targets_mean": 6985.1, + "valid_targets_min": 3372 + }, + { + "epoch": 1.940677966101695, + "grad_norm": 0.43931009194342957, + "learning_rate": 3.6302653835187366e-05, + "loss": 0.3165, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21983201801776886, + "step": 1145, + "valid_targets_mean": 8416.0, + "valid_targets_min": 4348 + }, + { + "epoch": 1.9491525423728815, + "grad_norm": 0.3635875294493641, + "learning_rate": 3.625354839043377e-05, + "loss": 0.2881, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21698161959648132, + "step": 1150, + "valid_targets_mean": 12197.4, + "valid_targets_min": 4223 + }, + { + "epoch": 1.957627118644068, + "grad_norm": 0.44740401195878726, + "learning_rate": 3.620415267566601e-05, + "loss": 0.3325, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12177214026451111, + "step": 1155, + "valid_targets_mean": 7473.8, + "valid_targets_min": 3188 + }, + { + "epoch": 1.9661016949152543, + "grad_norm": 0.43597591281289183, + "learning_rate": 3.615446757303575e-05, + "loss": 0.324, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17450034618377686, + "step": 1160, + "valid_targets_mean": 8725.9, + "valid_targets_min": 2938 + }, + { + "epoch": 1.9745762711864407, + "grad_norm": 0.4088861911343299, + "learning_rate": 3.610449396986281e-05, + "loss": 0.3102, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09415778517723083, + "step": 1165, + "valid_targets_mean": 5801.5, + "valid_targets_min": 2078 + }, + { + "epoch": 1.9830508474576272, + "grad_norm": 0.4528914357560769, + "learning_rate": 3.6054232758619274e-05, + "loss": 0.3178, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19558541476726532, + "step": 1170, + "valid_targets_mean": 7415.4, + "valid_targets_min": 2414 + }, + { + "epoch": 1.9915254237288136, + "grad_norm": 0.41879929933975685, + "learning_rate": 3.600368483691361e-05, + "loss": 0.3299, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1337018609046936, + "step": 1175, + "valid_targets_mean": 7641.0, + "valid_targets_min": 2471 + }, + { + "epoch": 2.0, + "grad_norm": 0.4744894393948504, + "learning_rate": 3.59528511074746e-05, + "loss": 0.3012, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12567296624183655, + "step": 1180, + "valid_targets_mean": 5105.1, + "valid_targets_min": 1983 + }, + { + "epoch": 2.0084745762711864, + "grad_norm": 0.49589699132282694, + "learning_rate": 3.5901732478135235e-05, + "loss": 0.3156, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15305688977241516, + "step": 1185, + "valid_targets_mean": 6279.4, + "valid_targets_min": 2218 + }, + { + "epoch": 2.016949152542373, + "grad_norm": 0.4564322353401835, + "learning_rate": 3.5850329861816475e-05, + "loss": 0.2836, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13018441200256348, + "step": 1190, + "valid_targets_mean": 6881.9, + "valid_targets_min": 2751 + }, + { + "epoch": 2.0254237288135593, + "grad_norm": 0.413522941384289, + "learning_rate": 3.5798644176511e-05, + "loss": 0.2993, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09803292155265808, + "step": 1195, + "valid_targets_mean": 6203.8, + "valid_targets_min": 3028 + }, + { + "epoch": 2.0338983050847457, + "grad_norm": 0.535723752145441, + "learning_rate": 3.574667634526676e-05, + "loss": 0.3142, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1461130678653717, + "step": 1200, + "valid_targets_mean": 7318.0, + "valid_targets_min": 2705 + }, + { + "epoch": 2.042372881355932, + "grad_norm": 0.43264272243192015, + "learning_rate": 3.5694427296170514e-05, + "loss": 0.3068, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16311097145080566, + "step": 1205, + "valid_targets_mean": 8026.2, + "valid_targets_min": 4437 + }, + { + "epoch": 2.0508474576271185, + "grad_norm": 0.4587276888039481, + "learning_rate": 3.564189796233125e-05, + "loss": 0.316, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23014996945858002, + "step": 1210, + "valid_targets_mean": 9588.2, + "valid_targets_min": 3071 + }, + { + "epoch": 2.059322033898305, + "grad_norm": 0.4098705980870115, + "learning_rate": 3.5589089281863547e-05, + "loss": 0.2988, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10289862751960754, + "step": 1215, + "valid_targets_mean": 7517.5, + "valid_targets_min": 2735 + }, + { + "epoch": 2.0677966101694913, + "grad_norm": 0.45081745163120795, + "learning_rate": 3.553600219787079e-05, + "loss": 0.2926, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19421172142028809, + "step": 1220, + "valid_targets_mean": 8558.2, + "valid_targets_min": 3613 + }, + { + "epoch": 2.0762711864406778, + "grad_norm": 0.4669320032293127, + "learning_rate": 3.5482637658428315e-05, + "loss": 0.2965, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2341485321521759, + "step": 1225, + "valid_targets_mean": 9261.1, + "valid_targets_min": 1918 + }, + { + "epoch": 2.084745762711864, + "grad_norm": 0.4165523464643688, + "learning_rate": 3.542899661656653e-05, + "loss": 0.3234, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18077555298805237, + "step": 1230, + "valid_targets_mean": 9665.5, + "valid_targets_min": 4859 + }, + { + "epoch": 2.093220338983051, + "grad_norm": 0.4533647677488091, + "learning_rate": 3.5375080030253826e-05, + "loss": 0.3451, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12412886321544647, + "step": 1235, + "valid_targets_mean": 6687.0, + "valid_targets_min": 3687 + }, + { + "epoch": 2.1016949152542375, + "grad_norm": 0.41077985487893737, + "learning_rate": 3.532088886237956e-05, + "loss": 0.304, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11022284626960754, + "step": 1240, + "valid_targets_mean": 6293.6, + "valid_targets_min": 3275 + }, + { + "epoch": 2.110169491525424, + "grad_norm": 0.44189587866698343, + "learning_rate": 3.5266424080736765e-05, + "loss": 0.3092, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1553216278553009, + "step": 1245, + "valid_targets_mean": 7496.8, + "valid_targets_min": 3668 + }, + { + "epoch": 2.1186440677966103, + "grad_norm": 0.38809274252786646, + "learning_rate": 3.521168665800491e-05, + "loss": 0.3236, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25186485052108765, + "step": 1250, + "valid_targets_mean": 13246.9, + "valid_targets_min": 5924 + }, + { + "epoch": 2.1271186440677967, + "grad_norm": 0.4288106747949976, + "learning_rate": 3.515667757173254e-05, + "loss": 0.2971, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13772429525852203, + "step": 1255, + "valid_targets_mean": 7095.8, + "valid_targets_min": 2393 + }, + { + "epoch": 2.135593220338983, + "grad_norm": 0.40602821449574994, + "learning_rate": 3.5101397804319794e-05, + "loss": 0.2878, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17868921160697937, + "step": 1260, + "valid_targets_mean": 8826.4, + "valid_targets_min": 3085 + }, + { + "epoch": 2.1440677966101696, + "grad_norm": 0.4605717829368249, + "learning_rate": 3.5045848343000876e-05, + "loss": 0.2871, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17283692955970764, + "step": 1265, + "valid_targets_mean": 7659.4, + "valid_targets_min": 2962 + }, + { + "epoch": 2.152542372881356, + "grad_norm": 0.41756095018949657, + "learning_rate": 3.4990030179826414e-05, + "loss": 0.3271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20672985911369324, + "step": 1270, + "valid_targets_mean": 10083.5, + "valid_targets_min": 4423 + }, + { + "epoch": 2.1610169491525424, + "grad_norm": 0.4083849930062024, + "learning_rate": 3.493394431164576e-05, + "loss": 0.2948, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.158214271068573, + "step": 1275, + "valid_targets_mean": 8329.9, + "valid_targets_min": 4882 + }, + { + "epoch": 2.169491525423729, + "grad_norm": 0.4318960098601155, + "learning_rate": 3.4877591740089144e-05, + "loss": 0.3051, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12314428389072418, + "step": 1280, + "valid_targets_mean": 6445.0, + "valid_targets_min": 2968 + }, + { + "epoch": 2.1779661016949152, + "grad_norm": 0.39896913203386186, + "learning_rate": 3.482097347154986e-05, + "loss": 0.2968, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15951798856258392, + "step": 1285, + "valid_targets_mean": 8640.9, + "valid_targets_min": 2702 + }, + { + "epoch": 2.1864406779661016, + "grad_norm": 0.37336280642419734, + "learning_rate": 3.476409051716621e-05, + "loss": 0.3091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19251972436904907, + "step": 1290, + "valid_targets_mean": 9940.2, + "valid_targets_min": 4335 + }, + { + "epoch": 2.194915254237288, + "grad_norm": 0.4561259005260555, + "learning_rate": 3.470694389280352e-05, + "loss": 0.2911, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13392677903175354, + "step": 1295, + "valid_targets_mean": 6471.1, + "valid_targets_min": 2576 + }, + { + "epoch": 2.2033898305084745, + "grad_norm": 0.4603341835246692, + "learning_rate": 3.464953461903593e-05, + "loss": 0.283, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16801318526268005, + "step": 1300, + "valid_targets_mean": 7596.4, + "valid_targets_min": 2538 + }, + { + "epoch": 2.211864406779661, + "grad_norm": 0.46161972023282255, + "learning_rate": 3.459186372112824e-05, + "loss": 0.308, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12896674871444702, + "step": 1305, + "valid_targets_mean": 6762.2, + "valid_targets_min": 3949 + }, + { + "epoch": 2.2203389830508473, + "grad_norm": 0.40321694072808506, + "learning_rate": 3.453393222901753e-05, + "loss": 0.3253, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09447690844535828, + "step": 1310, + "valid_targets_mean": 7179.5, + "valid_targets_min": 1996 + }, + { + "epoch": 2.2288135593220337, + "grad_norm": 0.3970364700148203, + "learning_rate": 3.44757411772948e-05, + "loss": 0.3152, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17469201982021332, + "step": 1315, + "valid_targets_mean": 8900.5, + "valid_targets_min": 3699 + }, + { + "epoch": 2.23728813559322, + "grad_norm": 0.4442152861861335, + "learning_rate": 3.441729160518652e-05, + "loss": 0.3282, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20383930206298828, + "step": 1320, + "valid_targets_mean": 8585.8, + "valid_targets_min": 3771 + }, + { + "epoch": 2.2457627118644066, + "grad_norm": 0.38677144835780297, + "learning_rate": 3.435858455653601e-05, + "loss": 0.3084, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1343754380941391, + "step": 1325, + "valid_targets_mean": 7899.6, + "valid_targets_min": 2683 + }, + { + "epoch": 2.2542372881355934, + "grad_norm": 0.4600895771290722, + "learning_rate": 3.429962107978485e-05, + "loss": 0.3119, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10500349849462509, + "step": 1330, + "valid_targets_mean": 5783.9, + "valid_targets_min": 2090 + }, + { + "epoch": 2.26271186440678, + "grad_norm": 0.44544228237488553, + "learning_rate": 3.424040222795412e-05, + "loss": 0.303, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10833808779716492, + "step": 1335, + "valid_targets_mean": 6813.5, + "valid_targets_min": 3555 + }, + { + "epoch": 2.2711864406779663, + "grad_norm": 0.5397040058239221, + "learning_rate": 3.418092905862563e-05, + "loss": 0.294, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12867417931556702, + "step": 1340, + "valid_targets_mean": 7460.1, + "valid_targets_min": 4134 + }, + { + "epoch": 2.2796610169491527, + "grad_norm": 0.42024068799982406, + "learning_rate": 3.412120263392301e-05, + "loss": 0.2951, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15891410410404205, + "step": 1345, + "valid_targets_mean": 8347.6, + "valid_targets_min": 5053 + }, + { + "epoch": 2.288135593220339, + "grad_norm": 0.3973157849670298, + "learning_rate": 3.406122402049272e-05, + "loss": 0.3212, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1449463963508606, + "step": 1350, + "valid_targets_mean": 8319.0, + "valid_targets_min": 3013 + }, + { + "epoch": 2.2966101694915255, + "grad_norm": 0.4098586276920284, + "learning_rate": 3.400099428948505e-05, + "loss": 0.2984, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2429947853088379, + "step": 1355, + "valid_targets_mean": 9789.1, + "valid_targets_min": 3908 + }, + { + "epoch": 2.305084745762712, + "grad_norm": 0.3970559339256101, + "learning_rate": 3.394051451653496e-05, + "loss": 0.3003, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12642662227153778, + "step": 1360, + "valid_targets_mean": 7892.9, + "valid_targets_min": 5071 + }, + { + "epoch": 2.3135593220338984, + "grad_norm": 0.4108955207902681, + "learning_rate": 3.3879785781742875e-05, + "loss": 0.3187, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18125218152999878, + "step": 1365, + "valid_targets_mean": 8801.2, + "valid_targets_min": 3025 + }, + { + "epoch": 2.3220338983050848, + "grad_norm": 0.4571326386467896, + "learning_rate": 3.381880916965541e-05, + "loss": 0.2955, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10770291835069656, + "step": 1370, + "valid_targets_mean": 5185.5, + "valid_targets_min": 2769 + }, + { + "epoch": 2.330508474576271, + "grad_norm": 0.4425655935422459, + "learning_rate": 3.3757585769245964e-05, + "loss": 0.3139, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1520739048719406, + "step": 1375, + "valid_targets_mean": 7233.2, + "valid_targets_min": 3917 + }, + { + "epoch": 2.3389830508474576, + "grad_norm": 0.42190473263887457, + "learning_rate": 3.3696116673895314e-05, + "loss": 0.3236, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20773333311080933, + "step": 1380, + "valid_targets_mean": 9542.8, + "valid_targets_min": 3725 + }, + { + "epoch": 2.347457627118644, + "grad_norm": 0.4493341593347805, + "learning_rate": 3.363440298137207e-05, + "loss": 0.3091, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12796102464199066, + "step": 1385, + "valid_targets_mean": 6003.8, + "valid_targets_min": 2280 + }, + { + "epoch": 2.3559322033898304, + "grad_norm": 0.458582309301439, + "learning_rate": 3.357244579381306e-05, + "loss": 0.2928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14768704771995544, + "step": 1390, + "valid_targets_mean": 7220.0, + "valid_targets_min": 1887 + }, + { + "epoch": 2.364406779661017, + "grad_norm": 0.4434794038080658, + "learning_rate": 3.351024621770369e-05, + "loss": 0.3004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09222928434610367, + "step": 1395, + "valid_targets_mean": 5011.0, + "valid_targets_min": 2432 + }, + { + "epoch": 2.3728813559322033, + "grad_norm": 0.459366599598582, + "learning_rate": 3.34478053638581e-05, + "loss": 0.2874, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1464822143316269, + "step": 1400, + "valid_targets_mean": 7131.9, + "valid_targets_min": 1920 + }, + { + "epoch": 2.3813559322033897, + "grad_norm": 0.4432040890783054, + "learning_rate": 3.3385124347399415e-05, + "loss": 0.2771, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11013621091842651, + "step": 1405, + "valid_targets_mean": 7265.8, + "valid_targets_min": 2334 + }, + { + "epoch": 2.389830508474576, + "grad_norm": 0.46103691686332354, + "learning_rate": 3.33222042877398e-05, + "loss": 0.3024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1722109466791153, + "step": 1410, + "valid_targets_mean": 7767.9, + "valid_targets_min": 2377 + }, + { + "epoch": 2.3983050847457625, + "grad_norm": 0.37676236353350995, + "learning_rate": 3.325904630856042e-05, + "loss": 0.2977, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18165141344070435, + "step": 1415, + "valid_targets_mean": 10603.2, + "valid_targets_min": 3445 + }, + { + "epoch": 2.406779661016949, + "grad_norm": 0.4812188982000893, + "learning_rate": 3.319565153779146e-05, + "loss": 0.3493, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2623385190963745, + "step": 1420, + "valid_targets_mean": 9967.9, + "valid_targets_min": 5092 + }, + { + "epoch": 2.415254237288136, + "grad_norm": 0.43821349980034335, + "learning_rate": 3.31320211075919e-05, + "loss": 0.3226, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.181278258562088, + "step": 1425, + "valid_targets_mean": 7767.0, + "valid_targets_min": 2223 + }, + { + "epoch": 2.423728813559322, + "grad_norm": 0.43105954825794324, + "learning_rate": 3.306815615432936e-05, + "loss": 0.3404, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15349501371383667, + "step": 1430, + "valid_targets_mean": 8044.6, + "valid_targets_min": 2971 + }, + { + "epoch": 2.4322033898305087, + "grad_norm": 0.4250749591327943, + "learning_rate": 3.300405781855975e-05, + "loss": 0.2958, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19317375123500824, + "step": 1435, + "valid_targets_mean": 9132.2, + "valid_targets_min": 5937 + }, + { + "epoch": 2.440677966101695, + "grad_norm": 0.40498564400648046, + "learning_rate": 3.2939727245006956e-05, + "loss": 0.3017, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16090430319309235, + "step": 1440, + "valid_targets_mean": 8930.4, + "valid_targets_min": 2974 + }, + { + "epoch": 2.4491525423728815, + "grad_norm": 0.48877090107273713, + "learning_rate": 3.2875165582542347e-05, + "loss": 0.3066, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09781038761138916, + "step": 1445, + "valid_targets_mean": 6088.9, + "valid_targets_min": 3606 + }, + { + "epoch": 2.457627118644068, + "grad_norm": 0.3862290985798537, + "learning_rate": 3.281037398416427e-05, + "loss": 0.2713, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12969066202640533, + "step": 1450, + "valid_targets_mean": 6886.4, + "valid_targets_min": 2380 + }, + { + "epoch": 2.4661016949152543, + "grad_norm": 0.44122203466271714, + "learning_rate": 3.2745353606977506e-05, + "loss": 0.3181, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11083347350358963, + "step": 1455, + "valid_targets_mean": 7865.2, + "valid_targets_min": 3044 + }, + { + "epoch": 2.4745762711864407, + "grad_norm": 0.7076361422745008, + "learning_rate": 3.268010561217253e-05, + "loss": 0.3097, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1400471329689026, + "step": 1460, + "valid_targets_mean": 8643.5, + "valid_targets_min": 4271 + }, + { + "epoch": 2.483050847457627, + "grad_norm": 0.4108590790686095, + "learning_rate": 3.261463116500483e-05, + "loss": 0.2961, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10852666199207306, + "step": 1465, + "valid_targets_mean": 6323.6, + "valid_targets_min": 2827 + }, + { + "epoch": 2.4915254237288136, + "grad_norm": 0.3937363346055893, + "learning_rate": 3.254893143477408e-05, + "loss": 0.2933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2064848393201828, + "step": 1470, + "valid_targets_mean": 10218.9, + "valid_targets_min": 2817 + }, + { + "epoch": 2.5, + "grad_norm": 0.4345245205233806, + "learning_rate": 3.248300759480323e-05, + "loss": 0.3205, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25157177448272705, + "step": 1475, + "valid_targets_mean": 10160.5, + "valid_targets_min": 5139 + }, + { + "epoch": 2.5084745762711864, + "grad_norm": 0.39745802334240105, + "learning_rate": 3.241686082241761e-05, + "loss": 0.2873, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14739897847175598, + "step": 1480, + "valid_targets_mean": 8801.1, + "valid_targets_min": 2475 + }, + { + "epoch": 2.516949152542373, + "grad_norm": 0.4100222960644638, + "learning_rate": 3.235049229892384e-05, + "loss": 0.2667, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13044124841690063, + "step": 1485, + "valid_targets_mean": 7302.6, + "valid_targets_min": 4819 + }, + { + "epoch": 2.5254237288135593, + "grad_norm": 0.3980396958505919, + "learning_rate": 3.2283903209588795e-05, + "loss": 0.3194, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13375581800937653, + "step": 1490, + "valid_targets_mean": 7664.4, + "valid_targets_min": 2365 + }, + { + "epoch": 2.5338983050847457, + "grad_norm": 0.41717765885227137, + "learning_rate": 3.221709474361838e-05, + "loss": 0.3028, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12687115371227264, + "step": 1495, + "valid_targets_mean": 6919.5, + "valid_targets_min": 2906 + }, + { + "epoch": 2.542372881355932, + "grad_norm": 0.44285595090520824, + "learning_rate": 3.215006809413633e-05, + "loss": 0.3209, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14034491777420044, + "step": 1500, + "valid_targets_mean": 6343.8, + "valid_targets_min": 2378 + }, + { + "epoch": 2.5508474576271185, + "grad_norm": 0.4539630383428859, + "learning_rate": 3.20828244581629e-05, + "loss": 0.2896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17994040250778198, + "step": 1505, + "valid_targets_mean": 7082.8, + "valid_targets_min": 1483 + }, + { + "epoch": 2.559322033898305, + "grad_norm": 0.37842399756128303, + "learning_rate": 3.2015365036593466e-05, + "loss": 0.3325, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14443901181221008, + "step": 1510, + "valid_targets_mean": 9318.1, + "valid_targets_min": 4422 + }, + { + "epoch": 2.5677966101694913, + "grad_norm": 0.4031984420404293, + "learning_rate": 3.194769103417709e-05, + "loss": 0.2919, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2048453390598297, + "step": 1515, + "valid_targets_mean": 8194.1, + "valid_targets_min": 2893 + }, + { + "epoch": 2.576271186440678, + "grad_norm": 0.4610141663712719, + "learning_rate": 3.1879803659495015e-05, + "loss": 0.2994, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0940924882888794, + "step": 1520, + "valid_targets_mean": 5814.2, + "valid_targets_min": 2611 + }, + { + "epoch": 2.584745762711864, + "grad_norm": 0.4852703147549595, + "learning_rate": 3.181170412493906e-05, + "loss": 0.2972, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14191953837871552, + "step": 1525, + "valid_targets_mean": 5749.0, + "valid_targets_min": 2852 + }, + { + "epoch": 2.593220338983051, + "grad_norm": 0.5507609460623812, + "learning_rate": 3.1743393646689996e-05, + "loss": 0.2709, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13722501695156097, + "step": 1530, + "valid_targets_mean": 6979.1, + "valid_targets_min": 3623 + }, + { + "epoch": 2.601694915254237, + "grad_norm": 0.4300925010809667, + "learning_rate": 3.1674873444695804e-05, + "loss": 0.2859, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1543220430612564, + "step": 1535, + "valid_targets_mean": 8165.1, + "valid_targets_min": 4529 + }, + { + "epoch": 2.610169491525424, + "grad_norm": 0.38467946654939505, + "learning_rate": 3.1606144742649875e-05, + "loss": 0.2991, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18148282170295715, + "step": 1540, + "valid_targets_mean": 9327.8, + "valid_targets_min": 3421 + }, + { + "epoch": 2.6186440677966103, + "grad_norm": 0.44073639139217463, + "learning_rate": 3.15372087679692e-05, + "loss": 0.271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1646798849105835, + "step": 1545, + "valid_targets_mean": 7244.9, + "valid_targets_min": 2862 + }, + { + "epoch": 2.6271186440677967, + "grad_norm": 0.40731009930415046, + "learning_rate": 3.1468066751772415e-05, + "loss": 0.3008, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2068018913269043, + "step": 1550, + "valid_targets_mean": 9822.1, + "valid_targets_min": 3866 + }, + { + "epoch": 2.635593220338983, + "grad_norm": 0.4071433314116324, + "learning_rate": 3.1398719928857834e-05, + "loss": 0.2868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1655319333076477, + "step": 1555, + "valid_targets_mean": 9227.1, + "valid_targets_min": 2545 + }, + { + "epoch": 2.6440677966101696, + "grad_norm": 0.4223741084175799, + "learning_rate": 3.132916953768137e-05, + "loss": 0.2862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16019152104854584, + "step": 1560, + "valid_targets_mean": 7759.1, + "valid_targets_min": 2677 + }, + { + "epoch": 2.652542372881356, + "grad_norm": 0.6251705169824648, + "learning_rate": 3.1259416820334446e-05, + "loss": 0.2933, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14683334529399872, + "step": 1565, + "valid_targets_mean": 6256.4, + "valid_targets_min": 2512 + }, + { + "epoch": 2.6610169491525424, + "grad_norm": 0.4217541993199595, + "learning_rate": 3.118946302252181e-05, + "loss": 0.2766, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11165773123502731, + "step": 1570, + "valid_targets_mean": 5845.6, + "valid_targets_min": 3682 + }, + { + "epoch": 2.669491525423729, + "grad_norm": 0.5203085397178071, + "learning_rate": 3.1119309393539256e-05, + "loss": 0.278, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10624048113822937, + "step": 1575, + "valid_targets_mean": 6041.9, + "valid_targets_min": 2597 + }, + { + "epoch": 2.6779661016949152, + "grad_norm": 0.43635851830952677, + "learning_rate": 3.104895718625136e-05, + "loss": 0.2893, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.197655588388443, + "step": 1580, + "valid_targets_mean": 9161.6, + "valid_targets_min": 3570 + }, + { + "epoch": 2.6864406779661016, + "grad_norm": 0.43131575202846884, + "learning_rate": 3.097840765706908e-05, + "loss": 0.3102, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16160637140274048, + "step": 1585, + "valid_targets_mean": 7732.6, + "valid_targets_min": 2676 + }, + { + "epoch": 2.694915254237288, + "grad_norm": 0.4339183055258037, + "learning_rate": 3.0907662065927297e-05, + "loss": 0.3017, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17281053960323334, + "step": 1590, + "valid_targets_mean": 7739.5, + "valid_targets_min": 4027 + }, + { + "epoch": 2.7033898305084745, + "grad_norm": 0.35870588624638666, + "learning_rate": 3.083672167626237e-05, + "loss": 0.308, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13526961207389832, + "step": 1595, + "valid_targets_mean": 8089.1, + "valid_targets_min": 3416 + }, + { + "epoch": 2.711864406779661, + "grad_norm": 0.4042611998689289, + "learning_rate": 3.0765587754989516e-05, + "loss": 0.2901, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20892217755317688, + "step": 1600, + "valid_targets_mean": 10267.0, + "valid_targets_min": 4718 + }, + { + "epoch": 2.7203389830508473, + "grad_norm": 0.4181924475190532, + "learning_rate": 3.069426157248022e-05, + "loss": 0.2903, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12476156651973724, + "step": 1605, + "valid_targets_mean": 7643.0, + "valid_targets_min": 2520 + }, + { + "epoch": 2.7288135593220337, + "grad_norm": 0.3917558783537411, + "learning_rate": 3.062274440253953e-05, + "loss": 0.3259, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16830119490623474, + "step": 1610, + "valid_targets_mean": 8693.2, + "valid_targets_min": 2490 + }, + { + "epoch": 2.7372881355932206, + "grad_norm": 0.46809935098902955, + "learning_rate": 3.0551037522383325e-05, + "loss": 0.3096, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14017032086849213, + "step": 1615, + "valid_targets_mean": 6046.9, + "valid_targets_min": 2945 + }, + { + "epoch": 2.7457627118644066, + "grad_norm": 0.37833917782164894, + "learning_rate": 3.0479142212615457e-05, + "loss": 0.3053, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15063466131687164, + "step": 1620, + "valid_targets_mean": 8857.0, + "valid_targets_min": 2289 + }, + { + "epoch": 2.7542372881355934, + "grad_norm": 0.3914212900105464, + "learning_rate": 3.0407059757204968e-05, + "loss": 0.3126, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12059652805328369, + "step": 1625, + "valid_targets_mean": 7383.4, + "valid_targets_min": 2769 + }, + { + "epoch": 2.7627118644067794, + "grad_norm": 0.3982809928364594, + "learning_rate": 3.033479144346308e-05, + "loss": 0.3394, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16488566994667053, + "step": 1630, + "valid_targets_mean": 8842.8, + "valid_targets_min": 3991 + }, + { + "epoch": 2.7711864406779663, + "grad_norm": 0.4054605993752062, + "learning_rate": 3.0262338562020246e-05, + "loss": 0.3046, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11971583962440491, + "step": 1635, + "valid_targets_mean": 6456.0, + "valid_targets_min": 2512 + }, + { + "epoch": 2.7796610169491527, + "grad_norm": 0.3737092120949275, + "learning_rate": 3.018970240680308e-05, + "loss": 0.3088, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1984914243221283, + "step": 1640, + "valid_targets_mean": 11606.8, + "valid_targets_min": 4314 + }, + { + "epoch": 2.788135593220339, + "grad_norm": 0.4122776261507556, + "learning_rate": 3.011688427501126e-05, + "loss": 0.3076, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19593146443367004, + "step": 1645, + "valid_targets_mean": 10147.1, + "valid_targets_min": 4324 + }, + { + "epoch": 2.7966101694915255, + "grad_norm": 0.4120936192506969, + "learning_rate": 3.0043885467094382e-05, + "loss": 0.314, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09063713997602463, + "step": 1650, + "valid_targets_mean": 5500.9, + "valid_targets_min": 2577 + }, + { + "epoch": 2.805084745762712, + "grad_norm": 0.41153868372446584, + "learning_rate": 2.9970707286728676e-05, + "loss": 0.3015, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16134697198867798, + "step": 1655, + "valid_targets_mean": 7913.1, + "valid_targets_min": 3032 + }, + { + "epoch": 2.8135593220338984, + "grad_norm": 0.4432301462525058, + "learning_rate": 2.9897351040793786e-05, + "loss": 0.2792, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1156914085149765, + "step": 1660, + "valid_targets_mean": 5392.9, + "valid_targets_min": 2803 + }, + { + "epoch": 2.8220338983050848, + "grad_norm": 0.37137395497879805, + "learning_rate": 2.9823818039349407e-05, + "loss": 0.2803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1545882374048233, + "step": 1665, + "valid_targets_mean": 8667.0, + "valid_targets_min": 2654 + }, + { + "epoch": 2.830508474576271, + "grad_norm": 0.3803691362199709, + "learning_rate": 2.9750109595611884e-05, + "loss": 0.29, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14982467889785767, + "step": 1670, + "valid_targets_mean": 8274.6, + "valid_targets_min": 2564 + }, + { + "epoch": 2.8389830508474576, + "grad_norm": 0.4223686202070055, + "learning_rate": 2.967622702593074e-05, + "loss": 0.287, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14334501326084137, + "step": 1675, + "valid_targets_mean": 6762.0, + "valid_targets_min": 2553 + }, + { + "epoch": 2.847457627118644, + "grad_norm": 0.3815613871971945, + "learning_rate": 2.9602171649765235e-05, + "loss": 0.3084, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14498014748096466, + "step": 1680, + "valid_targets_mean": 9098.2, + "valid_targets_min": 4838 + }, + { + "epoch": 2.8559322033898304, + "grad_norm": 0.3902627621269955, + "learning_rate": 2.9527944789660732e-05, + "loss": 0.2907, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13810594379901886, + "step": 1685, + "valid_targets_mean": 8409.4, + "valid_targets_min": 3393 + }, + { + "epoch": 2.864406779661017, + "grad_norm": 0.628093664119152, + "learning_rate": 2.9453547771225088e-05, + "loss": 0.2887, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13151387870311737, + "step": 1690, + "valid_targets_mean": 6601.6, + "valid_targets_min": 2044 + }, + { + "epoch": 2.8728813559322033, + "grad_norm": 0.427340927181757, + "learning_rate": 2.9378981923105026e-05, + "loss": 0.2823, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1253517121076584, + "step": 1695, + "valid_targets_mean": 6468.5, + "valid_targets_min": 3352 + }, + { + "epoch": 2.8813559322033897, + "grad_norm": 0.44354144947684965, + "learning_rate": 2.930424857696236e-05, + "loss": 0.3211, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11692166328430176, + "step": 1700, + "valid_targets_mean": 6686.8, + "valid_targets_min": 3658 + }, + { + "epoch": 2.889830508474576, + "grad_norm": 0.3887488246954865, + "learning_rate": 2.922934906745024e-05, + "loss": 0.3209, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17277604341506958, + "step": 1705, + "valid_targets_mean": 8767.5, + "valid_targets_min": 3271 + }, + { + "epoch": 2.898305084745763, + "grad_norm": 0.41082516803384356, + "learning_rate": 2.9154284732189285e-05, + "loss": 0.3164, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10397391021251678, + "step": 1710, + "valid_targets_mean": 7295.6, + "valid_targets_min": 3874 + }, + { + "epoch": 2.906779661016949, + "grad_norm": 0.520174678897419, + "learning_rate": 2.907905691174374e-05, + "loss": 0.2904, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15237274765968323, + "step": 1715, + "valid_targets_mean": 7268.4, + "valid_targets_min": 2388 + }, + { + "epoch": 2.915254237288136, + "grad_norm": 0.41374806575745976, + "learning_rate": 2.90036669495975e-05, + "loss": 0.3213, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.25504952669143677, + "step": 1720, + "valid_targets_mean": 10837.8, + "valid_targets_min": 4089 + }, + { + "epoch": 2.923728813559322, + "grad_norm": 0.3946990253757063, + "learning_rate": 2.8928116192130127e-05, + "loss": 0.3215, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15455634891986847, + "step": 1725, + "valid_targets_mean": 8932.4, + "valid_targets_min": 4566 + }, + { + "epoch": 2.9322033898305087, + "grad_norm": 0.39390529109371075, + "learning_rate": 2.8852405988592802e-05, + "loss": 0.2815, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12555049359798431, + "step": 1730, + "valid_targets_mean": 6847.9, + "valid_targets_min": 3663 + }, + { + "epoch": 2.940677966101695, + "grad_norm": 0.3977049857931394, + "learning_rate": 2.8776537691084247e-05, + "loss": 0.2986, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20120170712471008, + "step": 1735, + "valid_targets_mean": 9868.5, + "valid_targets_min": 5401 + }, + { + "epoch": 2.9491525423728815, + "grad_norm": 0.4093046968551465, + "learning_rate": 2.8700512654526543e-05, + "loss": 0.3046, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11536908149719238, + "step": 1740, + "valid_targets_mean": 6808.9, + "valid_targets_min": 3326 + }, + { + "epoch": 2.957627118644068, + "grad_norm": 0.432841030661367, + "learning_rate": 2.8624332236640975e-05, + "loss": 0.2869, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15368211269378662, + "step": 1745, + "valid_targets_mean": 8233.4, + "valid_targets_min": 3212 + }, + { + "epoch": 2.9661016949152543, + "grad_norm": 0.4102333184670828, + "learning_rate": 2.854799779792375e-05, + "loss": 0.2825, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12388305366039276, + "step": 1750, + "valid_targets_mean": 6095.2, + "valid_targets_min": 1408 + }, + { + "epoch": 2.9745762711864407, + "grad_norm": 0.4002623064164881, + "learning_rate": 2.8471510701621708e-05, + "loss": 0.2853, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1477382779121399, + "step": 1755, + "valid_targets_mean": 7968.2, + "valid_targets_min": 2866 + }, + { + "epoch": 2.983050847457627, + "grad_norm": 0.4377202503665735, + "learning_rate": 2.8394872313707998e-05, + "loss": 0.3144, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.26475733518600464, + "step": 1760, + "valid_targets_mean": 9149.2, + "valid_targets_min": 2294 + }, + { + "epoch": 2.9915254237288136, + "grad_norm": 0.44825538415638183, + "learning_rate": 2.8318084002857654e-05, + "loss": 0.3017, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14710551500320435, + "step": 1765, + "valid_targets_mean": 6853.5, + "valid_targets_min": 4405 + }, + { + "epoch": 3.0, + "grad_norm": 0.44970926433466, + "learning_rate": 2.824114714042317e-05, + "loss": 0.2819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1573638767004013, + "step": 1770, + "valid_targets_mean": 6817.6, + "valid_targets_min": 2624 + }, + { + "epoch": 3.0084745762711864, + "grad_norm": 0.4105908086436348, + "learning_rate": 2.8164063100409992e-05, + "loss": 0.3075, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17770880460739136, + "step": 1775, + "valid_targets_mean": 11150.9, + "valid_targets_min": 5217 + }, + { + "epoch": 3.016949152542373, + "grad_norm": 0.5067931517713864, + "learning_rate": 2.8086833259452006e-05, + "loss": 0.2576, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14972740411758423, + "step": 1780, + "valid_targets_mean": 7127.5, + "valid_targets_min": 2953 + }, + { + "epoch": 3.0254237288135593, + "grad_norm": 0.4489078904191542, + "learning_rate": 2.8009458996786934e-05, + "loss": 0.2891, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1222635805606842, + "step": 1785, + "valid_targets_mean": 7865.5, + "valid_targets_min": 4141 + }, + { + "epoch": 3.0338983050847457, + "grad_norm": 0.40723043787890206, + "learning_rate": 2.7931941694231705e-05, + "loss": 0.3207, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19112850725650787, + "step": 1790, + "valid_targets_mean": 9303.0, + "valid_targets_min": 3926 + }, + { + "epoch": 3.042372881355932, + "grad_norm": 0.4062904680169323, + "learning_rate": 2.7854282736157777e-05, + "loss": 0.2785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10347068309783936, + "step": 1795, + "valid_targets_mean": 6714.1, + "valid_targets_min": 3013 + }, + { + "epoch": 3.0508474576271185, + "grad_norm": 0.44122857506466184, + "learning_rate": 2.777648350946642e-05, + "loss": 0.3067, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14287069439888, + "step": 1800, + "valid_targets_mean": 7854.1, + "valid_targets_min": 4512 + }, + { + "epoch": 3.059322033898305, + "grad_norm": 0.404748824204509, + "learning_rate": 2.7698545403563934e-05, + "loss": 0.2858, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17268605530261993, + "step": 1805, + "valid_targets_mean": 10822.9, + "valid_targets_min": 4384 + }, + { + "epoch": 3.0677966101694913, + "grad_norm": 0.4196241669238361, + "learning_rate": 2.7620469810336854e-05, + "loss": 0.2608, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10796263068914413, + "step": 1810, + "valid_targets_mean": 7407.6, + "valid_targets_min": 3581 + }, + { + "epoch": 3.0762711864406778, + "grad_norm": 0.42127068984410215, + "learning_rate": 2.754225812412708e-05, + "loss": 0.2882, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24290607869625092, + "step": 1815, + "valid_targets_mean": 10485.5, + "valid_targets_min": 3288 + }, + { + "epoch": 3.084745762711864, + "grad_norm": 0.40150665185852374, + "learning_rate": 2.7463911741706982e-05, + "loss": 0.3186, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20899540185928345, + "step": 1820, + "valid_targets_mean": 12581.9, + "valid_targets_min": 4997 + }, + { + "epoch": 3.093220338983051, + "grad_norm": 0.43521402083428806, + "learning_rate": 2.738543206225445e-05, + "loss": 0.3186, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15953466296195984, + "step": 1825, + "valid_targets_mean": 7655.8, + "valid_targets_min": 3011 + }, + { + "epoch": 3.1016949152542375, + "grad_norm": 0.3962730874041192, + "learning_rate": 2.7306820487327906e-05, + "loss": 0.2562, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09205099940299988, + "step": 1830, + "valid_targets_mean": 6310.2, + "valid_targets_min": 3610 + }, + { + "epoch": 3.110169491525424, + "grad_norm": 0.4359479743437313, + "learning_rate": 2.7228078420841277e-05, + "loss": 0.2873, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1524626463651657, + "step": 1835, + "valid_targets_mean": 8744.8, + "valid_targets_min": 4485 + }, + { + "epoch": 3.1186440677966103, + "grad_norm": 0.40005383291971675, + "learning_rate": 2.714920726903892e-05, + "loss": 0.2612, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12869741022586823, + "step": 1840, + "valid_targets_mean": 9577.2, + "valid_targets_min": 2321 + }, + { + "epoch": 3.1271186440677967, + "grad_norm": 0.4308906520109213, + "learning_rate": 2.7070208440470525e-05, + "loss": 0.3008, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10741919279098511, + "step": 1845, + "valid_targets_mean": 6685.9, + "valid_targets_min": 1623 + }, + { + "epoch": 3.135593220338983, + "grad_norm": 0.4487801888845667, + "learning_rate": 2.699108334596592e-05, + "loss": 0.2956, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1341993808746338, + "step": 1850, + "valid_targets_mean": 6188.6, + "valid_targets_min": 1880 + }, + { + "epoch": 3.1440677966101696, + "grad_norm": 0.7011442094054635, + "learning_rate": 2.6911833398609923e-05, + "loss": 0.2518, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1395626664161682, + "step": 1855, + "valid_targets_mean": 7024.2, + "valid_targets_min": 3107 + }, + { + "epoch": 3.152542372881356, + "grad_norm": 0.3881444819150857, + "learning_rate": 2.683246001371706e-05, + "loss": 0.2862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15838421881198883, + "step": 1860, + "valid_targets_mean": 8513.5, + "valid_targets_min": 3722 + }, + { + "epoch": 3.1610169491525424, + "grad_norm": 0.4653199843483416, + "learning_rate": 2.6752964608806338e-05, + "loss": 0.2854, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16325102746486664, + "step": 1865, + "valid_targets_mean": 8335.9, + "valid_targets_min": 2760 + }, + { + "epoch": 3.169491525423729, + "grad_norm": 0.40908544408335373, + "learning_rate": 2.6673348603575884e-05, + "loss": 0.2775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11137696355581284, + "step": 1870, + "valid_targets_mean": 7132.8, + "valid_targets_min": 2195 + }, + { + "epoch": 3.1779661016949152, + "grad_norm": 0.41926613528884465, + "learning_rate": 2.6593613419877615e-05, + "loss": 0.2927, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1452498883008957, + "step": 1875, + "valid_targets_mean": 7861.4, + "valid_targets_min": 3541 + }, + { + "epoch": 3.1864406779661016, + "grad_norm": 0.4257068600277493, + "learning_rate": 2.6513760481691842e-05, + "loss": 0.3094, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2095491737127304, + "step": 1880, + "valid_targets_mean": 10550.8, + "valid_targets_min": 4405 + }, + { + "epoch": 3.194915254237288, + "grad_norm": 0.43015557115036207, + "learning_rate": 2.6433791215101848e-05, + "loss": 0.2881, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16623318195343018, + "step": 1885, + "valid_targets_mean": 8816.5, + "valid_targets_min": 2364 + }, + { + "epoch": 3.2033898305084745, + "grad_norm": 0.5857399597790577, + "learning_rate": 2.6353707048268397e-05, + "loss": 0.3341, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19451278448104858, + "step": 1890, + "valid_targets_mean": 6448.4, + "valid_targets_min": 2277 + }, + { + "epoch": 3.211864406779661, + "grad_norm": 0.44055645554466305, + "learning_rate": 2.6273509411404234e-05, + "loss": 0.2995, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1880090832710266, + "step": 1895, + "valid_targets_mean": 8197.2, + "valid_targets_min": 2485 + }, + { + "epoch": 3.2203389830508473, + "grad_norm": 0.42726136968985057, + "learning_rate": 2.6193199736748572e-05, + "loss": 0.2862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10744710266590118, + "step": 1900, + "valid_targets_mean": 6499.0, + "valid_targets_min": 3085 + }, + { + "epoch": 3.2288135593220337, + "grad_norm": 0.412861428529702, + "learning_rate": 2.611277945854148e-05, + "loss": 0.3022, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1249886304140091, + "step": 1905, + "valid_targets_mean": 7129.1, + "valid_targets_min": 2327 + }, + { + "epoch": 3.23728813559322, + "grad_norm": 0.4210756070607052, + "learning_rate": 2.603225001299829e-05, + "loss": 0.3021, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18664366006851196, + "step": 1910, + "valid_targets_mean": 9470.2, + "valid_targets_min": 2951 + }, + { + "epoch": 3.2457627118644066, + "grad_norm": 0.4716787626786254, + "learning_rate": 2.595161283828392e-05, + "loss": 0.3117, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10674425959587097, + "step": 1915, + "valid_targets_mean": 5877.9, + "valid_targets_min": 2610 + }, + { + "epoch": 3.2542372881355934, + "grad_norm": 0.4247269620871819, + "learning_rate": 2.5870869374487227e-05, + "loss": 0.2706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1867872029542923, + "step": 1920, + "valid_targets_mean": 8236.0, + "valid_targets_min": 4338 + }, + { + "epoch": 3.26271186440678, + "grad_norm": 1.5928683081810084, + "learning_rate": 2.579002106359527e-05, + "loss": 0.2969, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16273048520088196, + "step": 1925, + "valid_targets_mean": 6147.6, + "valid_targets_min": 2297 + }, + { + "epoch": 3.2711864406779663, + "grad_norm": 0.6792443743882338, + "learning_rate": 2.5709069349467562e-05, + "loss": 0.3404, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13474991917610168, + "step": 1930, + "valid_targets_mean": 7983.4, + "valid_targets_min": 2797 + }, + { + "epoch": 3.2796610169491527, + "grad_norm": 0.39487045934462767, + "learning_rate": 2.562801567781026e-05, + "loss": 0.3009, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.104096919298172, + "step": 1935, + "valid_targets_mean": 8069.4, + "valid_targets_min": 3613 + }, + { + "epoch": 3.288135593220339, + "grad_norm": 0.6447224029652622, + "learning_rate": 2.554686149615038e-05, + "loss": 0.2584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1299767941236496, + "step": 1940, + "valid_targets_mean": 8466.2, + "valid_targets_min": 5145 + }, + { + "epoch": 3.2966101694915255, + "grad_norm": 0.4022839549304916, + "learning_rate": 2.5465608253809944e-05, + "loss": 0.2659, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1030406802892685, + "step": 1945, + "valid_targets_mean": 7143.6, + "valid_targets_min": 1983 + }, + { + "epoch": 3.305084745762712, + "grad_norm": 0.48756887419396766, + "learning_rate": 2.5384257401880077e-05, + "loss": 0.2707, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13388022780418396, + "step": 1950, + "valid_targets_mean": 7445.9, + "valid_targets_min": 3532 + }, + { + "epoch": 3.3135593220338984, + "grad_norm": 0.47900051260276966, + "learning_rate": 2.530281039319509e-05, + "loss": 0.2552, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1074599027633667, + "step": 1955, + "valid_targets_mean": 5242.0, + "valid_targets_min": 2971 + }, + { + "epoch": 3.3220338983050848, + "grad_norm": 0.40376480776936347, + "learning_rate": 2.5221268682306565e-05, + "loss": 0.2629, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12285137176513672, + "step": 1960, + "valid_targets_mean": 8048.4, + "valid_targets_min": 2983 + }, + { + "epoch": 3.330508474576271, + "grad_norm": 0.4042419082898355, + "learning_rate": 2.5139633725457345e-05, + "loss": 0.3055, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1488046646118164, + "step": 1965, + "valid_targets_mean": 8609.8, + "valid_targets_min": 3878 + }, + { + "epoch": 3.3389830508474576, + "grad_norm": 0.42120896138472086, + "learning_rate": 2.505790698055554e-05, + "loss": 0.2804, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11100804805755615, + "step": 1970, + "valid_targets_mean": 7231.4, + "valid_targets_min": 5579 + }, + { + "epoch": 3.347457627118644, + "grad_norm": 0.4478852216820594, + "learning_rate": 2.49760899071485e-05, + "loss": 0.2928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1743670552968979, + "step": 1975, + "valid_targets_mean": 7310.9, + "valid_targets_min": 2377 + }, + { + "epoch": 3.3559322033898304, + "grad_norm": 0.4453850109809996, + "learning_rate": 2.4894183966396735e-05, + "loss": 0.2705, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09428214281797409, + "step": 1980, + "valid_targets_mean": 5442.4, + "valid_targets_min": 2760 + }, + { + "epoch": 3.364406779661017, + "grad_norm": 0.386834238314581, + "learning_rate": 2.4812190621047822e-05, + "loss": 0.2963, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10531061887741089, + "step": 1985, + "valid_targets_mean": 8721.5, + "valid_targets_min": 4360 + }, + { + "epoch": 3.3728813559322033, + "grad_norm": 0.41974129998150933, + "learning_rate": 2.4730111335410298e-05, + "loss": 0.3058, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17613358795642853, + "step": 1990, + "valid_targets_mean": 7780.2, + "valid_targets_min": 1613 + }, + { + "epoch": 3.3813559322033897, + "grad_norm": 0.3986453592630357, + "learning_rate": 2.4647947575327468e-05, + "loss": 0.2798, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09645844250917435, + "step": 1995, + "valid_targets_mean": 5781.1, + "valid_targets_min": 3468 + }, + { + "epoch": 3.389830508474576, + "grad_norm": 0.42160649350404233, + "learning_rate": 2.4565700808151297e-05, + "loss": 0.2525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10721124708652496, + "step": 2000, + "valid_targets_mean": 6691.4, + "valid_targets_min": 3904 + }, + { + "epoch": 3.3983050847457625, + "grad_norm": 0.395642493359529, + "learning_rate": 2.4483372502716124e-05, + "loss": 0.282, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08538131415843964, + "step": 2005, + "valid_targets_mean": 5464.9, + "valid_targets_min": 3818 + }, + { + "epoch": 3.406779661016949, + "grad_norm": 0.4377539465369687, + "learning_rate": 2.44009641293125e-05, + "loss": 0.29, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1419958472251892, + "step": 2010, + "valid_targets_mean": 7003.5, + "valid_targets_min": 2705 + }, + { + "epoch": 3.415254237288136, + "grad_norm": 0.390606658868245, + "learning_rate": 2.431847715966087e-05, + "loss": 0.2811, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09183446317911148, + "step": 2015, + "valid_targets_mean": 6089.4, + "valid_targets_min": 2693 + }, + { + "epoch": 3.423728813559322, + "grad_norm": 0.44601554122640463, + "learning_rate": 2.423591306688534e-05, + "loss": 0.2722, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11288672685623169, + "step": 2020, + "valid_targets_mean": 6390.1, + "valid_targets_min": 2918 + }, + { + "epoch": 3.4322033898305087, + "grad_norm": 0.45105339427901514, + "learning_rate": 2.4153273325487346e-05, + "loss": 0.2813, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17822352051734924, + "step": 2025, + "valid_targets_mean": 8611.0, + "valid_targets_min": 2586 + }, + { + "epoch": 3.440677966101695, + "grad_norm": 0.4580856128531317, + "learning_rate": 2.407055941131932e-05, + "loss": 0.2675, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18808065354824066, + "step": 2030, + "valid_targets_mean": 7461.8, + "valid_targets_min": 2938 + }, + { + "epoch": 3.4491525423728815, + "grad_norm": 0.42333388708901265, + "learning_rate": 2.3987772801558328e-05, + "loss": 0.2873, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22082702815532684, + "step": 2035, + "valid_targets_mean": 8332.9, + "valid_targets_min": 3046 + }, + { + "epoch": 3.457627118644068, + "grad_norm": 0.4249810503191986, + "learning_rate": 2.3904914974679705e-05, + "loss": 0.2845, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10297828912734985, + "step": 2040, + "valid_targets_mean": 6474.1, + "valid_targets_min": 3617 + }, + { + "epoch": 3.4661016949152543, + "grad_norm": 0.42287779451494156, + "learning_rate": 2.3821987410430646e-05, + "loss": 0.2671, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12721748650074005, + "step": 2045, + "valid_targets_mean": 7318.4, + "valid_targets_min": 2590 + }, + { + "epoch": 3.4745762711864407, + "grad_norm": 0.41487510241403797, + "learning_rate": 2.3738991589803783e-05, + "loss": 0.2978, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20614950358867645, + "step": 2050, + "valid_targets_mean": 12113.4, + "valid_targets_min": 7751 + }, + { + "epoch": 3.483050847457627, + "grad_norm": 0.40445186849790093, + "learning_rate": 2.36559289950107e-05, + "loss": 0.281, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18696916103363037, + "step": 2055, + "valid_targets_mean": 9497.6, + "valid_targets_min": 3046 + }, + { + "epoch": 3.4915254237288136, + "grad_norm": 0.3808081177692009, + "learning_rate": 2.357280110945552e-05, + "loss": 0.2985, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15332549810409546, + "step": 2060, + "valid_targets_mean": 9510.1, + "valid_targets_min": 4850 + }, + { + "epoch": 3.5, + "grad_norm": 0.3954559102872441, + "learning_rate": 2.3489609417708383e-05, + "loss": 0.3108, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07950934767723083, + "step": 2065, + "valid_targets_mean": 6016.5, + "valid_targets_min": 2468 + }, + { + "epoch": 3.5084745762711864, + "grad_norm": 0.47525617396789216, + "learning_rate": 2.340635540547892e-05, + "loss": 0.2611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09396298229694366, + "step": 2070, + "valid_targets_mean": 6578.6, + "valid_targets_min": 2441 + }, + { + "epoch": 3.516949152542373, + "grad_norm": 0.41700924280969, + "learning_rate": 2.3323040559589755e-05, + "loss": 0.2853, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15074723958969116, + "step": 2075, + "valid_targets_mean": 7548.5, + "valid_targets_min": 2942 + }, + { + "epoch": 3.5254237288135593, + "grad_norm": 0.4045155062865091, + "learning_rate": 2.3239666367949917e-05, + "loss": 0.2843, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1475975513458252, + "step": 2080, + "valid_targets_mean": 7759.4, + "valid_targets_min": 3593 + }, + { + "epoch": 3.5338983050847457, + "grad_norm": 0.45096675546516823, + "learning_rate": 2.3156234319528295e-05, + "loss": 0.2665, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09082523733377457, + "step": 2085, + "valid_targets_mean": 6154.9, + "valid_targets_min": 2823 + }, + { + "epoch": 3.542372881355932, + "grad_norm": 0.4350300656202291, + "learning_rate": 2.3072745904327015e-05, + "loss": 0.2958, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09289325028657913, + "step": 2090, + "valid_targets_mean": 6375.0, + "valid_targets_min": 5184 + }, + { + "epoch": 3.5508474576271185, + "grad_norm": 0.3910107816800009, + "learning_rate": 2.298920261335488e-05, + "loss": 0.2588, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09527669847011566, + "step": 2095, + "valid_targets_mean": 6311.5, + "valid_targets_min": 2557 + }, + { + "epoch": 3.559322033898305, + "grad_norm": 0.39669458488926956, + "learning_rate": 2.2905605938600685e-05, + "loss": 0.2911, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20767447352409363, + "step": 2100, + "valid_targets_mean": 12417.0, + "valid_targets_min": 6509 + }, + { + "epoch": 3.5677966101694913, + "grad_norm": 0.3901163518152081, + "learning_rate": 2.2821957373006617e-05, + "loss": 0.2634, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18332387506961823, + "step": 2105, + "valid_targets_mean": 10700.6, + "valid_targets_min": 6652 + }, + { + "epoch": 3.576271186440678, + "grad_norm": 0.4205397491800668, + "learning_rate": 2.273825841044156e-05, + "loss": 0.2958, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15848278999328613, + "step": 2110, + "valid_targets_mean": 8436.5, + "valid_targets_min": 2593 + }, + { + "epoch": 3.584745762711864, + "grad_norm": 0.39017980717827244, + "learning_rate": 2.2654510545674442e-05, + "loss": 0.2788, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14073039591312408, + "step": 2115, + "valid_targets_mean": 8208.6, + "valid_targets_min": 3676 + }, + { + "epoch": 3.593220338983051, + "grad_norm": 0.413374304774211, + "learning_rate": 2.257071527434753e-05, + "loss": 0.2984, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15054945647716522, + "step": 2120, + "valid_targets_mean": 7802.0, + "valid_targets_min": 3176 + }, + { + "epoch": 3.601694915254237, + "grad_norm": 0.3924930780824449, + "learning_rate": 2.2486874092949708e-05, + "loss": 0.2826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12603721022605896, + "step": 2125, + "valid_targets_mean": 7997.5, + "valid_targets_min": 3955 + }, + { + "epoch": 3.610169491525424, + "grad_norm": 0.49818479968661744, + "learning_rate": 2.2402988498789767e-05, + "loss": 0.2928, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10280188918113708, + "step": 2130, + "valid_targets_mean": 7086.6, + "valid_targets_min": 2653 + }, + { + "epoch": 3.6186440677966103, + "grad_norm": 0.46928295157294475, + "learning_rate": 2.2319059989969668e-05, + "loss": 0.2946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1649240404367447, + "step": 2135, + "valid_targets_mean": 7845.5, + "valid_targets_min": 3649 + }, + { + "epoch": 3.6271186440677967, + "grad_norm": 0.462776490971101, + "learning_rate": 2.2235090065357773e-05, + "loss": 0.296, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18557463586330414, + "step": 2140, + "valid_targets_mean": 8834.4, + "valid_targets_min": 4137 + }, + { + "epoch": 3.635593220338983, + "grad_norm": 0.41019380389792665, + "learning_rate": 2.215108022456208e-05, + "loss": 0.3039, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14846105873584747, + "step": 2145, + "valid_targets_mean": 8244.1, + "valid_targets_min": 4463 + }, + { + "epoch": 3.6440677966101696, + "grad_norm": 0.38387126303093333, + "learning_rate": 2.2067031967903443e-05, + "loss": 0.2561, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13300243020057678, + "step": 2150, + "valid_targets_mean": 9330.9, + "valid_targets_min": 4265 + }, + { + "epoch": 3.652542372881356, + "grad_norm": 0.3585813319451801, + "learning_rate": 2.1982946796388788e-05, + "loss": 0.3076, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17027175426483154, + "step": 2155, + "valid_targets_mean": 11430.1, + "valid_targets_min": 4582 + }, + { + "epoch": 3.6610169491525424, + "grad_norm": 0.41586972741470607, + "learning_rate": 2.1898826211684297e-05, + "loss": 0.2756, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15760239958763123, + "step": 2160, + "valid_targets_mean": 8135.5, + "valid_targets_min": 3723 + }, + { + "epoch": 3.669491525423729, + "grad_norm": 0.41201232045267855, + "learning_rate": 2.1814671716088593e-05, + "loss": 0.3082, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08165358006954193, + "step": 2165, + "valid_targets_mean": 6309.6, + "valid_targets_min": 2574 + }, + { + "epoch": 3.6779661016949152, + "grad_norm": 0.41165756708117474, + "learning_rate": 2.17304848125059e-05, + "loss": 0.294, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1658606231212616, + "step": 2170, + "valid_targets_mean": 8837.6, + "valid_targets_min": 4222 + }, + { + "epoch": 3.6864406779661016, + "grad_norm": 0.42523034204474275, + "learning_rate": 2.1646267004419218e-05, + "loss": 0.2758, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14023277163505554, + "step": 2175, + "valid_targets_mean": 7177.4, + "valid_targets_min": 2581 + }, + { + "epoch": 3.694915254237288, + "grad_norm": 0.4868073516276877, + "learning_rate": 2.1562019795863463e-05, + "loss": 0.2974, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2292966991662979, + "step": 2180, + "valid_targets_mean": 9235.9, + "valid_targets_min": 4157 + }, + { + "epoch": 3.7033898305084745, + "grad_norm": 0.4759427055257523, + "learning_rate": 2.147774469139863e-05, + "loss": 0.2534, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10483382642269135, + "step": 2185, + "valid_targets_mean": 6474.5, + "valid_targets_min": 3561 + }, + { + "epoch": 3.711864406779661, + "grad_norm": 0.40004361325929844, + "learning_rate": 2.1393443196082867e-05, + "loss": 0.2851, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22268085181713104, + "step": 2190, + "valid_targets_mean": 11323.0, + "valid_targets_min": 4111 + }, + { + "epoch": 3.7203389830508473, + "grad_norm": 0.45946059913819876, + "learning_rate": 2.1309116815445665e-05, + "loss": 0.2772, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09428419917821884, + "step": 2195, + "valid_targets_mean": 5618.4, + "valid_targets_min": 2974 + }, + { + "epoch": 3.7288135593220337, + "grad_norm": 0.41004051082840914, + "learning_rate": 2.1224767055460922e-05, + "loss": 0.2984, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1019151583313942, + "step": 2200, + "valid_targets_mean": 6402.8, + "valid_targets_min": 3145 + }, + { + "epoch": 3.7372881355932206, + "grad_norm": 0.43461723876845115, + "learning_rate": 2.114039542252008e-05, + "loss": 0.3004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15146639943122864, + "step": 2205, + "valid_targets_mean": 7431.5, + "valid_targets_min": 3175 + }, + { + "epoch": 3.7457627118644066, + "grad_norm": 0.4521829340840556, + "learning_rate": 2.1056003423405178e-05, + "loss": 0.3037, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2280888557434082, + "step": 2210, + "valid_targets_mean": 8751.5, + "valid_targets_min": 2953 + }, + { + "epoch": 3.7542372881355934, + "grad_norm": 0.3790459152189435, + "learning_rate": 2.0971592565262005e-05, + "loss": 0.2938, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19973929226398468, + "step": 2215, + "valid_targets_mean": 10629.4, + "valid_targets_min": 5965 + }, + { + "epoch": 3.7627118644067794, + "grad_norm": 0.41422926941177723, + "learning_rate": 2.088716435557313e-05, + "loss": 0.2809, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10134095698595047, + "step": 2220, + "valid_targets_mean": 7064.5, + "valid_targets_min": 3524 + }, + { + "epoch": 3.7711864406779663, + "grad_norm": 0.39304286386103016, + "learning_rate": 2.0802720302131016e-05, + "loss": 0.3082, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13361665606498718, + "step": 2225, + "valid_targets_mean": 8738.2, + "valid_targets_min": 1379 + }, + { + "epoch": 3.7796610169491527, + "grad_norm": 0.4221134423921705, + "learning_rate": 2.0718261913011055e-05, + "loss": 0.2891, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.155862957239151, + "step": 2230, + "valid_targets_mean": 7316.0, + "valid_targets_min": 3376 + }, + { + "epoch": 3.788135593220339, + "grad_norm": 0.41722855283264804, + "learning_rate": 2.063379069654468e-05, + "loss": 0.2434, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0939258486032486, + "step": 2235, + "valid_targets_mean": 5850.2, + "valid_targets_min": 1572 + }, + { + "epoch": 3.7966101694915255, + "grad_norm": 0.4165192968853601, + "learning_rate": 2.0549308161292405e-05, + "loss": 0.3023, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14758262038230896, + "step": 2240, + "valid_targets_mean": 7667.9, + "valid_targets_min": 2295 + }, + { + "epoch": 3.805084745762712, + "grad_norm": 0.4549005587942355, + "learning_rate": 2.0464815816016866e-05, + "loss": 0.2698, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11906784772872925, + "step": 2245, + "valid_targets_mean": 6164.1, + "valid_targets_min": 3415 + }, + { + "epoch": 3.8135593220338984, + "grad_norm": 0.5120597710854238, + "learning_rate": 2.0380315169655905e-05, + "loss": 0.3099, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15578162670135498, + "step": 2250, + "valid_targets_mean": 8488.0, + "valid_targets_min": 3736 + }, + { + "epoch": 3.8220338983050848, + "grad_norm": 0.4090425819904948, + "learning_rate": 2.0295807731295628e-05, + "loss": 0.2861, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16633877158164978, + "step": 2255, + "valid_targets_mean": 10121.9, + "valid_targets_min": 5852 + }, + { + "epoch": 3.830508474576271, + "grad_norm": 0.40631012505431235, + "learning_rate": 2.021129501014342e-05, + "loss": 0.2637, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20459291338920593, + "step": 2260, + "valid_targets_mean": 9120.1, + "valid_targets_min": 3127 + }, + { + "epoch": 3.8389830508474576, + "grad_norm": 0.4728245585759232, + "learning_rate": 2.0126778515501e-05, + "loss": 0.2859, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13413026928901672, + "step": 2265, + "valid_targets_mean": 6771.2, + "valid_targets_min": 2591 + }, + { + "epoch": 3.847457627118644, + "grad_norm": 0.41222402258457613, + "learning_rate": 2.0042259756737508e-05, + "loss": 0.3067, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12306500971317291, + "step": 2270, + "valid_targets_mean": 7164.5, + "valid_targets_min": 2683 + }, + { + "epoch": 3.8559322033898304, + "grad_norm": 0.41893653287844634, + "learning_rate": 1.99577402432625e-05, + "loss": 0.2925, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14483068883419037, + "step": 2275, + "valid_targets_mean": 7233.9, + "valid_targets_min": 3157 + }, + { + "epoch": 3.864406779661017, + "grad_norm": 0.4184775262000185, + "learning_rate": 1.9873221484499006e-05, + "loss": 0.2674, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17196547985076904, + "step": 2280, + "valid_targets_mean": 8610.6, + "valid_targets_min": 2654 + }, + { + "epoch": 3.8728813559322033, + "grad_norm": 0.3960402767223172, + "learning_rate": 1.978870498985659e-05, + "loss": 0.2946, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10620152950286865, + "step": 2285, + "valid_targets_mean": 6905.9, + "valid_targets_min": 2265 + }, + { + "epoch": 3.8813559322033897, + "grad_norm": 0.4615512806555675, + "learning_rate": 1.9704192268704372e-05, + "loss": 0.2703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1019475981593132, + "step": 2290, + "valid_targets_mean": 5965.6, + "valid_targets_min": 1695 + }, + { + "epoch": 3.889830508474576, + "grad_norm": 0.3745908989504767, + "learning_rate": 1.96196848303441e-05, + "loss": 0.2871, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1855563521385193, + "step": 2295, + "valid_targets_mean": 9307.8, + "valid_targets_min": 4413 + }, + { + "epoch": 3.898305084745763, + "grad_norm": 0.41468014631088473, + "learning_rate": 1.953518418398314e-05, + "loss": 0.2751, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14965686202049255, + "step": 2300, + "valid_targets_mean": 6959.0, + "valid_targets_min": 2596 + }, + { + "epoch": 3.906779661016949, + "grad_norm": 0.4100589095783291, + "learning_rate": 1.94506918387076e-05, + "loss": 0.284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15163038671016693, + "step": 2305, + "valid_targets_mean": 8789.0, + "valid_targets_min": 3569 + }, + { + "epoch": 3.915254237288136, + "grad_norm": 0.4296013561335367, + "learning_rate": 1.9366209303455322e-05, + "loss": 0.2901, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14941740036010742, + "step": 2310, + "valid_targets_mean": 8584.9, + "valid_targets_min": 5087 + }, + { + "epoch": 3.923728813559322, + "grad_norm": 0.3837812792576554, + "learning_rate": 1.928173808698895e-05, + "loss": 0.2884, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0891745388507843, + "step": 2315, + "valid_targets_mean": 6788.1, + "valid_targets_min": 4063 + }, + { + "epoch": 3.9322033898305087, + "grad_norm": 0.39210577389927076, + "learning_rate": 1.919727969786899e-05, + "loss": 0.2848, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14674127101898193, + "step": 2320, + "valid_targets_mean": 8767.0, + "valid_targets_min": 2817 + }, + { + "epoch": 3.940677966101695, + "grad_norm": 0.3850709276676864, + "learning_rate": 1.911283564442687e-05, + "loss": 0.2735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15203866362571716, + "step": 2325, + "valid_targets_mean": 8221.0, + "valid_targets_min": 2763 + }, + { + "epoch": 3.9491525423728815, + "grad_norm": 0.8402295961588697, + "learning_rate": 1.9028407434737998e-05, + "loss": 0.2789, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15760354697704315, + "step": 2330, + "valid_targets_mean": 6542.4, + "valid_targets_min": 708 + }, + { + "epoch": 3.957627118644068, + "grad_norm": 0.4045787479410837, + "learning_rate": 1.894399657659483e-05, + "loss": 0.2846, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13099080324172974, + "step": 2335, + "valid_targets_mean": 7977.9, + "valid_targets_min": 4438 + }, + { + "epoch": 3.9661016949152543, + "grad_norm": 0.4205585498329919, + "learning_rate": 1.8859604577479927e-05, + "loss": 0.2699, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1536920815706253, + "step": 2340, + "valid_targets_mean": 7394.5, + "valid_targets_min": 2784 + }, + { + "epoch": 3.9745762711864407, + "grad_norm": 0.4527956277921518, + "learning_rate": 1.877523294453908e-05, + "loss": 0.2759, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11789265275001526, + "step": 2345, + "valid_targets_mean": 5872.2, + "valid_targets_min": 2255 + }, + { + "epoch": 3.983050847457627, + "grad_norm": 0.41835202159742035, + "learning_rate": 1.869088318455434e-05, + "loss": 0.284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13039818406105042, + "step": 2350, + "valid_targets_mean": 7903.4, + "valid_targets_min": 4010 + }, + { + "epoch": 3.9915254237288136, + "grad_norm": 0.445312803177915, + "learning_rate": 1.8606556803917136e-05, + "loss": 0.2346, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12284279614686966, + "step": 2355, + "valid_targets_mean": 7089.2, + "valid_targets_min": 2543 + }, + { + "epoch": 4.0, + "grad_norm": 0.43140080733127073, + "learning_rate": 1.8522255308601378e-05, + "loss": 0.2733, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15027593076229095, + "step": 2360, + "valid_targets_mean": 6901.6, + "valid_targets_min": 1899 + }, + { + "epoch": 4.008474576271187, + "grad_norm": 0.4017056399777449, + "learning_rate": 1.8437980204136537e-05, + "loss": 0.2855, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1381210833787918, + "step": 2365, + "valid_targets_mean": 6844.1, + "valid_targets_min": 3599 + }, + { + "epoch": 4.016949152542373, + "grad_norm": 0.39834705448486263, + "learning_rate": 1.8353732995580786e-05, + "loss": 0.248, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13300484418869019, + "step": 2370, + "valid_targets_mean": 7599.9, + "valid_targets_min": 3499 + }, + { + "epoch": 4.02542372881356, + "grad_norm": 0.40105037270610566, + "learning_rate": 1.826951518749411e-05, + "loss": 0.2807, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15867316722869873, + "step": 2375, + "valid_targets_mean": 9657.4, + "valid_targets_min": 4560 + }, + { + "epoch": 4.033898305084746, + "grad_norm": 0.38773203722315996, + "learning_rate": 1.8185328283911417e-05, + "loss": 0.2649, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11912851780653, + "step": 2380, + "valid_targets_mean": 8314.6, + "valid_targets_min": 3415 + }, + { + "epoch": 4.0423728813559325, + "grad_norm": 0.452115428936143, + "learning_rate": 1.8101173788315707e-05, + "loss": 0.3013, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17245829105377197, + "step": 2385, + "valid_targets_mean": 8021.1, + "valid_targets_min": 2963 + }, + { + "epoch": 4.0508474576271185, + "grad_norm": 0.38361717703820286, + "learning_rate": 1.8017053203611215e-05, + "loss": 0.2583, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13646253943443298, + "step": 2390, + "valid_targets_mean": 8014.4, + "valid_targets_min": 3516 + }, + { + "epoch": 4.059322033898305, + "grad_norm": 0.4199999783766128, + "learning_rate": 1.7932968032096564e-05, + "loss": 0.2779, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07727745175361633, + "step": 2395, + "valid_targets_mean": 5372.4, + "valid_targets_min": 2652 + }, + { + "epoch": 4.067796610169491, + "grad_norm": 0.41474929838546026, + "learning_rate": 1.7848919775437924e-05, + "loss": 0.2679, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11556974798440933, + "step": 2400, + "valid_targets_mean": 7041.9, + "valid_targets_min": 2887 + }, + { + "epoch": 4.076271186440678, + "grad_norm": 0.4055672156105649, + "learning_rate": 1.776490993464223e-05, + "loss": 0.2886, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13929562270641327, + "step": 2405, + "valid_targets_mean": 8820.8, + "valid_targets_min": 4759 + }, + { + "epoch": 4.084745762711864, + "grad_norm": 0.4457946964522161, + "learning_rate": 1.768094001003033e-05, + "loss": 0.2592, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13029378652572632, + "step": 2410, + "valid_targets_mean": 7816.9, + "valid_targets_min": 4082 + }, + { + "epoch": 4.093220338983051, + "grad_norm": 0.41426137742366353, + "learning_rate": 1.759701150121024e-05, + "loss": 0.2474, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13159283995628357, + "step": 2415, + "valid_targets_mean": 7818.9, + "valid_targets_min": 2297 + }, + { + "epoch": 4.101694915254237, + "grad_norm": 2.09180034020413, + "learning_rate": 1.7513125907050302e-05, + "loss": 0.2916, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2276926338672638, + "step": 2420, + "valid_targets_mean": 6345.2, + "valid_targets_min": 2255 + }, + { + "epoch": 4.110169491525424, + "grad_norm": 0.4812271302697601, + "learning_rate": 1.742928472565248e-05, + "loss": 0.2718, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09323035925626755, + "step": 2425, + "valid_targets_mean": 5329.6, + "valid_targets_min": 2364 + }, + { + "epoch": 4.11864406779661, + "grad_norm": 0.4121616023593775, + "learning_rate": 1.7345489454325564e-05, + "loss": 0.2717, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07100459933280945, + "step": 2430, + "valid_targets_mean": 5565.1, + "valid_targets_min": 2627 + }, + { + "epoch": 4.127118644067797, + "grad_norm": 0.3960969727097476, + "learning_rate": 1.7261741589558448e-05, + "loss": 0.2892, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0982292965054512, + "step": 2435, + "valid_targets_mean": 6970.4, + "valid_targets_min": 3129 + }, + { + "epoch": 4.135593220338983, + "grad_norm": 0.44224599313603047, + "learning_rate": 1.717804262699339e-05, + "loss": 0.273, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15915395319461823, + "step": 2440, + "valid_targets_mean": 8128.4, + "valid_targets_min": 4652 + }, + { + "epoch": 4.1440677966101696, + "grad_norm": 0.5038339183568116, + "learning_rate": 1.7094394061399318e-05, + "loss": 0.2589, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11271384358406067, + "step": 2445, + "valid_targets_mean": 6489.4, + "valid_targets_min": 3508 + }, + { + "epoch": 4.1525423728813555, + "grad_norm": 0.41508836319034587, + "learning_rate": 1.701079738664512e-05, + "loss": 0.2622, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11254729330539703, + "step": 2450, + "valid_targets_mean": 7253.1, + "valid_targets_min": 2962 + }, + { + "epoch": 4.161016949152542, + "grad_norm": 0.4424755960913446, + "learning_rate": 1.6927254095672992e-05, + "loss": 0.266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10713516175746918, + "step": 2455, + "valid_targets_mean": 5768.9, + "valid_targets_min": 2435 + }, + { + "epoch": 4.169491525423728, + "grad_norm": 0.413886360698127, + "learning_rate": 1.6843765680471715e-05, + "loss": 0.2761, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11144141107797623, + "step": 2460, + "valid_targets_mean": 6971.0, + "valid_targets_min": 3025 + }, + { + "epoch": 4.177966101694915, + "grad_norm": 0.449283900078661, + "learning_rate": 1.6760333632050086e-05, + "loss": 0.2851, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1145852655172348, + "step": 2465, + "valid_targets_mean": 6577.0, + "valid_targets_min": 3004 + }, + { + "epoch": 4.186440677966102, + "grad_norm": 0.4130721972861414, + "learning_rate": 1.667695944041025e-05, + "loss": 0.2744, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11476103961467743, + "step": 2470, + "valid_targets_mean": 7053.8, + "valid_targets_min": 2920 + }, + { + "epoch": 4.194915254237288, + "grad_norm": 0.44076622259430287, + "learning_rate": 1.6593644594521082e-05, + "loss": 0.301, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.131443053483963, + "step": 2475, + "valid_targets_mean": 7432.2, + "valid_targets_min": 2321 + }, + { + "epoch": 4.203389830508475, + "grad_norm": 0.3966902009222672, + "learning_rate": 1.6510390582291624e-05, + "loss": 0.2627, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18674710392951965, + "step": 2480, + "valid_targets_mean": 9859.8, + "valid_targets_min": 4988 + }, + { + "epoch": 4.211864406779661, + "grad_norm": 0.40242400071919937, + "learning_rate": 1.6427198890544483e-05, + "loss": 0.27, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14379474520683289, + "step": 2485, + "valid_targets_mean": 8349.8, + "valid_targets_min": 2503 + }, + { + "epoch": 4.220338983050848, + "grad_norm": 0.404724712627933, + "learning_rate": 1.634407100498931e-05, + "loss": 0.2656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11465313285589218, + "step": 2490, + "valid_targets_mean": 7916.4, + "valid_targets_min": 4091 + }, + { + "epoch": 4.228813559322034, + "grad_norm": 0.44146534476895466, + "learning_rate": 1.626100841019623e-05, + "loss": 0.2649, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11277331411838531, + "step": 2495, + "valid_targets_mean": 6573.2, + "valid_targets_min": 3682 + }, + { + "epoch": 4.237288135593221, + "grad_norm": 0.4312186315459749, + "learning_rate": 1.6178012589569358e-05, + "loss": 0.29, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2054651379585266, + "step": 2500, + "valid_targets_mean": 9550.8, + "valid_targets_min": 4374 + }, + { + "epoch": 4.245762711864407, + "grad_norm": 0.40476837965859497, + "learning_rate": 1.60950850253203e-05, + "loss": 0.2792, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19124539196491241, + "step": 2505, + "valid_targets_mean": 9616.9, + "valid_targets_min": 4814 + }, + { + "epoch": 4.254237288135593, + "grad_norm": 0.3968716278946917, + "learning_rate": 1.601222719844168e-05, + "loss": 0.2614, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10168001055717468, + "step": 2510, + "valid_targets_mean": 8180.5, + "valid_targets_min": 4040 + }, + { + "epoch": 4.262711864406779, + "grad_norm": 0.40747689243864516, + "learning_rate": 1.5929440588680688e-05, + "loss": 0.2405, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1375788301229477, + "step": 2515, + "valid_targets_mean": 9741.0, + "valid_targets_min": 3275 + }, + { + "epoch": 4.271186440677966, + "grad_norm": 0.3850303265427952, + "learning_rate": 1.5846726674512654e-05, + "loss": 0.2487, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09620785713195801, + "step": 2520, + "valid_targets_mean": 7801.1, + "valid_targets_min": 3324 + }, + { + "epoch": 4.279661016949152, + "grad_norm": 0.4694382092602058, + "learning_rate": 1.576408693311466e-05, + "loss": 0.269, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18548263609409332, + "step": 2525, + "valid_targets_mean": 9196.9, + "valid_targets_min": 3320 + }, + { + "epoch": 4.288135593220339, + "grad_norm": 0.44415310400437963, + "learning_rate": 1.5681522840339143e-05, + "loss": 0.2741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1131313145160675, + "step": 2530, + "valid_targets_mean": 6663.9, + "valid_targets_min": 1599 + }, + { + "epoch": 4.296610169491525, + "grad_norm": 0.4711568751953429, + "learning_rate": 1.5599035870687515e-05, + "loss": 0.2872, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12717491388320923, + "step": 2535, + "valid_targets_mean": 6152.5, + "valid_targets_min": 2490 + }, + { + "epoch": 4.305084745762712, + "grad_norm": 0.42597584649898684, + "learning_rate": 1.5516627497283882e-05, + "loss": 0.2769, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19627636671066284, + "step": 2540, + "valid_targets_mean": 9630.6, + "valid_targets_min": 2618 + }, + { + "epoch": 4.313559322033898, + "grad_norm": 0.43880093890375366, + "learning_rate": 1.5434299191848713e-05, + "loss": 0.2819, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1505734920501709, + "step": 2545, + "valid_targets_mean": 7736.9, + "valid_targets_min": 2612 + }, + { + "epoch": 4.322033898305085, + "grad_norm": 0.424704792501036, + "learning_rate": 1.5352052424672535e-05, + "loss": 0.2662, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13303329050540924, + "step": 2550, + "valid_targets_mean": 6638.0, + "valid_targets_min": 1026 + }, + { + "epoch": 4.330508474576272, + "grad_norm": 0.43108950885838404, + "learning_rate": 1.5269888664589712e-05, + "loss": 0.2604, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09408214688301086, + "step": 2555, + "valid_targets_mean": 6848.9, + "valid_targets_min": 3869 + }, + { + "epoch": 4.338983050847458, + "grad_norm": 0.4331199795806166, + "learning_rate": 1.5187809378952182e-05, + "loss": 0.2572, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10691245645284653, + "step": 2560, + "valid_targets_mean": 6317.9, + "valid_targets_min": 1836 + }, + { + "epoch": 4.3474576271186445, + "grad_norm": 0.4080221972422781, + "learning_rate": 1.5105816033603266e-05, + "loss": 0.2446, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09583228826522827, + "step": 2565, + "valid_targets_mean": 6374.9, + "valid_targets_min": 3557 + }, + { + "epoch": 4.3559322033898304, + "grad_norm": 0.4591461541061156, + "learning_rate": 1.5023910092851509e-05, + "loss": 0.2713, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1480543315410614, + "step": 2570, + "valid_targets_mean": 7753.4, + "valid_targets_min": 4375 + }, + { + "epoch": 4.364406779661017, + "grad_norm": 0.4848625676758303, + "learning_rate": 1.4942093019444468e-05, + "loss": 0.2729, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21596203744411469, + "step": 2575, + "valid_targets_mean": 7900.9, + "valid_targets_min": 2254 + }, + { + "epoch": 4.372881355932203, + "grad_norm": 0.43994625173957885, + "learning_rate": 1.4860366274542663e-05, + "loss": 0.2594, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15222656726837158, + "step": 2580, + "valid_targets_mean": 6838.0, + "valid_targets_min": 3597 + }, + { + "epoch": 4.38135593220339, + "grad_norm": 0.3710706477412973, + "learning_rate": 1.4778731317693442e-05, + "loss": 0.249, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15013235807418823, + "step": 2585, + "valid_targets_mean": 10847.1, + "valid_targets_min": 2320 + }, + { + "epoch": 4.389830508474576, + "grad_norm": 0.4232170416787067, + "learning_rate": 1.4697189606804914e-05, + "loss": 0.2504, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11278915405273438, + "step": 2590, + "valid_targets_mean": 7221.4, + "valid_targets_min": 4252 + }, + { + "epoch": 4.398305084745763, + "grad_norm": 0.43244240188093136, + "learning_rate": 1.4615742598119927e-05, + "loss": 0.2638, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10896097123622894, + "step": 2595, + "valid_targets_mean": 7083.2, + "valid_targets_min": 2512 + }, + { + "epoch": 4.406779661016949, + "grad_norm": 0.45057508460672135, + "learning_rate": 1.4534391746190056e-05, + "loss": 0.2781, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.106690414249897, + "step": 2600, + "valid_targets_mean": 7374.6, + "valid_targets_min": 2558 + }, + { + "epoch": 4.415254237288136, + "grad_norm": 0.3987793446104239, + "learning_rate": 1.4453138503849622e-05, + "loss": 0.3089, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11705589294433594, + "step": 2605, + "valid_targets_mean": 7237.1, + "valid_targets_min": 2997 + }, + { + "epoch": 4.423728813559322, + "grad_norm": 0.40192297388199794, + "learning_rate": 1.4371984322189754e-05, + "loss": 0.266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14990919828414917, + "step": 2610, + "valid_targets_mean": 8691.2, + "valid_targets_min": 4250 + }, + { + "epoch": 4.432203389830509, + "grad_norm": 0.4301286768523898, + "learning_rate": 1.4290930650532448e-05, + "loss": 0.2693, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12087419629096985, + "step": 2615, + "valid_targets_mean": 6839.6, + "valid_targets_min": 3396 + }, + { + "epoch": 4.440677966101695, + "grad_norm": 0.4169197167794308, + "learning_rate": 1.4209978936404731e-05, + "loss": 0.2896, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12181919068098068, + "step": 2620, + "valid_targets_mean": 6649.1, + "valid_targets_min": 3252 + }, + { + "epoch": 4.4491525423728815, + "grad_norm": 0.3788917067773201, + "learning_rate": 1.4129130625512776e-05, + "loss": 0.2989, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12227948009967804, + "step": 2625, + "valid_targets_mean": 9027.2, + "valid_targets_min": 3075 + }, + { + "epoch": 4.4576271186440675, + "grad_norm": 0.3927073339165043, + "learning_rate": 1.4048387161716086e-05, + "loss": 0.2752, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13999193906784058, + "step": 2630, + "valid_targets_mean": 8721.6, + "valid_targets_min": 4154 + }, + { + "epoch": 4.466101694915254, + "grad_norm": 0.37925582901807475, + "learning_rate": 1.3967749987001717e-05, + "loss": 0.2661, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1542431116104126, + "step": 2635, + "valid_targets_mean": 8545.4, + "valid_targets_min": 2035 + }, + { + "epoch": 4.47457627118644, + "grad_norm": 0.42249464110337775, + "learning_rate": 1.3887220541458521e-05, + "loss": 0.2741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1404310166835785, + "step": 2640, + "valid_targets_mean": 6762.9, + "valid_targets_min": 1685 + }, + { + "epoch": 4.483050847457627, + "grad_norm": 0.4044514014020801, + "learning_rate": 1.380680026325143e-05, + "loss": 0.2373, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10981861501932144, + "step": 2645, + "valid_targets_mean": 7138.6, + "valid_targets_min": 1454 + }, + { + "epoch": 4.491525423728813, + "grad_norm": 0.43127865596098164, + "learning_rate": 1.3726490588595776e-05, + "loss": 0.2693, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16169732809066772, + "step": 2650, + "valid_targets_mean": 8033.4, + "valid_targets_min": 4169 + }, + { + "epoch": 4.5, + "grad_norm": 0.4126171780381458, + "learning_rate": 1.3646292951731615e-05, + "loss": 0.2637, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12301076948642731, + "step": 2655, + "valid_targets_mean": 7606.8, + "valid_targets_min": 2710 + }, + { + "epoch": 4.508474576271187, + "grad_norm": 0.44949688957992373, + "learning_rate": 1.3566208784898157e-05, + "loss": 0.2708, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10534683614969254, + "step": 2660, + "valid_targets_mean": 7020.6, + "valid_targets_min": 3889 + }, + { + "epoch": 4.516949152542373, + "grad_norm": 0.4357503321418902, + "learning_rate": 1.348623951830816e-05, + "loss": 0.2614, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12253161519765854, + "step": 2665, + "valid_targets_mean": 6581.1, + "valid_targets_min": 2321 + }, + { + "epoch": 4.52542372881356, + "grad_norm": 0.44286522859112215, + "learning_rate": 1.3406386580122389e-05, + "loss": 0.2893, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06985418498516083, + "step": 2670, + "valid_targets_mean": 4871.2, + "valid_targets_min": 2161 + }, + { + "epoch": 4.533898305084746, + "grad_norm": 0.3996632722680479, + "learning_rate": 1.332665139642412e-05, + "loss": 0.2752, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2208811491727829, + "step": 2675, + "valid_targets_mean": 10772.2, + "valid_targets_min": 3331 + }, + { + "epoch": 4.5423728813559325, + "grad_norm": 0.40122880549202283, + "learning_rate": 1.3247035391193664e-05, + "loss": 0.269, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08148348331451416, + "step": 2680, + "valid_targets_mean": 6943.9, + "valid_targets_min": 3318 + }, + { + "epoch": 4.5508474576271185, + "grad_norm": 1.4664085501189077, + "learning_rate": 1.3167539986282938e-05, + "loss": 0.2426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1282748281955719, + "step": 2685, + "valid_targets_mean": 7150.5, + "valid_targets_min": 3801 + }, + { + "epoch": 4.559322033898305, + "grad_norm": 0.41068133220034153, + "learning_rate": 1.3088166601390087e-05, + "loss": 0.2769, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13739727437496185, + "step": 2690, + "valid_targets_mean": 8443.8, + "valid_targets_min": 3915 + }, + { + "epoch": 4.567796610169491, + "grad_norm": 0.4358260643462685, + "learning_rate": 1.3008916654034085e-05, + "loss": 0.3127, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12440723925828934, + "step": 2695, + "valid_targets_mean": 6443.5, + "valid_targets_min": 2839 + }, + { + "epoch": 4.576271186440678, + "grad_norm": 0.4033459245522396, + "learning_rate": 1.2929791559529484e-05, + "loss": 0.2956, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1328117847442627, + "step": 2700, + "valid_targets_mean": 8286.0, + "valid_targets_min": 4292 + }, + { + "epoch": 4.584745762711864, + "grad_norm": 0.509860313674305, + "learning_rate": 1.2850792730961082e-05, + "loss": 0.2884, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1029135212302208, + "step": 2705, + "valid_targets_mean": 6252.9, + "valid_targets_min": 2382 + }, + { + "epoch": 4.593220338983051, + "grad_norm": 0.39399787378544626, + "learning_rate": 1.2771921579158728e-05, + "loss": 0.2869, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20253822207450867, + "step": 2710, + "valid_targets_mean": 11100.1, + "valid_targets_min": 5362 + }, + { + "epoch": 4.601694915254237, + "grad_norm": 0.43028156540452184, + "learning_rate": 1.26931795126721e-05, + "loss": 0.2668, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19609040021896362, + "step": 2715, + "valid_targets_mean": 8093.4, + "valid_targets_min": 2671 + }, + { + "epoch": 4.610169491525424, + "grad_norm": 0.4088693456437344, + "learning_rate": 1.2614567937745554e-05, + "loss": 0.276, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12928591668605804, + "step": 2720, + "valid_targets_mean": 8354.4, + "valid_targets_min": 3275 + }, + { + "epoch": 4.61864406779661, + "grad_norm": 0.4011900430152117, + "learning_rate": 1.2536088258293018e-05, + "loss": 0.2745, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12023374438285828, + "step": 2725, + "valid_targets_mean": 7815.4, + "valid_targets_min": 3321 + }, + { + "epoch": 4.627118644067797, + "grad_norm": 0.4216623663257302, + "learning_rate": 1.2457741875872928e-05, + "loss": 0.2664, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12958815693855286, + "step": 2730, + "valid_targets_mean": 8720.1, + "valid_targets_min": 3636 + }, + { + "epoch": 4.635593220338983, + "grad_norm": 0.42531144106872554, + "learning_rate": 1.2379530189663156e-05, + "loss": 0.2709, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1629747748374939, + "step": 2735, + "valid_targets_mean": 7630.5, + "valid_targets_min": 1955 + }, + { + "epoch": 4.6440677966101696, + "grad_norm": 0.39544228333495296, + "learning_rate": 1.2301454596436076e-05, + "loss": 0.2594, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15124405920505524, + "step": 2740, + "valid_targets_mean": 8636.4, + "valid_targets_min": 3610 + }, + { + "epoch": 4.652542372881356, + "grad_norm": 0.4116205020726074, + "learning_rate": 1.222351649053359e-05, + "loss": 0.2476, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08724246919155121, + "step": 2745, + "valid_targets_mean": 6308.4, + "valid_targets_min": 2704 + }, + { + "epoch": 4.661016949152542, + "grad_norm": 0.48209650657999265, + "learning_rate": 1.2145717263842228e-05, + "loss": 0.2767, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14682641625404358, + "step": 2750, + "valid_targets_mean": 10237.4, + "valid_targets_min": 4855 + }, + { + "epoch": 4.669491525423728, + "grad_norm": 0.4050539093958239, + "learning_rate": 1.2068058305768298e-05, + "loss": 0.2791, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17600831389427185, + "step": 2755, + "valid_targets_mean": 8910.4, + "valid_targets_min": 1547 + }, + { + "epoch": 4.677966101694915, + "grad_norm": 0.47359617695979656, + "learning_rate": 1.1990541003213072e-05, + "loss": 0.28, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1162843182682991, + "step": 2760, + "valid_targets_mean": 5670.5, + "valid_targets_min": 1780 + }, + { + "epoch": 4.686440677966102, + "grad_norm": 0.45098232315939285, + "learning_rate": 1.1913166740547999e-05, + "loss": 0.2634, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15938234329223633, + "step": 2765, + "valid_targets_mean": 8252.0, + "valid_targets_min": 2094 + }, + { + "epoch": 4.694915254237288, + "grad_norm": 0.42415528394204816, + "learning_rate": 1.1835936899590017e-05, + "loss": 0.2779, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1282048523426056, + "step": 2770, + "valid_targets_mean": 7349.8, + "valid_targets_min": 3722 + }, + { + "epoch": 4.703389830508475, + "grad_norm": 0.5501015060302535, + "learning_rate": 1.1758852859576842e-05, + "loss": 0.2816, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09744997322559357, + "step": 2775, + "valid_targets_mean": 6856.6, + "valid_targets_min": 4375 + }, + { + "epoch": 4.711864406779661, + "grad_norm": 0.41381090284821775, + "learning_rate": 1.1681915997142354e-05, + "loss": 0.2716, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14670927822589874, + "step": 2780, + "valid_targets_mean": 8840.9, + "valid_targets_min": 2936 + }, + { + "epoch": 4.720338983050848, + "grad_norm": 0.4122761510436273, + "learning_rate": 1.1605127686292009e-05, + "loss": 0.2706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1822691261768341, + "step": 2785, + "valid_targets_mean": 11162.5, + "valid_targets_min": 4307 + }, + { + "epoch": 4.728813559322034, + "grad_norm": 0.4002624945514117, + "learning_rate": 1.15284892983783e-05, + "loss": 0.2803, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11932104080915451, + "step": 2790, + "valid_targets_mean": 7222.4, + "valid_targets_min": 3333 + }, + { + "epoch": 4.737288135593221, + "grad_norm": 0.4230079310160459, + "learning_rate": 1.1452002202076256e-05, + "loss": 0.2907, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1321437805891037, + "step": 2795, + "valid_targets_mean": 8008.5, + "valid_targets_min": 2513 + }, + { + "epoch": 4.745762711864407, + "grad_norm": 0.42567673397305217, + "learning_rate": 1.1375667763359031e-05, + "loss": 0.2662, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1508551388978958, + "step": 2800, + "valid_targets_mean": 8618.0, + "valid_targets_min": 3894 + }, + { + "epoch": 4.754237288135593, + "grad_norm": 0.4067557863784692, + "learning_rate": 1.1299487345473457e-05, + "loss": 0.2793, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11135031282901764, + "step": 2805, + "valid_targets_mean": 7020.6, + "valid_targets_min": 3364 + }, + { + "epoch": 4.762711864406779, + "grad_norm": 0.4522576941628964, + "learning_rate": 1.1223462308915767e-05, + "loss": 0.2785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16139456629753113, + "step": 2810, + "valid_targets_mean": 7060.9, + "valid_targets_min": 2693 + }, + { + "epoch": 4.771186440677966, + "grad_norm": 0.5393715006124433, + "learning_rate": 1.1147594011407203e-05, + "loss": 0.2935, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1524142324924469, + "step": 2815, + "valid_targets_mean": 9858.4, + "valid_targets_min": 4735 + }, + { + "epoch": 4.779661016949152, + "grad_norm": 0.45800608337889726, + "learning_rate": 1.1071883807869886e-05, + "loss": 0.2785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17067819833755493, + "step": 2820, + "valid_targets_mean": 7938.4, + "valid_targets_min": 3822 + }, + { + "epoch": 4.788135593220339, + "grad_norm": 0.4408014445770491, + "learning_rate": 1.0996333050402503e-05, + "loss": 0.291, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11198888719081879, + "step": 2825, + "valid_targets_mean": 6231.6, + "valid_targets_min": 3402 + }, + { + "epoch": 4.796610169491525, + "grad_norm": 0.37288081152923613, + "learning_rate": 1.0920943088256266e-05, + "loss": 0.2941, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1768973469734192, + "step": 2830, + "valid_targets_mean": 11278.0, + "valid_targets_min": 3670 + }, + { + "epoch": 4.805084745762712, + "grad_norm": 0.4675055974588808, + "learning_rate": 1.0845715267810716e-05, + "loss": 0.2702, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1645127385854721, + "step": 2835, + "valid_targets_mean": 6527.4, + "valid_targets_min": 1805 + }, + { + "epoch": 4.813559322033898, + "grad_norm": 0.4249196921645124, + "learning_rate": 1.0770650932549768e-05, + "loss": 0.2785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16321763396263123, + "step": 2840, + "valid_targets_mean": 9060.0, + "valid_targets_min": 3082 + }, + { + "epoch": 4.822033898305085, + "grad_norm": 0.3797637838348183, + "learning_rate": 1.0695751423037639e-05, + "loss": 0.2612, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14523139595985413, + "step": 2845, + "valid_targets_mean": 9407.8, + "valid_targets_min": 5593 + }, + { + "epoch": 4.830508474576272, + "grad_norm": 0.3954316084644604, + "learning_rate": 1.0621018076894984e-05, + "loss": 0.271, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19016766548156738, + "step": 2850, + "valid_targets_mean": 9626.1, + "valid_targets_min": 2303 + }, + { + "epoch": 4.838983050847458, + "grad_norm": 0.40324943006421254, + "learning_rate": 1.0546452228774919e-05, + "loss": 0.2955, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20220312476158142, + "step": 2855, + "valid_targets_mean": 9551.8, + "valid_targets_min": 2084 + }, + { + "epoch": 4.847457627118644, + "grad_norm": 0.48347751893698887, + "learning_rate": 1.047205521033928e-05, + "loss": 0.2676, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0758141279220581, + "step": 2860, + "valid_targets_mean": 5134.4, + "valid_targets_min": 1779 + }, + { + "epoch": 4.8559322033898304, + "grad_norm": 0.436026077540386, + "learning_rate": 1.0397828350234765e-05, + "loss": 0.2733, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07567907124757767, + "step": 2865, + "valid_targets_mean": 5584.2, + "valid_targets_min": 3528 + }, + { + "epoch": 4.864406779661017, + "grad_norm": 0.4049340454560818, + "learning_rate": 1.032377297406926e-05, + "loss": 0.2686, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15704092383384705, + "step": 2870, + "valid_targets_mean": 7979.8, + "valid_targets_min": 3866 + }, + { + "epoch": 4.872881355932203, + "grad_norm": 0.4273483414490487, + "learning_rate": 1.0249890404388121e-05, + "loss": 0.2786, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13683602213859558, + "step": 2875, + "valid_targets_mean": 7722.0, + "valid_targets_min": 2714 + }, + { + "epoch": 4.88135593220339, + "grad_norm": 0.45261744678626115, + "learning_rate": 1.0176181960650595e-05, + "loss": 0.2617, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20523563027381897, + "step": 2880, + "valid_targets_mean": 10396.2, + "valid_targets_min": 4946 + }, + { + "epoch": 4.889830508474576, + "grad_norm": 0.4123299125145364, + "learning_rate": 1.0102648959206212e-05, + "loss": 0.2485, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09504781663417816, + "step": 2885, + "valid_targets_mean": 6910.4, + "valid_targets_min": 1489 + }, + { + "epoch": 4.898305084745763, + "grad_norm": 0.41252518581842956, + "learning_rate": 1.0029292713271334e-05, + "loss": 0.2661, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1620301753282547, + "step": 2890, + "valid_targets_mean": 7912.0, + "valid_targets_min": 3266 + }, + { + "epoch": 4.906779661016949, + "grad_norm": 0.4165804006140129, + "learning_rate": 9.956114532905625e-06, + "loss": 0.3071, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12451449036598206, + "step": 2895, + "valid_targets_mean": 7785.5, + "valid_targets_min": 4724 + }, + { + "epoch": 4.915254237288136, + "grad_norm": 0.451277868879136, + "learning_rate": 9.883115724988743e-06, + "loss": 0.2764, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12704214453697205, + "step": 2900, + "valid_targets_mean": 7152.0, + "valid_targets_min": 2405 + }, + { + "epoch": 4.923728813559322, + "grad_norm": 0.43488588524419447, + "learning_rate": 9.810297593196922e-06, + "loss": 0.2858, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11738678067922592, + "step": 2905, + "valid_targets_mean": 6999.5, + "valid_targets_min": 1506 + }, + { + "epoch": 4.932203389830509, + "grad_norm": 0.3902190977630454, + "learning_rate": 9.73766143797976e-06, + "loss": 0.2991, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18499219417572021, + "step": 2910, + "valid_targets_mean": 9337.6, + "valid_targets_min": 3041 + }, + { + "epoch": 4.940677966101695, + "grad_norm": 0.39113577433521207, + "learning_rate": 9.665208556536918e-06, + "loss": 0.2572, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1580629199743271, + "step": 2915, + "valid_targets_mean": 10305.0, + "valid_targets_min": 2760 + }, + { + "epoch": 4.9491525423728815, + "grad_norm": 0.4467637307215622, + "learning_rate": 9.592940242795035e-06, + "loss": 0.2705, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12799829244613647, + "step": 2920, + "valid_targets_mean": 7904.2, + "valid_targets_min": 4206 + }, + { + "epoch": 4.9576271186440675, + "grad_norm": 0.4012050592034206, + "learning_rate": 9.520857787384548e-06, + "loss": 0.2931, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.22492660582065582, + "step": 2925, + "valid_targets_mean": 10071.6, + "valid_targets_min": 4056 + }, + { + "epoch": 4.966101694915254, + "grad_norm": 0.41909262364226657, + "learning_rate": 9.44896247761669e-06, + "loss": 0.288, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11922876536846161, + "step": 2930, + "valid_targets_mean": 7519.2, + "valid_targets_min": 2369 + }, + { + "epoch": 4.97457627118644, + "grad_norm": 0.4550620897888851, + "learning_rate": 9.377255597460469e-06, + "loss": 0.2711, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14701971411705017, + "step": 2935, + "valid_targets_mean": 6977.8, + "valid_targets_min": 2432 + }, + { + "epoch": 4.983050847457627, + "grad_norm": 0.4859597855399959, + "learning_rate": 9.305738427519782e-06, + "loss": 0.2963, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10581295937299728, + "step": 2940, + "valid_targets_mean": 5638.9, + "valid_targets_min": 2058 + }, + { + "epoch": 4.991525423728813, + "grad_norm": 0.39353155286685454, + "learning_rate": 9.234412245010482e-06, + "loss": 0.2432, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16748303174972534, + "step": 2945, + "valid_targets_mean": 9775.4, + "valid_targets_min": 2320 + }, + { + "epoch": 5.0, + "grad_norm": 0.4039798684599604, + "learning_rate": 9.163278323737635e-06, + "loss": 0.2877, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1804552525281906, + "step": 2950, + "valid_targets_mean": 9501.8, + "valid_targets_min": 3224 + }, + { + "epoch": 5.008474576271187, + "grad_norm": 0.38321872526038875, + "learning_rate": 9.092337934072703e-06, + "loss": 0.2718, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1229909211397171, + "step": 2955, + "valid_targets_mean": 8311.6, + "valid_targets_min": 3501 + }, + { + "epoch": 5.016949152542373, + "grad_norm": 0.3910600282604654, + "learning_rate": 9.021592342930928e-06, + "loss": 0.2325, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10238224267959595, + "step": 2960, + "valid_targets_mean": 8050.2, + "valid_targets_min": 4143 + }, + { + "epoch": 5.02542372881356, + "grad_norm": 0.4128668247607092, + "learning_rate": 8.951042813748645e-06, + "loss": 0.2686, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1838410198688507, + "step": 2965, + "valid_targets_mean": 10353.2, + "valid_targets_min": 5165 + }, + { + "epoch": 5.033898305084746, + "grad_norm": 0.42280176381204043, + "learning_rate": 8.880690606460754e-06, + "loss": 0.2938, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1057187020778656, + "step": 2970, + "valid_targets_mean": 6608.2, + "valid_targets_min": 2466 + }, + { + "epoch": 5.0423728813559325, + "grad_norm": 0.41604038060167553, + "learning_rate": 8.8105369774782e-06, + "loss": 0.2736, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15914693474769592, + "step": 2975, + "valid_targets_mean": 8547.6, + "valid_targets_min": 2961 + }, + { + "epoch": 5.0508474576271185, + "grad_norm": 0.4448543436434134, + "learning_rate": 8.74058317966556e-06, + "loss": 0.2502, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0849103257060051, + "step": 2980, + "valid_targets_mean": 5587.1, + "valid_targets_min": 3852 + }, + { + "epoch": 5.059322033898305, + "grad_norm": 0.4520071230103935, + "learning_rate": 8.670830462318633e-06, + "loss": 0.2536, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06706386059522629, + "step": 2985, + "valid_targets_mean": 5294.6, + "valid_targets_min": 2342 + }, + { + "epoch": 5.067796610169491, + "grad_norm": 0.4253115930316411, + "learning_rate": 8.60128007114217e-06, + "loss": 0.2721, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14023102819919586, + "step": 2990, + "valid_targets_mean": 8529.0, + "valid_targets_min": 1747 + }, + { + "epoch": 5.076271186440678, + "grad_norm": 0.40617917977385865, + "learning_rate": 8.531933248227582e-06, + "loss": 0.2402, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10234525054693222, + "step": 2995, + "valid_targets_mean": 5504.6, + "valid_targets_min": 2211 + }, + { + "epoch": 5.084745762711864, + "grad_norm": 0.43359659053869215, + "learning_rate": 8.462791232030803e-06, + "loss": 0.2775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13905169069766998, + "step": 3000, + "valid_targets_mean": 8243.2, + "valid_targets_min": 4344 + }, + { + "epoch": 5.093220338983051, + "grad_norm": 0.40896650404071416, + "learning_rate": 8.393855257350132e-06, + "loss": 0.253, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1154506728053093, + "step": 3005, + "valid_targets_mean": 8325.2, + "valid_targets_min": 4186 + }, + { + "epoch": 5.101694915254237, + "grad_norm": 0.3946324268905825, + "learning_rate": 8.325126555304208e-06, + "loss": 0.2628, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17978814244270325, + "step": 3010, + "valid_targets_mean": 9363.6, + "valid_targets_min": 4084 + }, + { + "epoch": 5.110169491525424, + "grad_norm": 0.40443769374892974, + "learning_rate": 8.256606353310003e-06, + "loss": 0.2487, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14391589164733887, + "step": 3015, + "valid_targets_mean": 9450.2, + "valid_targets_min": 3354 + }, + { + "epoch": 5.11864406779661, + "grad_norm": 0.4146020029739842, + "learning_rate": 8.188295875060943e-06, + "loss": 0.2767, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07763639092445374, + "step": 3020, + "valid_targets_mean": 4914.9, + "valid_targets_min": 3264 + }, + { + "epoch": 5.127118644067797, + "grad_norm": 0.482213604261791, + "learning_rate": 8.12019634050499e-06, + "loss": 0.2284, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08091288059949875, + "step": 3025, + "valid_targets_mean": 7860.5, + "valid_targets_min": 4646 + }, + { + "epoch": 5.135593220338983, + "grad_norm": 0.434415926653262, + "learning_rate": 8.052308965822916e-06, + "loss": 0.2484, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13998384773731232, + "step": 3030, + "valid_targets_mean": 7786.8, + "valid_targets_min": 4468 + }, + { + "epoch": 5.1440677966101696, + "grad_norm": 0.4171807235049148, + "learning_rate": 7.984634963406537e-06, + "loss": 0.2513, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14208590984344482, + "step": 3035, + "valid_targets_mean": 8149.5, + "valid_targets_min": 4512 + }, + { + "epoch": 5.1525423728813555, + "grad_norm": 0.4459854996850316, + "learning_rate": 7.917175541837101e-06, + "loss": 0.2862, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09813421964645386, + "step": 3040, + "valid_targets_mean": 5958.5, + "valid_targets_min": 2438 + }, + { + "epoch": 5.161016949152542, + "grad_norm": 0.5331008651404957, + "learning_rate": 7.849931905863672e-06, + "loss": 0.29, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1268807053565979, + "step": 3045, + "valid_targets_mean": 6381.6, + "valid_targets_min": 3763 + }, + { + "epoch": 5.169491525423728, + "grad_norm": 0.41566933713800536, + "learning_rate": 7.78290525638163e-06, + "loss": 0.259, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18568013608455658, + "step": 3050, + "valid_targets_mean": 10965.6, + "valid_targets_min": 3702 + }, + { + "epoch": 5.177966101694915, + "grad_norm": 0.4537423909342909, + "learning_rate": 7.71609679041121e-06, + "loss": 0.252, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1523778736591339, + "step": 3055, + "valid_targets_mean": 7001.8, + "valid_targets_min": 3472 + }, + { + "epoch": 5.186440677966102, + "grad_norm": 0.39972642247618045, + "learning_rate": 7.649507701076164e-06, + "loss": 0.264, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1793755441904068, + "step": 3060, + "valid_targets_mean": 10015.9, + "valid_targets_min": 2070 + }, + { + "epoch": 5.194915254237288, + "grad_norm": 0.5131893981021648, + "learning_rate": 7.583139177582395e-06, + "loss": 0.2651, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20769305527210236, + "step": 3065, + "valid_targets_mean": 7827.6, + "valid_targets_min": 3091 + }, + { + "epoch": 5.203389830508475, + "grad_norm": 0.4237772198595128, + "learning_rate": 7.516992405196772e-06, + "loss": 0.2671, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13641873002052307, + "step": 3070, + "valid_targets_mean": 8499.1, + "valid_targets_min": 4822 + }, + { + "epoch": 5.211864406779661, + "grad_norm": 0.43454029280053935, + "learning_rate": 7.45106856522593e-06, + "loss": 0.2698, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10716796666383743, + "step": 3075, + "valid_targets_mean": 5610.5, + "valid_targets_min": 2809 + }, + { + "epoch": 5.220338983050848, + "grad_norm": 0.542776623150691, + "learning_rate": 7.385368834995168e-06, + "loss": 0.2312, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07319167256355286, + "step": 3080, + "valid_targets_mean": 5343.0, + "valid_targets_min": 2880 + }, + { + "epoch": 5.228813559322034, + "grad_norm": 0.4405988491699508, + "learning_rate": 7.319894387827473e-06, + "loss": 0.2531, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12842324376106262, + "step": 3085, + "valid_targets_mean": 7002.0, + "valid_targets_min": 3852 + }, + { + "epoch": 5.237288135593221, + "grad_norm": 0.43538162775303374, + "learning_rate": 7.254646393022502e-06, + "loss": 0.2826, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14618468284606934, + "step": 3090, + "valid_targets_mean": 8944.5, + "valid_targets_min": 2222 + }, + { + "epoch": 5.245762711864407, + "grad_norm": 0.3967557229503152, + "learning_rate": 7.189626015835733e-06, + "loss": 0.2851, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.146665558218956, + "step": 3095, + "valid_targets_mean": 9732.2, + "valid_targets_min": 2484 + }, + { + "epoch": 5.254237288135593, + "grad_norm": 0.43473594078288474, + "learning_rate": 7.124834417457664e-06, + "loss": 0.2511, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1037987545132637, + "step": 3100, + "valid_targets_mean": 6524.2, + "valid_targets_min": 2612 + }, + { + "epoch": 5.262711864406779, + "grad_norm": 0.41536450764949134, + "learning_rate": 7.060272754993051e-06, + "loss": 0.2538, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10834446549415588, + "step": 3105, + "valid_targets_mean": 7642.6, + "valid_targets_min": 3845 + }, + { + "epoch": 5.271186440677966, + "grad_norm": 0.4473707722528023, + "learning_rate": 6.9959421814402494e-06, + "loss": 0.2703, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09653830528259277, + "step": 3110, + "valid_targets_mean": 6060.6, + "valid_targets_min": 2940 + }, + { + "epoch": 5.279661016949152, + "grad_norm": 0.3828977830174924, + "learning_rate": 6.931843845670647e-06, + "loss": 0.2802, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18717047572135925, + "step": 3115, + "valid_targets_mean": 11332.6, + "valid_targets_min": 6662 + }, + { + "epoch": 5.288135593220339, + "grad_norm": 0.4156040861199304, + "learning_rate": 6.867978892408101e-06, + "loss": 0.2827, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06572428345680237, + "step": 3120, + "valid_targets_mean": 5527.4, + "valid_targets_min": 1893 + }, + { + "epoch": 5.296610169491525, + "grad_norm": 0.3849476640530131, + "learning_rate": 6.804348462208548e-06, + "loss": 0.2517, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14777298271656036, + "step": 3125, + "valid_targets_mean": 10066.9, + "valid_targets_min": 4775 + }, + { + "epoch": 5.305084745762712, + "grad_norm": 0.39696946143084255, + "learning_rate": 6.7409536914395866e-06, + "loss": 0.2565, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12509366869926453, + "step": 3130, + "valid_targets_mean": 8591.4, + "valid_targets_min": 6400 + }, + { + "epoch": 5.313559322033898, + "grad_norm": 0.4408601497824042, + "learning_rate": 6.677795712260206e-06, + "loss": 0.2706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10040117055177689, + "step": 3135, + "valid_targets_mean": 6520.1, + "valid_targets_min": 3393 + }, + { + "epoch": 5.322033898305085, + "grad_norm": 0.44317433702470693, + "learning_rate": 6.6148756526005855e-06, + "loss": 0.2646, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17337459325790405, + "step": 3140, + "valid_targets_mean": 7650.0, + "valid_targets_min": 2685 + }, + { + "epoch": 5.330508474576272, + "grad_norm": 0.40286799690359776, + "learning_rate": 6.552194636141909e-06, + "loss": 0.2458, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15087980031967163, + "step": 3145, + "valid_targets_mean": 8771.0, + "valid_targets_min": 3597 + }, + { + "epoch": 5.338983050847458, + "grad_norm": 0.4909543331964684, + "learning_rate": 6.489753782296315e-06, + "loss": 0.2616, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1323254108428955, + "step": 3150, + "valid_targets_mean": 6834.9, + "valid_targets_min": 1998 + }, + { + "epoch": 5.3474576271186445, + "grad_norm": 0.40375931958252953, + "learning_rate": 6.427554206186939e-06, + "loss": 0.2611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12938186526298523, + "step": 3155, + "valid_targets_mean": 8728.6, + "valid_targets_min": 3046 + }, + { + "epoch": 5.3559322033898304, + "grad_norm": 1.0021730618277505, + "learning_rate": 6.3655970186279314e-06, + "loss": 0.2804, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.155011385679245, + "step": 3160, + "valid_targets_mean": 7922.6, + "valid_targets_min": 4455 + }, + { + "epoch": 5.364406779661017, + "grad_norm": 0.4921109248318851, + "learning_rate": 6.30388332610469e-06, + "loss": 0.2494, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09664289653301239, + "step": 3165, + "valid_targets_mean": 6155.6, + "valid_targets_min": 2938 + }, + { + "epoch": 5.372881355932203, + "grad_norm": 0.43530974592088145, + "learning_rate": 6.242414230754044e-06, + "loss": 0.2668, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1272495836019516, + "step": 3170, + "valid_targets_mean": 7464.1, + "valid_targets_min": 2619 + }, + { + "epoch": 5.38135593220339, + "grad_norm": 0.4578893158386509, + "learning_rate": 6.181190830344601e-06, + "loss": 0.2883, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1435915231704712, + "step": 3175, + "valid_targets_mean": 6372.2, + "valid_targets_min": 3271 + }, + { + "epoch": 5.389830508474576, + "grad_norm": 0.4462141692455487, + "learning_rate": 6.120214218257128e-06, + "loss": 0.2572, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21038474142551422, + "step": 3180, + "valid_targets_mean": 9814.6, + "valid_targets_min": 4536 + }, + { + "epoch": 5.398305084745763, + "grad_norm": 0.4299686838750093, + "learning_rate": 6.059485483465048e-06, + "loss": 0.2501, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12408433854579926, + "step": 3185, + "valid_targets_mean": 7589.8, + "valid_targets_min": 3704 + }, + { + "epoch": 5.406779661016949, + "grad_norm": 0.47292848955183747, + "learning_rate": 5.999005710514956e-06, + "loss": 0.2464, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0928158164024353, + "step": 3190, + "valid_targets_mean": 5277.9, + "valid_targets_min": 1623 + }, + { + "epoch": 5.415254237288136, + "grad_norm": 0.45861849447302966, + "learning_rate": 5.938775979507287e-06, + "loss": 0.2645, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1349688470363617, + "step": 3195, + "valid_targets_mean": 7671.6, + "valid_targets_min": 2887 + }, + { + "epoch": 5.423728813559322, + "grad_norm": 0.4420789679504025, + "learning_rate": 5.878797366076994e-06, + "loss": 0.2437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1322856843471527, + "step": 3200, + "valid_targets_mean": 6415.5, + "valid_targets_min": 2320 + }, + { + "epoch": 5.432203389830509, + "grad_norm": 0.4391713104757039, + "learning_rate": 5.819070941374368e-06, + "loss": 0.2493, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1420782208442688, + "step": 3205, + "valid_targets_mean": 8150.8, + "valid_targets_min": 3889 + }, + { + "epoch": 5.440677966101695, + "grad_norm": 0.4251743635191406, + "learning_rate": 5.759597772045882e-06, + "loss": 0.2822, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18934570252895355, + "step": 3210, + "valid_targets_mean": 9116.6, + "valid_targets_min": 5251 + }, + { + "epoch": 5.4491525423728815, + "grad_norm": 0.4830379009128657, + "learning_rate": 5.700378920215159e-06, + "loss": 0.2749, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.21354344487190247, + "step": 3215, + "valid_targets_mean": 7386.4, + "valid_targets_min": 2457 + }, + { + "epoch": 5.4576271186440675, + "grad_norm": 0.43378238596603985, + "learning_rate": 5.641415443463994e-06, + "loss": 0.263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13009977340698242, + "step": 3220, + "valid_targets_mean": 7686.8, + "valid_targets_min": 1572 + }, + { + "epoch": 5.466101694915254, + "grad_norm": 0.4487432362569402, + "learning_rate": 5.5827083948134876e-06, + "loss": 0.2835, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11775340139865875, + "step": 3225, + "valid_targets_mean": 6541.5, + "valid_targets_min": 1230 + }, + { + "epoch": 5.47457627118644, + "grad_norm": 0.4417168658501371, + "learning_rate": 5.524258822705202e-06, + "loss": 0.231, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12141908705234528, + "step": 3230, + "valid_targets_mean": 7137.5, + "valid_targets_min": 4420 + }, + { + "epoch": 5.483050847457627, + "grad_norm": 0.4011977565905363, + "learning_rate": 5.466067770982482e-06, + "loss": 0.2954, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10719483345746994, + "step": 3235, + "valid_targets_mean": 6831.9, + "valid_targets_min": 2783 + }, + { + "epoch": 5.491525423728813, + "grad_norm": 0.4703209208499074, + "learning_rate": 5.4081362788717625e-06, + "loss": 0.2926, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09573965519666672, + "step": 3240, + "valid_targets_mean": 5653.0, + "valid_targets_min": 2811 + }, + { + "epoch": 5.5, + "grad_norm": 0.4100676442731336, + "learning_rate": 5.350465380964076e-06, + "loss": 0.2708, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1281810849905014, + "step": 3245, + "valid_targets_mean": 8242.9, + "valid_targets_min": 3222 + }, + { + "epoch": 5.508474576271187, + "grad_norm": 0.4017599475787098, + "learning_rate": 5.293056107196488e-06, + "loss": 0.2832, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10683394968509674, + "step": 3250, + "valid_targets_mean": 9198.4, + "valid_targets_min": 3032 + }, + { + "epoch": 5.516949152542373, + "grad_norm": 0.42952570667674034, + "learning_rate": 5.235909482833797e-06, + "loss": 0.2734, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07673219591379166, + "step": 3255, + "valid_targets_mean": 5805.6, + "valid_targets_min": 3517 + }, + { + "epoch": 5.52542372881356, + "grad_norm": 0.42578258546227, + "learning_rate": 5.179026528450146e-06, + "loss": 0.2656, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10920145362615585, + "step": 3260, + "valid_targets_mean": 6678.8, + "valid_targets_min": 3108 + }, + { + "epoch": 5.533898305084746, + "grad_norm": 0.4070334818927427, + "learning_rate": 5.1224082599108584e-06, + "loss": 0.2453, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1201343908905983, + "step": 3265, + "valid_targets_mean": 7852.5, + "valid_targets_min": 4216 + }, + { + "epoch": 5.5423728813559325, + "grad_norm": 0.42784838946022163, + "learning_rate": 5.066055688354246e-06, + "loss": 0.2778, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13358819484710693, + "step": 3270, + "valid_targets_mean": 7367.1, + "valid_targets_min": 3831 + }, + { + "epoch": 5.5508474576271185, + "grad_norm": 0.4172219694937724, + "learning_rate": 5.0099698201735855e-06, + "loss": 0.2627, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18104565143585205, + "step": 3275, + "valid_targets_mean": 9405.1, + "valid_targets_min": 2403 + }, + { + "epoch": 5.559322033898305, + "grad_norm": 0.4340880296880951, + "learning_rate": 4.954151656999122e-06, + "loss": 0.2614, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19006019830703735, + "step": 3280, + "valid_targets_mean": 8843.9, + "valid_targets_min": 2489 + }, + { + "epoch": 5.567796610169491, + "grad_norm": 0.4114568126592623, + "learning_rate": 4.898602195680214e-06, + "loss": 0.2586, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17298346757888794, + "step": 3285, + "valid_targets_mean": 10129.8, + "valid_targets_min": 2553 + }, + { + "epoch": 5.576271186440678, + "grad_norm": 0.4548674456421557, + "learning_rate": 4.843322428267465e-06, + "loss": 0.2595, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12810729444026947, + "step": 3290, + "valid_targets_mean": 7174.0, + "valid_targets_min": 2127 + }, + { + "epoch": 5.584745762711864, + "grad_norm": 0.4210044479771755, + "learning_rate": 4.788313341995096e-06, + "loss": 0.3024, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12651950120925903, + "step": 3295, + "valid_targets_mean": 7814.5, + "valid_targets_min": 3445 + }, + { + "epoch": 5.593220338983051, + "grad_norm": 0.4327273994886145, + "learning_rate": 4.733575919263238e-06, + "loss": 0.2738, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19881585240364075, + "step": 3300, + "valid_targets_mean": 9607.2, + "valid_targets_min": 2512 + }, + { + "epoch": 5.601694915254237, + "grad_norm": 0.6147338864346399, + "learning_rate": 4.679111137620442e-06, + "loss": 0.2778, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09989364445209503, + "step": 3305, + "valid_targets_mean": 7179.1, + "valid_targets_min": 2804 + }, + { + "epoch": 5.610169491525424, + "grad_norm": 0.47192581181781135, + "learning_rate": 4.624919969746171e-06, + "loss": 0.2771, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14433559775352478, + "step": 3310, + "valid_targets_mean": 7068.8, + "valid_targets_min": 3230 + }, + { + "epoch": 5.61864406779661, + "grad_norm": 0.444945216633516, + "learning_rate": 4.571003383433481e-06, + "loss": 0.2915, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15350055694580078, + "step": 3315, + "valid_targets_mean": 7867.9, + "valid_targets_min": 2732 + }, + { + "epoch": 5.627118644067797, + "grad_norm": 0.45738565401127745, + "learning_rate": 4.517362341571687e-06, + "loss": 0.2775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13060398399829865, + "step": 3320, + "valid_targets_mean": 7557.6, + "valid_targets_min": 2344 + }, + { + "epoch": 5.635593220338983, + "grad_norm": 0.43571689131803126, + "learning_rate": 4.463997802129221e-06, + "loss": 0.3019, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1402197778224945, + "step": 3325, + "valid_targets_mean": 8348.2, + "valid_targets_min": 4427 + }, + { + "epoch": 5.6440677966101696, + "grad_norm": 0.4055925761510759, + "learning_rate": 4.410910718136454e-06, + "loss": 0.2666, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15211747586727142, + "step": 3330, + "valid_targets_mean": 8970.2, + "valid_targets_min": 2698 + }, + { + "epoch": 5.652542372881356, + "grad_norm": 0.43189781407439476, + "learning_rate": 4.3581020376687566e-06, + "loss": 0.2634, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10128441452980042, + "step": 3335, + "valid_targets_mean": 7696.0, + "valid_targets_min": 3473 + }, + { + "epoch": 5.661016949152542, + "grad_norm": 0.4087620898274565, + "learning_rate": 4.305572703829495e-06, + "loss": 0.2525, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12019095569849014, + "step": 3340, + "valid_targets_mean": 8314.0, + "valid_targets_min": 4075 + }, + { + "epoch": 5.669491525423728, + "grad_norm": 0.4466182954085041, + "learning_rate": 4.253323654733248e-06, + "loss": 0.268, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10398929566144943, + "step": 3345, + "valid_targets_mean": 6251.1, + "valid_targets_min": 2741 + }, + { + "epoch": 5.677966101694915, + "grad_norm": 0.4102492357885536, + "learning_rate": 4.201355823488999e-06, + "loss": 0.2693, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14839188754558563, + "step": 3350, + "valid_targets_mean": 7566.4, + "valid_targets_min": 2820 + }, + { + "epoch": 5.686440677966102, + "grad_norm": 0.429597461830153, + "learning_rate": 4.149670138183526e-06, + "loss": 0.252, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0880933552980423, + "step": 3355, + "valid_targets_mean": 6382.2, + "valid_targets_min": 1402 + }, + { + "epoch": 5.694915254237288, + "grad_norm": 0.4527656431537339, + "learning_rate": 4.098267521864772e-06, + "loss": 0.2404, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18761232495307922, + "step": 3360, + "valid_targets_mean": 9463.6, + "valid_targets_min": 2262 + }, + { + "epoch": 5.703389830508475, + "grad_norm": 0.4736146006705307, + "learning_rate": 4.047148892525403e-06, + "loss": 0.2303, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13051925599575043, + "step": 3365, + "valid_targets_mean": 6238.5, + "valid_targets_min": 2453 + }, + { + "epoch": 5.711864406779661, + "grad_norm": 0.4308541611459727, + "learning_rate": 3.996315163086391e-06, + "loss": 0.2597, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1547306478023529, + "step": 3370, + "valid_targets_mean": 8064.6, + "valid_targets_min": 1780 + }, + { + "epoch": 5.720338983050848, + "grad_norm": 0.4535807208619825, + "learning_rate": 3.945767241380732e-06, + "loss": 0.2911, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18569964170455933, + "step": 3375, + "valid_targets_mean": 7775.9, + "valid_targets_min": 1638 + }, + { + "epoch": 5.728813559322034, + "grad_norm": 0.42463019132135493, + "learning_rate": 3.895506030137195e-06, + "loss": 0.2489, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12168467789888382, + "step": 3380, + "valid_targets_mean": 7172.5, + "valid_targets_min": 3905 + }, + { + "epoch": 5.737288135593221, + "grad_norm": 0.4268898814915035, + "learning_rate": 3.845532426964252e-06, + "loss": 0.2523, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10343629121780396, + "step": 3385, + "valid_targets_mean": 6808.6, + "valid_targets_min": 3001 + }, + { + "epoch": 5.745762711864407, + "grad_norm": 0.43399202163312156, + "learning_rate": 3.79584732433399e-06, + "loss": 0.2838, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1263071745634079, + "step": 3390, + "valid_targets_mean": 7826.1, + "valid_targets_min": 2558 + }, + { + "epoch": 5.754237288135593, + "grad_norm": 0.46299702282573624, + "learning_rate": 3.746451609566233e-06, + "loss": 0.2682, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.24947628378868103, + "step": 3395, + "valid_targets_mean": 9686.5, + "valid_targets_min": 5352 + }, + { + "epoch": 5.762711864406779, + "grad_norm": 0.4300572543032316, + "learning_rate": 3.697346164812643e-06, + "loss": 0.2436, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11409101635217667, + "step": 3400, + "valid_targets_mean": 7340.2, + "valid_targets_min": 3685 + }, + { + "epoch": 5.771186440677966, + "grad_norm": 0.43095238086040094, + "learning_rate": 3.6485318670409896e-06, + "loss": 0.263, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10468216240406036, + "step": 3405, + "valid_targets_mean": 6811.2, + "valid_targets_min": 2561 + }, + { + "epoch": 5.779661016949152, + "grad_norm": 0.4350339633592442, + "learning_rate": 3.6000095880194905e-06, + "loss": 0.2997, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11025195568799973, + "step": 3410, + "valid_targets_mean": 7529.4, + "valid_targets_min": 4237 + }, + { + "epoch": 5.788135593220339, + "grad_norm": 0.43082785906321686, + "learning_rate": 3.5517801943012443e-06, + "loss": 0.2655, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19818180799484253, + "step": 3415, + "valid_targets_mean": 9537.0, + "valid_targets_min": 4451 + }, + { + "epoch": 5.796610169491525, + "grad_norm": 0.4430112945153063, + "learning_rate": 3.5038445472087324e-06, + "loss": 0.2769, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13184988498687744, + "step": 3420, + "valid_targets_mean": 7243.6, + "valid_targets_min": 2380 + }, + { + "epoch": 5.805084745762712, + "grad_norm": 0.42639499607341547, + "learning_rate": 3.456203502818476e-06, + "loss": 0.2603, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12061448395252228, + "step": 3425, + "valid_targets_mean": 7819.5, + "valid_targets_min": 3869 + }, + { + "epoch": 5.813559322033898, + "grad_norm": 0.4498140348859522, + "learning_rate": 3.408857911945702e-06, + "loss": 0.292, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1824401319026947, + "step": 3430, + "valid_targets_mean": 9697.2, + "valid_targets_min": 2306 + }, + { + "epoch": 5.822033898305085, + "grad_norm": 0.4234852685317045, + "learning_rate": 3.36180862012919e-06, + "loss": 0.2665, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20986056327819824, + "step": 3435, + "valid_targets_mean": 11106.4, + "valid_targets_min": 3275 + }, + { + "epoch": 5.830508474576272, + "grad_norm": 0.4763529248983901, + "learning_rate": 3.3150564676161402e-06, + "loss": 0.2371, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10942620038986206, + "step": 3440, + "valid_targets_mean": 5915.1, + "valid_targets_min": 3315 + }, + { + "epoch": 5.838983050847458, + "grad_norm": 0.4110335170013424, + "learning_rate": 3.268602289347185e-06, + "loss": 0.2585, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19674979150295258, + "step": 3445, + "valid_targets_mean": 11239.5, + "valid_targets_min": 7700 + }, + { + "epoch": 5.847457627118644, + "grad_norm": 0.42926140283179787, + "learning_rate": 3.222446914941468e-06, + "loss": 0.2765, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17311033606529236, + "step": 3450, + "valid_targets_mean": 8367.1, + "valid_targets_min": 3209 + }, + { + "epoch": 5.8559322033898304, + "grad_norm": 0.4417294457886267, + "learning_rate": 3.176591168681851e-06, + "loss": 0.2596, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1375441551208496, + "step": 3455, + "valid_targets_mean": 7012.8, + "valid_targets_min": 2454 + }, + { + "epoch": 5.864406779661017, + "grad_norm": 0.4797537621478515, + "learning_rate": 3.131035869500152e-06, + "loss": 0.2477, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09695297479629517, + "step": 3460, + "valid_targets_mean": 5561.2, + "valid_targets_min": 3636 + }, + { + "epoch": 5.872881355932203, + "grad_norm": 0.4313456632664968, + "learning_rate": 3.085781830962564e-06, + "loss": 0.2613, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1107541099190712, + "step": 3465, + "valid_targets_mean": 6444.1, + "valid_targets_min": 2763 + }, + { + "epoch": 5.88135593220339, + "grad_norm": 0.4690289929958173, + "learning_rate": 3.0408298612550878e-06, + "loss": 0.2359, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17820698022842407, + "step": 3470, + "valid_targets_mean": 8475.8, + "valid_targets_min": 2853 + }, + { + "epoch": 5.889830508474576, + "grad_norm": 0.44581363318990813, + "learning_rate": 2.996180763169132e-06, + "loss": 0.2515, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0970480740070343, + "step": 3475, + "valid_targets_mean": 5847.0, + "valid_targets_min": 1685 + }, + { + "epoch": 5.898305084745763, + "grad_norm": 0.4517525975721219, + "learning_rate": 2.95183533408715e-06, + "loss": 0.2575, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09896974265575409, + "step": 3480, + "valid_targets_mean": 6291.0, + "valid_targets_min": 4223 + }, + { + "epoch": 5.906779661016949, + "grad_norm": 0.4231547916145084, + "learning_rate": 2.907794365968413e-06, + "loss": 0.2625, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1453133374452591, + "step": 3485, + "valid_targets_mean": 7656.8, + "valid_targets_min": 4438 + }, + { + "epoch": 5.915254237288136, + "grad_norm": 0.3783064656836768, + "learning_rate": 2.864058645334853e-06, + "loss": 0.2456, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1604553908109665, + "step": 3490, + "valid_targets_mean": 9655.2, + "valid_targets_min": 4413 + }, + { + "epoch": 5.923728813559322, + "grad_norm": 0.45849359516987925, + "learning_rate": 2.820628953257052e-06, + "loss": 0.2739, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11417422443628311, + "step": 3495, + "valid_targets_mean": 6660.8, + "valid_targets_min": 1428 + }, + { + "epoch": 5.932203389830509, + "grad_norm": 0.48560447287935155, + "learning_rate": 2.7775060653402387e-06, + "loss": 0.2747, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15297681093215942, + "step": 3500, + "valid_targets_mean": 9799.2, + "valid_targets_min": 4018 + }, + { + "epoch": 5.940677966101695, + "grad_norm": 0.4406222156886379, + "learning_rate": 2.7346907517104894e-06, + "loss": 0.2957, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.2106126993894577, + "step": 3505, + "valid_targets_mean": 10975.9, + "valid_targets_min": 5409 + }, + { + "epoch": 5.9491525423728815, + "grad_norm": 0.4076101967228653, + "learning_rate": 2.692183777000932e-06, + "loss": 0.2679, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12442032247781754, + "step": 3510, + "valid_targets_mean": 9008.0, + "valid_targets_min": 2858 + }, + { + "epoch": 5.9576271186440675, + "grad_norm": 0.4522416731205185, + "learning_rate": 2.64998590033813e-06, + "loss": 0.27, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09120093286037445, + "step": 3515, + "valid_targets_mean": 5499.0, + "valid_targets_min": 1332 + }, + { + "epoch": 5.966101694915254, + "grad_norm": 0.4686138209396213, + "learning_rate": 2.608097875328488e-06, + "loss": 0.2487, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08969251066446304, + "step": 3520, + "valid_targets_mean": 4958.8, + "valid_targets_min": 1880 + }, + { + "epoch": 5.97457627118644, + "grad_norm": 0.4500868944959106, + "learning_rate": 2.5665204500448137e-06, + "loss": 0.2494, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.06993599236011505, + "step": 3525, + "valid_targets_mean": 4305.2, + "valid_targets_min": 1955 + }, + { + "epoch": 5.983050847457627, + "grad_norm": 0.4525767839640078, + "learning_rate": 2.525254367012955e-06, + "loss": 0.2695, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17192375659942627, + "step": 3530, + "valid_targets_mean": 8511.8, + "valid_targets_min": 3580 + }, + { + "epoch": 5.991525423728813, + "grad_norm": 0.46060568530098406, + "learning_rate": 2.4843003631985486e-06, + "loss": 0.2968, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07557795196771622, + "step": 3535, + "valid_targets_mean": 5208.4, + "valid_targets_min": 3361 + }, + { + "epoch": 6.0, + "grad_norm": 0.3876894703853514, + "learning_rate": 2.4436591699938305e-06, + "loss": 0.2477, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10472303628921509, + "step": 3540, + "valid_targets_mean": 8267.6, + "valid_targets_min": 3516 + }, + { + "epoch": 6.008474576271187, + "grad_norm": 0.4072742789805857, + "learning_rate": 2.403331513204612e-06, + "loss": 0.2467, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13361892104148865, + "step": 3545, + "valid_targets_mean": 8191.2, + "valid_targets_min": 3116 + }, + { + "epoch": 6.016949152542373, + "grad_norm": 0.4197929261254614, + "learning_rate": 2.363318113037283e-06, + "loss": 0.2437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.19385768473148346, + "step": 3550, + "valid_targets_mean": 8499.4, + "valid_targets_min": 2320 + }, + { + "epoch": 6.02542372881356, + "grad_norm": 0.40312238215292895, + "learning_rate": 2.323619684085976e-06, + "loss": 0.2427, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10295233875513077, + "step": 3555, + "valid_targets_mean": 8334.4, + "valid_targets_min": 2999 + }, + { + "epoch": 6.033898305084746, + "grad_norm": 0.4303181949085986, + "learning_rate": 2.2842369353197858e-06, + "loss": 0.2412, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10104535520076752, + "step": 3560, + "valid_targets_mean": 6406.8, + "valid_targets_min": 3180 + }, + { + "epoch": 6.0423728813559325, + "grad_norm": 0.4117633582753061, + "learning_rate": 2.2451705700701185e-06, + "loss": 0.2457, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09206748008728027, + "step": 3565, + "valid_targets_mean": 6353.2, + "valid_targets_min": 1955 + }, + { + "epoch": 6.0508474576271185, + "grad_norm": 0.41630505109565613, + "learning_rate": 2.2064212860181258e-06, + "loss": 0.2364, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11224710941314697, + "step": 3570, + "valid_targets_mean": 6526.1, + "valid_targets_min": 2942 + }, + { + "epoch": 6.059322033898305, + "grad_norm": 0.43320385167493386, + "learning_rate": 2.1679897751822532e-06, + "loss": 0.2462, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14041966199874878, + "step": 3575, + "valid_targets_mean": 8613.5, + "valid_targets_min": 3013 + }, + { + "epoch": 6.067796610169491, + "grad_norm": 0.3981423363842407, + "learning_rate": 2.1298767239058684e-06, + "loss": 0.2797, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1767849326133728, + "step": 3580, + "valid_targets_mean": 10284.9, + "valid_targets_min": 3145 + }, + { + "epoch": 6.076271186440678, + "grad_norm": 0.43296591350601293, + "learning_rate": 2.0920828128450197e-06, + "loss": 0.2426, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16536885499954224, + "step": 3585, + "valid_targets_mean": 8499.6, + "valid_targets_min": 4263 + }, + { + "epoch": 6.084745762711864, + "grad_norm": 0.41880853588474576, + "learning_rate": 2.054608716956259e-06, + "loss": 0.2509, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10676504671573639, + "step": 3590, + "valid_targets_mean": 7362.6, + "valid_targets_min": 1904 + }, + { + "epoch": 6.093220338983051, + "grad_norm": 0.4669953694349539, + "learning_rate": 2.0174551054846158e-06, + "loss": 0.2423, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11701343953609467, + "step": 3595, + "valid_targets_mean": 6644.5, + "valid_targets_min": 2520 + }, + { + "epoch": 6.101694915254237, + "grad_norm": 0.3975138860321249, + "learning_rate": 1.9806226419516195e-06, + "loss": 0.286, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1058916300535202, + "step": 3600, + "valid_targets_mean": 7707.5, + "valid_targets_min": 5584 + }, + { + "epoch": 6.110169491525424, + "grad_norm": 0.3784624088657458, + "learning_rate": 1.94411198414346e-06, + "loss": 0.2674, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14342477917671204, + "step": 3605, + "valid_targets_mean": 10405.4, + "valid_targets_min": 2678 + }, + { + "epoch": 6.11864406779661, + "grad_norm": 0.4409579331355334, + "learning_rate": 1.9079237840992416e-06, + "loss": 0.2341, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09242921322584152, + "step": 3610, + "valid_targets_mean": 6214.9, + "valid_targets_min": 2570 + }, + { + "epoch": 6.127118644067797, + "grad_norm": 0.4274606131359512, + "learning_rate": 1.8720586880993452e-06, + "loss": 0.2768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13959690928459167, + "step": 3615, + "valid_targets_mean": 8091.1, + "valid_targets_min": 1864 + }, + { + "epoch": 6.135593220338983, + "grad_norm": 0.46627588534538766, + "learning_rate": 1.8365173366538647e-06, + "loss": 0.3004, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17453311383724213, + "step": 3620, + "valid_targets_mean": 8541.9, + "valid_targets_min": 2926 + }, + { + "epoch": 6.1440677966101696, + "grad_norm": 0.42107709331487897, + "learning_rate": 1.8013003644911987e-06, + "loss": 0.2891, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16110199689865112, + "step": 3625, + "valid_targets_mean": 7763.9, + "valid_targets_min": 2914 + }, + { + "epoch": 6.1525423728813555, + "grad_norm": 0.4480699024314328, + "learning_rate": 1.7664084005466796e-06, + "loss": 0.2599, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1471949815750122, + "step": 3630, + "valid_targets_mean": 7371.1, + "valid_targets_min": 2475 + }, + { + "epoch": 6.161016949152542, + "grad_norm": 0.41092743558492617, + "learning_rate": 1.7318420679513725e-06, + "loss": 0.2685, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0986928790807724, + "step": 3635, + "valid_targets_mean": 7593.5, + "valid_targets_min": 5251 + }, + { + "epoch": 6.169491525423728, + "grad_norm": 0.46749871764465295, + "learning_rate": 1.6976019840209334e-06, + "loss": 0.2503, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13527244329452515, + "step": 3640, + "valid_targets_mean": 6722.5, + "valid_targets_min": 2997 + }, + { + "epoch": 6.177966101694915, + "grad_norm": 0.4178021620516435, + "learning_rate": 1.6636887602445839e-06, + "loss": 0.2503, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.145928755402565, + "step": 3645, + "valid_targets_mean": 8564.9, + "valid_targets_min": 4021 + }, + { + "epoch": 6.186440677966102, + "grad_norm": 0.4360686656344327, + "learning_rate": 1.6301030022741837e-06, + "loss": 0.2667, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1285943239927292, + "step": 3650, + "valid_targets_mean": 6744.0, + "valid_targets_min": 2676 + }, + { + "epoch": 6.194915254237288, + "grad_norm": 0.45989463629934857, + "learning_rate": 1.596845309913444e-06, + "loss": 0.2781, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13629180192947388, + "step": 3655, + "valid_targets_mean": 6774.1, + "valid_targets_min": 1806 + }, + { + "epoch": 6.203389830508475, + "grad_norm": 0.47108516701203995, + "learning_rate": 1.5639162771071736e-06, + "loss": 0.2598, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17500731348991394, + "step": 3660, + "valid_targets_mean": 8830.6, + "valid_targets_min": 3863 + }, + { + "epoch": 6.211864406779661, + "grad_norm": 0.4318354716431225, + "learning_rate": 1.5313164919307033e-06, + "loss": 0.2644, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.115082286298275, + "step": 3665, + "valid_targets_mean": 7130.6, + "valid_targets_min": 4331 + }, + { + "epoch": 6.220338983050848, + "grad_norm": 0.45277354650406365, + "learning_rate": 1.499046536579374e-06, + "loss": 0.2837, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17140638828277588, + "step": 3670, + "valid_targets_mean": 8954.1, + "valid_targets_min": 4589 + }, + { + "epoch": 6.228813559322034, + "grad_norm": 0.4323260035459006, + "learning_rate": 1.4671069873581312e-06, + "loss": 0.2868, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15814843773841858, + "step": 3675, + "valid_targets_mean": 9400.6, + "valid_targets_min": 5038 + }, + { + "epoch": 6.237288135593221, + "grad_norm": 0.4951365278391396, + "learning_rate": 1.4354984146712503e-06, + "loss": 0.2643, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11142455041408539, + "step": 3680, + "valid_targets_mean": 6961.5, + "valid_targets_min": 2195 + }, + { + "epoch": 6.245762711864407, + "grad_norm": 0.4681915800014983, + "learning_rate": 1.4042213830121344e-06, + "loss": 0.2519, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13957908749580383, + "step": 3685, + "valid_targets_mean": 6601.8, + "valid_targets_min": 3222 + }, + { + "epoch": 6.254237288135593, + "grad_norm": 0.39827114052731316, + "learning_rate": 1.3732764509532316e-06, + "loss": 0.2585, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13600842654705048, + "step": 3690, + "valid_targets_mean": 9952.6, + "valid_targets_min": 5541 + }, + { + "epoch": 6.262711864406779, + "grad_norm": 0.43483607655963946, + "learning_rate": 1.3426641711360788e-06, + "loss": 0.2557, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12304791808128357, + "step": 3695, + "valid_targets_mean": 7189.5, + "valid_targets_min": 3441 + }, + { + "epoch": 6.271186440677966, + "grad_norm": 0.4875466298631045, + "learning_rate": 1.3123850902614143e-06, + "loss": 0.2447, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09608499705791473, + "step": 3700, + "valid_targets_mean": 5324.5, + "valid_targets_min": 2321 + }, + { + "epoch": 6.279661016949152, + "grad_norm": 0.457368172054018, + "learning_rate": 1.2824397490794115e-06, + "loss": 0.2811, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12899936735630035, + "step": 3705, + "valid_targets_mean": 8255.5, + "valid_targets_min": 2920 + }, + { + "epoch": 6.288135593220339, + "grad_norm": 0.43355278373700595, + "learning_rate": 1.2528286823800495e-06, + "loss": 0.2507, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12112070620059967, + "step": 3710, + "valid_targets_mean": 6698.4, + "valid_targets_min": 2533 + }, + { + "epoch": 6.296610169491525, + "grad_norm": 0.40685153441048544, + "learning_rate": 1.223552418983518e-06, + "loss": 0.2665, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16152483224868774, + "step": 3715, + "valid_targets_mean": 8516.4, + "valid_targets_min": 2412 + }, + { + "epoch": 6.305084745762712, + "grad_norm": 0.4505350099831515, + "learning_rate": 1.1946114817308207e-06, + "loss": 0.2466, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12660183012485504, + "step": 3720, + "valid_targets_mean": 6932.0, + "valid_targets_min": 2418 + }, + { + "epoch": 6.313559322033898, + "grad_norm": 0.3939000677040436, + "learning_rate": 1.166006387474401e-06, + "loss": 0.2446, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1851355880498886, + "step": 3725, + "valid_targets_mean": 10651.5, + "valid_targets_min": 3329 + }, + { + "epoch": 6.322033898305085, + "grad_norm": 0.45424467174671546, + "learning_rate": 1.1377376470689328e-06, + "loss": 0.2735, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12378216534852982, + "step": 3730, + "valid_targets_mean": 6624.9, + "valid_targets_min": 3343 + }, + { + "epoch": 6.330508474576272, + "grad_norm": 0.4906347349043102, + "learning_rate": 1.1098057653621775e-06, + "loss": 0.2554, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16246655583381653, + "step": 3735, + "valid_targets_mean": 7427.4, + "valid_targets_min": 3773 + }, + { + "epoch": 6.338983050847458, + "grad_norm": 0.4419793742239489, + "learning_rate": 1.0822112411860042e-06, + "loss": 0.2669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09111709892749786, + "step": 3740, + "valid_targets_mean": 5823.4, + "valid_targets_min": 2388 + }, + { + "epoch": 6.3474576271186445, + "grad_norm": 0.4119491087505782, + "learning_rate": 1.0549545673474304e-06, + "loss": 0.2473, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0849679559469223, + "step": 3745, + "valid_targets_mean": 6402.6, + "valid_targets_min": 3380 + }, + { + "epoch": 6.3559322033898304, + "grad_norm": 0.43180351169822373, + "learning_rate": 1.0280362306198732e-06, + "loss": 0.2618, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16606824100017548, + "step": 3750, + "valid_targets_mean": 9833.5, + "valid_targets_min": 3799 + }, + { + "epoch": 6.364406779661017, + "grad_norm": 0.4414573273424959, + "learning_rate": 1.0014567117344121e-06, + "loss": 0.2702, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15046200156211853, + "step": 3755, + "valid_targets_mean": 6806.9, + "valid_targets_min": 2581 + }, + { + "epoch": 6.372881355932203, + "grad_norm": 0.49931100192630407, + "learning_rate": 9.752164853712355e-07, + "loss": 0.2445, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11329838633537292, + "step": 3760, + "valid_targets_mean": 7291.8, + "valid_targets_min": 3396 + }, + { + "epoch": 6.38135593220339, + "grad_norm": 0.44833955478818444, + "learning_rate": 9.493160201511409e-07, + "loss": 0.2669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11439744383096695, + "step": 3765, + "valid_targets_mean": 6983.8, + "valid_targets_min": 1695 + }, + { + "epoch": 6.389830508474576, + "grad_norm": 0.4599867420960204, + "learning_rate": 9.2375577862718e-07, + "loss": 0.2653, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09336736798286438, + "step": 3770, + "valid_targets_mean": 6872.4, + "valid_targets_min": 2334 + }, + { + "epoch": 6.398305084745763, + "grad_norm": 0.42332962451369605, + "learning_rate": 8.985362172763933e-07, + "loss": 0.2463, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09449931979179382, + "step": 3775, + "valid_targets_mean": 6810.9, + "valid_targets_min": 2346 + }, + { + "epoch": 6.406779661016949, + "grad_norm": 0.425561451589507, + "learning_rate": 8.736577864916617e-07, + "loss": 0.2513, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12058237195014954, + "step": 3780, + "valid_targets_mean": 7390.8, + "valid_targets_min": 2397 + }, + { + "epoch": 6.415254237288136, + "grad_norm": 0.45418010893807925, + "learning_rate": 8.491209305736525e-07, + "loss": 0.2546, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13464850187301636, + "step": 3785, + "valid_targets_mean": 7586.5, + "valid_targets_min": 2281 + }, + { + "epoch": 6.423728813559322, + "grad_norm": 0.47095225870799795, + "learning_rate": 8.249260877228993e-07, + "loss": 0.2876, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14261358976364136, + "step": 3790, + "valid_targets_mean": 8037.9, + "valid_targets_min": 2980 + }, + { + "epoch": 6.432203389830509, + "grad_norm": 0.47226974874730265, + "learning_rate": 8.010736900319615e-07, + "loss": 0.274, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09179393947124481, + "step": 3795, + "valid_targets_mean": 6363.5, + "valid_targets_min": 2705 + }, + { + "epoch": 6.440677966101695, + "grad_norm": 0.4137999596538719, + "learning_rate": 7.775641634777198e-07, + "loss": 0.2621, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11750559508800507, + "step": 3800, + "valid_targets_mean": 7810.5, + "valid_targets_min": 3108 + }, + { + "epoch": 6.4491525423728815, + "grad_norm": 0.43416780782235825, + "learning_rate": 7.543979279137592e-07, + "loss": 0.2596, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11820831149816513, + "step": 3805, + "valid_targets_mean": 6331.1, + "valid_targets_min": 2060 + }, + { + "epoch": 6.4576271186440675, + "grad_norm": 0.39459543683267795, + "learning_rate": 7.315753970628825e-07, + "loss": 0.2499, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09400268644094467, + "step": 3810, + "valid_targets_mean": 6584.6, + "valid_targets_min": 2612 + }, + { + "epoch": 6.466101694915254, + "grad_norm": 0.3991025138107861, + "learning_rate": 7.090969785097046e-07, + "loss": 0.2659, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1371954381465912, + "step": 3815, + "valid_targets_mean": 9368.8, + "valid_targets_min": 5943 + }, + { + "epoch": 6.47457627118644, + "grad_norm": 0.41902884093001064, + "learning_rate": 6.869630736933963e-07, + "loss": 0.2437, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07621373981237411, + "step": 3820, + "valid_targets_mean": 6019.4, + "valid_targets_min": 3804 + }, + { + "epoch": 6.483050847457627, + "grad_norm": 0.43114784690809344, + "learning_rate": 6.651740779004878e-07, + "loss": 0.2665, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14362475275993347, + "step": 3825, + "valid_targets_mean": 8519.8, + "valid_targets_min": 2553 + }, + { + "epoch": 6.491525423728813, + "grad_norm": 0.4970910956256326, + "learning_rate": 6.437303802578365e-07, + "loss": 0.2636, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09743285179138184, + "step": 3830, + "valid_targets_mean": 5775.5, + "valid_targets_min": 2753 + }, + { + "epoch": 6.5, + "grad_norm": 0.42624386382800444, + "learning_rate": 6.226323637256592e-07, + "loss": 0.2584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09431523084640503, + "step": 3835, + "valid_targets_mean": 6937.0, + "valid_targets_min": 2820 + }, + { + "epoch": 6.508474576271187, + "grad_norm": 0.4154172821087517, + "learning_rate": 6.018804050906957e-07, + "loss": 0.2607, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17171484231948853, + "step": 3840, + "valid_targets_mean": 9928.1, + "valid_targets_min": 3685 + }, + { + "epoch": 6.516949152542373, + "grad_norm": 0.44825920178188705, + "learning_rate": 5.814748749594845e-07, + "loss": 0.2934, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15356290340423584, + "step": 3845, + "valid_targets_mean": 7801.4, + "valid_targets_min": 2367 + }, + { + "epoch": 6.52542372881356, + "grad_norm": 0.408924077476942, + "learning_rate": 5.614161377517491e-07, + "loss": 0.2407, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11111468821763992, + "step": 3850, + "valid_targets_mean": 7373.9, + "valid_targets_min": 2314 + }, + { + "epoch": 6.533898305084746, + "grad_norm": 0.40618813884257093, + "learning_rate": 5.417045516938712e-07, + "loss": 0.262, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1602497547864914, + "step": 3855, + "valid_targets_mean": 11765.6, + "valid_targets_min": 3534 + }, + { + "epoch": 6.5423728813559325, + "grad_norm": 0.43314959447209905, + "learning_rate": 5.223404688125189e-07, + "loss": 0.266, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08646964281797409, + "step": 3860, + "valid_targets_mean": 5335.6, + "valid_targets_min": 1489 + }, + { + "epoch": 6.5508474576271185, + "grad_norm": 0.4574683919645152, + "learning_rate": 5.033242349283307e-07, + "loss": 0.2479, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13719329237937927, + "step": 3865, + "valid_targets_mean": 7237.4, + "valid_targets_min": 2632 + }, + { + "epoch": 6.559322033898305, + "grad_norm": 0.4811190220109463, + "learning_rate": 4.846561896497682e-07, + "loss": 0.2898, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11043386906385422, + "step": 3870, + "valid_targets_mean": 6984.5, + "valid_targets_min": 4574 + }, + { + "epoch": 6.567796610169491, + "grad_norm": 0.44943984706465684, + "learning_rate": 4.6633666636703325e-07, + "loss": 0.2574, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11624139547348022, + "step": 3875, + "valid_targets_mean": 7636.9, + "valid_targets_min": 4092 + }, + { + "epoch": 6.576271186440678, + "grad_norm": 0.4337255198830074, + "learning_rate": 4.4836599224611586e-07, + "loss": 0.2534, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09535051882266998, + "step": 3880, + "valid_targets_mean": 7153.5, + "valid_targets_min": 708 + }, + { + "epoch": 6.584745762711864, + "grad_norm": 0.5137590453948664, + "learning_rate": 4.3074448822295165e-07, + "loss": 0.2677, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.23429949581623077, + "step": 3885, + "valid_targets_mean": 11476.6, + "valid_targets_min": 1995 + }, + { + "epoch": 6.593220338983051, + "grad_norm": 0.44066724993264006, + "learning_rate": 4.1347246899770435e-07, + "loss": 0.2531, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1679355502128601, + "step": 3890, + "valid_targets_mean": 7367.9, + "valid_targets_min": 3472 + }, + { + "epoch": 6.601694915254237, + "grad_norm": 0.38946709881244124, + "learning_rate": 3.965502430291235e-07, + "loss": 0.2603, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1628887951374054, + "step": 3895, + "valid_targets_mean": 11073.8, + "valid_targets_min": 5088 + }, + { + "epoch": 6.610169491525424, + "grad_norm": 0.45495617220170287, + "learning_rate": 3.7997811252905135e-07, + "loss": 0.2469, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10322394967079163, + "step": 3900, + "valid_targets_mean": 6399.1, + "valid_targets_min": 2490 + }, + { + "epoch": 6.61864406779661, + "grad_norm": 0.4469868169141226, + "learning_rate": 3.6375637345701376e-07, + "loss": 0.2932, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13132116198539734, + "step": 3905, + "valid_targets_mean": 8170.0, + "valid_targets_min": 3419 + }, + { + "epoch": 6.627118644067797, + "grad_norm": 0.4334716085841391, + "learning_rate": 3.4788531551495307e-07, + "loss": 0.2353, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1662118136882782, + "step": 3910, + "valid_targets_mean": 9730.6, + "valid_targets_min": 3618 + }, + { + "epoch": 6.635593220338983, + "grad_norm": 0.45238587707581673, + "learning_rate": 3.323652221420326e-07, + "loss": 0.2673, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11458948254585266, + "step": 3915, + "valid_targets_mean": 7180.0, + "valid_targets_min": 3770 + }, + { + "epoch": 6.6440677966101696, + "grad_norm": 0.43546318359687897, + "learning_rate": 3.1719637050958706e-07, + "loss": 0.2459, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16936321556568146, + "step": 3920, + "valid_targets_mean": 7698.1, + "valid_targets_min": 2697 + }, + { + "epoch": 6.652542372881356, + "grad_norm": 0.47978180866821984, + "learning_rate": 3.023790315161734e-07, + "loss": 0.2355, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08654442429542542, + "step": 3925, + "valid_targets_mean": 5161.9, + "valid_targets_min": 2722 + }, + { + "epoch": 6.661016949152542, + "grad_norm": 0.4943904808254496, + "learning_rate": 2.8791346978273015e-07, + "loss": 0.2474, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1414184719324112, + "step": 3930, + "valid_targets_mean": 7625.6, + "valid_targets_min": 2919 + }, + { + "epoch": 6.669491525423728, + "grad_norm": 0.37408844476292086, + "learning_rate": 2.7379994364784556e-07, + "loss": 0.2761, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15006043016910553, + "step": 3935, + "valid_targets_mean": 10405.0, + "valid_targets_min": 6385 + }, + { + "epoch": 6.677966101694915, + "grad_norm": 0.5916409188859896, + "learning_rate": 2.60038705163157e-07, + "loss": 0.247, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08202853798866272, + "step": 3940, + "valid_targets_mean": 5563.0, + "valid_targets_min": 3361 + }, + { + "epoch": 6.686440677966102, + "grad_norm": 0.44034011345714885, + "learning_rate": 2.4663000008883664e-07, + "loss": 0.2969, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16353031992912292, + "step": 3945, + "valid_targets_mean": 9585.6, + "valid_targets_min": 3373 + }, + { + "epoch": 6.694915254237288, + "grad_norm": 0.4385202919657095, + "learning_rate": 2.335740678892129e-07, + "loss": 0.2584, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14852274954319, + "step": 3950, + "valid_targets_mean": 7825.4, + "valid_targets_min": 2417 + }, + { + "epoch": 6.703389830508475, + "grad_norm": 0.40864405535666826, + "learning_rate": 2.20871141728487e-07, + "loss": 0.2768, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.20317432284355164, + "step": 3955, + "valid_targets_mean": 10872.1, + "valid_targets_min": 3224 + }, + { + "epoch": 6.711864406779661, + "grad_norm": 0.4125200729598343, + "learning_rate": 2.0852144846656764e-07, + "loss": 0.2552, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11529850959777832, + "step": 3960, + "valid_targets_mean": 7395.0, + "valid_targets_min": 1786 + }, + { + "epoch": 6.720338983050848, + "grad_norm": 0.43566364650736705, + "learning_rate": 1.9652520865502734e-07, + "loss": 0.268, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11976931989192963, + "step": 3965, + "valid_targets_mean": 6824.1, + "valid_targets_min": 2624 + }, + { + "epoch": 6.728813559322034, + "grad_norm": 0.4071322712174556, + "learning_rate": 1.848826365331613e-07, + "loss": 0.2741, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.09024620801210403, + "step": 3970, + "valid_targets_mean": 6544.0, + "valid_targets_min": 3445 + }, + { + "epoch": 6.737288135593221, + "grad_norm": 0.39631248821410153, + "learning_rate": 1.7359394002415265e-07, + "loss": 0.255, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.08111607283353806, + "step": 3975, + "valid_targets_mean": 7050.0, + "valid_targets_min": 2954 + }, + { + "epoch": 6.745762711864407, + "grad_norm": 0.4759489882354291, + "learning_rate": 1.626593207313709e-07, + "loss": 0.2504, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0827079638838768, + "step": 3980, + "valid_targets_mean": 5926.6, + "valid_targets_min": 2590 + }, + { + "epoch": 6.754237288135593, + "grad_norm": 0.4341505959824497, + "learning_rate": 1.520789739347617e-07, + "loss": 0.2742, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14950904250144958, + "step": 3985, + "valid_targets_mean": 8768.5, + "valid_targets_min": 3997 + }, + { + "epoch": 6.762711864406779, + "grad_norm": 0.41517718572757417, + "learning_rate": 1.4185308858737367e-07, + "loss": 0.2872, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16266342997550964, + "step": 3990, + "valid_targets_mean": 7884.4, + "valid_targets_min": 2837 + }, + { + "epoch": 6.771186440677966, + "grad_norm": 0.4325233265851526, + "learning_rate": 1.3198184731196162e-07, + "loss": 0.2565, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11044515669345856, + "step": 3995, + "valid_targets_mean": 7797.4, + "valid_targets_min": 3860 + }, + { + "epoch": 6.779661016949152, + "grad_norm": 0.44277488862547426, + "learning_rate": 1.2246542639774871e-07, + "loss": 0.2911, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.0890350490808487, + "step": 4000, + "valid_targets_mean": 5685.8, + "valid_targets_min": 2485 + }, + { + "epoch": 6.788135593220339, + "grad_norm": 0.40915928129602963, + "learning_rate": 1.1330399579725814e-07, + "loss": 0.2516, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12274955958127975, + "step": 4005, + "valid_targets_mean": 7330.8, + "valid_targets_min": 2502 + }, + { + "epoch": 6.796610169491525, + "grad_norm": 0.4481096216804316, + "learning_rate": 1.0449771912329543e-07, + "loss": 0.2692, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13692034780979156, + "step": 4010, + "valid_targets_mean": 6854.6, + "valid_targets_min": 3286 + }, + { + "epoch": 6.805084745762712, + "grad_norm": 0.4980756119613032, + "learning_rate": 9.604675364601079e-08, + "loss": 0.2814, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.16621151566505432, + "step": 4015, + "valid_targets_mean": 8521.9, + "valid_targets_min": 2531 + }, + { + "epoch": 6.813559322033898, + "grad_norm": 0.4414209894178357, + "learning_rate": 8.795125029010587e-08, + "loss": 0.2358, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14769510924816132, + "step": 4020, + "valid_targets_mean": 7776.1, + "valid_targets_min": 1922 + }, + { + "epoch": 6.822033898305085, + "grad_norm": 0.4566413826484506, + "learning_rate": 8.021135363212252e-08, + "loss": 0.2683, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13060231506824493, + "step": 4025, + "valid_targets_mean": 7528.2, + "valid_targets_min": 3087 + }, + { + "epoch": 6.830508474576272, + "grad_norm": 0.4334680150880916, + "learning_rate": 7.282720189787374e-08, + "loss": 0.2589, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11092951148748398, + "step": 4030, + "valid_targets_mean": 6980.1, + "valid_targets_min": 2151 + }, + { + "epoch": 6.838983050847458, + "grad_norm": 0.5527431160252887, + "learning_rate": 6.579892695996571e-08, + "loss": 0.2547, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1120155081152916, + "step": 4035, + "valid_targets_mean": 7117.2, + "valid_targets_min": 4506 + }, + { + "epoch": 6.847457627118644, + "grad_norm": 0.45476280194826413, + "learning_rate": 5.912665433545517e-08, + "loss": 0.2601, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14059172570705414, + "step": 4040, + "valid_targets_mean": 7315.8, + "valid_targets_min": 2297 + }, + { + "epoch": 6.8559322033898304, + "grad_norm": 0.4045006325867193, + "learning_rate": 5.281050318358683e-08, + "loss": 0.2419, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.12083014845848083, + "step": 4045, + "valid_targets_mean": 9475.5, + "valid_targets_min": 3188 + }, + { + "epoch": 6.864406779661017, + "grad_norm": 0.37089524552087677, + "learning_rate": 4.685058630369055e-08, + "loss": 0.2775, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14551851153373718, + "step": 4050, + "valid_targets_mean": 9946.2, + "valid_targets_min": 3696 + }, + { + "epoch": 6.872881355932203, + "grad_norm": 0.4516798472346292, + "learning_rate": 4.124701013314969e-08, + "loss": 0.2836, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1390407681465149, + "step": 4055, + "valid_targets_mean": 8014.2, + "valid_targets_min": 1547 + }, + { + "epoch": 6.88135593220339, + "grad_norm": 0.7576344140395941, + "learning_rate": 3.599987474550259e-08, + "loss": 0.2677, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17070727050304413, + "step": 4060, + "valid_targets_mean": 8283.6, + "valid_targets_min": 2605 + }, + { + "epoch": 6.889830508474576, + "grad_norm": 0.4028968163886396, + "learning_rate": 3.110927384865958e-08, + "loss": 0.2785, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.18179358541965485, + "step": 4065, + "valid_targets_mean": 10546.0, + "valid_targets_min": 4354 + }, + { + "epoch": 6.898305084745763, + "grad_norm": 0.4085313254945271, + "learning_rate": 2.6575294783230954e-08, + "loss": 0.2329, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07890482246875763, + "step": 4070, + "valid_targets_mean": 5749.8, + "valid_targets_min": 2419 + }, + { + "epoch": 6.906779661016949, + "grad_norm": 0.4373124817628321, + "learning_rate": 2.239801852095935e-08, + "loss": 0.2457, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1607382893562317, + "step": 4075, + "valid_targets_mean": 10599.2, + "valid_targets_min": 4414 + }, + { + "epoch": 6.915254237288136, + "grad_norm": 0.4514490229067648, + "learning_rate": 1.8577519663278697e-08, + "loss": 0.2409, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.11941658705472946, + "step": 4080, + "valid_targets_mean": 8650.9, + "valid_targets_min": 4959 + }, + { + "epoch": 6.923728813559322, + "grad_norm": 0.4589628551398868, + "learning_rate": 1.5113866439986357e-08, + "loss": 0.285, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1670331358909607, + "step": 4085, + "valid_targets_mean": 8175.0, + "valid_targets_min": 2563 + }, + { + "epoch": 6.932203389830509, + "grad_norm": 0.4694870576161979, + "learning_rate": 1.2007120708015241e-08, + "loss": 0.2731, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.13149237632751465, + "step": 4090, + "valid_targets_mean": 6409.9, + "valid_targets_min": 2424 + }, + { + "epoch": 6.940677966101695, + "grad_norm": 0.3956983726547488, + "learning_rate": 9.257337950332456e-09, + "loss": 0.2669, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1256042718887329, + "step": 4095, + "valid_targets_mean": 8328.4, + "valid_targets_min": 3379 + }, + { + "epoch": 6.9491525423728815, + "grad_norm": 0.4443933784478533, + "learning_rate": 6.864567274955658e-09, + "loss": 0.2706, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17330250144004822, + "step": 4100, + "valid_targets_mean": 8270.1, + "valid_targets_min": 2415 + }, + { + "epoch": 6.9576271186440675, + "grad_norm": 0.4977604162909399, + "learning_rate": 4.828851414062641e-09, + "loss": 0.2787, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.15118731558322906, + "step": 4105, + "valid_targets_mean": 9061.0, + "valid_targets_min": 4376 + }, + { + "epoch": 6.966101694915254, + "grad_norm": 0.4183573072699651, + "learning_rate": 3.150226723243055e-09, + "loss": 0.2701, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.1289590746164322, + "step": 4110, + "valid_targets_mean": 8290.0, + "valid_targets_min": 2677 + }, + { + "epoch": 6.97457627118644, + "grad_norm": 0.4187228386039913, + "learning_rate": 1.8287231808322702e-09, + "loss": 0.2611, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.14881855249404907, + "step": 4115, + "valid_targets_mean": 10096.6, + "valid_targets_min": 6877 + }, + { + "epoch": 6.983050847457627, + "grad_norm": 0.43938825018899513, + "learning_rate": 8.643643873962326e-10, + "loss": 0.246, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.10421787202358246, + "step": 4120, + "valid_targets_mean": 7506.1, + "valid_targets_min": 4050 + }, + { + "epoch": 6.991525423728813, + "grad_norm": 0.49406724429429383, + "learning_rate": 2.5716756528959553e-10, + "loss": 0.2559, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.17047889530658722, + "step": 4125, + "valid_targets_mean": 9332.6, + "valid_targets_min": 3606 + }, + { + "epoch": 7.0, + "grad_norm": 0.42283740256260893, + "learning_rate": 7.143558364841597e-12, + "loss": 0.2482, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07740560173988342, + "step": 4130, + "valid_targets_mean": 6206.0, + "valid_targets_min": 3223 + }, + { + "epoch": 7.0, + "loss_nan_ranks": 0, + "loss_rank_avg": 0.07740560173988342, + "step": 4130, + "total_flos": 2.6502668329167094e+18, + "train_loss": 0.3043324818911333, + "train_runtime": 64701.6572, + "train_samples_per_second": 1.021, + "train_steps_per_second": 0.064, + "valid_targets_mean": 6206.0, + "valid_targets_min": 3223 + } + ], + "logging_steps": 5, + "max_steps": 4130, + "num_input_tokens_seen": 0, + "num_train_epochs": 7, + "save_steps": 1500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 2.6502668329167094e+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..1cbd4fe --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:97d37e47ed5eac05f9f6a646a54c3a5a323532c34828ca9b08de854750a34393 +size 8785 diff --git a/training_loss.png b/training_loss.png new file mode 100644 index 0000000..ac37cd5 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