Upload model

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,61 @@
+---
+library_name: transformers
+license: other
+base_model: Qwen/Qwen2.5-7B-Instruct
+tags:
+- llama-factory
+- full
+- generated_from_trainer
+model-index:
+- name: c1_math_0d_4s_3k
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# c1_math_0d_4s_3k
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) on the mlfoundations-dev/c1_math_0d_4s_3k 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: 2e-05
+- train_batch_size: 1
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 16
+- gradient_accumulation_steps: 6
+- total_train_batch_size: 96
+- total_eval_batch_size: 128
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) 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 4.46.1
+- Pytorch 2.6.0a0+ecf3bae40a.nv25.01
+- Datasets 3.5.0
+- Tokenizers 0.20.3

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|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
+}

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 7.0,
+    "total_flos": 7.881049861301207e+17,
+    "train_loss": 0.3857915287281012,
+    "train_runtime": 8985.043,
+    "train_samples_per_second": 2.462,
+    "train_steps_per_second": 0.026
+}

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "/p/data1/mmlaion/dcft/hub/models--Qwen--Qwen2.5-7B-Instruct/snapshots/a09a35458c702b33eeacc393d103063234e8bc28",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.46.1",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.46.1"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8ced9245643e487ca1ef3a1fec4498b82bc40bea911aa3ca6dd5c111f2ab056
+size 4877660776

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5821864f6368eae1313c8c6eb731810e1bb600968ebd4dbb783767667278bcc1
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:002552f81a9cfa57ca7b02cbcf56d12aa979bf65e7f7989eb5fd415854b456d8
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cb8cfc942407bbd6c2c11b3a33c5d1989eed8a05615227d801efb9238ec15b5
+size 1089994880

model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15231233024
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "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.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "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.bias": "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-00003-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00003-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-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "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.bias": "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-00003-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "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.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "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.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "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.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "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.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "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.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "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.bias": "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-00002-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "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.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00003-of-00004.safetensors"
+  }
+}

runs/Apr26_23-44-05_jpbot-001-02.jupiter.internal/events.out.tfevents.1745703893.jpbot-001-02.jupiter.internal.1457864.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ac4d76ce27006d6c9c9714eaca81f9feb3a660059313c34f3896dce56fc00a2
+size 54101

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": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

start_end.json

@@ -0,0 +1 @@
+{"start_time": "2025-04-26 23:44:05", "end_time": "2025-04-27 02:14:46"}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "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": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "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
+    }
+  },
+  "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,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 7.0,
+    "total_flos": 7.881049861301207e+17,
+    "train_loss": 0.3857915287281012,
+    "train_runtime": 8985.043,
+    "train_samples_per_second": 2.462,
+    "train_steps_per_second": 0.026
+}

trainer_log.jsonl

@@ -0,0 +1,232 @@
+{"current_steps": 1, "total_steps": 231, "loss": 0.8216, "lr": 8.333333333333333e-07, "epoch": 0.030303030303030304, "percentage": 0.43, "elapsed_time": "0:00:47", "remaining_time": "3:00:59"}
+{"current_steps": 2, "total_steps": 231, "loss": 0.794, "lr": 1.6666666666666667e-06, "epoch": 0.06060606060606061, "percentage": 0.87, "elapsed_time": "0:01:28", "remaining_time": "2:48:31"}
+{"current_steps": 3, "total_steps": 231, "loss": 0.8213, "lr": 2.5e-06, "epoch": 0.09090909090909091, "percentage": 1.3, "elapsed_time": "0:02:07", "remaining_time": "2:41:42"}
+{"current_steps": 4, "total_steps": 231, "loss": 0.7913, "lr": 3.3333333333333333e-06, "epoch": 0.12121212121212122, "percentage": 1.73, "elapsed_time": "0:02:44", "remaining_time": "2:35:45"}
+{"current_steps": 5, "total_steps": 231, "loss": 0.7675, "lr": 4.166666666666667e-06, "epoch": 0.15151515151515152, "percentage": 2.16, "elapsed_time": "0:03:19", "remaining_time": "2:29:55"}
+{"current_steps": 6, "total_steps": 231, "loss": 0.7282, "lr": 5e-06, "epoch": 0.18181818181818182, "percentage": 2.6, "elapsed_time": "0:03:58", "remaining_time": "2:28:49"}
+{"current_steps": 7, "total_steps": 231, "loss": 0.7226, "lr": 5.833333333333334e-06, "epoch": 0.21212121212121213, "percentage": 3.03, "elapsed_time": "0:04:37", "remaining_time": "2:27:51"}
+{"current_steps": 8, "total_steps": 231, "loss": 0.7141, "lr": 6.666666666666667e-06, "epoch": 0.24242424242424243, "percentage": 3.46, "elapsed_time": "0:05:15", "remaining_time": "2:26:46"}
+{"current_steps": 9, "total_steps": 231, "loss": 0.7258, "lr": 7.500000000000001e-06, "epoch": 0.2727272727272727, "percentage": 3.9, "elapsed_time": "0:05:53", "remaining_time": "2:25:13"}
+{"current_steps": 10, "total_steps": 231, "loss": 0.6918, "lr": 8.333333333333334e-06, "epoch": 0.30303030303030304, "percentage": 4.33, "elapsed_time": "0:06:28", "remaining_time": "2:22:57"}
+{"current_steps": 11, "total_steps": 231, "loss": 0.6673, "lr": 9.166666666666666e-06, "epoch": 0.3333333333333333, "percentage": 4.76, "elapsed_time": "0:07:06", "remaining_time": "2:22:03"}
+{"current_steps": 12, "total_steps": 231, "loss": 0.6527, "lr": 1e-05, "epoch": 0.36363636363636365, "percentage": 5.19, "elapsed_time": "0:07:43", "remaining_time": "2:21:02"}
+{"current_steps": 13, "total_steps": 231, "loss": 0.6604, "lr": 1.0833333333333334e-05, "epoch": 0.3939393939393939, "percentage": 5.63, "elapsed_time": "0:08:18", "remaining_time": "2:19:13"}
+{"current_steps": 14, "total_steps": 231, "loss": 0.6126, "lr": 1.1666666666666668e-05, "epoch": 0.42424242424242425, "percentage": 6.06, "elapsed_time": "0:08:53", "remaining_time": "2:17:42"}
+{"current_steps": 15, "total_steps": 231, "loss": 0.6031, "lr": 1.25e-05, "epoch": 0.45454545454545453, "percentage": 6.49, "elapsed_time": "0:09:32", "remaining_time": "2:17:24"}
+{"current_steps": 16, "total_steps": 231, "loss": 0.6213, "lr": 1.3333333333333333e-05, "epoch": 0.48484848484848486, "percentage": 6.93, "elapsed_time": "0:10:10", "remaining_time": "2:16:38"}
+{"current_steps": 17, "total_steps": 231, "loss": 0.5851, "lr": 1.416666666666667e-05, "epoch": 0.5151515151515151, "percentage": 7.36, "elapsed_time": "0:10:47", "remaining_time": "2:15:54"}
+{"current_steps": 18, "total_steps": 231, "loss": 0.5827, "lr": 1.5000000000000002e-05, "epoch": 0.5454545454545454, "percentage": 7.79, "elapsed_time": "0:11:27", "remaining_time": "2:15:33"}
+{"current_steps": 19, "total_steps": 231, "loss": 0.5626, "lr": 1.5833333333333333e-05, "epoch": 0.5757575757575758, "percentage": 8.23, "elapsed_time": "0:12:06", "remaining_time": "2:15:00"}
+{"current_steps": 20, "total_steps": 231, "loss": 0.5622, "lr": 1.6666666666666667e-05, "epoch": 0.6060606060606061, "percentage": 8.66, "elapsed_time": "0:12:45", "remaining_time": "2:14:31"}
+{"current_steps": 21, "total_steps": 231, "loss": 0.5481, "lr": 1.7500000000000002e-05, "epoch": 0.6363636363636364, "percentage": 9.09, "elapsed_time": "0:13:20", "remaining_time": "2:13:25"}
+{"current_steps": 22, "total_steps": 231, "loss": 0.5595, "lr": 1.8333333333333333e-05, "epoch": 0.6666666666666666, "percentage": 9.52, "elapsed_time": "0:13:58", "remaining_time": "2:12:42"}
+{"current_steps": 23, "total_steps": 231, "loss": 0.5446, "lr": 1.916666666666667e-05, "epoch": 0.696969696969697, "percentage": 9.96, "elapsed_time": "0:14:35", "remaining_time": "2:12:00"}
+{"current_steps": 24, "total_steps": 231, "loss": 0.5457, "lr": 2e-05, "epoch": 0.7272727272727273, "percentage": 10.39, "elapsed_time": "0:15:16", "remaining_time": "2:11:46"}
+{"current_steps": 25, "total_steps": 231, "loss": 0.5346, "lr": 1.999884834944106e-05, "epoch": 0.7575757575757576, "percentage": 10.82, "elapsed_time": "0:15:53", "remaining_time": "2:10:53"}
+{"current_steps": 26, "total_steps": 231, "loss": 0.5221, "lr": 1.9995393663024054e-05, "epoch": 0.7878787878787878, "percentage": 11.26, "elapsed_time": "0:16:29", "remaining_time": "2:09:59"}
+{"current_steps": 27, "total_steps": 231, "loss": 0.5274, "lr": 1.9989636736467278e-05, "epoch": 0.8181818181818182, "percentage": 11.69, "elapsed_time": "0:17:06", "remaining_time": "2:09:16"}
+{"current_steps": 28, "total_steps": 231, "loss": 0.5231, "lr": 1.9981578895764272e-05, "epoch": 0.8484848484848485, "percentage": 12.12, "elapsed_time": "0:17:47", "remaining_time": "2:08:56"}
+{"current_steps": 29, "total_steps": 231, "loss": 0.5221, "lr": 1.9971221996878395e-05, "epoch": 0.8787878787878788, "percentage": 12.55, "elapsed_time": "0:18:25", "remaining_time": "2:08:19"}
+{"current_steps": 30, "total_steps": 231, "loss": 0.5332, "lr": 1.9958568425315316e-05, "epoch": 0.9090909090909091, "percentage": 12.99, "elapsed_time": "0:19:05", "remaining_time": "2:07:57"}
+{"current_steps": 31, "total_steps": 231, "loss": 0.5204, "lr": 1.9943621095573588e-05, "epoch": 0.9393939393939394, "percentage": 13.42, "elapsed_time": "0:19:41", "remaining_time": "2:07:02"}
+{"current_steps": 32, "total_steps": 231, "loss": 0.5183, "lr": 1.9926383450473344e-05, "epoch": 0.9696969696969697, "percentage": 13.85, "elapsed_time": "0:20:15", "remaining_time": "2:06:00"}
+{"current_steps": 33, "total_steps": 231, "loss": 0.5098, "lr": 1.9906859460363307e-05, "epoch": 1.0, "percentage": 14.29, "elapsed_time": "0:20:53", "remaining_time": "2:05:21"}
+{"current_steps": 34, "total_steps": 231, "loss": 0.4833, "lr": 1.9885053622206305e-05, "epoch": 1.0303030303030303, "percentage": 14.72, "elapsed_time": "0:22:07", "remaining_time": "2:08:12"}
+{"current_steps": 35, "total_steps": 231, "loss": 0.4789, "lr": 1.986097095854347e-05, "epoch": 1.0606060606060606, "percentage": 15.15, "elapsed_time": "0:22:47", "remaining_time": "2:07:38"}
+{"current_steps": 36, "total_steps": 231, "loss": 0.481, "lr": 1.9834617016337424e-05, "epoch": 1.0909090909090908, "percentage": 15.58, "elapsed_time": "0:23:27", "remaining_time": "2:07:02"}
+{"current_steps": 37, "total_steps": 231, "loss": 0.4686, "lr": 1.9805997865694616e-05, "epoch": 1.121212121212121, "percentage": 16.02, "elapsed_time": "0:24:04", "remaining_time": "2:06:12"}
+{"current_steps": 38, "total_steps": 231, "loss": 0.488, "lr": 1.9775120098467212e-05, "epoch": 1.1515151515151516, "percentage": 16.45, "elapsed_time": "0:24:39", "remaining_time": "2:05:15"}
+{"current_steps": 39, "total_steps": 231, "loss": 0.4706, "lr": 1.9741990826734793e-05, "epoch": 1.1818181818181819, "percentage": 16.88, "elapsed_time": "0:25:18", "remaining_time": "2:04:34"}
+{"current_steps": 40, "total_steps": 231, "loss": 0.4909, "lr": 1.970661768116622e-05, "epoch": 1.2121212121212122, "percentage": 17.32, "elapsed_time": "0:25:54", "remaining_time": "2:03:45"}
+{"current_steps": 41, "total_steps": 231, "loss": 0.4816, "lr": 1.9669008809262064e-05, "epoch": 1.2424242424242424, "percentage": 17.75, "elapsed_time": "0:26:32", "remaining_time": "2:03:00"}
+{"current_steps": 42, "total_steps": 231, "loss": 0.4856, "lr": 1.9629172873477995e-05, "epoch": 1.2727272727272727, "percentage": 18.18, "elapsed_time": "0:27:11", "remaining_time": "2:02:23"}
+{"current_steps": 43, "total_steps": 231, "loss": 0.4848, "lr": 1.9587119049229558e-05, "epoch": 1.303030303030303, "percentage": 18.61, "elapsed_time": "0:27:50", "remaining_time": "2:01:44"}
+{"current_steps": 44, "total_steps": 231, "loss": 0.4734, "lr": 1.954285702277879e-05, "epoch": 1.3333333333333333, "percentage": 19.05, "elapsed_time": "0:28:28", "remaining_time": "2:01:00"}
+{"current_steps": 45, "total_steps": 231, "loss": 0.4844, "lr": 1.9496396989003195e-05, "epoch": 1.3636363636363638, "percentage": 19.48, "elapsed_time": "0:29:05", "remaining_time": "2:00:14"}
+{"current_steps": 46, "total_steps": 231, "loss": 0.4597, "lr": 1.944774964904754e-05, "epoch": 1.393939393939394, "percentage": 19.91, "elapsed_time": "0:29:43", "remaining_time": "1:59:31"}
+{"current_steps": 47, "total_steps": 231, "loss": 0.4662, "lr": 1.9396926207859085e-05, "epoch": 1.4242424242424243, "percentage": 20.35, "elapsed_time": "0:30:19", "remaining_time": "1:58:43"}
+{"current_steps": 48, "total_steps": 231, "loss": 0.4398, "lr": 1.9343938371606714e-05, "epoch": 1.4545454545454546, "percentage": 20.78, "elapsed_time": "0:30:54", "remaining_time": "1:57:51"}
+{"current_steps": 49, "total_steps": 231, "loss": 0.4594, "lr": 1.9288798344984673e-05, "epoch": 1.4848484848484849, "percentage": 21.21, "elapsed_time": "0:31:34", "remaining_time": "1:57:16"}
+{"current_steps": 50, "total_steps": 231, "loss": 0.4627, "lr": 1.9231518828401458e-05, "epoch": 1.5151515151515151, "percentage": 21.65, "elapsed_time": "0:32:14", "remaining_time": "1:56:42"}
+{"current_steps": 51, "total_steps": 231, "loss": 0.4531, "lr": 1.917211301505453e-05, "epoch": 1.5454545454545454, "percentage": 22.08, "elapsed_time": "0:32:54", "remaining_time": "1:56:07"}
+{"current_steps": 52, "total_steps": 231, "loss": 0.451, "lr": 1.911059458789152e-05, "epoch": 1.5757575757575757, "percentage": 22.51, "elapsed_time": "0:33:29", "remaining_time": "1:55:16"}
+{"current_steps": 53, "total_steps": 231, "loss": 0.4679, "lr": 1.9046977716458627e-05, "epoch": 1.606060606060606, "percentage": 22.94, "elapsed_time": "0:34:06", "remaining_time": "1:54:34"}
+{"current_steps": 54, "total_steps": 231, "loss": 0.4649, "lr": 1.8981277053636963e-05, "epoch": 1.6363636363636362, "percentage": 23.38, "elapsed_time": "0:34:45", "remaining_time": "1:53:54"}
+{"current_steps": 55, "total_steps": 231, "loss": 0.4432, "lr": 1.891350773226754e-05, "epoch": 1.6666666666666665, "percentage": 23.81, "elapsed_time": "0:35:22", "remaining_time": "1:53:10"}
+{"current_steps": 56, "total_steps": 231, "loss": 0.454, "lr": 1.8843685361665724e-05, "epoch": 1.696969696969697, "percentage": 24.24, "elapsed_time": "0:35:59", "remaining_time": "1:52:28"}
+{"current_steps": 57, "total_steps": 231, "loss": 0.4644, "lr": 1.8771826024025944e-05, "epoch": 1.7272727272727273, "percentage": 24.68, "elapsed_time": "0:36:36", "remaining_time": "1:51:45"}
+{"current_steps": 58, "total_steps": 231, "loss": 0.4628, "lr": 1.8697946270717468e-05, "epoch": 1.7575757575757576, "percentage": 25.11, "elapsed_time": "0:37:12", "remaining_time": "1:50:59"}
+{"current_steps": 59, "total_steps": 231, "loss": 0.4599, "lr": 1.8622063118472135e-05, "epoch": 1.7878787878787878, "percentage": 25.54, "elapsed_time": "0:37:51", "remaining_time": "1:50:21"}
+{"current_steps": 60, "total_steps": 231, "loss": 0.4597, "lr": 1.8544194045464888e-05, "epoch": 1.8181818181818183, "percentage": 25.97, "elapsed_time": "0:38:29", "remaining_time": "1:49:41"}
+{"current_steps": 61, "total_steps": 231, "loss": 0.4659, "lr": 1.8464356987288012e-05, "epoch": 1.8484848484848486, "percentage": 26.41, "elapsed_time": "0:39:06", "remaining_time": "1:48:59"}
+{"current_steps": 62, "total_steps": 231, "loss": 0.4653, "lr": 1.8382570332820045e-05, "epoch": 1.878787878787879, "percentage": 26.84, "elapsed_time": "0:39:40", "remaining_time": "1:48:08"}
+{"current_steps": 63, "total_steps": 231, "loss": 0.4712, "lr": 1.8298852919990254e-05, "epoch": 1.9090909090909092, "percentage": 27.27, "elapsed_time": "0:40:18", "remaining_time": "1:47:30"}
+{"current_steps": 64, "total_steps": 231, "loss": 0.4659, "lr": 1.821322403143969e-05, "epoch": 1.9393939393939394, "percentage": 27.71, "elapsed_time": "0:40:58", "remaining_time": "1:46:54"}
+{"current_steps": 65, "total_steps": 231, "loss": 0.4617, "lr": 1.812570339007983e-05, "epoch": 1.9696969696969697, "percentage": 28.14, "elapsed_time": "0:41:34", "remaining_time": "1:46:10"}
+{"current_steps": 66, "total_steps": 231, "loss": 0.4556, "lr": 1.8036311154549783e-05, "epoch": 2.0, "percentage": 28.57, "elapsed_time": "0:42:15", "remaining_time": "1:45:37"}
+{"current_steps": 67, "total_steps": 231, "loss": 0.4028, "lr": 1.7945067914573147e-05, "epoch": 2.0303030303030303, "percentage": 29.0, "elapsed_time": "0:43:24", "remaining_time": "1:46:16"}
+{"current_steps": 68, "total_steps": 231, "loss": 0.4072, "lr": 1.7851994686215592e-05, "epoch": 2.0606060606060606, "percentage": 29.44, "elapsed_time": "0:44:03", "remaining_time": "1:45:36"}
+{"current_steps": 69, "total_steps": 231, "loss": 0.4112, "lr": 1.77571129070442e-05, "epoch": 2.090909090909091, "percentage": 29.87, "elapsed_time": "0:44:40", "remaining_time": "1:44:54"}
+{"current_steps": 70, "total_steps": 231, "loss": 0.424, "lr": 1.766044443118978e-05, "epoch": 2.121212121212121, "percentage": 30.3, "elapsed_time": "0:45:20", "remaining_time": "1:44:16"}
+{"current_steps": 71, "total_steps": 231, "loss": 0.4197, "lr": 1.7562011524313187e-05, "epoch": 2.1515151515151514, "percentage": 30.74, "elapsed_time": "0:45:58", "remaining_time": "1:43:35"}
+{"current_steps": 72, "total_steps": 231, "loss": 0.412, "lr": 1.7461836858476858e-05, "epoch": 2.1818181818181817, "percentage": 31.17, "elapsed_time": "0:46:35", "remaining_time": "1:42:52"}
+{"current_steps": 73, "total_steps": 231, "loss": 0.4163, "lr": 1.7359943506922775e-05, "epoch": 2.212121212121212, "percentage": 31.6, "elapsed_time": "0:47:13", "remaining_time": "1:42:12"}
+{"current_steps": 74, "total_steps": 231, "loss": 0.3957, "lr": 1.725635493875799e-05, "epoch": 2.242424242424242, "percentage": 32.03, "elapsed_time": "0:47:49", "remaining_time": "1:41:28"}
+{"current_steps": 75, "total_steps": 231, "loss": 0.3962, "lr": 1.7151095013548996e-05, "epoch": 2.2727272727272725, "percentage": 32.47, "elapsed_time": "0:48:28", "remaining_time": "1:40:48"}
+{"current_steps": 76, "total_steps": 231, "loss": 0.4197, "lr": 1.7044187975826126e-05, "epoch": 2.303030303030303, "percentage": 32.9, "elapsed_time": "0:49:04", "remaining_time": "1:40:05"}
+{"current_steps": 77, "total_steps": 231, "loss": 0.4174, "lr": 1.693565844949933e-05, "epoch": 2.3333333333333335, "percentage": 33.33, "elapsed_time": "0:49:43", "remaining_time": "1:39:27"}
+{"current_steps": 78, "total_steps": 231, "loss": 0.4096, "lr": 1.6825531432186545e-05, "epoch": 2.3636363636363638, "percentage": 33.77, "elapsed_time": "0:50:21", "remaining_time": "1:38:46"}
+{"current_steps": 79, "total_steps": 231, "loss": 0.4013, "lr": 1.671383228945597e-05, "epoch": 2.393939393939394, "percentage": 34.2, "elapsed_time": "0:50:58", "remaining_time": "1:38:04"}
+{"current_steps": 80, "total_steps": 231, "loss": 0.4135, "lr": 1.6600586748983642e-05, "epoch": 2.4242424242424243, "percentage": 34.63, "elapsed_time": "0:51:35", "remaining_time": "1:37:21"}
+{"current_steps": 81, "total_steps": 231, "loss": 0.4058, "lr": 1.648582089462756e-05, "epoch": 2.4545454545454546, "percentage": 35.06, "elapsed_time": "0:52:10", "remaining_time": "1:36:37"}
+{"current_steps": 82, "total_steps": 231, "loss": 0.3965, "lr": 1.6369561160419783e-05, "epoch": 2.484848484848485, "percentage": 35.5, "elapsed_time": "0:52:50", "remaining_time": "1:36:01"}
+{"current_steps": 83, "total_steps": 231, "loss": 0.3977, "lr": 1.625183432447789e-05, "epoch": 2.515151515151515, "percentage": 35.93, "elapsed_time": "0:53:26", "remaining_time": "1:35:18"}
+{"current_steps": 84, "total_steps": 231, "loss": 0.4097, "lr": 1.6132667502837164e-05, "epoch": 2.5454545454545454, "percentage": 36.36, "elapsed_time": "0:54:03", "remaining_time": "1:34:36"}
+{"current_steps": 85, "total_steps": 231, "loss": 0.4036, "lr": 1.6012088143204953e-05, "epoch": 2.5757575757575757, "percentage": 36.8, "elapsed_time": "0:54:42", "remaining_time": "1:33:57"}
+{"current_steps": 86, "total_steps": 231, "loss": 0.4025, "lr": 1.589012401863864e-05, "epoch": 2.606060606060606, "percentage": 37.23, "elapsed_time": "0:55:20", "remaining_time": "1:33:17"}
+{"current_steps": 87, "total_steps": 231, "loss": 0.3929, "lr": 1.5766803221148676e-05, "epoch": 2.6363636363636362, "percentage": 37.66, "elapsed_time": "0:55:58", "remaining_time": "1:32:38"}
+{"current_steps": 88, "total_steps": 231, "loss": 0.4087, "lr": 1.5642154155228124e-05, "epoch": 2.6666666666666665, "percentage": 38.1, "elapsed_time": "0:56:36", "remaining_time": "1:31:59"}
+{"current_steps": 89, "total_steps": 231, "loss": 0.4036, "lr": 1.5516205531310272e-05, "epoch": 2.6969696969696972, "percentage": 38.53, "elapsed_time": "0:57:14", "remaining_time": "1:31:19"}
+{"current_steps": 90, "total_steps": 231, "loss": 0.4014, "lr": 1.538898635915576e-05, "epoch": 2.7272727272727275, "percentage": 38.96, "elapsed_time": "0:57:53", "remaining_time": "1:30:41"}
+{"current_steps": 91, "total_steps": 231, "loss": 0.3989, "lr": 1.526052594117071e-05, "epoch": 2.757575757575758, "percentage": 39.39, "elapsed_time": "0:58:30", "remaining_time": "1:30:00"}
+{"current_steps": 92, "total_steps": 231, "loss": 0.3953, "lr": 1.513085386565758e-05, "epoch": 2.787878787878788, "percentage": 39.83, "elapsed_time": "0:59:07", "remaining_time": "1:29:20"}
+{"current_steps": 93, "total_steps": 231, "loss": 0.3982, "lr": 1.5000000000000002e-05, "epoch": 2.8181818181818183, "percentage": 40.26, "elapsed_time": "0:59:46", "remaining_time": "1:28:41"}
+{"current_steps": 94, "total_steps": 231, "loss": 0.4105, "lr": 1.4867994483783485e-05, "epoch": 2.8484848484848486, "percentage": 40.69, "elapsed_time": "1:00:23", "remaining_time": "1:28:01"}
+{"current_steps": 95, "total_steps": 231, "loss": 0.397, "lr": 1.4734867721853341e-05, "epoch": 2.878787878787879, "percentage": 41.13, "elapsed_time": "1:01:01", "remaining_time": "1:27:21"}
+{"current_steps": 96, "total_steps": 231, "loss": 0.4092, "lr": 1.4600650377311523e-05, "epoch": 2.909090909090909, "percentage": 41.56, "elapsed_time": "1:01:37", "remaining_time": "1:26:40"}
+{"current_steps": 97, "total_steps": 231, "loss": 0.4152, "lr": 1.4465373364454001e-05, "epoch": 2.9393939393939394, "percentage": 41.99, "elapsed_time": "1:02:16", "remaining_time": "1:26:01"}
+{"current_steps": 98, "total_steps": 231, "loss": 0.4021, "lr": 1.4329067841650274e-05, "epoch": 2.9696969696969697, "percentage": 42.42, "elapsed_time": "1:02:52", "remaining_time": "1:25:20"}
+{"current_steps": 99, "total_steps": 231, "loss": 0.3881, "lr": 1.4191765204166643e-05, "epoch": 3.0, "percentage": 42.86, "elapsed_time": "1:03:29", "remaining_time": "1:24:39"}
+{"current_steps": 100, "total_steps": 231, "loss": 0.3615, "lr": 1.4053497076934948e-05, "epoch": 3.0303030303030303, "percentage": 43.29, "elapsed_time": "1:04:45", "remaining_time": "1:24:49"}
+{"current_steps": 101, "total_steps": 231, "loss": 0.3647, "lr": 1.3914295307268396e-05, "epoch": 3.0606060606060606, "percentage": 43.72, "elapsed_time": "1:05:23", "remaining_time": "1:24:09"}
+{"current_steps": 102, "total_steps": 231, "loss": 0.3607, "lr": 1.3774191957526144e-05, "epoch": 3.090909090909091, "percentage": 44.16, "elapsed_time": "1:06:02", "remaining_time": "1:23:31"}
+{"current_steps": 103, "total_steps": 231, "loss": 0.3542, "lr": 1.3633219297728415e-05, "epoch": 3.121212121212121, "percentage": 44.59, "elapsed_time": "1:06:41", "remaining_time": "1:22:52"}
+{"current_steps": 104, "total_steps": 231, "loss": 0.3622, "lr": 1.3491409798123687e-05, "epoch": 3.1515151515151514, "percentage": 45.02, "elapsed_time": "1:07:19", "remaining_time": "1:22:12"}
+{"current_steps": 105, "total_steps": 231, "loss": 0.3641, "lr": 1.3348796121709862e-05, "epoch": 3.1818181818181817, "percentage": 45.45, "elapsed_time": "1:07:56", "remaining_time": "1:21:31"}
+{"current_steps": 106, "total_steps": 231, "loss": 0.3357, "lr": 1.3205411116710973e-05, "epoch": 3.212121212121212, "percentage": 45.89, "elapsed_time": "1:08:30", "remaining_time": "1:20:47"}
+{"current_steps": 107, "total_steps": 231, "loss": 0.3528, "lr": 1.3061287809011243e-05, "epoch": 3.242424242424242, "percentage": 46.32, "elapsed_time": "1:09:10", "remaining_time": "1:20:09"}
+{"current_steps": 108, "total_steps": 231, "loss": 0.3467, "lr": 1.291645939454825e-05, "epoch": 3.2727272727272725, "percentage": 46.75, "elapsed_time": "1:09:48", "remaining_time": "1:19:30"}
+{"current_steps": 109, "total_steps": 231, "loss": 0.3548, "lr": 1.277095923166689e-05, "epoch": 3.303030303030303, "percentage": 47.19, "elapsed_time": "1:10:26", "remaining_time": "1:18:50"}
+{"current_steps": 110, "total_steps": 231, "loss": 0.3628, "lr": 1.2624820833435939e-05, "epoch": 3.3333333333333335, "percentage": 47.62, "elapsed_time": "1:11:01", "remaining_time": "1:18:07"}
+{"current_steps": 111, "total_steps": 231, "loss": 0.3445, "lr": 1.2478077859929e-05, "epoch": 3.3636363636363638, "percentage": 48.05, "elapsed_time": "1:11:39", "remaining_time": "1:17:27"}
+{"current_steps": 112, "total_steps": 231, "loss": 0.3518, "lr": 1.2330764110471567e-05, "epoch": 3.393939393939394, "percentage": 48.48, "elapsed_time": "1:12:15", "remaining_time": "1:16:46"}
+{"current_steps": 113, "total_steps": 231, "loss": 0.349, "lr": 1.2182913515856016e-05, "epoch": 3.4242424242424243, "percentage": 48.92, "elapsed_time": "1:12:55", "remaining_time": "1:16:08"}
+{"current_steps": 114, "total_steps": 231, "loss": 0.3568, "lr": 1.2034560130526341e-05, "epoch": 3.4545454545454546, "percentage": 49.35, "elapsed_time": "1:13:32", "remaining_time": "1:15:29"}
+{"current_steps": 115, "total_steps": 231, "loss": 0.3621, "lr": 1.1885738124734359e-05, "epoch": 3.484848484848485, "percentage": 49.78, "elapsed_time": "1:14:11", "remaining_time": "1:14:50"}
+{"current_steps": 116, "total_steps": 231, "loss": 0.3653, "lr": 1.1736481776669307e-05, "epoch": 3.515151515151515, "percentage": 50.22, "elapsed_time": "1:14:49", "remaining_time": "1:14:11"}
+{"current_steps": 117, "total_steps": 231, "loss": 0.3434, "lr": 1.1586825464562515e-05, "epoch": 3.5454545454545454, "percentage": 50.65, "elapsed_time": "1:15:26", "remaining_time": "1:13:30"}
+{"current_steps": 118, "total_steps": 231, "loss": 0.3474, "lr": 1.1436803658769082e-05, "epoch": 3.5757575757575757, "percentage": 51.08, "elapsed_time": "1:16:05", "remaining_time": "1:12:51"}
+{"current_steps": 119, "total_steps": 231, "loss": 0.3526, "lr": 1.1286450913828313e-05, "epoch": 3.606060606060606, "percentage": 51.52, "elapsed_time": "1:16:43", "remaining_time": "1:12:12"}
+{"current_steps": 120, "total_steps": 231, "loss": 0.3551, "lr": 1.113580186050475e-05, "epoch": 3.6363636363636362, "percentage": 51.95, "elapsed_time": "1:17:19", "remaining_time": "1:11:31"}
+{"current_steps": 121, "total_steps": 231, "loss": 0.3568, "lr": 1.0984891197811686e-05, "epoch": 3.6666666666666665, "percentage": 52.38, "elapsed_time": "1:17:57", "remaining_time": "1:10:52"}
+{"current_steps": 122, "total_steps": 231, "loss": 0.3449, "lr": 1.0833753685018935e-05, "epoch": 3.6969696969696972, "percentage": 52.81, "elapsed_time": "1:18:35", "remaining_time": "1:10:13"}
+{"current_steps": 123, "total_steps": 231, "loss": 0.362, "lr": 1.0682424133646712e-05, "epoch": 3.7272727272727275, "percentage": 53.25, "elapsed_time": "1:19:14", "remaining_time": "1:09:34"}
+{"current_steps": 124, "total_steps": 231, "loss": 0.3625, "lr": 1.0530937399447496e-05, "epoch": 3.757575757575758, "percentage": 53.68, "elapsed_time": "1:19:49", "remaining_time": "1:08:52"}
+{"current_steps": 125, "total_steps": 231, "loss": 0.3571, "lr": 1.0379328374377715e-05, "epoch": 3.787878787878788, "percentage": 54.11, "elapsed_time": "1:20:26", "remaining_time": "1:08:12"}
+{"current_steps": 126, "total_steps": 231, "loss": 0.3547, "lr": 1.0227631978561057e-05, "epoch": 3.8181818181818183, "percentage": 54.55, "elapsed_time": "1:21:04", "remaining_time": "1:07:34"}
+{"current_steps": 127, "total_steps": 231, "loss": 0.3623, "lr": 1.0075883152245334e-05, "epoch": 3.8484848484848486, "percentage": 54.98, "elapsed_time": "1:21:41", "remaining_time": "1:06:54"}
+{"current_steps": 128, "total_steps": 231, "loss": 0.342, "lr": 9.92411684775467e-06, "epoch": 3.878787878787879, "percentage": 55.41, "elapsed_time": "1:22:18", "remaining_time": "1:06:14"}
+{"current_steps": 129, "total_steps": 231, "loss": 0.3482, "lr": 9.772368021438943e-06, "epoch": 3.909090909090909, "percentage": 55.84, "elapsed_time": "1:22:55", "remaining_time": "1:05:34"}
+{"current_steps": 130, "total_steps": 231, "loss": 0.3522, "lr": 9.620671625622287e-06, "epoch": 3.9393939393939394, "percentage": 56.28, "elapsed_time": "1:23:32", "remaining_time": "1:04:54"}
+{"current_steps": 131, "total_steps": 231, "loss": 0.349, "lr": 9.469062600552509e-06, "epoch": 3.9696969696969697, "percentage": 56.71, "elapsed_time": "1:24:09", "remaining_time": "1:04:14"}
+{"current_steps": 132, "total_steps": 231, "loss": 0.3557, "lr": 9.317575866353293e-06, "epoch": 4.0, "percentage": 57.14, "elapsed_time": "1:24:47", "remaining_time": "1:03:35"}
+{"current_steps": 133, "total_steps": 231, "loss": 0.3101, "lr": 9.166246314981066e-06, "epoch": 4.03030303030303, "percentage": 57.58, "elapsed_time": "1:26:01", "remaining_time": "1:03:23"}
+{"current_steps": 134, "total_steps": 231, "loss": 0.309, "lr": 9.015108802188314e-06, "epoch": 4.0606060606060606, "percentage": 58.01, "elapsed_time": "1:26:39", "remaining_time": "1:02:44"}
+{"current_steps": 135, "total_steps": 231, "loss": 0.3245, "lr": 8.86419813949525e-06, "epoch": 4.090909090909091, "percentage": 58.44, "elapsed_time": "1:27:15", "remaining_time": "1:02:03"}
+{"current_steps": 136, "total_steps": 231, "loss": 0.3191, "lr": 8.71354908617169e-06, "epoch": 4.121212121212121, "percentage": 58.87, "elapsed_time": "1:27:53", "remaining_time": "1:01:24"}
+{"current_steps": 137, "total_steps": 231, "loss": 0.3044, "lr": 8.56319634123092e-06, "epoch": 4.151515151515151, "percentage": 59.31, "elapsed_time": "1:28:30", "remaining_time": "1:00:43"}
+{"current_steps": 138, "total_steps": 231, "loss": 0.3031, "lr": 8.413174535437486e-06, "epoch": 4.181818181818182, "percentage": 59.74, "elapsed_time": "1:29:08", "remaining_time": "1:00:04"}
+{"current_steps": 139, "total_steps": 231, "loss": 0.3132, "lr": 8.263518223330698e-06, "epoch": 4.212121212121212, "percentage": 60.17, "elapsed_time": "1:29:43", "remaining_time": "0:59:23"}
+{"current_steps": 140, "total_steps": 231, "loss": 0.309, "lr": 8.114261875265643e-06, "epoch": 4.242424242424242, "percentage": 60.61, "elapsed_time": "1:30:21", "remaining_time": "0:58:44"}
+{"current_steps": 141, "total_steps": 231, "loss": 0.3141, "lr": 7.965439869473664e-06, "epoch": 4.2727272727272725, "percentage": 61.04, "elapsed_time": "1:30:59", "remaining_time": "0:58:05"}
+{"current_steps": 142, "total_steps": 231, "loss": 0.3123, "lr": 7.817086484143987e-06, "epoch": 4.303030303030303, "percentage": 61.47, "elapsed_time": "1:31:37", "remaining_time": "0:57:25"}
+{"current_steps": 143, "total_steps": 231, "loss": 0.299, "lr": 7.669235889528436e-06, "epoch": 4.333333333333333, "percentage": 61.9, "elapsed_time": "1:32:16", "remaining_time": "0:56:47"}
+{"current_steps": 144, "total_steps": 231, "loss": 0.3163, "lr": 7.521922140071003e-06, "epoch": 4.363636363636363, "percentage": 62.34, "elapsed_time": "1:32:55", "remaining_time": "0:56:08"}
+{"current_steps": 145, "total_steps": 231, "loss": 0.3012, "lr": 7.375179166564062e-06, "epoch": 4.393939393939394, "percentage": 62.77, "elapsed_time": "1:33:32", "remaining_time": "0:55:28"}
+{"current_steps": 146, "total_steps": 231, "loss": 0.3124, "lr": 7.2290407683331154e-06, "epoch": 4.424242424242424, "percentage": 63.2, "elapsed_time": "1:34:10", "remaining_time": "0:54:49"}
+{"current_steps": 147, "total_steps": 231, "loss": 0.2999, "lr": 7.0835406054517505e-06, "epoch": 4.454545454545454, "percentage": 63.64, "elapsed_time": "1:34:48", "remaining_time": "0:54:10"}
+{"current_steps": 148, "total_steps": 231, "loss": 0.3166, "lr": 6.93871219098876e-06, "epoch": 4.484848484848484, "percentage": 64.07, "elapsed_time": "1:35:25", "remaining_time": "0:53:31"}
+{"current_steps": 149, "total_steps": 231, "loss": 0.3134, "lr": 6.79458888328903e-06, "epoch": 4.515151515151516, "percentage": 64.5, "elapsed_time": "1:36:04", "remaining_time": "0:52:52"}
+{"current_steps": 150, "total_steps": 231, "loss": 0.3171, "lr": 6.651203878290139e-06, "epoch": 4.545454545454545, "percentage": 64.94, "elapsed_time": "1:36:43", "remaining_time": "0:52:13"}
+{"current_steps": 151, "total_steps": 231, "loss": 0.3101, "lr": 6.508590201876317e-06, "epoch": 4.575757575757576, "percentage": 65.37, "elapsed_time": "1:37:19", "remaining_time": "0:51:33"}
+{"current_steps": 152, "total_steps": 231, "loss": 0.3043, "lr": 6.366780702271589e-06, "epoch": 4.606060606060606, "percentage": 65.8, "elapsed_time": "1:37:55", "remaining_time": "0:50:53"}
+{"current_steps": 153, "total_steps": 231, "loss": 0.31, "lr": 6.225808042473857e-06, "epoch": 4.636363636363637, "percentage": 66.23, "elapsed_time": "1:38:33", "remaining_time": "0:50:14"}
+{"current_steps": 154, "total_steps": 231, "loss": 0.2977, "lr": 6.085704692731609e-06, "epoch": 4.666666666666667, "percentage": 66.67, "elapsed_time": "1:39:09", "remaining_time": "0:49:34"}
+{"current_steps": 155, "total_steps": 231, "loss": 0.3174, "lr": 5.946502923065054e-06, "epoch": 4.696969696969697, "percentage": 67.1, "elapsed_time": "1:39:45", "remaining_time": "0:48:54"}
+{"current_steps": 156, "total_steps": 231, "loss": 0.3138, "lr": 5.8082347958333625e-06, "epoch": 4.7272727272727275, "percentage": 67.53, "elapsed_time": "1:40:24", "remaining_time": "0:48:16"}
+{"current_steps": 157, "total_steps": 231, "loss": 0.3201, "lr": 5.670932158349732e-06, "epoch": 4.757575757575758, "percentage": 67.97, "elapsed_time": "1:41:01", "remaining_time": "0:47:37"}
+{"current_steps": 158, "total_steps": 231, "loss": 0.3055, "lr": 5.534626635546e-06, "epoch": 4.787878787878788, "percentage": 68.4, "elapsed_time": "1:41:39", "remaining_time": "0:46:57"}
+{"current_steps": 159, "total_steps": 231, "loss": 0.3047, "lr": 5.399349622688479e-06, "epoch": 4.818181818181818, "percentage": 68.83, "elapsed_time": "1:42:16", "remaining_time": "0:46:18"}
+{"current_steps": 160, "total_steps": 231, "loss": 0.3079, "lr": 5.2651322781466606e-06, "epoch": 4.848484848484849, "percentage": 69.26, "elapsed_time": "1:42:55", "remaining_time": "0:45:40"}
+{"current_steps": 161, "total_steps": 231, "loss": 0.3043, "lr": 5.132005516216512e-06, "epoch": 4.878787878787879, "percentage": 69.7, "elapsed_time": "1:43:33", "remaining_time": "0:45:01"}
+{"current_steps": 162, "total_steps": 231, "loss": 0.3132, "lr": 5.000000000000003e-06, "epoch": 4.909090909090909, "percentage": 70.13, "elapsed_time": "1:44:07", "remaining_time": "0:44:21"}
+{"current_steps": 163, "total_steps": 231, "loss": 0.3133, "lr": 4.869146134342426e-06, "epoch": 4.9393939393939394, "percentage": 70.56, "elapsed_time": "1:44:45", "remaining_time": "0:43:42"}
+{"current_steps": 164, "total_steps": 231, "loss": 0.3133, "lr": 4.739474058829288e-06, "epoch": 4.96969696969697, "percentage": 71.0, "elapsed_time": "1:45:22", "remaining_time": "0:43:03"}
+{"current_steps": 165, "total_steps": 231, "loss": 0.3062, "lr": 4.611013640844245e-06, "epoch": 5.0, "percentage": 71.43, "elapsed_time": "1:46:00", "remaining_time": "0:42:24"}
+{"current_steps": 166, "total_steps": 231, "loss": 0.2825, "lr": 4.483794468689728e-06, "epoch": 5.03030303030303, "percentage": 71.86, "elapsed_time": "1:47:21", "remaining_time": "0:42:02"}
+{"current_steps": 167, "total_steps": 231, "loss": 0.2881, "lr": 4.357845844771881e-06, "epoch": 5.0606060606060606, "percentage": 72.29, "elapsed_time": "1:47:59", "remaining_time": "0:41:23"}
+{"current_steps": 168, "total_steps": 231, "loss": 0.2864, "lr": 4.2331967788513295e-06, "epoch": 5.090909090909091, "percentage": 72.73, "elapsed_time": "1:48:37", "remaining_time": "0:40:44"}
+{"current_steps": 169, "total_steps": 231, "loss": 0.2857, "lr": 4.109875981361363e-06, "epoch": 5.121212121212121, "percentage": 73.16, "elapsed_time": "1:49:16", "remaining_time": "0:40:05"}
+{"current_steps": 170, "total_steps": 231, "loss": 0.285, "lr": 3.987911856795047e-06, "epoch": 5.151515151515151, "percentage": 73.59, "elapsed_time": "1:49:55", "remaining_time": "0:39:26"}
+{"current_steps": 171, "total_steps": 231, "loss": 0.2824, "lr": 3.867332497162836e-06, "epoch": 5.181818181818182, "percentage": 74.03, "elapsed_time": "1:50:31", "remaining_time": "0:38:46"}
+{"current_steps": 172, "total_steps": 231, "loss": 0.2711, "lr": 3.748165675522113e-06, "epoch": 5.212121212121212, "percentage": 74.46, "elapsed_time": "1:51:10", "remaining_time": "0:38:08"}
+{"current_steps": 173, "total_steps": 231, "loss": 0.2724, "lr": 3.630438839580217e-06, "epoch": 5.242424242424242, "percentage": 74.89, "elapsed_time": "1:51:46", "remaining_time": "0:37:28"}
+{"current_steps": 174, "total_steps": 231, "loss": 0.2873, "lr": 3.5141791053724405e-06, "epoch": 5.2727272727272725, "percentage": 75.32, "elapsed_time": "1:52:23", "remaining_time": "0:36:49"}
+{"current_steps": 175, "total_steps": 231, "loss": 0.2735, "lr": 3.399413251016359e-06, "epoch": 5.303030303030303, "percentage": 75.76, "elapsed_time": "1:53:00", "remaining_time": "0:36:09"}
+{"current_steps": 176, "total_steps": 231, "loss": 0.2687, "lr": 3.2861677105440335e-06, "epoch": 5.333333333333333, "percentage": 76.19, "elapsed_time": "1:53:37", "remaining_time": "0:35:30"}
+{"current_steps": 177, "total_steps": 231, "loss": 0.2745, "lr": 3.174468567813461e-06, "epoch": 5.363636363636363, "percentage": 76.62, "elapsed_time": "1:54:12", "remaining_time": "0:34:50"}
+{"current_steps": 178, "total_steps": 231, "loss": 0.2758, "lr": 3.0643415505006733e-06, "epoch": 5.393939393939394, "percentage": 77.06, "elapsed_time": "1:54:49", "remaining_time": "0:34:11"}
+{"current_steps": 179, "total_steps": 231, "loss": 0.278, "lr": 2.9558120241738786e-06, "epoch": 5.424242424242424, "percentage": 77.49, "elapsed_time": "1:55:27", "remaining_time": "0:33:32"}
+{"current_steps": 180, "total_steps": 231, "loss": 0.2713, "lr": 2.8489049864510053e-06, "epoch": 5.454545454545454, "percentage": 77.92, "elapsed_time": "1:56:05", "remaining_time": "0:32:53"}
+{"current_steps": 181, "total_steps": 231, "loss": 0.2758, "lr": 2.7436450612420098e-06, "epoch": 5.484848484848484, "percentage": 78.35, "elapsed_time": "1:56:42", "remaining_time": "0:32:14"}
+{"current_steps": 182, "total_steps": 231, "loss": 0.2709, "lr": 2.640056493077231e-06, "epoch": 5.515151515151516, "percentage": 78.79, "elapsed_time": "1:57:18", "remaining_time": "0:31:35"}
+{"current_steps": 183, "total_steps": 231, "loss": 0.2793, "lr": 2.5381631415231455e-06, "epoch": 5.545454545454545, "percentage": 79.22, "elapsed_time": "1:57:56", "remaining_time": "0:30:56"}
+{"current_steps": 184, "total_steps": 231, "loss": 0.2733, "lr": 2.4379884756868167e-06, "epoch": 5.575757575757576, "percentage": 79.65, "elapsed_time": "1:58:34", "remaining_time": "0:30:17"}
+{"current_steps": 185, "total_steps": 231, "loss": 0.2745, "lr": 2.339555568810221e-06, "epoch": 5.606060606060606, "percentage": 80.09, "elapsed_time": "1:59:11", "remaining_time": "0:29:38"}
+{"current_steps": 186, "total_steps": 231, "loss": 0.2646, "lr": 2.2428870929558012e-06, "epoch": 5.636363636363637, "percentage": 80.52, "elapsed_time": "1:59:46", "remaining_time": "0:28:58"}
+{"current_steps": 187, "total_steps": 231, "loss": 0.268, "lr": 2.1480053137844115e-06, "epoch": 5.666666666666667, "percentage": 80.95, "elapsed_time": "2:00:23", "remaining_time": "0:28:19"}
+{"current_steps": 188, "total_steps": 231, "loss": 0.2747, "lr": 2.054932085426856e-06, "epoch": 5.696969696969697, "percentage": 81.39, "elapsed_time": "2:01:02", "remaining_time": "0:27:41"}
+{"current_steps": 189, "total_steps": 231, "loss": 0.2784, "lr": 1.963688845450218e-06, "epoch": 5.7272727272727275, "percentage": 81.82, "elapsed_time": "2:01:39", "remaining_time": "0:27:02"}
+{"current_steps": 190, "total_steps": 231, "loss": 0.2743, "lr": 1.8742966099201699e-06, "epoch": 5.757575757575758, "percentage": 82.25, "elapsed_time": "2:02:19", "remaining_time": "0:26:23"}
+{"current_steps": 191, "total_steps": 231, "loss": 0.2755, "lr": 1.7867759685603115e-06, "epoch": 5.787878787878788, "percentage": 82.68, "elapsed_time": "2:02:58", "remaining_time": "0:25:45"}
+{"current_steps": 192, "total_steps": 231, "loss": 0.2802, "lr": 1.7011470800097496e-06, "epoch": 5.818181818181818, "percentage": 83.12, "elapsed_time": "2:03:35", "remaining_time": "0:25:06"}
+{"current_steps": 193, "total_steps": 231, "loss": 0.2783, "lr": 1.6174296671799571e-06, "epoch": 5.848484848484849, "percentage": 83.55, "elapsed_time": "2:04:14", "remaining_time": "0:24:27"}
+{"current_steps": 194, "total_steps": 231, "loss": 0.2794, "lr": 1.5356430127119915e-06, "epoch": 5.878787878787879, "percentage": 83.98, "elapsed_time": "2:04:52", "remaining_time": "0:23:49"}
+{"current_steps": 195, "total_steps": 231, "loss": 0.2922, "lr": 1.4558059545351144e-06, "epoch": 5.909090909090909, "percentage": 84.42, "elapsed_time": "2:05:30", "remaining_time": "0:23:10"}
+{"current_steps": 196, "total_steps": 231, "loss": 0.2665, "lr": 1.3779368815278648e-06, "epoch": 5.9393939393939394, "percentage": 84.85, "elapsed_time": "2:06:09", "remaining_time": "0:22:31"}
+{"current_steps": 197, "total_steps": 231, "loss": 0.2768, "lr": 1.302053729282533e-06, "epoch": 5.96969696969697, "percentage": 85.28, "elapsed_time": "2:06:48", "remaining_time": "0:21:53"}
+{"current_steps": 198, "total_steps": 231, "loss": 0.2786, "lr": 1.2281739759740575e-06, "epoch": 6.0, "percentage": 85.71, "elapsed_time": "2:07:23", "remaining_time": "0:21:13"}
+{"current_steps": 199, "total_steps": 231, "loss": 0.2628, "lr": 1.156314638334277e-06, "epoch": 6.03030303030303, "percentage": 86.15, "elapsed_time": "2:08:35", "remaining_time": "0:20:40"}
+{"current_steps": 200, "total_steps": 231, "loss": 0.269, "lr": 1.086492267732462e-06, "epoch": 6.0606060606060606, "percentage": 86.58, "elapsed_time": "2:09:14", "remaining_time": "0:20:01"}
+{"current_steps": 201, "total_steps": 231, "loss": 0.2532, "lr": 1.01872294636304e-06, "epoch": 6.090909090909091, "percentage": 87.01, "elapsed_time": "2:09:52", "remaining_time": "0:19:23"}
+{"current_steps": 202, "total_steps": 231, "loss": 0.263, "lr": 9.530222835413739e-07, "epoch": 6.121212121212121, "percentage": 87.45, "elapsed_time": "2:10:32", "remaining_time": "0:18:44"}
+{"current_steps": 203, "total_steps": 231, "loss": 0.2652, "lr": 8.894054121084839e-07, "epoch": 6.151515151515151, "percentage": 87.88, "elapsed_time": "2:11:12", "remaining_time": "0:18:05"}
+{"current_steps": 204, "total_steps": 231, "loss": 0.2495, "lr": 8.278869849454718e-07, "epoch": 6.181818181818182, "percentage": 88.31, "elapsed_time": "2:11:49", "remaining_time": "0:17:26"}
+{"current_steps": 205, "total_steps": 231, "loss": 0.2635, "lr": 7.684811715985429e-07, "epoch": 6.212121212121212, "percentage": 88.74, "elapsed_time": "2:12:27", "remaining_time": "0:16:47"}
+{"current_steps": 206, "total_steps": 231, "loss": 0.2691, "lr": 7.1120165501533e-07, "epoch": 6.242424242424242, "percentage": 89.18, "elapsed_time": "2:13:06", "remaining_time": "0:16:09"}
+{"current_steps": 207, "total_steps": 231, "loss": 0.2705, "lr": 6.560616283932897e-07, "epoch": 6.2727272727272725, "percentage": 89.61, "elapsed_time": "2:13:44", "remaining_time": "0:15:30"}
+{"current_steps": 208, "total_steps": 231, "loss": 0.2672, "lr": 6.030737921409169e-07, "epoch": 6.303030303030303, "percentage": 90.04, "elapsed_time": "2:14:21", "remaining_time": "0:14:51"}
+{"current_steps": 209, "total_steps": 231, "loss": 0.2619, "lr": 5.522503509524591e-07, "epoch": 6.333333333333333, "percentage": 90.48, "elapsed_time": "2:14:56", "remaining_time": "0:14:12"}
+{"current_steps": 210, "total_steps": 231, "loss": 0.2707, "lr": 5.036030109968082e-07, "epoch": 6.363636363636363, "percentage": 90.91, "elapsed_time": "2:15:33", "remaining_time": "0:13:33"}
+{"current_steps": 211, "total_steps": 231, "loss": 0.2582, "lr": 4.5714297722121105e-07, "epoch": 6.393939393939394, "percentage": 91.34, "elapsed_time": "2:16:12", "remaining_time": "0:12:54"}
+{"current_steps": 212, "total_steps": 231, "loss": 0.2684, "lr": 4.128809507704445e-07, "epoch": 6.424242424242424, "percentage": 91.77, "elapsed_time": "2:16:48", "remaining_time": "0:12:15"}
+{"current_steps": 213, "total_steps": 231, "loss": 0.2573, "lr": 3.708271265220087e-07, "epoch": 6.454545454545454, "percentage": 92.21, "elapsed_time": "2:17:27", "remaining_time": "0:11:36"}
+{"current_steps": 214, "total_steps": 231, "loss": 0.2557, "lr": 3.309911907379393e-07, "epoch": 6.484848484848484, "percentage": 92.64, "elapsed_time": "2:18:05", "remaining_time": "0:10:58"}
+{"current_steps": 215, "total_steps": 231, "loss": 0.2543, "lr": 2.9338231883378365e-07, "epoch": 6.515151515151516, "percentage": 93.07, "elapsed_time": "2:18:42", "remaining_time": "0:10:19"}
+{"current_steps": 216, "total_steps": 231, "loss": 0.2662, "lr": 2.5800917326521013e-07, "epoch": 6.545454545454545, "percentage": 93.51, "elapsed_time": "2:19:18", "remaining_time": "0:09:40"}
+{"current_steps": 217, "total_steps": 231, "loss": 0.256, "lr": 2.248799015327907e-07, "epoch": 6.575757575757576, "percentage": 93.94, "elapsed_time": "2:19:55", "remaining_time": "0:09:01"}
+{"current_steps": 218, "total_steps": 231, "loss": 0.2561, "lr": 1.9400213430538773e-07, "epoch": 6.606060606060606, "percentage": 94.37, "elapsed_time": "2:20:33", "remaining_time": "0:08:22"}
+{"current_steps": 219, "total_steps": 231, "loss": 0.263, "lr": 1.6538298366257975e-07, "epoch": 6.636363636363637, "percentage": 94.81, "elapsed_time": "2:21:10", "remaining_time": "0:07:44"}
+{"current_steps": 220, "total_steps": 231, "loss": 0.2535, "lr": 1.3902904145653094e-07, "epoch": 6.666666666666667, "percentage": 95.24, "elapsed_time": "2:21:46", "remaining_time": "0:07:05"}
+{"current_steps": 221, "total_steps": 231, "loss": 0.2555, "lr": 1.1494637779369766e-07, "epoch": 6.696969696969697, "percentage": 95.67, "elapsed_time": "2:22:25", "remaining_time": "0:06:26"}
+{"current_steps": 222, "total_steps": 231, "loss": 0.2644, "lr": 9.314053963669245e-08, "epoch": 6.7272727272727275, "percentage": 96.1, "elapsed_time": "2:23:00", "remaining_time": "0:05:47"}
+{"current_steps": 223, "total_steps": 231, "loss": 0.2707, "lr": 7.361654952665608e-08, "epoch": 6.757575757575758, "percentage": 96.54, "elapsed_time": "2:23:37", "remaining_time": "0:05:09"}
+{"current_steps": 224, "total_steps": 231, "loss": 0.265, "lr": 5.637890442641403e-08, "epoch": 6.787878787878788, "percentage": 96.97, "elapsed_time": "2:24:15", "remaining_time": "0:04:30"}
+{"current_steps": 225, "total_steps": 231, "loss": 0.2608, "lr": 4.143157468468717e-08, "epoch": 6.818181818181818, "percentage": 97.4, "elapsed_time": "2:24:55", "remaining_time": "0:03:51"}
+{"current_steps": 226, "total_steps": 231, "loss": 0.2551, "lr": 2.8778003121607834e-08, "epoch": 6.848484848484849, "percentage": 97.84, "elapsed_time": "2:25:30", "remaining_time": "0:03:13"}
+{"current_steps": 227, "total_steps": 231, "loss": 0.2647, "lr": 1.8421104235727406e-08, "epoch": 6.878787878787879, "percentage": 98.27, "elapsed_time": "2:26:07", "remaining_time": "0:02:34"}
+{"current_steps": 228, "total_steps": 231, "loss": 0.2552, "lr": 1.0363263532724433e-08, "epoch": 6.909090909090909, "percentage": 98.7, "elapsed_time": "2:26:45", "remaining_time": "0:01:55"}
+{"current_steps": 229, "total_steps": 231, "loss": 0.2584, "lr": 4.606336975948589e-09, "epoch": 6.9393939393939394, "percentage": 99.13, "elapsed_time": "2:27:21", "remaining_time": "0:01:17"}
+{"current_steps": 230, "total_steps": 231, "loss": 0.2668, "lr": 1.1516505589381777e-09, "epoch": 6.96969696969697, "percentage": 99.57, "elapsed_time": "2:28:00", "remaining_time": "0:00:38"}
+{"current_steps": 231, "total_steps": 231, "loss": 0.2646, "lr": 0.0, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "2:28:39", "remaining_time": "0:00:00"}
+{"current_steps": 231, "total_steps": 231, "epoch": 7.0, "percentage": 100.0, "elapsed_time": "2:29:43", "remaining_time": "0:00:00"}

trainer_state.json

@@ -0,0 +1,1659 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 231,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.030303030303030304,
+      "grad_norm": 6.389679464143325,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 0.8216,
+      "step": 1
+    },
+    {
+      "epoch": 0.06060606060606061,
+      "grad_norm": 6.0870029465685995,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 0.794,
+      "step": 2
+    },
+    {
+      "epoch": 0.09090909090909091,
+      "grad_norm": 6.33447423819787,
+      "learning_rate": 2.5e-06,
+      "loss": 0.8213,
+      "step": 3
+    },
+    {
+      "epoch": 0.12121212121212122,
+      "grad_norm": 5.620528894398241,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 0.7913,
+      "step": 4
+    },
+    {
+      "epoch": 0.15151515151515152,
+      "grad_norm": 4.316178832830539,
+      "learning_rate": 4.166666666666667e-06,
+      "loss": 0.7675,
+      "step": 5
+    },
+    {
+      "epoch": 0.18181818181818182,
+      "grad_norm": 2.339462965558909,
+      "learning_rate": 5e-06,
+      "loss": 0.7282,
+      "step": 6
+    },
+    {
+      "epoch": 0.21212121212121213,
+      "grad_norm": 1.9581073288072453,
+      "learning_rate": 5.833333333333334e-06,
+      "loss": 0.7226,
+      "step": 7
+    },
+    {
+      "epoch": 0.24242424242424243,
+      "grad_norm": 3.6598363210046525,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 0.7141,
+      "step": 8
+    },
+    {
+      "epoch": 0.2727272727272727,
+      "grad_norm": 4.008832019540338,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.7258,
+      "step": 9
+    },
+    {
+      "epoch": 0.30303030303030304,
+      "grad_norm": 3.8023588513331323,
+      "learning_rate": 8.333333333333334e-06,
+      "loss": 0.6918,
+      "step": 10
+    },
+    {
+      "epoch": 0.3333333333333333,
+      "grad_norm": 3.60329148202554,
+      "learning_rate": 9.166666666666666e-06,
+      "loss": 0.6673,
+      "step": 11
+    },
+    {
+      "epoch": 0.36363636363636365,
+      "grad_norm": 3.0412468601087648,
+      "learning_rate": 1e-05,
+      "loss": 0.6527,
+      "step": 12
+    },
+    {
+      "epoch": 0.3939393939393939,
+      "grad_norm": 1.7821663971794524,
+      "learning_rate": 1.0833333333333334e-05,
+      "loss": 0.6604,
+      "step": 13
+    },
+    {
+      "epoch": 0.42424242424242425,
+      "grad_norm": 1.469929373790711,
+      "learning_rate": 1.1666666666666668e-05,
+      "loss": 0.6126,
+      "step": 14
+    },
+    {
+      "epoch": 0.45454545454545453,
+      "grad_norm": 1.7088130414515645,
+      "learning_rate": 1.25e-05,
+      "loss": 0.6031,
+      "step": 15
+    },
+    {
+      "epoch": 0.48484848484848486,
+      "grad_norm": 1.4345672873378874,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 0.6213,
+      "step": 16
+    },
+    {
+      "epoch": 0.5151515151515151,
+      "grad_norm": 0.9537217329798097,
+      "learning_rate": 1.416666666666667e-05,
+      "loss": 0.5851,
+      "step": 17
+    },
+    {
+      "epoch": 0.5454545454545454,
+      "grad_norm": 1.0359334660741335,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.5827,
+      "step": 18
+    },
+    {
+      "epoch": 0.5757575757575758,
+      "grad_norm": 0.8612991371038484,
+      "learning_rate": 1.5833333333333333e-05,
+      "loss": 0.5626,
+      "step": 19
+    },
+    {
+      "epoch": 0.6060606060606061,
+      "grad_norm": 0.8568176137969006,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.5622,
+      "step": 20
+    },
+    {
+      "epoch": 0.6363636363636364,
+      "grad_norm": 0.8522777322963402,
+      "learning_rate": 1.7500000000000002e-05,
+      "loss": 0.5481,
+      "step": 21
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.7453158889642924,
+      "learning_rate": 1.8333333333333333e-05,
+      "loss": 0.5595,
+      "step": 22
+    },
+    {
+      "epoch": 0.696969696969697,
+      "grad_norm": 0.9322760926453303,
+      "learning_rate": 1.916666666666667e-05,
+      "loss": 0.5446,
+      "step": 23
+    },
+    {
+      "epoch": 0.7272727272727273,
+      "grad_norm": 0.7311347675851437,
+      "learning_rate": 2e-05,
+      "loss": 0.5457,
+      "step": 24
+    },
+    {
+      "epoch": 0.7575757575757576,
+      "grad_norm": 0.7724040724529722,
+      "learning_rate": 1.999884834944106e-05,
+      "loss": 0.5346,
+      "step": 25
+    },
+    {
+      "epoch": 0.7878787878787878,
+      "grad_norm": 0.7913425312973258,
+      "learning_rate": 1.9995393663024054e-05,
+      "loss": 0.5221,
+      "step": 26
+    },
+    {
+      "epoch": 0.8181818181818182,
+      "grad_norm": 0.5435818443653805,
+      "learning_rate": 1.9989636736467278e-05,
+      "loss": 0.5274,
+      "step": 27
+    },
+    {
+      "epoch": 0.8484848484848485,
+      "grad_norm": 0.7246903580516652,
+      "learning_rate": 1.9981578895764272e-05,
+      "loss": 0.5231,
+      "step": 28
+    },
+    {
+      "epoch": 0.8787878787878788,
+      "grad_norm": 0.5172030157917001,
+      "learning_rate": 1.9971221996878395e-05,
+      "loss": 0.5221,
+      "step": 29
+    },
+    {
+      "epoch": 0.9090909090909091,
+      "grad_norm": 0.7810391897625707,
+      "learning_rate": 1.9958568425315316e-05,
+      "loss": 0.5332,
+      "step": 30
+    },
+    {
+      "epoch": 0.9393939393939394,
+      "grad_norm": 0.5497698223387213,
+      "learning_rate": 1.9943621095573588e-05,
+      "loss": 0.5204,
+      "step": 31
+    },
+    {
+      "epoch": 0.9696969696969697,
+      "grad_norm": 0.5947629897020668,
+      "learning_rate": 1.9926383450473344e-05,
+      "loss": 0.5183,
+      "step": 32
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.5330072862604534,
+      "learning_rate": 1.9906859460363307e-05,
+      "loss": 0.5098,
+      "step": 33
+    },
+    {
+      "epoch": 1.0303030303030303,
+      "grad_norm": 0.4721557957467901,
+      "learning_rate": 1.9885053622206305e-05,
+      "loss": 0.4833,
+      "step": 34
+    },
+    {
+      "epoch": 1.0606060606060606,
+      "grad_norm": 0.44696690995060356,
+      "learning_rate": 1.986097095854347e-05,
+      "loss": 0.4789,
+      "step": 35
+    },
+    {
+      "epoch": 1.0909090909090908,
+      "grad_norm": 0.4715290430220381,
+      "learning_rate": 1.9834617016337424e-05,
+      "loss": 0.481,
+      "step": 36
+    },
+    {
+      "epoch": 1.121212121212121,
+      "grad_norm": 0.484211942279594,
+      "learning_rate": 1.9805997865694616e-05,
+      "loss": 0.4686,
+      "step": 37
+    },
+    {
+      "epoch": 1.1515151515151516,
+      "grad_norm": 0.4224809049475823,
+      "learning_rate": 1.9775120098467212e-05,
+      "loss": 0.488,
+      "step": 38
+    },
+    {
+      "epoch": 1.1818181818181819,
+      "grad_norm": 0.4767478991829516,
+      "learning_rate": 1.9741990826734793e-05,
+      "loss": 0.4706,
+      "step": 39
+    },
+    {
+      "epoch": 1.2121212121212122,
+      "grad_norm": 0.41475242363404724,
+      "learning_rate": 1.970661768116622e-05,
+      "loss": 0.4909,
+      "step": 40
+    },
+    {
+      "epoch": 1.2424242424242424,
+      "grad_norm": 0.4493742206439678,
+      "learning_rate": 1.9669008809262064e-05,
+      "loss": 0.4816,
+      "step": 41
+    },
+    {
+      "epoch": 1.2727272727272727,
+      "grad_norm": 0.4858174428866086,
+      "learning_rate": 1.9629172873477995e-05,
+      "loss": 0.4856,
+      "step": 42
+    },
+    {
+      "epoch": 1.303030303030303,
+      "grad_norm": 0.5370196524998,
+      "learning_rate": 1.9587119049229558e-05,
+      "loss": 0.4848,
+      "step": 43
+    },
+    {
+      "epoch": 1.3333333333333333,
+      "grad_norm": 0.3865029753757071,
+      "learning_rate": 1.954285702277879e-05,
+      "loss": 0.4734,
+      "step": 44
+    },
+    {
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.5428346734484963,
+      "learning_rate": 1.9496396989003195e-05,
+      "loss": 0.4844,
+      "step": 45
+    },
+    {
+      "epoch": 1.393939393939394,
+      "grad_norm": 0.49276277542024816,
+      "learning_rate": 1.944774964904754e-05,
+      "loss": 0.4597,
+      "step": 46
+    },
+    {
+      "epoch": 1.4242424242424243,
+      "grad_norm": 0.5019994974591285,
+      "learning_rate": 1.9396926207859085e-05,
+      "loss": 0.4662,
+      "step": 47
+    },
+    {
+      "epoch": 1.4545454545454546,
+      "grad_norm": 0.46095938340959086,
+      "learning_rate": 1.9343938371606714e-05,
+      "loss": 0.4398,
+      "step": 48
+    },
+    {
+      "epoch": 1.4848484848484849,
+      "grad_norm": 0.5436027541613718,
+      "learning_rate": 1.9288798344984673e-05,
+      "loss": 0.4594,
+      "step": 49
+    },
+    {
+      "epoch": 1.5151515151515151,
+      "grad_norm": 0.512382069946856,
+      "learning_rate": 1.9231518828401458e-05,
+      "loss": 0.4627,
+      "step": 50
+    },
+    {
+      "epoch": 1.5454545454545454,
+      "grad_norm": 0.3993417396610489,
+      "learning_rate": 1.917211301505453e-05,
+      "loss": 0.4531,
+      "step": 51
+    },
+    {
+      "epoch": 1.5757575757575757,
+      "grad_norm": 0.4553369444771597,
+      "learning_rate": 1.911059458789152e-05,
+      "loss": 0.451,
+      "step": 52
+    },
+    {
+      "epoch": 1.606060606060606,
+      "grad_norm": 0.4296801768417054,
+      "learning_rate": 1.9046977716458627e-05,
+      "loss": 0.4679,
+      "step": 53
+    },
+    {
+      "epoch": 1.6363636363636362,
+      "grad_norm": 0.39723792447393597,
+      "learning_rate": 1.8981277053636963e-05,
+      "loss": 0.4649,
+      "step": 54
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.43610761040124835,
+      "learning_rate": 1.891350773226754e-05,
+      "loss": 0.4432,
+      "step": 55
+    },
+    {
+      "epoch": 1.696969696969697,
+      "grad_norm": 0.368349776002198,
+      "learning_rate": 1.8843685361665724e-05,
+      "loss": 0.454,
+      "step": 56
+    },
+    {
+      "epoch": 1.7272727272727273,
+      "grad_norm": 0.4389169614250592,
+      "learning_rate": 1.8771826024025944e-05,
+      "loss": 0.4644,
+      "step": 57
+    },
+    {
+      "epoch": 1.7575757575757576,
+      "grad_norm": 0.39930247566261223,
+      "learning_rate": 1.8697946270717468e-05,
+      "loss": 0.4628,
+      "step": 58
+    },
+    {
+      "epoch": 1.7878787878787878,
+      "grad_norm": 0.42246137547786905,
+      "learning_rate": 1.8622063118472135e-05,
+      "loss": 0.4599,
+      "step": 59
+    },
+    {
+      "epoch": 1.8181818181818183,
+      "grad_norm": 0.4991758233281764,
+      "learning_rate": 1.8544194045464888e-05,
+      "loss": 0.4597,
+      "step": 60
+    },
+    {
+      "epoch": 1.8484848484848486,
+      "grad_norm": 0.3500385967202033,
+      "learning_rate": 1.8464356987288012e-05,
+      "loss": 0.4659,
+      "step": 61
+    },
+    {
+      "epoch": 1.878787878787879,
+      "grad_norm": 0.5338413731056074,
+      "learning_rate": 1.8382570332820045e-05,
+      "loss": 0.4653,
+      "step": 62
+    },
+    {
+      "epoch": 1.9090909090909092,
+      "grad_norm": 0.39315339859730924,
+      "learning_rate": 1.8298852919990254e-05,
+      "loss": 0.4712,
+      "step": 63
+    },
+    {
+      "epoch": 1.9393939393939394,
+      "grad_norm": 0.49055759389219106,
+      "learning_rate": 1.821322403143969e-05,
+      "loss": 0.4659,
+      "step": 64
+    },
+    {
+      "epoch": 1.9696969696969697,
+      "grad_norm": 0.4998865137451953,
+      "learning_rate": 1.812570339007983e-05,
+      "loss": 0.4617,
+      "step": 65
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.5281523703208337,
+      "learning_rate": 1.8036311154549783e-05,
+      "loss": 0.4556,
+      "step": 66
+    },
+    {
+      "epoch": 2.0303030303030303,
+      "grad_norm": 0.5183352728640479,
+      "learning_rate": 1.7945067914573147e-05,
+      "loss": 0.4028,
+      "step": 67
+    },
+    {
+      "epoch": 2.0606060606060606,
+      "grad_norm": 0.43323389056033274,
+      "learning_rate": 1.7851994686215592e-05,
+      "loss": 0.4072,
+      "step": 68
+    },
+    {
+      "epoch": 2.090909090909091,
+      "grad_norm": 0.490538354979083,
+      "learning_rate": 1.77571129070442e-05,
+      "loss": 0.4112,
+      "step": 69
+    },
+    {
+      "epoch": 2.121212121212121,
+      "grad_norm": 0.48720992951592584,
+      "learning_rate": 1.766044443118978e-05,
+      "loss": 0.424,
+      "step": 70
+    },
+    {
+      "epoch": 2.1515151515151514,
+      "grad_norm": 0.51892933006533,
+      "learning_rate": 1.7562011524313187e-05,
+      "loss": 0.4197,
+      "step": 71
+    },
+    {
+      "epoch": 2.1818181818181817,
+      "grad_norm": 0.43316356012449,
+      "learning_rate": 1.7461836858476858e-05,
+      "loss": 0.412,
+      "step": 72
+    },
+    {
+      "epoch": 2.212121212121212,
+      "grad_norm": 0.5011583477381106,
+      "learning_rate": 1.7359943506922775e-05,
+      "loss": 0.4163,
+      "step": 73
+    },
+    {
+      "epoch": 2.242424242424242,
+      "grad_norm": 0.37391566522456676,
+      "learning_rate": 1.725635493875799e-05,
+      "loss": 0.3957,
+      "step": 74
+    },
+    {
+      "epoch": 2.2727272727272725,
+      "grad_norm": 0.46126379346949903,
+      "learning_rate": 1.7151095013548996e-05,
+      "loss": 0.3962,
+      "step": 75
+    },
+    {
+      "epoch": 2.303030303030303,
+      "grad_norm": 0.5138856949785989,
+      "learning_rate": 1.7044187975826126e-05,
+      "loss": 0.4197,
+      "step": 76
+    },
+    {
+      "epoch": 2.3333333333333335,
+      "grad_norm": 0.5122522819282725,
+      "learning_rate": 1.693565844949933e-05,
+      "loss": 0.4174,
+      "step": 77
+    },
+    {
+      "epoch": 2.3636363636363638,
+      "grad_norm": 0.4269939619340365,
+      "learning_rate": 1.6825531432186545e-05,
+      "loss": 0.4096,
+      "step": 78
+    },
+    {
+      "epoch": 2.393939393939394,
+      "grad_norm": 0.3884930259980605,
+      "learning_rate": 1.671383228945597e-05,
+      "loss": 0.4013,
+      "step": 79
+    },
+    {
+      "epoch": 2.4242424242424243,
+      "grad_norm": 0.4330282401428612,
+      "learning_rate": 1.6600586748983642e-05,
+      "loss": 0.4135,
+      "step": 80
+    },
+    {
+      "epoch": 2.4545454545454546,
+      "grad_norm": 0.49867447378146007,
+      "learning_rate": 1.648582089462756e-05,
+      "loss": 0.4058,
+      "step": 81
+    },
+    {
+      "epoch": 2.484848484848485,
+      "grad_norm": 0.4000948733390901,
+      "learning_rate": 1.6369561160419783e-05,
+      "loss": 0.3965,
+      "step": 82
+    },
+    {
+      "epoch": 2.515151515151515,
+      "grad_norm": 0.40127339644535653,
+      "learning_rate": 1.625183432447789e-05,
+      "loss": 0.3977,
+      "step": 83
+    },
+    {
+      "epoch": 2.5454545454545454,
+      "grad_norm": 0.4163281768705279,
+      "learning_rate": 1.6132667502837164e-05,
+      "loss": 0.4097,
+      "step": 84
+    },
+    {
+      "epoch": 2.5757575757575757,
+      "grad_norm": 0.4130162141154564,
+      "learning_rate": 1.6012088143204953e-05,
+      "loss": 0.4036,
+      "step": 85
+    },
+    {
+      "epoch": 2.606060606060606,
+      "grad_norm": 0.38323571037782045,
+      "learning_rate": 1.589012401863864e-05,
+      "loss": 0.4025,
+      "step": 86
+    },
+    {
+      "epoch": 2.6363636363636362,
+      "grad_norm": 0.4359390222033282,
+      "learning_rate": 1.5766803221148676e-05,
+      "loss": 0.3929,
+      "step": 87
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "grad_norm": 0.390478372661262,
+      "learning_rate": 1.5642154155228124e-05,
+      "loss": 0.4087,
+      "step": 88
+    },
+    {
+      "epoch": 2.6969696969696972,
+      "grad_norm": 0.3847582060805819,
+      "learning_rate": 1.5516205531310272e-05,
+      "loss": 0.4036,
+      "step": 89
+    },
+    {
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.4010141318395045,
+      "learning_rate": 1.538898635915576e-05,
+      "loss": 0.4014,
+      "step": 90
+    },
+    {
+      "epoch": 2.757575757575758,
+      "grad_norm": 0.39998137279337886,
+      "learning_rate": 1.526052594117071e-05,
+      "loss": 0.3989,
+      "step": 91
+    },
+    {
+      "epoch": 2.787878787878788,
+      "grad_norm": 0.383782592700116,
+      "learning_rate": 1.513085386565758e-05,
+      "loss": 0.3953,
+      "step": 92
+    },
+    {
+      "epoch": 2.8181818181818183,
+      "grad_norm": 0.33484545487270734,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.3982,
+      "step": 93
+    },
+    {
+      "epoch": 2.8484848484848486,
+      "grad_norm": 0.4174189055080555,
+      "learning_rate": 1.4867994483783485e-05,
+      "loss": 0.4105,
+      "step": 94
+    },
+    {
+      "epoch": 2.878787878787879,
+      "grad_norm": 0.3688575408583084,
+      "learning_rate": 1.4734867721853341e-05,
+      "loss": 0.397,
+      "step": 95
+    },
+    {
+      "epoch": 2.909090909090909,
+      "grad_norm": 0.3911223572397203,
+      "learning_rate": 1.4600650377311523e-05,
+      "loss": 0.4092,
+      "step": 96
+    },
+    {
+      "epoch": 2.9393939393939394,
+      "grad_norm": 0.393205250447062,
+      "learning_rate": 1.4465373364454001e-05,
+      "loss": 0.4152,
+      "step": 97
+    },
+    {
+      "epoch": 2.9696969696969697,
+      "grad_norm": 0.33163654704896206,
+      "learning_rate": 1.4329067841650274e-05,
+      "loss": 0.4021,
+      "step": 98
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.35852210007782426,
+      "learning_rate": 1.4191765204166643e-05,
+      "loss": 0.3881,
+      "step": 99
+    },
+    {
+      "epoch": 3.0303030303030303,
+      "grad_norm": 0.4276929900256796,
+      "learning_rate": 1.4053497076934948e-05,
+      "loss": 0.3615,
+      "step": 100
+    },
+    {
+      "epoch": 3.0606060606060606,
+      "grad_norm": 0.36943088418479125,
+      "learning_rate": 1.3914295307268396e-05,
+      "loss": 0.3647,
+      "step": 101
+    },
+    {
+      "epoch": 3.090909090909091,
+      "grad_norm": 0.416864423936772,
+      "learning_rate": 1.3774191957526144e-05,
+      "loss": 0.3607,
+      "step": 102
+    },
+    {
+      "epoch": 3.121212121212121,
+      "grad_norm": 0.454946907994254,
+      "learning_rate": 1.3633219297728415e-05,
+      "loss": 0.3542,
+      "step": 103
+    },
+    {
+      "epoch": 3.1515151515151514,
+      "grad_norm": 0.4105877635581912,
+      "learning_rate": 1.3491409798123687e-05,
+      "loss": 0.3622,
+      "step": 104
+    },
+    {
+      "epoch": 3.1818181818181817,
+      "grad_norm": 0.39877498537617606,
+      "learning_rate": 1.3348796121709862e-05,
+      "loss": 0.3641,
+      "step": 105
+    },
+    {
+      "epoch": 3.212121212121212,
+      "grad_norm": 0.40857691652727696,
+      "learning_rate": 1.3205411116710973e-05,
+      "loss": 0.3357,
+      "step": 106
+    },
+    {
+      "epoch": 3.242424242424242,
+      "grad_norm": 0.3774293129860823,
+      "learning_rate": 1.3061287809011243e-05,
+      "loss": 0.3528,
+      "step": 107
+    },
+    {
+      "epoch": 3.2727272727272725,
+      "grad_norm": 0.36457845009271533,
+      "learning_rate": 1.291645939454825e-05,
+      "loss": 0.3467,
+      "step": 108
+    },
+    {
+      "epoch": 3.303030303030303,
+      "grad_norm": 0.43697595376571186,
+      "learning_rate": 1.277095923166689e-05,
+      "loss": 0.3548,
+      "step": 109
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 0.3279772156769785,
+      "learning_rate": 1.2624820833435939e-05,
+      "loss": 0.3628,
+      "step": 110
+    },
+    {
+      "epoch": 3.3636363636363638,
+      "grad_norm": 0.3674180825534724,
+      "learning_rate": 1.2478077859929e-05,
+      "loss": 0.3445,
+      "step": 111
+    },
+    {
+      "epoch": 3.393939393939394,
+      "grad_norm": 0.36677402712205043,
+      "learning_rate": 1.2330764110471567e-05,
+      "loss": 0.3518,
+      "step": 112
+    },
+    {
+      "epoch": 3.4242424242424243,
+      "grad_norm": 0.32484641950884563,
+      "learning_rate": 1.2182913515856016e-05,
+      "loss": 0.349,
+      "step": 113
+    },
+    {
+      "epoch": 3.4545454545454546,
+      "grad_norm": 0.32762570869276525,
+      "learning_rate": 1.2034560130526341e-05,
+      "loss": 0.3568,
+      "step": 114
+    },
+    {
+      "epoch": 3.484848484848485,
+      "grad_norm": 0.3732488392977816,
+      "learning_rate": 1.1885738124734359e-05,
+      "loss": 0.3621,
+      "step": 115
+    },
+    {
+      "epoch": 3.515151515151515,
+      "grad_norm": 0.3085942762155234,
+      "learning_rate": 1.1736481776669307e-05,
+      "loss": 0.3653,
+      "step": 116
+    },
+    {
+      "epoch": 3.5454545454545454,
+      "grad_norm": 0.33005914795893343,
+      "learning_rate": 1.1586825464562515e-05,
+      "loss": 0.3434,
+      "step": 117
+    },
+    {
+      "epoch": 3.5757575757575757,
+      "grad_norm": 0.37800192242609815,
+      "learning_rate": 1.1436803658769082e-05,
+      "loss": 0.3474,
+      "step": 118
+    },
+    {
+      "epoch": 3.606060606060606,
+      "grad_norm": 0.2913441889881699,
+      "learning_rate": 1.1286450913828313e-05,
+      "loss": 0.3526,
+      "step": 119
+    },
+    {
+      "epoch": 3.6363636363636362,
+      "grad_norm": 0.32804396327735513,
+      "learning_rate": 1.113580186050475e-05,
+      "loss": 0.3551,
+      "step": 120
+    },
+    {
+      "epoch": 3.6666666666666665,
+      "grad_norm": 0.33516462116193874,
+      "learning_rate": 1.0984891197811686e-05,
+      "loss": 0.3568,
+      "step": 121
+    },
+    {
+      "epoch": 3.6969696969696972,
+      "grad_norm": 0.32908995740085434,
+      "learning_rate": 1.0833753685018935e-05,
+      "loss": 0.3449,
+      "step": 122
+    },
+    {
+      "epoch": 3.7272727272727275,
+      "grad_norm": 0.32147000359122696,
+      "learning_rate": 1.0682424133646712e-05,
+      "loss": 0.362,
+      "step": 123
+    },
+    {
+      "epoch": 3.757575757575758,
+      "grad_norm": 0.3096197644123865,
+      "learning_rate": 1.0530937399447496e-05,
+      "loss": 0.3625,
+      "step": 124
+    },
+    {
+      "epoch": 3.787878787878788,
+      "grad_norm": 0.31487973380195133,
+      "learning_rate": 1.0379328374377715e-05,
+      "loss": 0.3571,
+      "step": 125
+    },
+    {
+      "epoch": 3.8181818181818183,
+      "grad_norm": 0.2896993260466313,
+      "learning_rate": 1.0227631978561057e-05,
+      "loss": 0.3547,
+      "step": 126
+    },
+    {
+      "epoch": 3.8484848484848486,
+      "grad_norm": 0.30987572421685594,
+      "learning_rate": 1.0075883152245334e-05,
+      "loss": 0.3623,
+      "step": 127
+    },
+    {
+      "epoch": 3.878787878787879,
+      "grad_norm": 0.2989010699414137,
+      "learning_rate": 9.92411684775467e-06,
+      "loss": 0.342,
+      "step": 128
+    },
+    {
+      "epoch": 3.909090909090909,
+      "grad_norm": 0.33270560821661216,
+      "learning_rate": 9.772368021438943e-06,
+      "loss": 0.3482,
+      "step": 129
+    },
+    {
+      "epoch": 3.9393939393939394,
+      "grad_norm": 0.30590884996875034,
+      "learning_rate": 9.620671625622287e-06,
+      "loss": 0.3522,
+      "step": 130
+    },
+    {
+      "epoch": 3.9696969696969697,
+      "grad_norm": 0.3413263440780111,
+      "learning_rate": 9.469062600552509e-06,
+      "loss": 0.349,
+      "step": 131
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.3221210837028033,
+      "learning_rate": 9.317575866353293e-06,
+      "loss": 0.3557,
+      "step": 132
+    },
+    {
+      "epoch": 4.03030303030303,
+      "grad_norm": 0.4210306757150854,
+      "learning_rate": 9.166246314981066e-06,
+      "loss": 0.3101,
+      "step": 133
+    },
+    {
+      "epoch": 4.0606060606060606,
+      "grad_norm": 0.35741966288239124,
+      "learning_rate": 9.015108802188314e-06,
+      "loss": 0.309,
+      "step": 134
+    },
+    {
+      "epoch": 4.090909090909091,
+      "grad_norm": 0.4083522228049602,
+      "learning_rate": 8.86419813949525e-06,
+      "loss": 0.3245,
+      "step": 135
+    },
+    {
+      "epoch": 4.121212121212121,
+      "grad_norm": 0.43594044169127666,
+      "learning_rate": 8.71354908617169e-06,
+      "loss": 0.3191,
+      "step": 136
+    },
+    {
+      "epoch": 4.151515151515151,
+      "grad_norm": 0.4148276318838158,
+      "learning_rate": 8.56319634123092e-06,
+      "loss": 0.3044,
+      "step": 137
+    },
+    {
+      "epoch": 4.181818181818182,
+      "grad_norm": 0.3631741823106515,
+      "learning_rate": 8.413174535437486e-06,
+      "loss": 0.3031,
+      "step": 138
+    },
+    {
+      "epoch": 4.212121212121212,
+      "grad_norm": 0.3539162628389827,
+      "learning_rate": 8.263518223330698e-06,
+      "loss": 0.3132,
+      "step": 139
+    },
+    {
+      "epoch": 4.242424242424242,
+      "grad_norm": 0.39308778047792964,
+      "learning_rate": 8.114261875265643e-06,
+      "loss": 0.309,
+      "step": 140
+    },
+    {
+      "epoch": 4.2727272727272725,
+      "grad_norm": 0.4137786335768427,
+      "learning_rate": 7.965439869473664e-06,
+      "loss": 0.3141,
+      "step": 141
+    },
+    {
+      "epoch": 4.303030303030303,
+      "grad_norm": 0.33053991835814944,
+      "learning_rate": 7.817086484143987e-06,
+      "loss": 0.3123,
+      "step": 142
+    },
+    {
+      "epoch": 4.333333333333333,
+      "grad_norm": 0.360604137311516,
+      "learning_rate": 7.669235889528436e-06,
+      "loss": 0.299,
+      "step": 143
+    },
+    {
+      "epoch": 4.363636363636363,
+      "grad_norm": 0.3886434471042827,
+      "learning_rate": 7.521922140071003e-06,
+      "loss": 0.3163,
+      "step": 144
+    },
+    {
+      "epoch": 4.393939393939394,
+      "grad_norm": 0.3369677782711241,
+      "learning_rate": 7.375179166564062e-06,
+      "loss": 0.3012,
+      "step": 145
+    },
+    {
+      "epoch": 4.424242424242424,
+      "grad_norm": 0.3192004662092612,
+      "learning_rate": 7.2290407683331154e-06,
+      "loss": 0.3124,
+      "step": 146
+    },
+    {
+      "epoch": 4.454545454545454,
+      "grad_norm": 0.31553914305354386,
+      "learning_rate": 7.0835406054517505e-06,
+      "loss": 0.2999,
+      "step": 147
+    },
+    {
+      "epoch": 4.484848484848484,
+      "grad_norm": 0.32409683962722596,
+      "learning_rate": 6.93871219098876e-06,
+      "loss": 0.3166,
+      "step": 148
+    },
+    {
+      "epoch": 4.515151515151516,
+      "grad_norm": 0.29279884654404215,
+      "learning_rate": 6.79458888328903e-06,
+      "loss": 0.3134,
+      "step": 149
+    },
+    {
+      "epoch": 4.545454545454545,
+      "grad_norm": 0.3272960594603835,
+      "learning_rate": 6.651203878290139e-06,
+      "loss": 0.3171,
+      "step": 150
+    },
+    {
+      "epoch": 4.575757575757576,
+      "grad_norm": 0.3186069661608066,
+      "learning_rate": 6.508590201876317e-06,
+      "loss": 0.3101,
+      "step": 151
+    },
+    {
+      "epoch": 4.606060606060606,
+      "grad_norm": 0.2648439505655438,
+      "learning_rate": 6.366780702271589e-06,
+      "loss": 0.3043,
+      "step": 152
+    },
+    {
+      "epoch": 4.636363636363637,
+      "grad_norm": 0.2788087197565788,
+      "learning_rate": 6.225808042473857e-06,
+      "loss": 0.31,
+      "step": 153
+    },
+    {
+      "epoch": 4.666666666666667,
+      "grad_norm": 0.2503203391643529,
+      "learning_rate": 6.085704692731609e-06,
+      "loss": 0.2977,
+      "step": 154
+    },
+    {
+      "epoch": 4.696969696969697,
+      "grad_norm": 0.30809314794877196,
+      "learning_rate": 5.946502923065054e-06,
+      "loss": 0.3174,
+      "step": 155
+    },
+    {
+      "epoch": 4.7272727272727275,
+      "grad_norm": 0.3045080386845063,
+      "learning_rate": 5.8082347958333625e-06,
+      "loss": 0.3138,
+      "step": 156
+    },
+    {
+      "epoch": 4.757575757575758,
+      "grad_norm": 0.2797657437245348,
+      "learning_rate": 5.670932158349732e-06,
+      "loss": 0.3201,
+      "step": 157
+    },
+    {
+      "epoch": 4.787878787878788,
+      "grad_norm": 0.2638418194723567,
+      "learning_rate": 5.534626635546e-06,
+      "loss": 0.3055,
+      "step": 158
+    },
+    {
+      "epoch": 4.818181818181818,
+      "grad_norm": 0.27116179637904714,
+      "learning_rate": 5.399349622688479e-06,
+      "loss": 0.3047,
+      "step": 159
+    },
+    {
+      "epoch": 4.848484848484849,
+      "grad_norm": 0.2609842736087905,
+      "learning_rate": 5.2651322781466606e-06,
+      "loss": 0.3079,
+      "step": 160
+    },
+    {
+      "epoch": 4.878787878787879,
+      "grad_norm": 0.2637332780667072,
+      "learning_rate": 5.132005516216512e-06,
+      "loss": 0.3043,
+      "step": 161
+    },
+    {
+      "epoch": 4.909090909090909,
+      "grad_norm": 0.2548209552892343,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 0.3132,
+      "step": 162
+    },
+    {
+      "epoch": 4.9393939393939394,
+      "grad_norm": 0.26407785000721573,
+      "learning_rate": 4.869146134342426e-06,
+      "loss": 0.3133,
+      "step": 163
+    },
+    {
+      "epoch": 4.96969696969697,
+      "grad_norm": 0.26320423865848713,
+      "learning_rate": 4.739474058829288e-06,
+      "loss": 0.3133,
+      "step": 164
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.25260839130565943,
+      "learning_rate": 4.611013640844245e-06,
+      "loss": 0.3062,
+      "step": 165
+    },
+    {
+      "epoch": 5.03030303030303,
+      "grad_norm": 0.35045364828201614,
+      "learning_rate": 4.483794468689728e-06,
+      "loss": 0.2825,
+      "step": 166
+    },
+    {
+      "epoch": 5.0606060606060606,
+      "grad_norm": 0.2878497108222042,
+      "learning_rate": 4.357845844771881e-06,
+      "loss": 0.2881,
+      "step": 167
+    },
+    {
+      "epoch": 5.090909090909091,
+      "grad_norm": 0.2692364179021947,
+      "learning_rate": 4.2331967788513295e-06,
+      "loss": 0.2864,
+      "step": 168
+    },
+    {
+      "epoch": 5.121212121212121,
+      "grad_norm": 0.3397140513642043,
+      "learning_rate": 4.109875981361363e-06,
+      "loss": 0.2857,
+      "step": 169
+    },
+    {
+      "epoch": 5.151515151515151,
+      "grad_norm": 0.33923809826546575,
+      "learning_rate": 3.987911856795047e-06,
+      "loss": 0.285,
+      "step": 170
+    },
+    {
+      "epoch": 5.181818181818182,
+      "grad_norm": 0.2797575682833663,
+      "learning_rate": 3.867332497162836e-06,
+      "loss": 0.2824,
+      "step": 171
+    },
+    {
+      "epoch": 5.212121212121212,
+      "grad_norm": 0.2653710601957683,
+      "learning_rate": 3.748165675522113e-06,
+      "loss": 0.2711,
+      "step": 172
+    },
+    {
+      "epoch": 5.242424242424242,
+      "grad_norm": 0.27260257926718856,
+      "learning_rate": 3.630438839580217e-06,
+      "loss": 0.2724,
+      "step": 173
+    },
+    {
+      "epoch": 5.2727272727272725,
+      "grad_norm": 0.27852051116246085,
+      "learning_rate": 3.5141791053724405e-06,
+      "loss": 0.2873,
+      "step": 174
+    },
+    {
+      "epoch": 5.303030303030303,
+      "grad_norm": 0.2856413126687512,
+      "learning_rate": 3.399413251016359e-06,
+      "loss": 0.2735,
+      "step": 175
+    },
+    {
+      "epoch": 5.333333333333333,
+      "grad_norm": 0.2593430557542982,
+      "learning_rate": 3.2861677105440335e-06,
+      "loss": 0.2687,
+      "step": 176
+    },
+    {
+      "epoch": 5.363636363636363,
+      "grad_norm": 0.27770413574926345,
+      "learning_rate": 3.174468567813461e-06,
+      "loss": 0.2745,
+      "step": 177
+    },
+    {
+      "epoch": 5.393939393939394,
+      "grad_norm": 0.28604505936188646,
+      "learning_rate": 3.0643415505006733e-06,
+      "loss": 0.2758,
+      "step": 178
+    },
+    {
+      "epoch": 5.424242424242424,
+      "grad_norm": 0.2670316366650927,
+      "learning_rate": 2.9558120241738786e-06,
+      "loss": 0.278,
+      "step": 179
+    },
+    {
+      "epoch": 5.454545454545454,
+      "grad_norm": 0.26192988499195347,
+      "learning_rate": 2.8489049864510053e-06,
+      "loss": 0.2713,
+      "step": 180
+    },
+    {
+      "epoch": 5.484848484848484,
+      "grad_norm": 0.2579041025278341,
+      "learning_rate": 2.7436450612420098e-06,
+      "loss": 0.2758,
+      "step": 181
+    },
+    {
+      "epoch": 5.515151515151516,
+      "grad_norm": 0.2581571699654079,
+      "learning_rate": 2.640056493077231e-06,
+      "loss": 0.2709,
+      "step": 182
+    },
+    {
+      "epoch": 5.545454545454545,
+      "grad_norm": 0.26358679708819666,
+      "learning_rate": 2.5381631415231455e-06,
+      "loss": 0.2793,
+      "step": 183
+    },
+    {
+      "epoch": 5.575757575757576,
+      "grad_norm": 0.26036766758260393,
+      "learning_rate": 2.4379884756868167e-06,
+      "loss": 0.2733,
+      "step": 184
+    },
+    {
+      "epoch": 5.606060606060606,
+      "grad_norm": 0.24900549026134164,
+      "learning_rate": 2.339555568810221e-06,
+      "loss": 0.2745,
+      "step": 185
+    },
+    {
+      "epoch": 5.636363636363637,
+      "grad_norm": 0.24263829872831869,
+      "learning_rate": 2.2428870929558012e-06,
+      "loss": 0.2646,
+      "step": 186
+    },
+    {
+      "epoch": 5.666666666666667,
+      "grad_norm": 0.23173769595787047,
+      "learning_rate": 2.1480053137844115e-06,
+      "loss": 0.268,
+      "step": 187
+    },
+    {
+      "epoch": 5.696969696969697,
+      "grad_norm": 0.24743207197013373,
+      "learning_rate": 2.054932085426856e-06,
+      "loss": 0.2747,
+      "step": 188
+    },
+    {
+      "epoch": 5.7272727272727275,
+      "grad_norm": 0.24678627068896922,
+      "learning_rate": 1.963688845450218e-06,
+      "loss": 0.2784,
+      "step": 189
+    },
+    {
+      "epoch": 5.757575757575758,
+      "grad_norm": 0.24292786245623466,
+      "learning_rate": 1.8742966099201699e-06,
+      "loss": 0.2743,
+      "step": 190
+    },
+    {
+      "epoch": 5.787878787878788,
+      "grad_norm": 0.2301052425255644,
+      "learning_rate": 1.7867759685603115e-06,
+      "loss": 0.2755,
+      "step": 191
+    },
+    {
+      "epoch": 5.818181818181818,
+      "grad_norm": 0.24436621048715285,
+      "learning_rate": 1.7011470800097496e-06,
+      "loss": 0.2802,
+      "step": 192
+    },
+    {
+      "epoch": 5.848484848484849,
+      "grad_norm": 0.24233892802925538,
+      "learning_rate": 1.6174296671799571e-06,
+      "loss": 0.2783,
+      "step": 193
+    },
+    {
+      "epoch": 5.878787878787879,
+      "grad_norm": 0.2326648485384654,
+      "learning_rate": 1.5356430127119915e-06,
+      "loss": 0.2794,
+      "step": 194
+    },
+    {
+      "epoch": 5.909090909090909,
+      "grad_norm": 0.23967329818495606,
+      "learning_rate": 1.4558059545351144e-06,
+      "loss": 0.2922,
+      "step": 195
+    },
+    {
+      "epoch": 5.9393939393939394,
+      "grad_norm": 0.23136254304772033,
+      "learning_rate": 1.3779368815278648e-06,
+      "loss": 0.2665,
+      "step": 196
+    },
+    {
+      "epoch": 5.96969696969697,
+      "grad_norm": 0.22442234807503147,
+      "learning_rate": 1.302053729282533e-06,
+      "loss": 0.2768,
+      "step": 197
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 0.2248897102418934,
+      "learning_rate": 1.2281739759740575e-06,
+      "loss": 0.2786,
+      "step": 198
+    },
+    {
+      "epoch": 6.03030303030303,
+      "grad_norm": 0.27551238016074,
+      "learning_rate": 1.156314638334277e-06,
+      "loss": 0.2628,
+      "step": 199
+    },
+    {
+      "epoch": 6.0606060606060606,
+      "grad_norm": 0.30409342282998103,
+      "learning_rate": 1.086492267732462e-06,
+      "loss": 0.269,
+      "step": 200
+    },
+    {
+      "epoch": 6.090909090909091,
+      "grad_norm": 0.24172010021563448,
+      "learning_rate": 1.01872294636304e-06,
+      "loss": 0.2532,
+      "step": 201
+    },
+    {
+      "epoch": 6.121212121212121,
+      "grad_norm": 0.22075170600714575,
+      "learning_rate": 9.530222835413739e-07,
+      "loss": 0.263,
+      "step": 202
+    },
+    {
+      "epoch": 6.151515151515151,
+      "grad_norm": 0.23075253241023455,
+      "learning_rate": 8.894054121084839e-07,
+      "loss": 0.2652,
+      "step": 203
+    },
+    {
+      "epoch": 6.181818181818182,
+      "grad_norm": 0.22700574128593443,
+      "learning_rate": 8.278869849454718e-07,
+      "loss": 0.2495,
+      "step": 204
+    },
+    {
+      "epoch": 6.212121212121212,
+      "grad_norm": 0.23962162206606205,
+      "learning_rate": 7.684811715985429e-07,
+      "loss": 0.2635,
+      "step": 205
+    },
+    {
+      "epoch": 6.242424242424242,
+      "grad_norm": 0.33390979469076315,
+      "learning_rate": 7.1120165501533e-07,
+      "loss": 0.2691,
+      "step": 206
+    },
+    {
+      "epoch": 6.2727272727272725,
+      "grad_norm": 0.24819559014018444,
+      "learning_rate": 6.560616283932897e-07,
+      "loss": 0.2705,
+      "step": 207
+    },
+    {
+      "epoch": 6.303030303030303,
+      "grad_norm": 0.2600890340128434,
+      "learning_rate": 6.030737921409169e-07,
+      "loss": 0.2672,
+      "step": 208
+    },
+    {
+      "epoch": 6.333333333333333,
+      "grad_norm": 0.3911203830207476,
+      "learning_rate": 5.522503509524591e-07,
+      "loss": 0.2619,
+      "step": 209
+    },
+    {
+      "epoch": 6.363636363636363,
+      "grad_norm": 0.22909261281290824,
+      "learning_rate": 5.036030109968082e-07,
+      "loss": 0.2707,
+      "step": 210
+    },
+    {
+      "epoch": 6.393939393939394,
+      "grad_norm": 0.21607504208098888,
+      "learning_rate": 4.5714297722121105e-07,
+      "loss": 0.2582,
+      "step": 211
+    },
+    {
+      "epoch": 6.424242424242424,
+      "grad_norm": 0.2306256207445463,
+      "learning_rate": 4.128809507704445e-07,
+      "loss": 0.2684,
+      "step": 212
+    },
+    {
+      "epoch": 6.454545454545454,
+      "grad_norm": 0.22395608903910025,
+      "learning_rate": 3.708271265220087e-07,
+      "loss": 0.2573,
+      "step": 213
+    },
+    {
+      "epoch": 6.484848484848484,
+      "grad_norm": 0.22022957019820769,
+      "learning_rate": 3.309911907379393e-07,
+      "loss": 0.2557,
+      "step": 214
+    },
+    {
+      "epoch": 6.515151515151516,
+      "grad_norm": 0.22210833297225904,
+      "learning_rate": 2.9338231883378365e-07,
+      "loss": 0.2543,
+      "step": 215
+    },
+    {
+      "epoch": 6.545454545454545,
+      "grad_norm": 0.23110809543643812,
+      "learning_rate": 2.5800917326521013e-07,
+      "loss": 0.2662,
+      "step": 216
+    },
+    {
+      "epoch": 6.575757575757576,
+      "grad_norm": 0.22296248460503695,
+      "learning_rate": 2.248799015327907e-07,
+      "loss": 0.256,
+      "step": 217
+    },
+    {
+      "epoch": 6.606060606060606,
+      "grad_norm": 0.20990335785125108,
+      "learning_rate": 1.9400213430538773e-07,
+      "loss": 0.2561,
+      "step": 218
+    },
+    {
+      "epoch": 6.636363636363637,
+      "grad_norm": 0.2165769544705027,
+      "learning_rate": 1.6538298366257975e-07,
+      "loss": 0.263,
+      "step": 219
+    },
+    {
+      "epoch": 6.666666666666667,
+      "grad_norm": 0.21781489536538548,
+      "learning_rate": 1.3902904145653094e-07,
+      "loss": 0.2535,
+      "step": 220
+    },
+    {
+      "epoch": 6.696969696969697,
+      "grad_norm": 0.20910433160702532,
+      "learning_rate": 1.1494637779369766e-07,
+      "loss": 0.2555,
+      "step": 221
+    },
+    {
+      "epoch": 6.7272727272727275,
+      "grad_norm": 0.2236679596736996,
+      "learning_rate": 9.314053963669245e-08,
+      "loss": 0.2644,
+      "step": 222
+    },
+    {
+      "epoch": 6.757575757575758,
+      "grad_norm": 0.27223019451133684,
+      "learning_rate": 7.361654952665608e-08,
+      "loss": 0.2707,
+      "step": 223
+    },
+    {
+      "epoch": 6.787878787878788,
+      "grad_norm": 0.21385109093961496,
+      "learning_rate": 5.637890442641403e-08,
+      "loss": 0.265,
+      "step": 224
+    },
+    {
+      "epoch": 6.818181818181818,
+      "grad_norm": 0.22145945419878782,
+      "learning_rate": 4.143157468468717e-08,
+      "loss": 0.2608,
+      "step": 225
+    },
+    {
+      "epoch": 6.848484848484849,
+      "grad_norm": 0.21308961257858408,
+      "learning_rate": 2.8778003121607834e-08,
+      "loss": 0.2551,
+      "step": 226
+    },
+    {
+      "epoch": 6.878787878787879,
+      "grad_norm": 0.2135751235389027,
+      "learning_rate": 1.8421104235727406e-08,
+      "loss": 0.2647,
+      "step": 227
+    },
+    {
+      "epoch": 6.909090909090909,
+      "grad_norm": 0.21280976371458274,
+      "learning_rate": 1.0363263532724433e-08,
+      "loss": 0.2552,
+      "step": 228
+    },
+    {
+      "epoch": 6.9393939393939394,
+      "grad_norm": 0.2128839202522525,
+      "learning_rate": 4.606336975948589e-09,
+      "loss": 0.2584,
+      "step": 229
+    },
+    {
+      "epoch": 6.96969696969697,
+      "grad_norm": 0.2185104258110578,
+      "learning_rate": 1.1516505589381777e-09,
+      "loss": 0.2668,
+      "step": 230
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 0.21354802243051257,
+      "learning_rate": 0.0,
+      "loss": 0.2646,
+      "step": 231
+    },
+    {
+      "epoch": 7.0,
+      "step": 231,
+      "total_flos": 7.881049861301207e+17,
+      "train_loss": 0.3857915287281012,
+      "train_runtime": 8985.043,
+      "train_samples_per_second": 2.462,
+      "train_steps_per_second": 0.026
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 231,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 7,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.881049861301207e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a41c763b93aa88f248e3760798ed0c80bf59a23e47bd7b220fa1849ac67d80bc
+size 7352