Delete Qwen3-4B-Math-R1-CoT-SFT

Qwen3-4B-Math-R1-CoT-SFT/README.md

@@ -1,59 +0,0 @@
----
-library_name: transformers
-license: other
-base_model: models/Qwen3-4B-Base
-tags:
-- llama-factory
-- full
-- generated_from_trainer
-model-index:
-- name: sft
-  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. -->
-
-# sft
-
-This model is a fine-tuned version of [models/Qwen3-4B-Base](https://huggingface.co/models/Qwen3-4B-Base) on the open_thoughts 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: 5e-05
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- gradient_accumulation_steps: 32
-- total_train_batch_size: 64
-- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 0.05
-- num_epochs: 4.0
-
-### Training results
-
-
-
-### Framework versions
-
-- Transformers 4.52.4
-- Pytorch 2.8.0+cu128
-- Datasets 3.6.0
-- Tokenizers 0.21.4

Qwen3-4B-Math-R1-CoT-SFT/added_tokens.json

@@ -1,28 +0,0 @@
-{
-  "</think>": 151668,
-  "</tool_call>": 151658,
-  "</tool_response>": 151666,
-  "<think>": 151667,
-  "<tool_call>": 151657,
-  "<tool_response>": 151665,
-  "<|box_end|>": 151649,
-  "<|box_start|>": 151648,
-  "<|endoftext|>": 151643,
-  "<|file_sep|>": 151664,
-  "<|fim_middle|>": 151660,
-  "<|fim_pad|>": 151662,
-  "<|fim_prefix|>": 151659,
-  "<|fim_suffix|>": 151661,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644,
-  "<|image_pad|>": 151655,
-  "<|object_ref_end|>": 151647,
-  "<|object_ref_start|>": 151646,
-  "<|quad_end|>": 151651,
-  "<|quad_start|>": 151650,
-  "<|repo_name|>": 151663,
-  "<|video_pad|>": 151656,
-  "<|vision_end|>": 151653,
-  "<|vision_pad|>": 151654,
-  "<|vision_start|>": 151652
-}

Qwen3-4B-Math-R1-CoT-SFT/all_results.json

@@ -1,9 +0,0 @@
-{
-    "effective_tokens_per_sec": 24290.389169769056,
-    "epoch": 4.0,
-    "total_flos": 4.3530238414695694e+19,
-    "train_loss": 0.15123671168837785,
-    "train_runtime": 82201.3053,
-    "train_samples_per_second": 1.977,
-    "train_steps_per_second": 0.031
-}

Qwen3-4B-Math-R1-CoT-SFT/chat_template.jinja

@@ -1,85 +0,0 @@
-{%- if tools %}
-    {{- '<|im_start|>system\n' }}
-    {%- if messages[0].role == 'system' %}
-        {{- messages[0].content + '\n\n' }}
-    {%- endif %}
-    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
-    {%- for tool in tools %}
-        {{- "\n" }}
-        {{- tool | tojson }}
-    {%- endfor %}
-    {{- "\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" }}
-{%- else %}
-    {%- if messages[0].role == 'system' %}
-        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
-    {%- endif %}
-{%- endif %}
-{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
-{%- for message in messages[::-1] %}
-    {%- set index = (messages|length - 1) - loop.index0 %}
-    {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
-        {%- set ns.multi_step_tool = false %}
-        {%- set ns.last_query_index = index %}
-    {%- endif %}
-{%- endfor %}
-{%- for message in messages %}
-    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
-        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
-    {%- elif message.role == "assistant" %}
-        {%- set content = message.content %}
-        {%- set reasoning_content = '' %}
-        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
-            {%- set reasoning_content = message.reasoning_content %}
-        {%- else %}
-            {%- if '</think>' in message.content %}
-                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
-                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
-            {%- endif %}
-        {%- endif %}
-        {%- if loop.index0 > ns.last_query_index %}
-            {%- if loop.last or (not loop.last and reasoning_content) %}
-                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
-            {%- else %}
-                {{- '<|im_start|>' + message.role + '\n' + content }}
-            {%- endif %}
-        {%- else %}
-            {{- '<|im_start|>' + message.role + '\n' + content }}
-        {%- endif %}
-        {%- if message.tool_calls %}
-            {%- for tool_call in message.tool_calls %}
-                {%- if (loop.first and content) or (not loop.first) %}
-                    {{- '\n' }}
-                {%- endif %}
-                {%- if tool_call.function %}
-                    {%- set tool_call = tool_call.function %}
-                {%- endif %}
-                {{- '<tool_call>\n{"name": "' }}
-                {{- tool_call.name }}
-                {{- '", "arguments": ' }}
-                {%- if tool_call.arguments is string %}
-                    {{- tool_call.arguments }}
-                {%- else %}
-                    {{- tool_call.arguments | tojson }}
-                {%- endif %}
-                {{- '}\n</tool_call>' }}
-            {%- endfor %}
-        {%- endif %}
-        {{- '<|im_end|>\n' }}
-    {%- elif message.role == "tool" %}
-        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
-            {{- '<|im_start|>user' }}
-        {%- endif %}
-        {{- '\n<tool_response>\n' }}
-        {{- message.content }}
-        {{- '\n</tool_response>' }}
-        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
-            {{- '<|im_end|>\n' }}
-        {%- endif %}
-    {%- endif %}
-{%- endfor %}
-{%- if add_generation_prompt %}
-    {{- '<|im_start|>assistant\n' }}
-    {%- if enable_thinking is defined and enable_thinking is false %}
-        {{- '<think>\n\n</think>\n\n' }}
-    {%- endif %}
-{%- endif %}

Qwen3-4B-Math-R1-CoT-SFT/config.json

@@ -1,30 +0,0 @@
-{
-  "architectures": [
-    "Qwen3ForCausalLM"
-  ],
-  "attention_bias": false,
-  "attention_dropout": 0.0,
-  "bos_token_id": 151643,
-  "eos_token_id": 151643,
-  "head_dim": 128,
-  "hidden_act": "silu",
-  "hidden_size": 2560,
-  "initializer_range": 0.02,
-  "intermediate_size": 9728,
-  "max_position_embeddings": 32768,
-  "max_window_layers": 36,
-  "model_type": "qwen3",
-  "num_attention_heads": 32,
-  "num_hidden_layers": 36,
-  "num_key_value_heads": 8,
-  "rms_norm_eps": 1e-06,
-  "rope_scaling": null,
-  "rope_theta": 1000000,
-  "sliding_window": null,
-  "tie_word_embeddings": true,
-  "torch_dtype": "bfloat16",
-  "transformers_version": "4.52.4",
-  "use_cache": false,
-  "use_sliding_window": false,
-  "vocab_size": 151936
-}

Qwen3-4B-Math-R1-CoT-SFT/generation_config.json

@@ -1,6 +0,0 @@
-{
-  "bos_token_id": 151643,
-  "eos_token_id": 151643,
-  "max_new_tokens": 2048,
-  "transformers_version": "4.52.4"
-}

Qwen3-4B-Math-R1-CoT-SFT/model-00001-of-00002.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4b79f2dc490d7dd405409615b2970b2913972c694d45d30c37c59e61e5fb427f
-size 4967215360

Qwen3-4B-Math-R1-CoT-SFT/model-00002-of-00002.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:dfbef910dba4b28ae32760f46c94c7ecc11cfdca910cee94c61fcb102a1ac409
-size 3855679144

Qwen3-4B-Math-R1-CoT-SFT/model.safetensors.index.json

@@ -1,406 +0,0 @@
-{
-  "metadata": {
-    "total_size": 8822848512
-  },
-  "weight_map": {
-    "lm_head.weight": "model-00002-of-00002.safetensors",
-    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.20.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.28.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.29.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.30.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.30.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.31.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.32.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.33.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.34.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.input_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.self_attn.k_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.self_attn.q_norm.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.35.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
-    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
-    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
-    "model.norm.weight": "model-00002-of-00002.safetensors"
-  }
-}

Qwen3-4B-Math-R1-CoT-SFT/special_tokens_map.json

@@ -1,31 +0,0 @@
-{
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "eos_token": {
-    "content": "<|im_end|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  }
-}

Qwen3-4B-Math-R1-CoT-SFT/tokenizer.json

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
-size 11422654

Qwen3-4B-Math-R1-CoT-SFT/tokenizer_config.json

@@ -1,240 +0,0 @@
-{
-  "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
-    },
-    "151665": {
-      "content": "<tool_response>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151666": {
-      "content": "</tool_response>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151667": {
-      "content": "<think>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151668": {
-      "content": "</think>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    }
-  },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "bos_token": null,
-  "clean_up_tokenization_spaces": false,
-  "eos_token": "<|im_end|>",
-  "errors": "replace",
-  "extra_special_tokens": {},
-  "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
-  "padding_side": "right",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
-}

Qwen3-4B-Math-R1-CoT-SFT/train_results.json

@@ -1,9 +0,0 @@
-{
-    "effective_tokens_per_sec": 24290.389169769056,
-    "epoch": 4.0,
-    "total_flos": 4.3530238414695694e+19,
-    "train_loss": 0.15123671168837785,
-    "train_runtime": 82201.3053,
-    "train_samples_per_second": 1.977,
-    "train_steps_per_second": 0.031
-}

Qwen3-4B-Math-R1-CoT-SFT/trainer_state.json

@@ -1,3592 +0,0 @@
-{
-  "best_global_step": null,
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 4.0,
-  "eval_steps": 500,
-  "global_step": 2539,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.007877116975187082,
-      "grad_norm": 3.437255382537842,
-      "learning_rate": 1.5748031496062992e-06,
-      "loss": 0.672,
-      "step": 5
-    },
-    {
-      "epoch": 0.015754233950374164,
-      "grad_norm": 1.9117826223373413,
-      "learning_rate": 3.5433070866141735e-06,
-      "loss": 0.6371,
-      "step": 10
-    },
-    {
-      "epoch": 0.023631350925561245,
-      "grad_norm": 1.74402916431427,
-      "learning_rate": 5.511811023622048e-06,
-      "loss": 0.6065,
-      "step": 15
-    },
-    {
-      "epoch": 0.03150846790074833,
-      "grad_norm": 0.9878258109092712,
-      "learning_rate": 7.4803149606299226e-06,
-      "loss": 0.5662,
-      "step": 20
-    },
-    {
-      "epoch": 0.03938558487593541,
-      "grad_norm": 0.8288599252700806,
-      "learning_rate": 9.448818897637795e-06,
-      "loss": 0.5486,
-      "step": 25
-    },
-    {
-      "epoch": 0.04726270185112249,
-      "grad_norm": 0.6100791096687317,
-      "learning_rate": 1.141732283464567e-05,
-      "loss": 0.5297,
-      "step": 30
-    },
-    {
-      "epoch": 0.05513981882630957,
-      "grad_norm": 0.4823300540447235,
-      "learning_rate": 1.3385826771653545e-05,
-      "loss": 0.5211,
-      "step": 35
-    },
-    {
-      "epoch": 0.06301693580149666,
-      "grad_norm": 0.4592387080192566,
-      "learning_rate": 1.535433070866142e-05,
-      "loss": 0.5176,
-      "step": 40
-    },
-    {
-      "epoch": 0.07089405277668373,
-      "grad_norm": 0.3606854975223541,
-      "learning_rate": 1.732283464566929e-05,
-      "loss": 0.5071,
-      "step": 45
-    },
-    {
-      "epoch": 0.07877116975187082,
-      "grad_norm": 0.4212338328361511,
-      "learning_rate": 1.9291338582677166e-05,
-      "loss": 0.4941,
-      "step": 50
-    },
-    {
-      "epoch": 0.08664828672705789,
-      "grad_norm": 0.33868658542633057,
-      "learning_rate": 2.125984251968504e-05,
-      "loss": 0.4888,
-      "step": 55
-    },
-    {
-      "epoch": 0.09452540370224498,
-      "grad_norm": 0.3226923942565918,
-      "learning_rate": 2.3228346456692916e-05,
-      "loss": 0.4819,
-      "step": 60
-    },
-    {
-      "epoch": 0.10240252067743207,
-      "grad_norm": 0.43986427783966064,
-      "learning_rate": 2.5196850393700788e-05,
-      "loss": 0.4881,
-      "step": 65
-    },
-    {
-      "epoch": 0.11027963765261914,
-      "grad_norm": 0.40936681628227234,
-      "learning_rate": 2.7165354330708666e-05,
-      "loss": 0.4871,
-      "step": 70
-    },
-    {
-      "epoch": 0.11815675462780623,
-      "grad_norm": 0.3969690799713135,
-      "learning_rate": 2.9133858267716534e-05,
-      "loss": 0.4721,
-      "step": 75
-    },
-    {
-      "epoch": 0.1260338716029933,
-      "grad_norm": 0.4242527186870575,
-      "learning_rate": 3.110236220472441e-05,
-      "loss": 0.4841,
-      "step": 80
-    },
-    {
-      "epoch": 0.13391098857818037,
-      "grad_norm": 0.48484691977500916,
-      "learning_rate": 3.3070866141732284e-05,
-      "loss": 0.4771,
-      "step": 85
-    },
-    {
-      "epoch": 0.14178810555336746,
-      "grad_norm": 0.5061110854148865,
-      "learning_rate": 3.5039370078740156e-05,
-      "loss": 0.4718,
-      "step": 90
-    },
-    {
-      "epoch": 0.14966522252855455,
-      "grad_norm": 0.4648071527481079,
-      "learning_rate": 3.7007874015748034e-05,
-      "loss": 0.474,
-      "step": 95
-    },
-    {
-      "epoch": 0.15754233950374164,
-      "grad_norm": 0.5444895625114441,
-      "learning_rate": 3.8976377952755905e-05,
-      "loss": 0.4761,
-      "step": 100
-    },
-    {
-      "epoch": 0.16541945647892872,
-      "grad_norm": 0.5805575251579285,
-      "learning_rate": 4.0944881889763784e-05,
-      "loss": 0.4668,
-      "step": 105
-    },
-    {
-      "epoch": 0.17329657345411578,
-      "grad_norm": 0.40570878982543945,
-      "learning_rate": 4.2913385826771655e-05,
-      "loss": 0.467,
-      "step": 110
-    },
-    {
-      "epoch": 0.18117369042930287,
-      "grad_norm": 0.44745564460754395,
-      "learning_rate": 4.488188976377953e-05,
-      "loss": 0.4801,
-      "step": 115
-    },
-    {
-      "epoch": 0.18905080740448996,
-      "grad_norm": 0.5328453183174133,
-      "learning_rate": 4.6850393700787405e-05,
-      "loss": 0.4663,
-      "step": 120
-    },
-    {
-      "epoch": 0.19692792437967704,
-      "grad_norm": 0.3628937602043152,
-      "learning_rate": 4.881889763779528e-05,
-      "loss": 0.4644,
-      "step": 125
-    },
-    {
-      "epoch": 0.20480504135486413,
-      "grad_norm": 0.44744938611984253,
-      "learning_rate": 4.999991524704281e-05,
-      "loss": 0.4695,
-      "step": 130
-    },
-    {
-      "epoch": 0.2126821583300512,
-      "grad_norm": 0.6168819069862366,
-      "learning_rate": 4.999896178287378e-05,
-      "loss": 0.4636,
-      "step": 135
-    },
-    {
-      "epoch": 0.22055927530523828,
-      "grad_norm": 0.4544437527656555,
-      "learning_rate": 4.999694895387824e-05,
-      "loss": 0.4728,
-      "step": 140
-    },
-    {
-      "epoch": 0.22843639228042537,
-      "grad_norm": 0.3462134003639221,
-      "learning_rate": 4.999387684535255e-05,
-      "loss": 0.4704,
-      "step": 145
-    },
-    {
-      "epoch": 0.23631350925561245,
-      "grad_norm": 0.40045836567878723,
-      "learning_rate": 4.998974558748146e-05,
-      "loss": 0.4688,
-      "step": 150
-    },
-    {
-      "epoch": 0.24419062623079954,
-      "grad_norm": 0.46955132484436035,
-      "learning_rate": 4.9984555355332604e-05,
-      "loss": 0.467,
-      "step": 155
-    },
-    {
-      "epoch": 0.2520677432059866,
-      "grad_norm": 0.4244947135448456,
-      "learning_rate": 4.997830636884911e-05,
-      "loss": 0.4656,
-      "step": 160
-    },
-    {
-      "epoch": 0.2599448601811737,
-      "grad_norm": 0.41033992171287537,
-      "learning_rate": 4.997099889284021e-05,
-      "loss": 0.4679,
-      "step": 165
-    },
-    {
-      "epoch": 0.26782197715636075,
-      "grad_norm": 0.39735573530197144,
-      "learning_rate": 4.996263323697011e-05,
-      "loss": 0.4686,
-      "step": 170
-    },
-    {
-      "epoch": 0.27569909413154786,
-      "grad_norm": 0.3747316300868988,
-      "learning_rate": 4.9953209755744784e-05,
-      "loss": 0.4701,
-      "step": 175
-    },
-    {
-      "epoch": 0.2835762111067349,
-      "grad_norm": 0.39879927039146423,
-      "learning_rate": 4.9942728848496995e-05,
-      "loss": 0.4629,
-      "step": 180
-    },
-    {
-      "epoch": 0.29145332808192204,
-      "grad_norm": 0.3395926356315613,
-      "learning_rate": 4.993119095936937e-05,
-      "loss": 0.4676,
-      "step": 185
-    },
-    {
-      "epoch": 0.2993304450571091,
-      "grad_norm": 0.35948389768600464,
-      "learning_rate": 4.991859657729557e-05,
-      "loss": 0.4468,
-      "step": 190
-    },
-    {
-      "epoch": 0.30720756203229616,
-      "grad_norm": 0.33689770102500916,
-      "learning_rate": 4.9904946235979574e-05,
-      "loss": 0.4624,
-      "step": 195
-    },
-    {
-      "epoch": 0.31508467900748327,
-      "grad_norm": 0.38664934039115906,
-      "learning_rate": 4.9890240513873064e-05,
-      "loss": 0.4611,
-      "step": 200
-    },
-    {
-      "epoch": 0.32296179598267033,
-      "grad_norm": 0.3196653723716736,
-      "learning_rate": 4.9874480034150915e-05,
-      "loss": 0.4604,
-      "step": 205
-    },
-    {
-      "epoch": 0.33083891295785745,
-      "grad_norm": 0.38587889075279236,
-      "learning_rate": 4.985766546468477e-05,
-      "loss": 0.4597,
-      "step": 210
-    },
-    {
-      "epoch": 0.3387160299330445,
-      "grad_norm": 0.3755643963813782,
-      "learning_rate": 4.983979751801475e-05,
-      "loss": 0.4638,
-      "step": 215
-    },
-    {
-      "epoch": 0.34659314690823156,
-      "grad_norm": 0.3800589442253113,
-      "learning_rate": 4.982087695131928e-05,
-      "loss": 0.4579,
-      "step": 220
-    },
-    {
-      "epoch": 0.3544702638834187,
-      "grad_norm": 0.3112618327140808,
-      "learning_rate": 4.9800904566382984e-05,
-      "loss": 0.445,
-      "step": 225
-    },
-    {
-      "epoch": 0.36234738085860574,
-      "grad_norm": 0.33558526635169983,
-      "learning_rate": 4.9779881209562675e-05,
-      "loss": 0.4552,
-      "step": 230
-    },
-    {
-      "epoch": 0.37022449783379285,
-      "grad_norm": 0.35156235098838806,
-      "learning_rate": 4.975780777175154e-05,
-      "loss": 0.4642,
-      "step": 235
-    },
-    {
-      "epoch": 0.3781016148089799,
-      "grad_norm": 0.34751108288764954,
-      "learning_rate": 4.9734685188341376e-05,
-      "loss": 0.454,
-      "step": 240
-    },
-    {
-      "epoch": 0.385978731784167,
-      "grad_norm": 0.3409808874130249,
-      "learning_rate": 4.971051443918294e-05,
-      "loss": 0.4605,
-      "step": 245
-    },
-    {
-      "epoch": 0.3938558487593541,
-      "grad_norm": 0.3360814154148102,
-      "learning_rate": 4.9685296548544424e-05,
-      "loss": 0.4642,
-      "step": 250
-    },
-    {
-      "epoch": 0.40173296573454115,
-      "grad_norm": 0.3765697777271271,
-      "learning_rate": 4.965903258506806e-05,
-      "loss": 0.4516,
-      "step": 255
-    },
-    {
-      "epoch": 0.40961008270972826,
-      "grad_norm": 0.3913894295692444,
-      "learning_rate": 4.9631723661724836e-05,
-      "loss": 0.4598,
-      "step": 260
-    },
-    {
-      "epoch": 0.4174871996849153,
-      "grad_norm": 0.3312707841396332,
-      "learning_rate": 4.9603370935767324e-05,
-      "loss": 0.4554,
-      "step": 265
-    },
-    {
-      "epoch": 0.4253643166601024,
-      "grad_norm": 0.2816666066646576,
-      "learning_rate": 4.957397560868064e-05,
-      "loss": 0.4565,
-      "step": 270
-    },
-    {
-      "epoch": 0.4332414336352895,
-      "grad_norm": 0.3484485447406769,
-      "learning_rate": 4.954353892613154e-05,
-      "loss": 0.4589,
-      "step": 275
-    },
-    {
-      "epoch": 0.44111855061047656,
-      "grad_norm": 0.3825587332248688,
-      "learning_rate": 4.951206217791564e-05,
-      "loss": 0.4598,
-      "step": 280
-    },
-    {
-      "epoch": 0.44899566758566367,
-      "grad_norm": 0.3513055443763733,
-      "learning_rate": 4.94795466979027e-05,
-      "loss": 0.4495,
-      "step": 285
-    },
-    {
-      "epoch": 0.45687278456085073,
-      "grad_norm": 0.38703492283821106,
-      "learning_rate": 4.94459938639802e-05,
-      "loss": 0.442,
-      "step": 290
-    },
-    {
-      "epoch": 0.4647499015360378,
-      "grad_norm": 0.28819456696510315,
-      "learning_rate": 4.941140509799487e-05,
-      "loss": 0.4462,
-      "step": 295
-    },
-    {
-      "epoch": 0.4726270185112249,
-      "grad_norm": 0.2926427125930786,
-      "learning_rate": 4.937578186569244e-05,
-      "loss": 0.4513,
-      "step": 300
-    },
-    {
-      "epoch": 0.48050413548641197,
-      "grad_norm": 0.28402814269065857,
-      "learning_rate": 4.933912567665561e-05,
-      "loss": 0.4553,
-      "step": 305
-    },
-    {
-      "epoch": 0.4883812524615991,
-      "grad_norm": 0.31271660327911377,
-      "learning_rate": 4.930143808423995e-05,
-      "loss": 0.4389,
-      "step": 310
-    },
-    {
-      "epoch": 0.49625836943678614,
-      "grad_norm": 0.39874470233917236,
-      "learning_rate": 4.926272068550819e-05,
-      "loss": 0.4583,
-      "step": 315
-    },
-    {
-      "epoch": 0.5041354864119733,
-      "grad_norm": 0.3334454298019409,
-      "learning_rate": 4.922297512116248e-05,
-      "loss": 0.4603,
-      "step": 320
-    },
-    {
-      "epoch": 0.5120126033871603,
-      "grad_norm": 0.28501057624816895,
-      "learning_rate": 4.9182203075474884e-05,
-      "loss": 0.4539,
-      "step": 325
-    },
-    {
-      "epoch": 0.5198897203623474,
-      "grad_norm": 0.3027922809123993,
-      "learning_rate": 4.914040627621599e-05,
-      "loss": 0.4573,
-      "step": 330
-    },
-    {
-      "epoch": 0.5277668373375345,
-      "grad_norm": 0.31392157077789307,
-      "learning_rate": 4.909758649458172e-05,
-      "loss": 0.4547,
-      "step": 335
-    },
-    {
-      "epoch": 0.5356439543127215,
-      "grad_norm": 0.30944597721099854,
-      "learning_rate": 4.905374554511825e-05,
-      "loss": 0.443,
-      "step": 340
-    },
-    {
-      "epoch": 0.5435210712879086,
-      "grad_norm": 0.2820224463939667,
-      "learning_rate": 4.9008885285645134e-05,
-      "loss": 0.4535,
-      "step": 345
-    },
-    {
-      "epoch": 0.5513981882630957,
-      "grad_norm": 0.2703908681869507,
-      "learning_rate": 4.8963007617176535e-05,
-      "loss": 0.4456,
-      "step": 350
-    },
-    {
-      "epoch": 0.5592753052382828,
-      "grad_norm": 0.2866289019584656,
-      "learning_rate": 4.8916114483840746e-05,
-      "loss": 0.4468,
-      "step": 355
-    },
-    {
-      "epoch": 0.5671524222134698,
-      "grad_norm": 0.30911338329315186,
-      "learning_rate": 4.886820787279773e-05,
-      "loss": 0.4419,
-      "step": 360
-    },
-    {
-      "epoch": 0.575029539188657,
-      "grad_norm": 0.2732813060283661,
-      "learning_rate": 4.8819289814154935e-05,
-      "loss": 0.4333,
-      "step": 365
-    },
-    {
-      "epoch": 0.5829066561638441,
-      "grad_norm": 0.32516050338745117,
-      "learning_rate": 4.876936238088127e-05,
-      "loss": 0.4466,
-      "step": 370
-    },
-    {
-      "epoch": 0.5907837731390311,
-      "grad_norm": 0.2863224446773529,
-      "learning_rate": 4.871842768871928e-05,
-      "loss": 0.4408,
-      "step": 375
-    },
-    {
-      "epoch": 0.5986608901142182,
-      "grad_norm": 0.27347731590270996,
-      "learning_rate": 4.866648789609545e-05,
-      "loss": 0.4422,
-      "step": 380
-    },
-    {
-      "epoch": 0.6065380070894053,
-      "grad_norm": 0.27762892842292786,
-      "learning_rate": 4.8613545204028764e-05,
-      "loss": 0.4425,
-      "step": 385
-    },
-    {
-      "epoch": 0.6144151240645923,
-      "grad_norm": 0.3145371675491333,
-      "learning_rate": 4.855960185603742e-05,
-      "loss": 0.4451,
-      "step": 390
-    },
-    {
-      "epoch": 0.6222922410397794,
-      "grad_norm": 0.2932167649269104,
-      "learning_rate": 4.850466013804377e-05,
-      "loss": 0.4434,
-      "step": 395
-    },
-    {
-      "epoch": 0.6301693580149665,
-      "grad_norm": 0.2880720794200897,
-      "learning_rate": 4.844872237827745e-05,
-      "loss": 0.4475,
-      "step": 400
-    },
-    {
-      "epoch": 0.6380464749901537,
-      "grad_norm": 0.3783859610557556,
-      "learning_rate": 4.8391790947176696e-05,
-      "loss": 0.446,
-      "step": 405
-    },
-    {
-      "epoch": 0.6459235919653407,
-      "grad_norm": 0.31301021575927734,
-      "learning_rate": 4.833386825728794e-05,
-      "loss": 0.4462,
-      "step": 410
-    },
-    {
-      "epoch": 0.6538007089405278,
-      "grad_norm": 0.28316691517829895,
-      "learning_rate": 4.8274956763163506e-05,
-      "loss": 0.4485,
-      "step": 415
-    },
-    {
-      "epoch": 0.6616778259157149,
-      "grad_norm": 0.3179584741592407,
-      "learning_rate": 4.821505896125769e-05,
-      "loss": 0.4372,
-      "step": 420
-    },
-    {
-      "epoch": 0.6695549428909019,
-      "grad_norm": 0.3088538646697998,
-      "learning_rate": 4.815417738982085e-05,
-      "loss": 0.4377,
-      "step": 425
-    },
-    {
-      "epoch": 0.677432059866089,
-      "grad_norm": 0.31807076930999756,
-      "learning_rate": 4.809231462879196e-05,
-      "loss": 0.4373,
-      "step": 430
-    },
-    {
-      "epoch": 0.6853091768412761,
-      "grad_norm": 0.2823260724544525,
-      "learning_rate": 4.802947329968921e-05,
-      "loss": 0.4375,
-      "step": 435
-    },
-    {
-      "epoch": 0.6931862938164631,
-      "grad_norm": 0.30851051211357117,
-      "learning_rate": 4.7965656065498944e-05,
-      "loss": 0.4498,
-      "step": 440
-    },
-    {
-      "epoch": 0.7010634107916502,
-      "grad_norm": 0.28332722187042236,
-      "learning_rate": 4.7900865630562786e-05,
-      "loss": 0.4342,
-      "step": 445
-    },
-    {
-      "epoch": 0.7089405277668374,
-      "grad_norm": 0.26585566997528076,
-      "learning_rate": 4.783510474046308e-05,
-      "loss": 0.4402,
-      "step": 450
-    },
-    {
-      "epoch": 0.7168176447420245,
-      "grad_norm": 0.26203709840774536,
-      "learning_rate": 4.77683761819065e-05,
-      "loss": 0.4393,
-      "step": 455
-    },
-    {
-      "epoch": 0.7246947617172115,
-      "grad_norm": 0.27073633670806885,
-      "learning_rate": 4.7700682782606015e-05,
-      "loss": 0.4378,
-      "step": 460
-    },
-    {
-      "epoch": 0.7325718786923986,
-      "grad_norm": 0.2871417999267578,
-      "learning_rate": 4.763202741116098e-05,
-      "loss": 0.4362,
-      "step": 465
-    },
-    {
-      "epoch": 0.7404489956675857,
-      "grad_norm": 0.27669113874435425,
-      "learning_rate": 4.756241297693566e-05,
-      "loss": 0.4269,
-      "step": 470
-    },
-    {
-      "epoch": 0.7483261126427727,
-      "grad_norm": 0.26247602701187134,
-      "learning_rate": 4.749184242993587e-05,
-      "loss": 0.442,
-      "step": 475
-    },
-    {
-      "epoch": 0.7562032296179598,
-      "grad_norm": 0.2640887498855591,
-      "learning_rate": 4.742031876068402e-05,
-      "loss": 0.4435,
-      "step": 480
-    },
-    {
-      "epoch": 0.7640803465931469,
-      "grad_norm": 0.2536977231502533,
-      "learning_rate": 4.734784500009237e-05,
-      "loss": 0.4317,
-      "step": 485
-    },
-    {
-      "epoch": 0.771957463568334,
-      "grad_norm": 0.2515846788883209,
-      "learning_rate": 4.727442421933456e-05,
-      "loss": 0.4326,
-      "step": 490
-    },
-    {
-      "epoch": 0.7798345805435211,
-      "grad_norm": 0.2645900547504425,
-      "learning_rate": 4.720005952971551e-05,
-      "loss": 0.4336,
-      "step": 495
-    },
-    {
-      "epoch": 0.7877116975187082,
-      "grad_norm": 0.2777593433856964,
-      "learning_rate": 4.7124754082539546e-05,
-      "loss": 0.4402,
-      "step": 500
-    },
-    {
-      "epoch": 0.7955888144938952,
-      "grad_norm": 0.29428908228874207,
-      "learning_rate": 4.704851106897689e-05,
-      "loss": 0.4292,
-      "step": 505
-    },
-    {
-      "epoch": 0.8034659314690823,
-      "grad_norm": 0.28475651144981384,
-      "learning_rate": 4.6971333719928354e-05,
-      "loss": 0.4307,
-      "step": 510
-    },
-    {
-      "epoch": 0.8113430484442694,
-      "grad_norm": 0.2999843955039978,
-      "learning_rate": 4.6893225305888556e-05,
-      "loss": 0.4408,
-      "step": 515
-    },
-    {
-      "epoch": 0.8192201654194565,
-      "grad_norm": 0.28412893414497375,
-      "learning_rate": 4.68141891368072e-05,
-      "loss": 0.4375,
-      "step": 520
-    },
-    {
-      "epoch": 0.8270972823946435,
-      "grad_norm": 0.24423280358314514,
-      "learning_rate": 4.6734228561948864e-05,
-      "loss": 0.4413,
-      "step": 525
-    },
-    {
-      "epoch": 0.8349743993698306,
-      "grad_norm": 0.3219039738178253,
-      "learning_rate": 4.6653346969751096e-05,
-      "loss": 0.4402,
-      "step": 530
-    },
-    {
-      "epoch": 0.8428515163450178,
-      "grad_norm": 0.2778218984603882,
-      "learning_rate": 4.65715477876808e-05,
-      "loss": 0.4299,
-      "step": 535
-    },
-    {
-      "epoch": 0.8507286333202048,
-      "grad_norm": 0.2895450294017792,
-      "learning_rate": 4.6488834482088965e-05,
-      "loss": 0.4351,
-      "step": 540
-    },
-    {
-      "epoch": 0.8586057502953919,
-      "grad_norm": 0.28678175806999207,
-      "learning_rate": 4.640521055806383e-05,
-      "loss": 0.4467,
-      "step": 545
-    },
-    {
-      "epoch": 0.866482867270579,
-      "grad_norm": 0.28237128257751465,
-      "learning_rate": 4.632067955928231e-05,
-      "loss": 0.4256,
-      "step": 550
-    },
-    {
-      "epoch": 0.874359984245766,
-      "grad_norm": 0.2734551727771759,
-      "learning_rate": 4.6235245067859825e-05,
-      "loss": 0.4346,
-      "step": 555
-    },
-    {
-      "epoch": 0.8822371012209531,
-      "grad_norm": 0.25271713733673096,
-      "learning_rate": 4.614891070419855e-05,
-      "loss": 0.43,
-      "step": 560
-    },
-    {
-      "epoch": 0.8901142181961402,
-      "grad_norm": 0.2673031687736511,
-      "learning_rate": 4.606168012683394e-05,
-      "loss": 0.4334,
-      "step": 565
-    },
-    {
-      "epoch": 0.8979913351713273,
-      "grad_norm": 0.24039021134376526,
-      "learning_rate": 4.5973557032279726e-05,
-      "loss": 0.4354,
-      "step": 570
-    },
-    {
-      "epoch": 0.9058684521465143,
-      "grad_norm": 0.2770460546016693,
-      "learning_rate": 4.588454515487125e-05,
-      "loss": 0.4277,
-      "step": 575
-    },
-    {
-      "epoch": 0.9137455691217015,
-      "grad_norm": 0.2645154893398285,
-      "learning_rate": 4.5794648266607254e-05,
-      "loss": 0.4242,
-      "step": 580
-    },
-    {
-      "epoch": 0.9216226860968886,
-      "grad_norm": 0.2790268659591675,
-      "learning_rate": 4.570387017699e-05,
-      "loss": 0.4276,
-      "step": 585
-    },
-    {
-      "epoch": 0.9294998030720756,
-      "grad_norm": 0.2759014070034027,
-      "learning_rate": 4.561221473286384e-05,
-      "loss": 0.4365,
-      "step": 590
-    },
-    {
-      "epoch": 0.9373769200472627,
-      "grad_norm": 0.27472007274627686,
-      "learning_rate": 4.551968581825222e-05,
-      "loss": 0.4318,
-      "step": 595
-    },
-    {
-      "epoch": 0.9452540370224498,
-      "grad_norm": 0.27848538756370544,
-      "learning_rate": 4.5426287354193095e-05,
-      "loss": 0.4302,
-      "step": 600
-    },
-    {
-      "epoch": 0.9531311539976368,
-      "grad_norm": 0.25084635615348816,
-      "learning_rate": 4.533202329857271e-05,
-      "loss": 0.4398,
-      "step": 605
-    },
-    {
-      "epoch": 0.9610082709728239,
-      "grad_norm": 0.2645019590854645,
-      "learning_rate": 4.523689764595797e-05,
-      "loss": 0.4352,
-      "step": 610
-    },
-    {
-      "epoch": 0.968885387948011,
-      "grad_norm": 0.2712864875793457,
-      "learning_rate": 4.5140914427427094e-05,
-      "loss": 0.4266,
-      "step": 615
-    },
-    {
-      "epoch": 0.9767625049231982,
-      "grad_norm": 0.2347964495420456,
-      "learning_rate": 4.504407771039883e-05,
-      "loss": 0.4249,
-      "step": 620
-    },
-    {
-      "epoch": 0.9846396218983852,
-      "grad_norm": 0.23019365966320038,
-      "learning_rate": 4.494639159846006e-05,
-      "loss": 0.4292,
-      "step": 625
-    },
-    {
-      "epoch": 0.9925167388735723,
-      "grad_norm": 0.22928465902805328,
-      "learning_rate": 4.484786023119194e-05,
-      "loss": 0.4309,
-      "step": 630
-    },
-    {
-      "epoch": 1.0,
-      "grad_norm": 0.253717303276062,
-      "learning_rate": 4.474848778399446e-05,
-      "loss": 0.4261,
-      "step": 635
-    },
-    {
-      "epoch": 1.007877116975187,
-      "grad_norm": 0.35631492733955383,
-      "learning_rate": 4.464827846790952e-05,
-      "loss": 0.388,
-      "step": 640
-    },
-    {
-      "epoch": 1.0157542339503742,
-      "grad_norm": 0.3446408212184906,
-      "learning_rate": 4.454723652944246e-05,
-      "loss": 0.3889,
-      "step": 645
-    },
-    {
-      "epoch": 1.0236313509255612,
-      "grad_norm": 0.29462265968322754,
-      "learning_rate": 4.4445366250382106e-05,
-      "loss": 0.3927,
-      "step": 650
-    },
-    {
-      "epoch": 1.0315084679007482,
-      "grad_norm": 0.26057687401771545,
-      "learning_rate": 4.434267194761936e-05,
-      "loss": 0.3883,
-      "step": 655
-    },
-    {
-      "epoch": 1.0393855848759355,
-      "grad_norm": 0.2740859091281891,
-      "learning_rate": 4.423915797296425e-05,
-      "loss": 0.394,
-      "step": 660
-    },
-    {
-      "epoch": 1.0472627018511225,
-      "grad_norm": 0.2503489553928375,
-      "learning_rate": 4.413482871296148e-05,
-      "loss": 0.3769,
-      "step": 665
-    },
-    {
-      "epoch": 1.0551398188263095,
-      "grad_norm": 0.2788833677768707,
-      "learning_rate": 4.402968858870459e-05,
-      "loss": 0.3866,
-      "step": 670
-    },
-    {
-      "epoch": 1.0630169358014967,
-      "grad_norm": 0.24859720468521118,
-      "learning_rate": 4.39237420556486e-05,
-      "loss": 0.3856,
-      "step": 675
-    },
-    {
-      "epoch": 1.0708940527766837,
-      "grad_norm": 0.26013121008872986,
-      "learning_rate": 4.381699360342116e-05,
-      "loss": 0.3929,
-      "step": 680
-    },
-    {
-      "epoch": 1.078771169751871,
-      "grad_norm": 0.2571970522403717,
-      "learning_rate": 4.370944775563239e-05,
-      "loss": 0.3664,
-      "step": 685
-    },
-    {
-      "epoch": 1.086648286727058,
-      "grad_norm": 0.2991076111793518,
-      "learning_rate": 4.360110906968305e-05,
-      "loss": 0.379,
-      "step": 690
-    },
-    {
-      "epoch": 1.094525403702245,
-      "grad_norm": 0.7817800045013428,
-      "learning_rate": 4.3491982136571554e-05,
-      "loss": 0.3871,
-      "step": 695
-    },
-    {
-      "epoch": 1.1024025206774322,
-      "grad_norm": 0.31027108430862427,
-      "learning_rate": 4.3382071580699314e-05,
-      "loss": 0.39,
-      "step": 700
-    },
-    {
-      "epoch": 1.1102796376526192,
-      "grad_norm": 0.27426421642303467,
-      "learning_rate": 4.327138205967484e-05,
-      "loss": 0.3832,
-      "step": 705
-    },
-    {
-      "epoch": 1.1181567546278062,
-      "grad_norm": 0.27434617280960083,
-      "learning_rate": 4.3159918264116334e-05,
-      "loss": 0.3937,
-      "step": 710
-    },
-    {
-      "epoch": 1.1260338716029934,
-      "grad_norm": 0.2739649713039398,
-      "learning_rate": 4.3047684917452936e-05,
-      "loss": 0.3971,
-      "step": 715
-    },
-    {
-      "epoch": 1.1339109885781804,
-      "grad_norm": 0.2901259958744049,
-      "learning_rate": 4.293468677572454e-05,
-      "loss": 0.387,
-      "step": 720
-    },
-    {
-      "epoch": 1.1417881055533674,
-      "grad_norm": 0.30935901403427124,
-      "learning_rate": 4.282092862738029e-05,
-      "loss": 0.3787,
-      "step": 725
-    },
-    {
-      "epoch": 1.1496652225285546,
-      "grad_norm": 0.27035894989967346,
-      "learning_rate": 4.270641529307562e-05,
-      "loss": 0.3727,
-      "step": 730
-    },
-    {
-      "epoch": 1.1575423395037416,
-      "grad_norm": 0.262545645236969,
-      "learning_rate": 4.2591151625467986e-05,
-      "loss": 0.3866,
-      "step": 735
-    },
-    {
-      "epoch": 1.1654194564789286,
-      "grad_norm": 0.2450433075428009,
-      "learning_rate": 4.2475142509011254e-05,
-      "loss": 0.3836,
-      "step": 740
-    },
-    {
-      "epoch": 1.1732965734541159,
-      "grad_norm": 0.24818231165409088,
-      "learning_rate": 4.235839285974868e-05,
-      "loss": 0.3854,
-      "step": 745
-    },
-    {
-      "epoch": 1.1811736904293029,
-      "grad_norm": 0.241587832570076,
-      "learning_rate": 4.2240907625104606e-05,
-      "loss": 0.3862,
-      "step": 750
-    },
-    {
-      "epoch": 1.1890508074044899,
-      "grad_norm": 0.24221687018871307,
-      "learning_rate": 4.2122691783674786e-05,
-      "loss": 0.382,
-      "step": 755
-    },
-    {
-      "epoch": 1.196927924379677,
-      "grad_norm": 0.24578852951526642,
-      "learning_rate": 4.200375034501543e-05,
-      "loss": 0.3812,
-      "step": 760
-    },
-    {
-      "epoch": 1.204805041354864,
-      "grad_norm": 0.27418452501296997,
-      "learning_rate": 4.188408834943093e-05,
-      "loss": 0.3894,
-      "step": 765
-    },
-    {
-      "epoch": 1.212682158330051,
-      "grad_norm": 0.28048843145370483,
-      "learning_rate": 4.176371086776023e-05,
-      "loss": 0.3695,
-      "step": 770
-    },
-    {
-      "epoch": 1.2205592753052383,
-      "grad_norm": 0.2632172703742981,
-      "learning_rate": 4.1642623001161976e-05,
-      "loss": 0.3846,
-      "step": 775
-    },
-    {
-      "epoch": 1.2284363922804253,
-      "grad_norm": 0.25057244300842285,
-      "learning_rate": 4.152082988089834e-05,
-      "loss": 0.3892,
-      "step": 780
-    },
-    {
-      "epoch": 1.2363135092556123,
-      "grad_norm": 0.28263312578201294,
-      "learning_rate": 4.139833666811756e-05,
-      "loss": 0.3832,
-      "step": 785
-    },
-    {
-      "epoch": 1.2441906262307996,
-      "grad_norm": 0.26317495107650757,
-      "learning_rate": 4.127514855363524e-05,
-      "loss": 0.3831,
-      "step": 790
-    },
-    {
-      "epoch": 1.2520677432059866,
-      "grad_norm": 0.28825777769088745,
-      "learning_rate": 4.1151270757714405e-05,
-      "loss": 0.3902,
-      "step": 795
-    },
-    {
-      "epoch": 1.2599448601811738,
-      "grad_norm": 0.23847556114196777,
-      "learning_rate": 4.102670852984425e-05,
-      "loss": 0.3871,
-      "step": 800
-    },
-    {
-      "epoch": 1.2678219771563608,
-      "grad_norm": 0.27880316972732544,
-      "learning_rate": 4.090146714851769e-05,
-      "loss": 0.3822,
-      "step": 805
-    },
-    {
-      "epoch": 1.2756990941315478,
-      "grad_norm": 0.25444021821022034,
-      "learning_rate": 4.077555192100772e-05,
-      "loss": 0.382,
-      "step": 810
-    },
-    {
-      "epoch": 1.2835762111067348,
-      "grad_norm": 0.2789304554462433,
-      "learning_rate": 4.0648968183142424e-05,
-      "loss": 0.3783,
-      "step": 815
-    },
-    {
-      "epoch": 1.291453328081922,
-      "grad_norm": 0.2782425880432129,
-      "learning_rate": 4.052172129907897e-05,
-      "loss": 0.3965,
-      "step": 820
-    },
-    {
-      "epoch": 1.299330445057109,
-      "grad_norm": 0.25539177656173706,
-      "learning_rate": 4.039381666107621e-05,
-      "loss": 0.3897,
-      "step": 825
-    },
-    {
-      "epoch": 1.3072075620322963,
-      "grad_norm": 0.26664361357688904,
-      "learning_rate": 4.026525968926624e-05,
-      "loss": 0.3816,
-      "step": 830
-    },
-    {
-      "epoch": 1.3150846790074833,
-      "grad_norm": 0.2982277274131775,
-      "learning_rate": 4.013605583142464e-05,
-      "loss": 0.3967,
-      "step": 835
-    },
-    {
-      "epoch": 1.3229617959826703,
-      "grad_norm": 0.24720807373523712,
-      "learning_rate": 4.000621056273971e-05,
-      "loss": 0.3859,
-      "step": 840
-    },
-    {
-      "epoch": 1.3308389129578575,
-      "grad_norm": 0.2673453986644745,
-      "learning_rate": 3.987572938558035e-05,
-      "loss": 0.3862,
-      "step": 845
-    },
-    {
-      "epoch": 1.3387160299330445,
-      "grad_norm": 0.242339089512825,
-      "learning_rate": 3.974461782926299e-05,
-      "loss": 0.3903,
-      "step": 850
-    },
-    {
-      "epoch": 1.3465931469082315,
-      "grad_norm": 0.24661767482757568,
-      "learning_rate": 3.96128814498172e-05,
-      "loss": 0.3831,
-      "step": 855
-    },
-    {
-      "epoch": 1.3544702638834187,
-      "grad_norm": 0.25107643008232117,
-      "learning_rate": 3.9480525829750295e-05,
-      "loss": 0.3815,
-      "step": 860
-    },
-    {
-      "epoch": 1.3623473808586057,
-      "grad_norm": 0.25703874230384827,
-      "learning_rate": 3.9347556577810724e-05,
-      "loss": 0.3784,
-      "step": 865
-    },
-    {
-      "epoch": 1.3702244978337927,
-      "grad_norm": 0.24667410552501678,
-      "learning_rate": 3.921397932875044e-05,
-      "loss": 0.3832,
-      "step": 870
-    },
-    {
-      "epoch": 1.37810161480898,
-      "grad_norm": 0.24698962271213531,
-      "learning_rate": 3.9079799743086076e-05,
-      "loss": 0.3807,
-      "step": 875
-    },
-    {
-      "epoch": 1.385978731784167,
-      "grad_norm": 0.2628852128982544,
-      "learning_rate": 3.8945023506859125e-05,
-      "loss": 0.3833,
-      "step": 880
-    },
-    {
-      "epoch": 1.3938558487593542,
-      "grad_norm": 0.24828363955020905,
-      "learning_rate": 3.880965633139493e-05,
-      "loss": 0.3893,
-      "step": 885
-    },
-    {
-      "epoch": 1.4017329657345412,
-      "grad_norm": 0.25798603892326355,
-      "learning_rate": 3.867370395306068e-05,
-      "loss": 0.3908,
-      "step": 890
-    },
-    {
-      "epoch": 1.4096100827097282,
-      "grad_norm": 0.24343079328536987,
-      "learning_rate": 3.8537172133022334e-05,
-      "loss": 0.3838,
-      "step": 895
-    },
-    {
-      "epoch": 1.4174871996849152,
-      "grad_norm": 0.25076451897621155,
-      "learning_rate": 3.840006665700049e-05,
-      "loss": 0.3865,
-      "step": 900
-    },
-    {
-      "epoch": 1.4253643166601024,
-      "grad_norm": 0.24691860377788544,
-      "learning_rate": 3.8262393335025175e-05,
-      "loss": 0.3785,
-      "step": 905
-    },
-    {
-      "epoch": 1.4332414336352894,
-      "grad_norm": 0.2401050329208374,
-      "learning_rate": 3.812415800118968e-05,
-      "loss": 0.3907,
-      "step": 910
-    },
-    {
-      "epoch": 1.4411185506104767,
-      "grad_norm": 0.24668364226818085,
-      "learning_rate": 3.7985366513403275e-05,
-      "loss": 0.3744,
-      "step": 915
-    },
-    {
-      "epoch": 1.4489956675856637,
-      "grad_norm": 0.24690192937850952,
-      "learning_rate": 3.784602475314303e-05,
-      "loss": 0.3841,
-      "step": 920
-    },
-    {
-      "epoch": 1.4568727845608507,
-      "grad_norm": 0.25426268577575684,
-      "learning_rate": 3.7706138625204565e-05,
-      "loss": 0.3867,
-      "step": 925
-    },
-    {
-      "epoch": 1.4647499015360377,
-      "grad_norm": 0.23657885193824768,
-      "learning_rate": 3.756571405745178e-05,
-      "loss": 0.3793,
-      "step": 930
-    },
-    {
-      "epoch": 1.472627018511225,
-      "grad_norm": 0.2615351974964142,
-      "learning_rate": 3.7424757000565734e-05,
-      "loss": 0.3904,
-      "step": 935
-    },
-    {
-      "epoch": 1.480504135486412,
-      "grad_norm": 0.30141061544418335,
-      "learning_rate": 3.7283273427792407e-05,
-      "loss": 0.3871,
-      "step": 940
-    },
-    {
-      "epoch": 1.4883812524615991,
-      "grad_norm": 0.23688402771949768,
-      "learning_rate": 3.714126933468959e-05,
-      "loss": 0.3846,
-      "step": 945
-    },
-    {
-      "epoch": 1.4962583694367861,
-      "grad_norm": 0.22304469347000122,
-      "learning_rate": 3.6998750738872876e-05,
-      "loss": 0.3819,
-      "step": 950
-    },
-    {
-      "epoch": 1.5041354864119731,
-      "grad_norm": 0.25306451320648193,
-      "learning_rate": 3.6855723679760545e-05,
-      "loss": 0.3884,
-      "step": 955
-    },
-    {
-      "epoch": 1.5120126033871601,
-      "grad_norm": 0.24044828116893768,
-      "learning_rate": 3.671219421831775e-05,
-      "loss": 0.3819,
-      "step": 960
-    },
-    {
-      "epoch": 1.5198897203623474,
-      "grad_norm": 0.2118435502052307,
-      "learning_rate": 3.656816843679959e-05,
-      "loss": 0.3794,
-      "step": 965
-    },
-    {
-      "epoch": 1.5277668373375346,
-      "grad_norm": 0.2519252598285675,
-      "learning_rate": 3.64236524384934e-05,
-      "loss": 0.3896,
-      "step": 970
-    },
-    {
-      "epoch": 1.5356439543127216,
-      "grad_norm": 0.24061793088912964,
-      "learning_rate": 3.627865234746014e-05,
-      "loss": 0.3874,
-      "step": 975
-    },
-    {
-      "epoch": 1.5435210712879086,
-      "grad_norm": 0.237691268324852,
-      "learning_rate": 3.6133174308274826e-05,
-      "loss": 0.3955,
-      "step": 980
-    },
-    {
-      "epoch": 1.5513981882630956,
-      "grad_norm": 0.2733243703842163,
-      "learning_rate": 3.598722448576619e-05,
-      "loss": 0.3935,
-      "step": 985
-    },
-    {
-      "epoch": 1.5592753052382828,
-      "grad_norm": 0.2568168044090271,
-      "learning_rate": 3.584080906475542e-05,
-      "loss": 0.3848,
-      "step": 990
-    },
-    {
-      "epoch": 1.5671524222134698,
-      "grad_norm": 0.2655527889728546,
-      "learning_rate": 3.569393424979405e-05,
-      "loss": 0.3901,
-      "step": 995
-    },
-    {
-      "epoch": 1.575029539188657,
-      "grad_norm": 0.25115740299224854,
-      "learning_rate": 3.554660626490109e-05,
-      "loss": 0.3814,
-      "step": 1000
-    },
-    {
-      "epoch": 1.582906656163844,
-      "grad_norm": 0.23420318961143494,
-      "learning_rate": 3.53988313532992e-05,
-      "loss": 0.386,
-      "step": 1005
-    },
-    {
-      "epoch": 1.590783773139031,
-      "grad_norm": 0.2673567235469818,
-      "learning_rate": 3.52506157771502e-05,
-      "loss": 0.3894,
-      "step": 1010
-    },
-    {
-      "epoch": 1.598660890114218,
-      "grad_norm": 0.2313881516456604,
-      "learning_rate": 3.510196581728963e-05,
-      "loss": 0.3861,
-      "step": 1015
-    },
-    {
-      "epoch": 1.6065380070894053,
-      "grad_norm": 0.23093390464782715,
-      "learning_rate": 3.495288777296067e-05,
-      "loss": 0.3882,
-      "step": 1020
-    },
-    {
-      "epoch": 1.6144151240645923,
-      "grad_norm": 0.2772485017776489,
-      "learning_rate": 3.480338796154712e-05,
-      "loss": 0.3776,
-      "step": 1025
-    },
-    {
-      "epoch": 1.6222922410397795,
-      "grad_norm": 0.2581574320793152,
-      "learning_rate": 3.4653472718305745e-05,
-      "loss": 0.3887,
-      "step": 1030
-    },
-    {
-      "epoch": 1.6301693580149665,
-      "grad_norm": 0.2368578016757965,
-      "learning_rate": 3.450314839609781e-05,
-      "loss": 0.3848,
-      "step": 1035
-    },
-    {
-      "epoch": 1.6380464749901535,
-      "grad_norm": 0.2310653030872345,
-      "learning_rate": 3.435242136511984e-05,
-      "loss": 0.372,
-      "step": 1040
-    },
-    {
-      "epoch": 1.6459235919653405,
-      "grad_norm": 0.23389846086502075,
-      "learning_rate": 3.420129801263367e-05,
-      "loss": 0.3958,
-      "step": 1045
-    },
-    {
-      "epoch": 1.6538007089405278,
-      "grad_norm": 0.23078757524490356,
-      "learning_rate": 3.404978474269583e-05,
-      "loss": 0.3807,
-      "step": 1050
-    },
-    {
-      "epoch": 1.661677825915715,
-      "grad_norm": 0.23647628724575043,
-      "learning_rate": 3.389788797588611e-05,
-      "loss": 0.3702,
-      "step": 1055
-    },
-    {
-      "epoch": 1.669554942890902,
-      "grad_norm": 0.23526796698570251,
-      "learning_rate": 3.374561414903552e-05,
-      "loss": 0.3795,
-      "step": 1060
-    },
-    {
-      "epoch": 1.677432059866089,
-      "grad_norm": 0.23299497365951538,
-      "learning_rate": 3.3592969714953464e-05,
-      "loss": 0.3829,
-      "step": 1065
-    },
-    {
-      "epoch": 1.685309176841276,
-      "grad_norm": 0.4668526351451874,
-      "learning_rate": 3.343996114215436e-05,
-      "loss": 0.3844,
-      "step": 1070
-    },
-    {
-      "epoch": 1.693186293816463,
-      "grad_norm": 0.23494206368923187,
-      "learning_rate": 3.328659491458348e-05,
-      "loss": 0.3851,
-      "step": 1075
-    },
-    {
-      "epoch": 1.7010634107916502,
-      "grad_norm": 0.250499427318573,
-      "learning_rate": 3.31328775313422e-05,
-      "loss": 0.3811,
-      "step": 1080
-    },
-    {
-      "epoch": 1.7089405277668375,
-      "grad_norm": 0.2643508315086365,
-      "learning_rate": 3.297881550641261e-05,
-      "loss": 0.3848,
-      "step": 1085
-    },
-    {
-      "epoch": 1.7168176447420245,
-      "grad_norm": 0.2502589225769043,
-      "learning_rate": 3.282441536838144e-05,
-      "loss": 0.389,
-      "step": 1090
-    },
-    {
-      "epoch": 1.7246947617172115,
-      "grad_norm": 0.2423262745141983,
-      "learning_rate": 3.266968366016342e-05,
-      "loss": 0.3784,
-      "step": 1095
-    },
-    {
-      "epoch": 1.7325718786923985,
-      "grad_norm": 0.2834143340587616,
-      "learning_rate": 3.251462693872403e-05,
-      "loss": 0.3775,
-      "step": 1100
-    },
-    {
-      "epoch": 1.7404489956675857,
-      "grad_norm": 0.24142897129058838,
-      "learning_rate": 3.2359251774801616e-05,
-      "loss": 0.3756,
-      "step": 1105
-    },
-    {
-      "epoch": 1.7483261126427727,
-      "grad_norm": 0.24487896263599396,
-      "learning_rate": 3.2203564752628956e-05,
-      "loss": 0.3851,
-      "step": 1110
-    },
-    {
-      "epoch": 1.75620322961796,
-      "grad_norm": 0.22995340824127197,
-      "learning_rate": 3.204757246965424e-05,
-      "loss": 0.3888,
-      "step": 1115
-    },
-    {
-      "epoch": 1.764080346593147,
-      "grad_norm": 0.21457414329051971,
-      "learning_rate": 3.18912815362615e-05,
-      "loss": 0.3858,
-      "step": 1120
-    },
-    {
-      "epoch": 1.771957463568334,
-      "grad_norm": 0.21458269655704498,
-      "learning_rate": 3.173469857549048e-05,
-      "loss": 0.3812,
-      "step": 1125
-    },
-    {
-      "epoch": 1.779834580543521,
-      "grad_norm": 0.23233187198638916,
-      "learning_rate": 3.157783022275599e-05,
-      "loss": 0.3849,
-      "step": 1130
-    },
-    {
-      "epoch": 1.7877116975187082,
-      "grad_norm": 0.23942947387695312,
-      "learning_rate": 3.14206831255667e-05,
-      "loss": 0.3863,
-      "step": 1135
-    },
-    {
-      "epoch": 1.7955888144938952,
-      "grad_norm": 0.22999627888202667,
-      "learning_rate": 3.126326394324346e-05,
-      "loss": 0.39,
-      "step": 1140
-    },
-    {
-      "epoch": 1.8034659314690824,
-      "grad_norm": 0.22376017272472382,
-      "learning_rate": 3.110557934663708e-05,
-      "loss": 0.3776,
-      "step": 1145
-    },
-    {
-      "epoch": 1.8113430484442694,
-      "grad_norm": 0.20644953846931458,
-      "learning_rate": 3.09476360178457e-05,
-      "loss": 0.3772,
-      "step": 1150
-    },
-    {
-      "epoch": 1.8192201654194564,
-      "grad_norm": 0.23101069033145905,
-      "learning_rate": 3.078944064993152e-05,
-      "loss": 0.3796,
-      "step": 1155
-    },
-    {
-      "epoch": 1.8270972823946434,
-      "grad_norm": 0.23300951719284058,
-      "learning_rate": 3.063099994663731e-05,
-      "loss": 0.3766,
-      "step": 1160
-    },
-    {
-      "epoch": 1.8349743993698306,
-      "grad_norm": 0.4477635324001312,
-      "learning_rate": 3.0472320622102207e-05,
-      "loss": 0.3864,
-      "step": 1165
-    },
-    {
-      "epoch": 1.8428515163450179,
-      "grad_norm": 0.25588539242744446,
-      "learning_rate": 3.0313409400577264e-05,
-      "loss": 0.3766,
-      "step": 1170
-    },
-    {
-      "epoch": 1.8507286333202049,
-      "grad_norm": 0.24095283448696136,
-      "learning_rate": 3.0154273016140504e-05,
-      "loss": 0.386,
-      "step": 1175
-    },
-    {
-      "epoch": 1.8586057502953919,
-      "grad_norm": 0.2307787984609604,
-      "learning_rate": 2.999491821241148e-05,
-      "loss": 0.3817,
-      "step": 1180
-    },
-    {
-      "epoch": 1.8664828672705789,
-      "grad_norm": 0.24299356341362,
-      "learning_rate": 2.9835351742265622e-05,
-      "loss": 0.3888,
-      "step": 1185
-    },
-    {
-      "epoch": 1.8743599842457659,
-      "grad_norm": 0.22085171937942505,
-      "learning_rate": 2.967558036754796e-05,
-      "loss": 0.376,
-      "step": 1190
-    },
-    {
-      "epoch": 1.8822371012209531,
-      "grad_norm": 0.22518432140350342,
-      "learning_rate": 2.9515610858786657e-05,
-      "loss": 0.3841,
-      "step": 1195
-    },
-    {
-      "epoch": 1.8901142181961403,
-      "grad_norm": 0.21670100092887878,
-      "learning_rate": 2.935544999490607e-05,
-      "loss": 0.3876,
-      "step": 1200
-    },
-    {
-      "epoch": 1.8979913351713273,
-      "grad_norm": 0.22420799732208252,
-      "learning_rate": 2.9195104562939474e-05,
-      "loss": 0.39,
-      "step": 1205
-    },
-    {
-      "epoch": 1.9058684521465143,
-      "grad_norm": 0.22058328986167908,
-      "learning_rate": 2.90345813577415e-05,
-      "loss": 0.3775,
-      "step": 1210
-    },
-    {
-      "epoch": 1.9137455691217014,
-      "grad_norm": 0.22855107486248016,
-      "learning_rate": 2.8873887181700127e-05,
-      "loss": 0.3821,
-      "step": 1215
-    },
-    {
-      "epoch": 1.9216226860968886,
-      "grad_norm": 0.2261258214712143,
-      "learning_rate": 2.8713028844448464e-05,
-      "loss": 0.3881,
-      "step": 1220
-    },
-    {
-      "epoch": 1.9294998030720756,
-      "grad_norm": 0.23165945708751678,
-      "learning_rate": 2.8552013162576186e-05,
-      "loss": 0.3813,
-      "step": 1225
-    },
-    {
-      "epoch": 1.9373769200472628,
-      "grad_norm": 0.2159801721572876,
-      "learning_rate": 2.8390846959340638e-05,
-      "loss": 0.3827,
-      "step": 1230
-    },
-    {
-      "epoch": 1.9452540370224498,
-      "grad_norm": 0.2017299383878708,
-      "learning_rate": 2.8229537064377735e-05,
-      "loss": 0.3784,
-      "step": 1235
-    },
-    {
-      "epoch": 1.9531311539976368,
-      "grad_norm": 0.21241925656795502,
-      "learning_rate": 2.806809031341251e-05,
-      "loss": 0.3763,
-      "step": 1240
-    },
-    {
-      "epoch": 1.9610082709728238,
-      "grad_norm": 0.22526709735393524,
-      "learning_rate": 2.790651354796945e-05,
-      "loss": 0.3762,
-      "step": 1245
-    },
-    {
-      "epoch": 1.968885387948011,
-      "grad_norm": 0.21094514429569244,
-      "learning_rate": 2.7744813615082588e-05,
-      "loss": 0.3792,
-      "step": 1250
-    },
-    {
-      "epoch": 1.9767625049231983,
-      "grad_norm": 0.2524140775203705,
-      "learning_rate": 2.7582997367005337e-05,
-      "loss": 0.3873,
-      "step": 1255
-    },
-    {
-      "epoch": 1.9846396218983853,
-      "grad_norm": 0.23768506944179535,
-      "learning_rate": 2.742107166092015e-05,
-      "loss": 0.3696,
-      "step": 1260
-    },
-    {
-      "epoch": 1.9925167388735723,
-      "grad_norm": 0.21980994939804077,
-      "learning_rate": 2.7259043358647863e-05,
-      "loss": 0.3852,
-      "step": 1265
-    },
-    {
-      "epoch": 2.0,
-      "grad_norm": 0.244730606675148,
-      "learning_rate": 2.709691932635701e-05,
-      "loss": 0.3773,
-      "step": 1270
-    },
-    {
-      "epoch": 2.007877116975187,
-      "grad_norm": 0.2792533338069916,
-      "learning_rate": 2.6934706434272794e-05,
-      "loss": 0.3618,
-      "step": 1275
-    },
-    {
-      "epoch": 2.015754233950374,
-      "grad_norm": 0.2545369565486908,
-      "learning_rate": 2.6772411556385966e-05,
-      "loss": 0.3541,
-      "step": 1280
-    },
-    {
-      "epoch": 2.0236313509255615,
-      "grad_norm": 0.2422221451997757,
-      "learning_rate": 2.6610041570161563e-05,
-      "loss": 0.3592,
-      "step": 1285
-    },
-    {
-      "epoch": 2.0315084679007485,
-      "grad_norm": 0.2332969456911087,
-      "learning_rate": 2.6447603356247396e-05,
-      "loss": 0.3516,
-      "step": 1290
-    },
-    {
-      "epoch": 2.0393855848759355,
-      "grad_norm": 0.27313655614852905,
-      "learning_rate": 2.628510379818256e-05,
-      "loss": 0.354,
-      "step": 1295
-    },
-    {
-      "epoch": 2.0472627018511225,
-      "grad_norm": 0.22538913786411285,
-      "learning_rate": 2.6122549782105666e-05,
-      "loss": 0.3487,
-      "step": 1300
-    },
-    {
-      "epoch": 2.0551398188263095,
-      "grad_norm": 0.24524974822998047,
-      "learning_rate": 2.5959948196463058e-05,
-      "loss": 0.3497,
-      "step": 1305
-    },
-    {
-      "epoch": 2.0630169358014965,
-      "grad_norm": 0.2330634593963623,
-      "learning_rate": 2.5797305931716924e-05,
-      "loss": 0.3539,
-      "step": 1310
-    },
-    {
-      "epoch": 2.070894052776684,
-      "grad_norm": 0.247636616230011,
-      "learning_rate": 2.5634629880053256e-05,
-      "loss": 0.3492,
-      "step": 1315
-    },
-    {
-      "epoch": 2.078771169751871,
-      "grad_norm": 0.21172009408473969,
-      "learning_rate": 2.5471926935089835e-05,
-      "loss": 0.343,
-      "step": 1320
-    },
-    {
-      "epoch": 2.086648286727058,
-      "grad_norm": 0.25462400913238525,
-      "learning_rate": 2.5309203991584073e-05,
-      "loss": 0.3417,
-      "step": 1325
-    },
-    {
-      "epoch": 2.094525403702245,
-      "grad_norm": 0.24423840641975403,
-      "learning_rate": 2.514646794514085e-05,
-      "loss": 0.3383,
-      "step": 1330
-    },
-    {
-      "epoch": 2.102402520677432,
-      "grad_norm": 0.23146118223667145,
-      "learning_rate": 2.4983725691920294e-05,
-      "loss": 0.344,
-      "step": 1335
-    },
-    {
-      "epoch": 2.110279637652619,
-      "grad_norm": 0.229843407869339,
-      "learning_rate": 2.482098412834556e-05,
-      "loss": 0.3429,
-      "step": 1340
-    },
-    {
-      "epoch": 2.1181567546278064,
-      "grad_norm": 0.24268518388271332,
-      "learning_rate": 2.4658250150810578e-05,
-      "loss": 0.3337,
-      "step": 1345
-    },
-    {
-      "epoch": 2.1260338716029934,
-      "grad_norm": 0.2576902210712433,
-      "learning_rate": 2.449553065538781e-05,
-      "loss": 0.344,
-      "step": 1350
-    },
-    {
-      "epoch": 2.1339109885781804,
-      "grad_norm": 0.2369920015335083,
-      "learning_rate": 2.4332832537536e-05,
-      "loss": 0.3362,
-      "step": 1355
-    },
-    {
-      "epoch": 2.1417881055533674,
-      "grad_norm": 0.2687665522098541,
-      "learning_rate": 2.4170162691808034e-05,
-      "loss": 0.3318,
-      "step": 1360
-    },
-    {
-      "epoch": 2.1496652225285544,
-      "grad_norm": 0.23115476965904236,
-      "learning_rate": 2.4007528011558683e-05,
-      "loss": 0.3338,
-      "step": 1365
-    },
-    {
-      "epoch": 2.157542339503742,
-      "grad_norm": 0.3254106342792511,
-      "learning_rate": 2.3844935388652545e-05,
-      "loss": 0.3337,
-      "step": 1370
-    },
-    {
-      "epoch": 2.165419456478929,
-      "grad_norm": 0.2389201521873474,
-      "learning_rate": 2.3682391713171998e-05,
-      "loss": 0.3269,
-      "step": 1375
-    },
-    {
-      "epoch": 2.173296573454116,
-      "grad_norm": 0.23049896955490112,
-      "learning_rate": 2.351990387312519e-05,
-      "loss": 0.3255,
-      "step": 1380
-    },
-    {
-      "epoch": 2.181173690429303,
-      "grad_norm": 0.24909381568431854,
-      "learning_rate": 2.3357478754154173e-05,
-      "loss": 0.3362,
-      "step": 1385
-    },
-    {
-      "epoch": 2.18905080740449,
-      "grad_norm": 0.25661030411720276,
-      "learning_rate": 2.3195123239243098e-05,
-      "loss": 0.3246,
-      "step": 1390
-    },
-    {
-      "epoch": 2.196927924379677,
-      "grad_norm": 0.225275456905365,
-      "learning_rate": 2.303284420842658e-05,
-      "loss": 0.3246,
-      "step": 1395
-    },
-    {
-      "epoch": 2.2048050413548643,
-      "grad_norm": 0.2919777035713196,
-      "learning_rate": 2.2870648538498085e-05,
-      "loss": 0.3277,
-      "step": 1400
-    },
-    {
-      "epoch": 2.2126821583300513,
-      "grad_norm": 0.25202763080596924,
-      "learning_rate": 2.2708543102718542e-05,
-      "loss": 0.3249,
-      "step": 1405
-    },
-    {
-      "epoch": 2.2205592753052383,
-      "grad_norm": 0.292521595954895,
-      "learning_rate": 2.254653477052514e-05,
-      "loss": 0.3302,
-      "step": 1410
-    },
-    {
-      "epoch": 2.2284363922804253,
-      "grad_norm": 0.2488042116165161,
-      "learning_rate": 2.238463040724011e-05,
-      "loss": 0.3302,
-      "step": 1415
-    },
-    {
-      "epoch": 2.2363135092556123,
-      "grad_norm": 0.24722595512866974,
-      "learning_rate": 2.2222836873779888e-05,
-      "loss": 0.3289,
-      "step": 1420
-    },
-    {
-      "epoch": 2.2441906262307993,
-      "grad_norm": 0.25138890743255615,
-      "learning_rate": 2.2061161026364353e-05,
-      "loss": 0.3286,
-      "step": 1425
-    },
-    {
-      "epoch": 2.252067743205987,
-      "grad_norm": 0.23644718527793884,
-      "learning_rate": 2.1899609716226263e-05,
-      "loss": 0.3278,
-      "step": 1430
-    },
-    {
-      "epoch": 2.259944860181174,
-      "grad_norm": 0.2490551620721817,
-      "learning_rate": 2.173818978932095e-05,
-      "loss": 0.3275,
-      "step": 1435
-    },
-    {
-      "epoch": 2.267821977156361,
-      "grad_norm": 0.22873945534229279,
-      "learning_rate": 2.1576908086036195e-05,
-      "loss": 0.3315,
-      "step": 1440
-    },
-    {
-      "epoch": 2.275699094131548,
-      "grad_norm": 0.24433307349681854,
-      "learning_rate": 2.1415771440902384e-05,
-      "loss": 0.3308,
-      "step": 1445
-    },
-    {
-      "epoch": 2.283576211106735,
-      "grad_norm": 0.24108295142650604,
-      "learning_rate": 2.1254786682302847e-05,
-      "loss": 0.325,
-      "step": 1450
-    },
-    {
-      "epoch": 2.2914533280819223,
-      "grad_norm": 0.275078147649765,
-      "learning_rate": 2.1093960632184513e-05,
-      "loss": 0.3317,
-      "step": 1455
-    },
-    {
-      "epoch": 2.2993304450571093,
-      "grad_norm": 0.24685098230838776,
-      "learning_rate": 2.0933300105768867e-05,
-      "loss": 0.3175,
-      "step": 1460
-    },
-    {
-      "epoch": 2.3072075620322963,
-      "grad_norm": 0.22858254611492157,
-      "learning_rate": 2.0772811911263048e-05,
-      "loss": 0.3268,
-      "step": 1465
-    },
-    {
-      "epoch": 2.3150846790074833,
-      "grad_norm": 0.25581493973731995,
-      "learning_rate": 2.0612502849571425e-05,
-      "loss": 0.3259,
-      "step": 1470
-    },
-    {
-      "epoch": 2.3229617959826703,
-      "grad_norm": 0.2207001894712448,
-      "learning_rate": 2.0452379714007375e-05,
-      "loss": 0.3268,
-      "step": 1475
-    },
-    {
-      "epoch": 2.3308389129578573,
-      "grad_norm": 0.2586366832256317,
-      "learning_rate": 2.0292449290005395e-05,
-      "loss": 0.3264,
-      "step": 1480
-    },
-    {
-      "epoch": 2.3387160299330443,
-      "grad_norm": 0.2496010810136795,
-      "learning_rate": 2.013271835483357e-05,
-      "loss": 0.3275,
-      "step": 1485
-    },
-    {
-      "epoch": 2.3465931469082317,
-      "grad_norm": 0.21064487099647522,
-      "learning_rate": 1.9973193677306377e-05,
-      "loss": 0.324,
-      "step": 1490
-    },
-    {
-      "epoch": 2.3544702638834187,
-      "grad_norm": 0.22213426232337952,
-      "learning_rate": 1.9813882017497848e-05,
-      "loss": 0.3172,
-      "step": 1495
-    },
-    {
-      "epoch": 2.3623473808586057,
-      "grad_norm": 0.29100117087364197,
-      "learning_rate": 1.9654790126455103e-05,
-      "loss": 0.3235,
-      "step": 1500
-    },
-    {
-      "epoch": 2.3702244978337927,
-      "grad_norm": 0.24606060981750488,
-      "learning_rate": 1.949592474591224e-05,
-      "loss": 0.3332,
-      "step": 1505
-    },
-    {
-      "epoch": 2.3781016148089797,
-      "grad_norm": 0.23023582994937897,
-      "learning_rate": 1.9337292608004715e-05,
-      "loss": 0.3238,
-      "step": 1510
-    },
-    {
-      "epoch": 2.385978731784167,
-      "grad_norm": 0.2493922859430313,
-      "learning_rate": 1.917890043498397e-05,
-      "loss": 0.3284,
-      "step": 1515
-    },
-    {
-      "epoch": 2.393855848759354,
-      "grad_norm": 0.2237468957901001,
-      "learning_rate": 1.9020754938932624e-05,
-      "loss": 0.3307,
-      "step": 1520
-    },
-    {
-      "epoch": 2.401732965734541,
-      "grad_norm": 0.24473683536052704,
-      "learning_rate": 1.8862862821480025e-05,
-      "loss": 0.3219,
-      "step": 1525
-    },
-    {
-      "epoch": 2.409610082709728,
-      "grad_norm": 0.2150821089744568,
-      "learning_rate": 1.8705230773518267e-05,
-      "loss": 0.328,
-      "step": 1530
-    },
-    {
-      "epoch": 2.417487199684915,
-      "grad_norm": 0.24754087626934052,
-      "learning_rate": 1.854786547491864e-05,
-      "loss": 0.3251,
-      "step": 1535
-    },
-    {
-      "epoch": 2.425364316660102,
-      "grad_norm": 0.23762992024421692,
-      "learning_rate": 1.8390773594248568e-05,
-      "loss": 0.3266,
-      "step": 1540
-    },
-    {
-      "epoch": 2.4332414336352897,
-      "grad_norm": 0.23614896833896637,
-      "learning_rate": 1.823396178848903e-05,
-      "loss": 0.3283,
-      "step": 1545
-    },
-    {
-      "epoch": 2.4411185506104767,
-      "grad_norm": 0.2320897877216339,
-      "learning_rate": 1.8077436702752448e-05,
-      "loss": 0.3286,
-      "step": 1550
-    },
-    {
-      "epoch": 2.4489956675856637,
-      "grad_norm": 0.22540679574012756,
-      "learning_rate": 1.7921204970001065e-05,
-      "loss": 0.3215,
-      "step": 1555
-    },
-    {
-      "epoch": 2.4568727845608507,
-      "grad_norm": 0.24557819962501526,
-      "learning_rate": 1.776527321076596e-05,
-      "loss": 0.3183,
-      "step": 1560
-    },
-    {
-      "epoch": 2.4647499015360377,
-      "grad_norm": 0.23252086341381073,
-      "learning_rate": 1.7609648032866362e-05,
-      "loss": 0.3209,
-      "step": 1565
-    },
-    {
-      "epoch": 2.4726270185112247,
-      "grad_norm": 0.27775880694389343,
-      "learning_rate": 1.7454336031129743e-05,
-      "loss": 0.324,
-      "step": 1570
-    },
-    {
-      "epoch": 2.480504135486412,
-      "grad_norm": 0.225099116563797,
-      "learning_rate": 1.729934378711228e-05,
-      "loss": 0.3263,
-      "step": 1575
-    },
-    {
-      "epoch": 2.488381252461599,
-      "grad_norm": 0.2224494367837906,
-      "learning_rate": 1.7144677868820026e-05,
-      "loss": 0.3128,
-      "step": 1580
-    },
-    {
-      "epoch": 2.496258369436786,
-      "grad_norm": 0.21441712975502014,
-      "learning_rate": 1.6990344830430498e-05,
-      "loss": 0.3294,
-      "step": 1585
-    },
-    {
-      "epoch": 2.504135486411973,
-      "grad_norm": 0.2346782237291336,
-      "learning_rate": 1.6836351212014996e-05,
-      "loss": 0.3302,
-      "step": 1590
-    },
-    {
-      "epoch": 2.51201260338716,
-      "grad_norm": 0.23030629754066467,
-      "learning_rate": 1.6682703539261442e-05,
-      "loss": 0.3268,
-      "step": 1595
-    },
-    {
-      "epoch": 2.5198897203623476,
-      "grad_norm": 0.2294086217880249,
-      "learning_rate": 1.652940832319785e-05,
-      "loss": 0.3288,
-      "step": 1600
-    },
-    {
-      "epoch": 2.5277668373375346,
-      "grad_norm": 0.32409849762916565,
-      "learning_rate": 1.6376472059916354e-05,
-      "loss": 0.3262,
-      "step": 1605
-    },
-    {
-      "epoch": 2.5356439543127216,
-      "grad_norm": 0.22377949953079224,
-      "learning_rate": 1.6223901230298062e-05,
-      "loss": 0.3171,
-      "step": 1610
-    },
-    {
-      "epoch": 2.5435210712879086,
-      "grad_norm": 0.2417742758989334,
-      "learning_rate": 1.6071702299738268e-05,
-      "loss": 0.3261,
-      "step": 1615
-    },
-    {
-      "epoch": 2.5513981882630956,
-      "grad_norm": 0.22446629405021667,
-      "learning_rate": 1.5919881717872564e-05,
-      "loss": 0.3199,
-      "step": 1620
-    },
-    {
-      "epoch": 2.559275305238283,
-      "grad_norm": 0.22529666125774384,
-      "learning_rate": 1.5768445918303505e-05,
-      "loss": 0.321,
-      "step": 1625
-    },
-    {
-      "epoch": 2.5671524222134696,
-      "grad_norm": 0.22095133364200592,
-      "learning_rate": 1.5617401318327985e-05,
-      "loss": 0.3185,
-      "step": 1630
-    },
-    {
-      "epoch": 2.575029539188657,
-      "grad_norm": 0.24521619081497192,
-      "learning_rate": 1.546675431866526e-05,
-      "loss": 0.3128,
-      "step": 1635
-    },
-    {
-      "epoch": 2.582906656163844,
-      "grad_norm": 0.23159317672252655,
-      "learning_rate": 1.5316511303185755e-05,
-      "loss": 0.3217,
-      "step": 1640
-    },
-    {
-      "epoch": 2.590783773139031,
-      "grad_norm": 0.22384974360466003,
-      "learning_rate": 1.5166678638640521e-05,
-      "loss": 0.3173,
-      "step": 1645
-    },
-    {
-      "epoch": 2.598660890114218,
-      "grad_norm": 0.22395044565200806,
-      "learning_rate": 1.5017262674391428e-05,
-      "loss": 0.3188,
-      "step": 1650
-    },
-    {
-      "epoch": 2.606538007089405,
-      "grad_norm": 0.24889998137950897,
-      "learning_rate": 1.486826974214209e-05,
-      "loss": 0.3193,
-      "step": 1655
-    },
-    {
-      "epoch": 2.6144151240645925,
-      "grad_norm": 0.25244370102882385,
-      "learning_rate": 1.4719706155669587e-05,
-      "loss": 0.3219,
-      "step": 1660
-    },
-    {
-      "epoch": 2.6222922410397795,
-      "grad_norm": 0.22579914331436157,
-      "learning_rate": 1.4571578210556896e-05,
-      "loss": 0.3209,
-      "step": 1665
-    },
-    {
-      "epoch": 2.6301693580149665,
-      "grad_norm": 0.21842360496520996,
-      "learning_rate": 1.4423892183926092e-05,
-      "loss": 0.3221,
-      "step": 1670
-    },
-    {
-      "epoch": 2.6380464749901535,
-      "grad_norm": 0.2209043651819229,
-      "learning_rate": 1.4276654334172332e-05,
-      "loss": 0.3235,
-      "step": 1675
-    },
-    {
-      "epoch": 2.6459235919653405,
-      "grad_norm": 0.24010902643203735,
-      "learning_rate": 1.4129870900698739e-05,
-      "loss": 0.3207,
-      "step": 1680
-    },
-    {
-      "epoch": 2.653800708940528,
-      "grad_norm": 0.2469053864479065,
-      "learning_rate": 1.3983548103651861e-05,
-      "loss": 0.3238,
-      "step": 1685
-    },
-    {
-      "epoch": 2.661677825915715,
-      "grad_norm": 0.21121683716773987,
-      "learning_rate": 1.3837692143658182e-05,
-      "loss": 0.317,
-      "step": 1690
-    },
-    {
-      "epoch": 2.669554942890902,
-      "grad_norm": 0.22918404638767242,
-      "learning_rate": 1.369230920156134e-05,
-      "loss": 0.3162,
-      "step": 1695
-    },
-    {
-      "epoch": 2.677432059866089,
-      "grad_norm": 0.22199533879756927,
-      "learning_rate": 1.3547405438160205e-05,
-      "loss": 0.3173,
-      "step": 1700
-    },
-    {
-      "epoch": 2.685309176841276,
-      "grad_norm": 0.22123722732067108,
-      "learning_rate": 1.340298699394777e-05,
-      "loss": 0.3156,
-      "step": 1705
-    },
-    {
-      "epoch": 2.693186293816463,
-      "grad_norm": 0.2306554913520813,
-      "learning_rate": 1.3259059988851e-05,
-      "loss": 0.3236,
-      "step": 1710
-    },
-    {
-      "epoch": 2.70106341079165,
-      "grad_norm": 0.21142704784870148,
-      "learning_rate": 1.3115630521971468e-05,
-      "loss": 0.3144,
-      "step": 1715
-    },
-    {
-      "epoch": 2.7089405277668375,
-      "grad_norm": 0.215803861618042,
-      "learning_rate": 1.2972704671326863e-05,
-      "loss": 0.3196,
-      "step": 1720
-    },
-    {
-      "epoch": 2.7168176447420245,
-      "grad_norm": 0.2311914712190628,
-      "learning_rate": 1.2830288493593451e-05,
-      "loss": 0.3179,
-      "step": 1725
-    },
-    {
-      "epoch": 2.7246947617172115,
-      "grad_norm": 0.23759211599826813,
-      "learning_rate": 1.2688388023849482e-05,
-      "loss": 0.3173,
-      "step": 1730
-    },
-    {
-      "epoch": 2.7325718786923985,
-      "grad_norm": 0.22561126947402954,
-      "learning_rate": 1.2547009275319313e-05,
-      "loss": 0.319,
-      "step": 1735
-    },
-    {
-      "epoch": 2.7404489956675855,
-      "grad_norm": 0.21002857387065887,
-      "learning_rate": 1.2406158239118677e-05,
-      "loss": 0.3104,
-      "step": 1740
-    },
-    {
-      "epoch": 2.748326112642773,
-      "grad_norm": 0.218175008893013,
-      "learning_rate": 1.2265840884000792e-05,
-      "loss": 0.3209,
-      "step": 1745
-    },
-    {
-      "epoch": 2.75620322961796,
-      "grad_norm": 0.27158647775650024,
-      "learning_rate": 1.2126063156103437e-05,
-      "loss": 0.3209,
-      "step": 1750
-    },
-    {
-      "epoch": 2.764080346593147,
-      "grad_norm": 0.2084042727947235,
-      "learning_rate": 1.1986830978696908e-05,
-      "loss": 0.3138,
-      "step": 1755
-    },
-    {
-      "epoch": 2.771957463568334,
-      "grad_norm": 0.20862416923046112,
-      "learning_rate": 1.1848150251933102e-05,
-      "loss": 0.3134,
-      "step": 1760
-    },
-    {
-      "epoch": 2.779834580543521,
-      "grad_norm": 0.19537927210330963,
-      "learning_rate": 1.1710026852595446e-05,
-      "loss": 0.3154,
-      "step": 1765
-    },
-    {
-      "epoch": 2.7877116975187084,
-      "grad_norm": 0.2035813182592392,
-      "learning_rate": 1.157246663384984e-05,
-      "loss": 0.3204,
-      "step": 1770
-    },
-    {
-      "epoch": 2.795588814493895,
-      "grad_norm": 0.20835839211940765,
-      "learning_rate": 1.1435475424996644e-05,
-      "loss": 0.3109,
-      "step": 1775
-    },
-    {
-      "epoch": 2.8034659314690824,
-      "grad_norm": 0.21280139684677124,
-      "learning_rate": 1.1299059031223706e-05,
-      "loss": 0.3127,
-      "step": 1780
-    },
-    {
-      "epoch": 2.8113430484442694,
-      "grad_norm": 0.21605828404426575,
-      "learning_rate": 1.1163223233360243e-05,
-      "loss": 0.3206,
-      "step": 1785
-    },
-    {
-      "epoch": 2.8192201654194564,
-      "grad_norm": 0.20913977921009064,
-      "learning_rate": 1.1027973787631956e-05,
-      "loss": 0.3187,
-      "step": 1790
-    },
-    {
-      "epoch": 2.8270972823946434,
-      "grad_norm": 0.20801085233688354,
-      "learning_rate": 1.0893316425417097e-05,
-      "loss": 0.3181,
-      "step": 1795
-    },
-    {
-      "epoch": 2.8349743993698304,
-      "grad_norm": 0.21318671107292175,
-      "learning_rate": 1.0759256853003578e-05,
-      "loss": 0.3181,
-      "step": 1800
-    },
-    {
-      "epoch": 2.842851516345018,
-      "grad_norm": 0.20348812639713287,
-      "learning_rate": 1.062580075134712e-05,
-      "loss": 0.3131,
-      "step": 1805
-    },
-    {
-      "epoch": 2.850728633320205,
-      "grad_norm": 0.22590206563472748,
-      "learning_rate": 1.0492953775830603e-05,
-      "loss": 0.3182,
-      "step": 1810
-    },
-    {
-      "epoch": 2.858605750295392,
-      "grad_norm": 0.2275339812040329,
-      "learning_rate": 1.0360721556024348e-05,
-      "loss": 0.3245,
-      "step": 1815
-    },
-    {
-      "epoch": 2.866482867270579,
-      "grad_norm": 0.20729248225688934,
-      "learning_rate": 1.0229109695447553e-05,
-      "loss": 0.3113,
-      "step": 1820
-    },
-    {
-      "epoch": 2.874359984245766,
-      "grad_norm": 0.21620720624923706,
-      "learning_rate": 1.0098123771330853e-05,
-      "loss": 0.3142,
-      "step": 1825
-    },
-    {
-      "epoch": 2.8822371012209533,
-      "grad_norm": 0.20153531432151794,
-      "learning_rate": 9.967769334380029e-06,
-      "loss": 0.3136,
-      "step": 1830
-    },
-    {
-      "epoch": 2.8901142181961403,
-      "grad_norm": 0.20845447480678558,
-      "learning_rate": 9.83805190854068e-06,
-      "loss": 0.3134,
-      "step": 1835
-    },
-    {
-      "epoch": 2.8979913351713273,
-      "grad_norm": 0.20220224559307098,
-      "learning_rate": 9.708976990764212e-06,
-      "loss": 0.317,
-      "step": 1840
-    },
-    {
-      "epoch": 2.9058684521465143,
-      "grad_norm": 0.22257234156131744,
-      "learning_rate": 9.580550050774906e-06,
-      "loss": 0.3121,
-      "step": 1845
-    },
-    {
-      "epoch": 2.9137455691217014,
-      "grad_norm": 0.2114904224872589,
-      "learning_rate": 9.452776530838096e-06,
-      "loss": 0.3092,
-      "step": 1850
-    },
-    {
-      "epoch": 2.921622686096889,
-      "grad_norm": 0.20534107089042664,
-      "learning_rate": 9.325661845529533e-06,
-      "loss": 0.3112,
-      "step": 1855
-    },
-    {
-      "epoch": 2.9294998030720754,
-      "grad_norm": 0.21934354305267334,
-      "learning_rate": 9.199211381505993e-06,
-      "loss": 0.3195,
-      "step": 1860
-    },
-    {
-      "epoch": 2.937376920047263,
-      "grad_norm": 0.24163486063480377,
-      "learning_rate": 9.073430497276969e-06,
-      "loss": 0.3135,
-      "step": 1865
-    },
-    {
-      "epoch": 2.94525403702245,
-      "grad_norm": 0.22086267173290253,
-      "learning_rate": 8.94832452297759e-06,
-      "loss": 0.3132,
-      "step": 1870
-    },
-    {
-      "epoch": 2.953131153997637,
-      "grad_norm": 0.20960792899131775,
-      "learning_rate": 8.823898760142759e-06,
-      "loss": 0.3218,
-      "step": 1875
-    },
-    {
-      "epoch": 2.961008270972824,
-      "grad_norm": 0.2152036726474762,
-      "learning_rate": 8.700158481482543e-06,
-      "loss": 0.3179,
-      "step": 1880
-    },
-    {
-      "epoch": 2.968885387948011,
-      "grad_norm": 0.2101101130247116,
-      "learning_rate": 8.577108930658636e-06,
-      "loss": 0.3123,
-      "step": 1885
-    },
-    {
-      "epoch": 2.9767625049231983,
-      "grad_norm": 0.21300119161605835,
-      "learning_rate": 8.454755322062228e-06,
-      "loss": 0.3119,
-      "step": 1890
-    },
-    {
-      "epoch": 2.9846396218983853,
-      "grad_norm": 0.20476487278938293,
-      "learning_rate": 8.333102840593015e-06,
-      "loss": 0.3139,
-      "step": 1895
-    },
-    {
-      "epoch": 2.9925167388735723,
-      "grad_norm": 0.20163965225219727,
-      "learning_rate": 8.212156641439495e-06,
-      "loss": 0.3159,
-      "step": 1900
-    },
-    {
-      "epoch": 3.0015754233950376,
-      "grad_norm": 0.3048245906829834,
-      "learning_rate": 8.091921849860466e-06,
-      "loss": 0.3674,
-      "step": 1905
-    },
-    {
-      "epoch": 3.0094525403702246,
-      "grad_norm": 0.26477697491645813,
-      "learning_rate": 7.972403560967895e-06,
-      "loss": 0.2759,
-      "step": 1910
-    },
-    {
-      "epoch": 3.0173296573454116,
-      "grad_norm": 0.25195425748825073,
-      "learning_rate": 7.853606839510976e-06,
-      "loss": 0.2867,
-      "step": 1915
-    },
-    {
-      "epoch": 3.0252067743205986,
-      "grad_norm": 0.2241075187921524,
-      "learning_rate": 7.735536719661495e-06,
-      "loss": 0.2845,
-      "step": 1920
-    },
-    {
-      "epoch": 3.0330838912957856,
-      "grad_norm": 0.2114437222480774,
-      "learning_rate": 7.618198204800509e-06,
-      "loss": 0.2803,
-      "step": 1925
-    },
-    {
-      "epoch": 3.0409610082709726,
-      "grad_norm": 0.21161282062530518,
-      "learning_rate": 7.5015962673063375e-06,
-      "loss": 0.2791,
-      "step": 1930
-    },
-    {
-      "epoch": 3.04883812524616,
-      "grad_norm": 0.21086104214191437,
-      "learning_rate": 7.3857358483438385e-06,
-      "loss": 0.2817,
-      "step": 1935
-    },
-    {
-      "epoch": 3.056715242221347,
-      "grad_norm": 0.21510441601276398,
-      "learning_rate": 7.270621857655002e-06,
-      "loss": 0.285,
-      "step": 1940
-    },
-    {
-      "epoch": 3.064592359196534,
-      "grad_norm": 0.20668771862983704,
-      "learning_rate": 7.156259173350926e-06,
-      "loss": 0.2789,
-      "step": 1945
-    },
-    {
-      "epoch": 3.072469476171721,
-      "grad_norm": 0.21624241769313812,
-      "learning_rate": 7.042652641705083e-06,
-      "loss": 0.2795,
-      "step": 1950
-    },
-    {
-      "epoch": 3.080346593146908,
-      "grad_norm": 0.22468064725399017,
-      "learning_rate": 6.9298070769479476e-06,
-      "loss": 0.2822,
-      "step": 1955
-    },
-    {
-      "epoch": 3.0882237101220955,
-      "grad_norm": 0.20948195457458496,
-      "learning_rate": 6.817727261062973e-06,
-      "loss": 0.2847,
-      "step": 1960
-    },
-    {
-      "epoch": 3.0961008270972825,
-      "grad_norm": 0.21020089089870453,
-      "learning_rate": 6.706417943584023e-06,
-      "loss": 0.2803,
-      "step": 1965
-    },
-    {
-      "epoch": 3.1039779440724695,
-      "grad_norm": 0.20508207380771637,
-      "learning_rate": 6.5958838413940016e-06,
-      "loss": 0.2902,
-      "step": 1970
-    },
-    {
-      "epoch": 3.1118550610476565,
-      "grad_norm": 0.21120624244213104,
-      "learning_rate": 6.486129638525021e-06,
-      "loss": 0.2858,
-      "step": 1975
-    },
-    {
-      "epoch": 3.1197321780228435,
-      "grad_norm": 0.2145949751138687,
-      "learning_rate": 6.377159985959921e-06,
-      "loss": 0.2805,
-      "step": 1980
-    },
-    {
-      "epoch": 3.1276092949980305,
-      "grad_norm": 0.20431272685527802,
-      "learning_rate": 6.2689795014351595e-06,
-      "loss": 0.2752,
-      "step": 1985
-    },
-    {
-      "epoch": 3.135486411973218,
-      "grad_norm": 0.20915834605693817,
-      "learning_rate": 6.161592769245114e-06,
-      "loss": 0.2774,
-      "step": 1990
-    },
-    {
-      "epoch": 3.143363528948405,
-      "grad_norm": 0.20422811806201935,
-      "learning_rate": 6.055004340047849e-06,
-      "loss": 0.28,
-      "step": 1995
-    },
-    {
-      "epoch": 3.151240645923592,
-      "grad_norm": 0.20162755250930786,
-      "learning_rate": 5.949218730672259e-06,
-      "loss": 0.2846,
-      "step": 2000
-    },
-    {
-      "epoch": 3.159117762898779,
-      "grad_norm": 0.2039792239665985,
-      "learning_rate": 5.844240423926659e-06,
-      "loss": 0.2826,
-      "step": 2005
-    },
-    {
-      "epoch": 3.166994879873966,
-      "grad_norm": 0.211518332362175,
-      "learning_rate": 5.740073868408799e-06,
-      "loss": 0.2796,
-      "step": 2010
-    },
-    {
-      "epoch": 3.174871996849153,
-      "grad_norm": 0.21034759283065796,
-      "learning_rate": 5.636723478317429e-06,
-      "loss": 0.2747,
-      "step": 2015
-    },
-    {
-      "epoch": 3.1827491138243404,
-      "grad_norm": 0.19976024329662323,
-      "learning_rate": 5.5341936332651336e-06,
-      "loss": 0.2794,
-      "step": 2020
-    },
-    {
-      "epoch": 3.1906262307995275,
-      "grad_norm": 0.19816116988658905,
-      "learning_rate": 5.432488678092807e-06,
-      "loss": 0.2743,
-      "step": 2025
-    },
-    {
-      "epoch": 3.1985033477747145,
-      "grad_norm": 0.20253697037696838,
-      "learning_rate": 5.331612922685522e-06,
-      "loss": 0.2787,
-      "step": 2030
-    },
-    {
-      "epoch": 3.2063804647499015,
-      "grad_norm": 0.2093832790851593,
-      "learning_rate": 5.231570641789899e-06,
-      "loss": 0.2787,
-      "step": 2035
-    },
-    {
-      "epoch": 3.2142575817250885,
-      "grad_norm": 0.2073364406824112,
-      "learning_rate": 5.13236607483292e-06,
-      "loss": 0.2806,
-      "step": 2040
-    },
-    {
-      "epoch": 3.222134698700276,
-      "grad_norm": 0.19635286927223206,
-      "learning_rate": 5.034003425742334e-06,
-      "loss": 0.2788,
-      "step": 2045
-    },
-    {
-      "epoch": 3.230011815675463,
-      "grad_norm": 0.20712953805923462,
-      "learning_rate": 4.9364868627684775e-06,
-      "loss": 0.2887,
-      "step": 2050
-    },
-    {
-      "epoch": 3.23788893265065,
-      "grad_norm": 0.2041015326976776,
-      "learning_rate": 4.839820518307628e-06,
-      "loss": 0.2821,
-      "step": 2055
-    },
-    {
-      "epoch": 3.245766049625837,
-      "grad_norm": 0.19976255297660828,
-      "learning_rate": 4.7440084887269035e-06,
-      "loss": 0.2823,
-      "step": 2060
-    },
-    {
-      "epoch": 3.253643166601024,
-      "grad_norm": 0.20047256350517273,
-      "learning_rate": 4.649054834190686e-06,
-      "loss": 0.2799,
-      "step": 2065
-    },
-    {
-      "epoch": 3.261520283576211,
-      "grad_norm": 0.20123493671417236,
-      "learning_rate": 4.554963578488561e-06,
-      "loss": 0.2825,
-      "step": 2070
-    },
-    {
-      "epoch": 3.2693974005513984,
-      "grad_norm": 0.19577762484550476,
-      "learning_rate": 4.461738708864768e-06,
-      "loss": 0.2813,
-      "step": 2075
-    },
-    {
-      "epoch": 3.2772745175265854,
-      "grad_norm": 0.20611698925495148,
-      "learning_rate": 4.369384175849287e-06,
-      "loss": 0.2825,
-      "step": 2080
-    },
-    {
-      "epoch": 3.2851516345017724,
-      "grad_norm": 0.20026405155658722,
-      "learning_rate": 4.277903893090407e-06,
-      "loss": 0.2766,
-      "step": 2085
-    },
-    {
-      "epoch": 3.2930287514769594,
-      "grad_norm": 0.19801442325115204,
-      "learning_rate": 4.187301737188859e-06,
-      "loss": 0.2756,
-      "step": 2090
-    },
-    {
-      "epoch": 3.3009058684521464,
-      "grad_norm": 0.19530606269836426,
-      "learning_rate": 4.0975815475335855e-06,
-      "loss": 0.2767,
-      "step": 2095
-    },
-    {
-      "epoch": 3.3087829854273334,
-      "grad_norm": 0.20650427043437958,
-      "learning_rate": 4.0087471261390094e-06,
-      "loss": 0.2752,
-      "step": 2100
-    },
-    {
-      "epoch": 3.316660102402521,
-      "grad_norm": 0.20899270474910736,
-      "learning_rate": 3.920802237483912e-06,
-      "loss": 0.2755,
-      "step": 2105
-    },
-    {
-      "epoch": 3.324537219377708,
-      "grad_norm": 0.21143369376659393,
-      "learning_rate": 3.833750608351924e-06,
-      "loss": 0.2805,
-      "step": 2110
-    },
-    {
-      "epoch": 3.332414336352895,
-      "grad_norm": 0.20321451127529144,
-      "learning_rate": 3.747595927673611e-06,
-      "loss": 0.2788,
-      "step": 2115
-    },
-    {
-      "epoch": 3.340291453328082,
-      "grad_norm": 0.1997046321630478,
-      "learning_rate": 3.6623418463701293e-06,
-      "loss": 0.2825,
-      "step": 2120
-    },
-    {
-      "epoch": 3.348168570303269,
-      "grad_norm": 0.4038223326206207,
-      "learning_rate": 3.5779919771985025e-06,
-      "loss": 0.2767,
-      "step": 2125
-    },
-    {
-      "epoch": 3.3560456872784563,
-      "grad_norm": 0.20923250913619995,
-      "learning_rate": 3.494549894598556e-06,
-      "loss": 0.28,
-      "step": 2130
-    },
-    {
-      "epoch": 3.3639228042536433,
-      "grad_norm": 0.1994774043560028,
-      "learning_rate": 3.4120191345414415e-06,
-      "loss": 0.2842,
-      "step": 2135
-    },
-    {
-      "epoch": 3.3717999212288303,
-      "grad_norm": 0.20118853449821472,
-      "learning_rate": 3.3304031943797527e-06,
-      "loss": 0.281,
-      "step": 2140
-    },
-    {
-      "epoch": 3.3796770382040173,
-      "grad_norm": 0.21133875846862793,
-      "learning_rate": 3.249705532699379e-06,
-      "loss": 0.2821,
-      "step": 2145
-    },
-    {
-      "epoch": 3.3875541551792043,
-      "grad_norm": 0.19587914645671844,
-      "learning_rate": 3.169929569172919e-06,
-      "loss": 0.2768,
-      "step": 2150
-    },
-    {
-      "epoch": 3.3954312721543913,
-      "grad_norm": 0.20176377892494202,
-      "learning_rate": 3.0910786844147473e-06,
-      "loss": 0.2765,
-      "step": 2155
-    },
-    {
-      "epoch": 3.4033083891295783,
-      "grad_norm": 0.19805395603179932,
-      "learning_rate": 3.013156219837776e-06,
-      "loss": 0.277,
-      "step": 2160
-    },
-    {
-      "epoch": 3.411185506104766,
-      "grad_norm": 0.2078244537115097,
-      "learning_rate": 2.936165477511868e-06,
-      "loss": 0.2784,
-      "step": 2165
-    },
-    {
-      "epoch": 3.419062623079953,
-      "grad_norm": 0.199508935213089,
-      "learning_rate": 2.8601097200238987e-06,
-      "loss": 0.2798,
-      "step": 2170
-    },
-    {
-      "epoch": 3.42693974005514,
-      "grad_norm": 0.2023046910762787,
-      "learning_rate": 2.7849921703394732e-06,
-      "loss": 0.2833,
-      "step": 2175
-    },
-    {
-      "epoch": 3.434816857030327,
-      "grad_norm": 0.20601923763751984,
-      "learning_rate": 2.7108160116663893e-06,
-      "loss": 0.2811,
-      "step": 2180
-    },
-    {
-      "epoch": 3.442693974005514,
-      "grad_norm": 0.20845656096935272,
-      "learning_rate": 2.637584387319733e-06,
-      "loss": 0.2844,
-      "step": 2185
-    },
-    {
-      "epoch": 3.4505710909807012,
-      "grad_norm": 0.19674457609653473,
-      "learning_rate": 2.565300400588647e-06,
-      "loss": 0.2846,
-      "step": 2190
-    },
-    {
-      "epoch": 3.4584482079558883,
-      "grad_norm": 0.197424054145813,
-      "learning_rate": 2.4939671146048727e-06,
-      "loss": 0.2792,
-      "step": 2195
-    },
-    {
-      "epoch": 3.4663253249310753,
-      "grad_norm": 0.19904725253582,
-      "learning_rate": 2.4235875522128997e-06,
-      "loss": 0.2903,
-      "step": 2200
-    },
-    {
-      "epoch": 3.4742024419062623,
-      "grad_norm": 0.19490273296833038,
-      "learning_rate": 2.3541646958419094e-06,
-      "loss": 0.2799,
-      "step": 2205
-    },
-    {
-      "epoch": 3.4820795588814493,
-      "grad_norm": 0.19701102375984192,
-      "learning_rate": 2.2857014873793577e-06,
-      "loss": 0.2815,
-      "step": 2210
-    },
-    {
-      "epoch": 3.4899566758566367,
-      "grad_norm": 0.1960253268480301,
-      "learning_rate": 2.218200828046324e-06,
-      "loss": 0.2767,
-      "step": 2215
-    },
-    {
-      "epoch": 3.4978337928318237,
-      "grad_norm": 0.1989275962114334,
-      "learning_rate": 2.151665578274581e-06,
-      "loss": 0.2795,
-      "step": 2220
-    },
-    {
-      "epoch": 3.5057109098070107,
-      "grad_norm": 0.19489678740501404,
-      "learning_rate": 2.0860985575853366e-06,
-      "loss": 0.2816,
-      "step": 2225
-    },
-    {
-      "epoch": 3.5135880267821977,
-      "grad_norm": 0.20096373558044434,
-      "learning_rate": 2.021502544469808e-06,
-      "loss": 0.2879,
-      "step": 2230
-    },
-    {
-      "epoch": 3.5214651437573847,
-      "grad_norm": 0.19433267414569855,
-      "learning_rate": 1.9578802762714427e-06,
-      "loss": 0.2814,
-      "step": 2235
-    },
-    {
-      "epoch": 3.5293422607325717,
-      "grad_norm": 0.20079940557479858,
-      "learning_rate": 1.8952344490699243e-06,
-      "loss": 0.2805,
-      "step": 2240
-    },
-    {
-      "epoch": 3.5372193777077587,
-      "grad_norm": 0.19769088923931122,
-      "learning_rate": 1.833567717566939e-06,
-      "loss": 0.2833,
-      "step": 2245
-    },
-    {
-      "epoch": 3.545096494682946,
-      "grad_norm": 0.20594307780265808,
-      "learning_rate": 1.7728826949736598e-06,
-      "loss": 0.2882,
-      "step": 2250
-    },
-    {
-      "epoch": 3.552973611658133,
-      "grad_norm": 0.19729258120059967,
-      "learning_rate": 1.7131819529000226e-06,
-      "loss": 0.2829,
-      "step": 2255
-    },
-    {
-      "epoch": 3.56085072863332,
-      "grad_norm": 0.19826674461364746,
-      "learning_rate": 1.6544680212457337e-06,
-      "loss": 0.2786,
-      "step": 2260
-    },
-    {
-      "epoch": 3.568727845608507,
-      "grad_norm": 0.19805769622325897,
-      "learning_rate": 1.5967433880930889e-06,
-      "loss": 0.2808,
-      "step": 2265
-    },
-    {
-      "epoch": 3.576604962583694,
-      "grad_norm": 0.20012123882770538,
-      "learning_rate": 1.540010499601513e-06,
-      "loss": 0.2825,
-      "step": 2270
-    },
-    {
-      "epoch": 3.5844820795588817,
-      "grad_norm": 0.20687070488929749,
-      "learning_rate": 1.4842717599039047e-06,
-      "loss": 0.2848,
-      "step": 2275
-    },
-    {
-      "epoch": 3.5923591965340687,
-      "grad_norm": 0.19811828434467316,
-      "learning_rate": 1.4295295310047668e-06,
-      "loss": 0.2785,
-      "step": 2280
-    },
-    {
-      "epoch": 3.6002363135092557,
-      "grad_norm": 0.19836102426052094,
-      "learning_rate": 1.375786132680118e-06,
-      "loss": 0.2808,
-      "step": 2285
-    },
-    {
-      "epoch": 3.6081134304844427,
-      "grad_norm": 0.19285066425800323,
-      "learning_rate": 1.3230438423791625e-06,
-      "loss": 0.2818,
-      "step": 2290
-    },
-    {
-      "epoch": 3.6159905474596297,
-      "grad_norm": 0.19814734160900116,
-      "learning_rate": 1.2713048951278144e-06,
-      "loss": 0.2817,
-      "step": 2295
-    },
-    {
-      "epoch": 3.623867664434817,
-      "grad_norm": 0.2252984642982483,
-      "learning_rate": 1.2205714834339521e-06,
-      "loss": 0.2822,
-      "step": 2300
-    },
-    {
-      "epoch": 3.6317447814100037,
-      "grad_norm": 0.19117222726345062,
-      "learning_rate": 1.1708457571945386e-06,
-      "loss": 0.2753,
-      "step": 2305
-    },
-    {
-      "epoch": 3.639621898385191,
-      "grad_norm": 0.19470417499542236,
-      "learning_rate": 1.1221298236044953e-06,
-      "loss": 0.2822,
-      "step": 2310
-    },
-    {
-      "epoch": 3.647499015360378,
-      "grad_norm": 0.2041274756193161,
-      "learning_rate": 1.0744257470674147e-06,
-      "loss": 0.2841,
-      "step": 2315
-    },
-    {
-      "epoch": 3.655376132335565,
-      "grad_norm": 0.19451037049293518,
-      "learning_rate": 1.0277355491080804e-06,
-      "loss": 0.2817,
-      "step": 2320
-    },
-    {
-      "epoch": 3.663253249310752,
-      "grad_norm": 0.1978326439857483,
-      "learning_rate": 9.820612082867969e-07,
-      "loss": 0.2797,
-      "step": 2325
-    },
-    {
-      "epoch": 3.671130366285939,
-      "grad_norm": 0.19272668659687042,
-      "learning_rate": 9.374046601155539e-07,
-      "loss": 0.2822,
-      "step": 2330
-    },
-    {
-      "epoch": 3.6790074832611266,
-      "grad_norm": 0.1939852088689804,
-      "learning_rate": 8.937677969759861e-07,
-      "loss": 0.2792,
-      "step": 2335
-    },
-    {
-      "epoch": 3.6868846002363136,
-      "grad_norm": 0.1928287148475647,
-      "learning_rate": 8.511524680392096e-07,
-      "loss": 0.2876,
-      "step": 2340
-    },
-    {
-      "epoch": 3.6947617172115006,
-      "grad_norm": 0.19896940886974335,
-      "learning_rate": 8.09560479187449e-07,
-      "loss": 0.2744,
-      "step": 2345
-    },
-    {
-      "epoch": 3.7026388341866876,
-      "grad_norm": 0.19617894291877747,
-      "learning_rate": 7.689935929374892e-07,
-      "loss": 0.2856,
-      "step": 2350
-    },
-    {
-      "epoch": 3.7105159511618746,
-      "grad_norm": 0.19769024848937988,
-      "learning_rate": 7.294535283660281e-07,
-      "loss": 0.279,
-      "step": 2355
-    },
-    {
-      "epoch": 3.718393068137062,
-      "grad_norm": 0.19515487551689148,
-      "learning_rate": 6.909419610367896e-07,
-      "loss": 0.2768,
-      "step": 2360
-    },
-    {
-      "epoch": 3.726270185112249,
-      "grad_norm": 0.19522498548030853,
-      "learning_rate": 6.534605229295415e-07,
-      "loss": 0.2839,
-      "step": 2365
-    },
-    {
-      "epoch": 3.734147302087436,
-      "grad_norm": 0.18874140083789825,
-      "learning_rate": 6.170108023709348e-07,
-      "loss": 0.2818,
-      "step": 2370
-    },
-    {
-      "epoch": 3.742024419062623,
-      "grad_norm": 0.19244030117988586,
-      "learning_rate": 5.815943439671901e-07,
-      "loss": 0.2779,
-      "step": 2375
-    },
-    {
-      "epoch": 3.74990153603781,
-      "grad_norm": 0.190862774848938,
-      "learning_rate": 5.472126485386475e-07,
-      "loss": 0.2784,
-      "step": 2380
-    },
-    {
-      "epoch": 3.7577786530129975,
-      "grad_norm": 0.197196364402771,
-      "learning_rate": 5.138671730561623e-07,
-      "loss": 0.2803,
-      "step": 2385
-    },
-    {
-      "epoch": 3.765655769988184,
-      "grad_norm": 0.19586575031280518,
-      "learning_rate": 4.815593305793703e-07,
-      "loss": 0.2776,
-      "step": 2390
-    },
-    {
-      "epoch": 3.7735328869633715,
-      "grad_norm": 0.19647184014320374,
-      "learning_rate": 4.502904901968141e-07,
-      "loss": 0.2824,
-      "step": 2395
-    },
-    {
-      "epoch": 3.7814100039385585,
-      "grad_norm": 0.20829348266124725,
-      "learning_rate": 4.200619769679004e-07,
-      "loss": 0.2821,
-      "step": 2400
-    },
-    {
-      "epoch": 3.7892871209137455,
-      "grad_norm": 0.19623522460460663,
-      "learning_rate": 3.908750718667753e-07,
-      "loss": 0.2816,
-      "step": 2405
-    },
-    {
-      "epoch": 3.7971642378889325,
-      "grad_norm": 0.19415368139743805,
-      "learning_rate": 3.6273101172802914e-07,
-      "loss": 0.2775,
-      "step": 2410
-    },
-    {
-      "epoch": 3.8050413548641195,
-      "grad_norm": 0.1985977739095688,
-      "learning_rate": 3.3563098919428827e-07,
-      "loss": 0.2787,
-      "step": 2415
-    },
-    {
-      "epoch": 3.812918471839307,
-      "grad_norm": 0.19479507207870483,
-      "learning_rate": 3.095761526656693e-07,
-      "loss": 0.2778,
-      "step": 2420
-    },
-    {
-      "epoch": 3.820795588814494,
-      "grad_norm": 0.2967446446418762,
-      "learning_rate": 2.8456760625111823e-07,
-      "loss": 0.2887,
-      "step": 2425
-    },
-    {
-      "epoch": 3.828672705789681,
-      "grad_norm": 0.19328995048999786,
-      "learning_rate": 2.6060640972162543e-07,
-      "loss": 0.2822,
-      "step": 2430
-    },
-    {
-      "epoch": 3.836549822764868,
-      "grad_norm": 0.19140738248825073,
-      "learning_rate": 2.3769357846531171e-07,
-      "loss": 0.2771,
-      "step": 2435
-    },
-    {
-      "epoch": 3.844426939740055,
-      "grad_norm": 0.194104865193367,
-      "learning_rate": 2.1583008344439603e-07,
-      "loss": 0.284,
-      "step": 2440
-    },
-    {
-      "epoch": 3.8523040567152425,
-      "grad_norm": 0.19096589088439941,
-      "learning_rate": 1.9501685115406176e-07,
-      "loss": 0.2793,
-      "step": 2445
-    },
-    {
-      "epoch": 3.8601811736904295,
-      "grad_norm": 0.19132515788078308,
-      "learning_rate": 1.7525476358318526e-07,
-      "loss": 0.2783,
-      "step": 2450
-    },
-    {
-      "epoch": 3.8680582906656165,
-      "grad_norm": 0.20566198229789734,
-      "learning_rate": 1.5654465817696584e-07,
-      "loss": 0.2791,
-      "step": 2455
-    },
-    {
-      "epoch": 3.8759354076408035,
-      "grad_norm": 0.20506292581558228,
-      "learning_rate": 1.3888732780142632e-07,
-      "loss": 0.2832,
-      "step": 2460
-    },
-    {
-      "epoch": 3.8838125246159905,
-      "grad_norm": 0.19119083881378174,
-      "learning_rate": 1.2228352070983719e-07,
-      "loss": 0.2808,
-      "step": 2465
-    },
-    {
-      "epoch": 3.8916896415911775,
-      "grad_norm": 0.19219276309013367,
-      "learning_rate": 1.0673394051098074e-07,
-      "loss": 0.2767,
-      "step": 2470
-    },
-    {
-      "epoch": 3.8995667585663645,
-      "grad_norm": 0.1935568004846573,
-      "learning_rate": 9.223924613935842e-08,
-      "loss": 0.2822,
-      "step": 2475
-    },
-    {
-      "epoch": 3.907443875541552,
-      "grad_norm": 0.19069112837314606,
-      "learning_rate": 7.880005182725469e-08,
-      "loss": 0.2862,
-      "step": 2480
-    },
-    {
-      "epoch": 3.915320992516739,
-      "grad_norm": 0.19846147298812866,
-      "learning_rate": 6.641692707870795e-08,
-      "loss": 0.2863,
-      "step": 2485
-    },
-    {
-      "epoch": 3.923198109491926,
-      "grad_norm": 0.19378872215747833,
-      "learning_rate": 5.509039664538251e-08,
-      "loss": 0.2827,
-      "step": 2490
-    },
-    {
-      "epoch": 3.931075226467113,
-      "grad_norm": 0.19526971876621246,
-      "learning_rate": 4.482094050433372e-08,
-      "loss": 0.2843,
-      "step": 2495
-    },
-    {
-      "epoch": 3.9389523434423,
-      "grad_norm": 0.32608553767204285,
-      "learning_rate": 3.5608993837651924e-08,
-      "loss": 0.2863,
-      "step": 2500
-    },
-    {
-      "epoch": 3.9468294604174874,
-      "grad_norm": 0.1961493343114853,
-      "learning_rate": 2.745494701404394e-08,
-      "loss": 0.2796,
-      "step": 2505
-    },
-    {
-      "epoch": 3.9547065773926744,
-      "grad_norm": 0.19112332165241241,
-      "learning_rate": 2.0359145572276804e-08,
-      "loss": 0.2821,
-      "step": 2510
-    },
-    {
-      "epoch": 3.9625836943678614,
-      "grad_norm": 0.19307857751846313,
-      "learning_rate": 1.4321890206533961e-08,
-      "loss": 0.2815,
-      "step": 2515
-    },
-    {
-      "epoch": 3.9704608113430484,
-      "grad_norm": 0.21154381334781647,
-      "learning_rate": 9.34343675368099e-09,
-      "loss": 0.2798,
-      "step": 2520
-    },
-    {
-      "epoch": 3.9783379283182354,
-      "grad_norm": 0.19480980932712555,
-      "learning_rate": 5.423996182415958e-09,
-      "loss": 0.2792,
-      "step": 2525
-    },
-    {
-      "epoch": 3.986215045293423,
-      "grad_norm": 0.1891375333070755,
-      "learning_rate": 2.5637345843348982e-09,
-      "loss": 0.2791,
-      "step": 2530
-    },
-    {
-      "epoch": 3.9940921622686094,
-      "grad_norm": 0.19509141147136688,
-      "learning_rate": 7.627731668985449e-10,
-      "loss": 0.2782,
-      "step": 2535
-    },
-    {
-      "epoch": 4.0,
-      "step": 2539,
-      "total_flos": 4.3530238414695694e+19,
-      "train_loss": 0.15123671168837785,
-      "train_runtime": 82201.3053,
-      "train_samples_per_second": 1.977,
-      "train_steps_per_second": 0.031
-    }
-  ],
-  "logging_steps": 5,
-  "max_steps": 2540,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
-  "save_steps": 318,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": false
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 4.3530238414695694e+19,
-  "train_batch_size": 2,
-  "trial_name": null,
-  "trial_params": null
-}

Qwen3-4B-Math-R1-CoT-SFT/training_args.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c08f76ab47b206e4b9c06d41cfab36618b2faaeae2f1c40b815ba9fb75cc2770
-size 7825