From e01595b1eb323c0bfc48cb36010f3b88756df06b Mon Sep 17 00:00:00 2001
From: ModelHub XC <noreply@modelhub.org.cn>
Date: Sat, 13 Jun 2026 17:29:16 +0800
Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?=
 =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?=
 =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Model: terasut/sft-qwen2.5-1.5b-instruct-eff32
Source: Original Platform
---
 .gitattributes                         |   39 +
 README.md                              |   57 +
 chat_template.jinja                    |   54 +
 checkpoint-1750/chat_template.jinja    |   54 +
 checkpoint-1750/config.json            |   61 +
 checkpoint-1750/generation_config.json |   14 +
 checkpoint-1750/model.safetensors      |    3 +
 checkpoint-1750/optimizer.pt           |    3 +
 checkpoint-1750/rng_state.pth          |    3 +
 checkpoint-1750/scheduler.pt           |    3 +
 checkpoint-1750/tokenizer.json         |    3 +
 checkpoint-1750/tokenizer_config.json  |   30 +
 checkpoint-1750/trainer_state.json     | 1784 ++++++++++++++++
 checkpoint-1750/training_args.bin      |    3 +
 checkpoint-2625/chat_template.jinja    |   54 +
 checkpoint-2625/config.json            |   61 +
 checkpoint-2625/generation_config.json |   14 +
 checkpoint-2625/model.safetensors      |    3 +
 checkpoint-2625/optimizer.pt           |    3 +
 checkpoint-2625/rng_state.pth          |    3 +
 checkpoint-2625/scheduler.pt           |    3 +
 checkpoint-2625/tokenizer.json         |    3 +
 checkpoint-2625/tokenizer_config.json  |   30 +
 checkpoint-2625/trainer_state.json     | 2654 ++++++++++++++++++++++++
 checkpoint-2625/training_args.bin      |    3 +
 checkpoint-875/chat_template.jinja     |   54 +
 checkpoint-875/config.json             |   61 +
 checkpoint-875/generation_config.json  |   14 +
 checkpoint-875/model.safetensors       |    3 +
 checkpoint-875/optimizer.pt            |    3 +
 checkpoint-875/rng_state.pth           |    3 +
 checkpoint-875/scheduler.pt            |    3 +
 checkpoint-875/tokenizer.json          |    3 +
 checkpoint-875/tokenizer_config.json   |   30 +
 checkpoint-875/trainer_state.json      |  904 ++++++++
 checkpoint-875/training_args.bin       |    3 +
 config.json                            |   61 +
 generation_config.json                 |   14 +
 model.safetensors                      |    3 +
 tokenizer.json                         |    3 +
 tokenizer_config.json                  |   30 +
 training_args.bin                      |    3 +
 42 files changed, 6137 insertions(+)
 create mode 100644 .gitattributes
 create mode 100644 README.md
 create mode 100644 chat_template.jinja
 create mode 100644 checkpoint-1750/chat_template.jinja
 create mode 100644 checkpoint-1750/config.json
 create mode 100644 checkpoint-1750/generation_config.json
 create mode 100644 checkpoint-1750/model.safetensors
 create mode 100644 checkpoint-1750/optimizer.pt
 create mode 100644 checkpoint-1750/rng_state.pth
 create mode 100644 checkpoint-1750/scheduler.pt
 create mode 100644 checkpoint-1750/tokenizer.json
 create mode 100644 checkpoint-1750/tokenizer_config.json
 create mode 100644 checkpoint-1750/trainer_state.json
 create mode 100644 checkpoint-1750/training_args.bin
 create mode 100644 checkpoint-2625/chat_template.jinja
 create mode 100644 checkpoint-2625/config.json
 create mode 100644 checkpoint-2625/generation_config.json
 create mode 100644 checkpoint-2625/model.safetensors
 create mode 100644 checkpoint-2625/optimizer.pt
 create mode 100644 checkpoint-2625/rng_state.pth
 create mode 100644 checkpoint-2625/scheduler.pt
 create mode 100644 checkpoint-2625/tokenizer.json
 create mode 100644 checkpoint-2625/tokenizer_config.json
 create mode 100644 checkpoint-2625/trainer_state.json
 create mode 100644 checkpoint-2625/training_args.bin
 create mode 100644 checkpoint-875/chat_template.jinja
 create mode 100644 checkpoint-875/config.json
 create mode 100644 checkpoint-875/generation_config.json
 create mode 100644 checkpoint-875/model.safetensors
 create mode 100644 checkpoint-875/optimizer.pt
 create mode 100644 checkpoint-875/rng_state.pth
 create mode 100644 checkpoint-875/scheduler.pt
 create mode 100644 checkpoint-875/tokenizer.json
 create mode 100644 checkpoint-875/tokenizer_config.json
 create mode 100644 checkpoint-875/trainer_state.json
 create mode 100644 checkpoint-875/training_args.bin
 create mode 100644 config.json
 create mode 100644 generation_config.json
 create mode 100644 model.safetensors
 create mode 100644 tokenizer.json
 create mode 100644 tokenizer_config.json
 create mode 100644 training_args.bin

diff --git a/.gitattributes b/.gitattributes
new file mode 100644
index 0000000..3b6fd86
--- /dev/null
+++ b/.gitattributes
@@ -0,0 +1,39 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoint-1750/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-2625/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-875/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..2195c47
--- /dev/null
+++ b/README.md
@@ -0,0 +1,57 @@
+---
+library_name: transformers
+model_name: sft-qwen2.5-1.5b-instruct-eff32
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for sft-qwen2.5-1.5b-instruct-eff32
+
+This model is a fine-tuned version of [None](https://huggingface.co/None).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/terasut-workspace/ai-patternrecog/runs/uafxmgbk) 
+
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 1.1.0
+- Transformers: 5.5.4
+- Pytorch: 2.10.0+cu128
+- Datasets: 4.8.4
+- Tokenizers: 0.22.2
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@software{vonwerra2020trl,
+  title   = {{TRL: Transformers Reinforcement Learning}},
+  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
+  license = {Apache-2.0},
+  url     = {https://github.com/huggingface/trl},
+  year    = {2020}
+}
+```
\ No newline at end of file
diff --git a/chat_template.jinja b/chat_template.jinja
new file mode 100644
index 0000000..bdf7919
--- /dev/null
+++ b/chat_template.jinja
@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- 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' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) 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' }}
+{%- endif %}
diff --git a/checkpoint-1750/chat_template.jinja b/checkpoint-1750/chat_template.jinja
new file mode 100644
index 0000000..bdf7919
--- /dev/null
+++ b/checkpoint-1750/chat_template.jinja
@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- 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' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) 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' }}
+{%- endif %}
diff --git a/checkpoint-1750/config.json b/checkpoint-1750/config.json
new file mode 100644
index 0000000..7ff1bce
--- /dev/null
+++ b/checkpoint-1750/config.json
@@ -0,0 +1,61 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "float32",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000.0,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.5.4",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/checkpoint-1750/generation_config.json b/checkpoint-1750/generation_config.json
new file mode 100644
index 0000000..846ba08
--- /dev/null
+++ b/checkpoint-1750/generation_config.json
@@ -0,0 +1,14 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "5.5.4",
+  "use_cache": false
+}
diff --git a/checkpoint-1750/model.safetensors b/checkpoint-1750/model.safetensors
new file mode 100644
index 0000000..7a016c1
--- /dev/null
+++ b/checkpoint-1750/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48b69ec3864095fcce9513b3ba341b6b0554f1dbd84c9e1393f1501b41140167
+size 6174895536
diff --git a/checkpoint-1750/optimizer.pt b/checkpoint-1750/optimizer.pt
new file mode 100644
index 0000000..686c6e4
--- /dev/null
+++ b/checkpoint-1750/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58d85847e17209c7e695400aaa9053f7244ca0755eebcaf61f982b682dcdf98f
+size 12350013801
diff --git a/checkpoint-1750/rng_state.pth b/checkpoint-1750/rng_state.pth
new file mode 100644
index 0000000..1feba1a
--- /dev/null
+++ b/checkpoint-1750/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:098b29492211804ab324a36f37466821d948280bb74fce4ba895c03f13ecd878
+size 14645
diff --git a/checkpoint-1750/scheduler.pt b/checkpoint-1750/scheduler.pt
new file mode 100644
index 0000000..9885f56
--- /dev/null
+++ b/checkpoint-1750/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc53ebd3c8ca395477da766ae7fa14a1cd87577954defcd87cfec2f1d6fd424a
+size 1465
diff --git a/checkpoint-1750/tokenizer.json b/checkpoint-1750/tokenizer.json
new file mode 100644
index 0000000..34510ff
--- /dev/null
+++ b/checkpoint-1750/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
+size 11421892
diff --git a/checkpoint-1750/tokenizer_config.json b/checkpoint-1750/tokenizer_config.json
new file mode 100644
index 0000000..0385b12
--- /dev/null
+++ b/checkpoint-1750/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_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|>"
+  ],
+  "is_local": true,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/checkpoint-1750/trainer_state.json b/checkpoint-1750/trainer_state.json
new file mode 100644
index 0000000..3eae4a5
--- /dev/null
+++ b/checkpoint-1750/trainer_state.json
@@ -0,0 +1,1784 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 1750,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.6696195736527444,
+      "epoch": 0.011428571428571429,
+      "grad_norm": 3.7833805084228516,
+      "learning_rate": 1.9931428571428572e-05,
+      "loss": 1.3203315734863281,
+      "mean_token_accuracy": 0.7056225836277008,
+      "num_tokens": 157266.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.7283091336488723,
+      "epoch": 0.022857142857142857,
+      "grad_norm": 3.4964022636413574,
+      "learning_rate": 1.9855238095238097e-05,
+      "loss": 1.4036809921264648,
+      "mean_token_accuracy": 0.6812524735927582,
+      "num_tokens": 300430.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.7663277059793472,
+      "epoch": 0.03428571428571429,
+      "grad_norm": 4.590687274932861,
+      "learning_rate": 1.977904761904762e-05,
+      "loss": 1.319040012359619,
+      "mean_token_accuracy": 0.6928375385701656,
+      "num_tokens": 456129.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.7699106186628342,
+      "epoch": 0.045714285714285714,
+      "grad_norm": 3.2531967163085938,
+      "learning_rate": 1.9702857142857144e-05,
+      "loss": 1.3776089668273925,
+      "mean_token_accuracy": 0.686311125010252,
+      "num_tokens": 615179.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.7856220349669456,
+      "epoch": 0.05714285714285714,
+      "grad_norm": 3.356832981109619,
+      "learning_rate": 1.9626666666666666e-05,
+      "loss": 1.32177791595459,
+      "mean_token_accuracy": 0.6952089086174965,
+      "num_tokens": 769672.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.7628433778882027,
+      "epoch": 0.06857142857142857,
+      "grad_norm": 3.965970754623413,
+      "learning_rate": 1.955047619047619e-05,
+      "loss": 1.338463306427002,
+      "mean_token_accuracy": 0.6953431364148855,
+      "num_tokens": 918402.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.7276440173387528,
+      "epoch": 0.08,
+      "grad_norm": 3.4119951725006104,
+      "learning_rate": 1.9474285714285717e-05,
+      "loss": 1.2838379859924316,
+      "mean_token_accuracy": 0.6987539391964674,
+      "num_tokens": 1076598.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.748648078739643,
+      "epoch": 0.09142857142857143,
+      "grad_norm": 3.6497740745544434,
+      "learning_rate": 1.9398095238095242e-05,
+      "loss": 1.3314701080322267,
+      "mean_token_accuracy": 0.695335328578949,
+      "num_tokens": 1228084.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.7111000671982766,
+      "epoch": 0.10285714285714286,
+      "grad_norm": 3.3122711181640625,
+      "learning_rate": 1.9321904761904764e-05,
+      "loss": 1.3027889251708984,
+      "mean_token_accuracy": 0.6987856134772301,
+      "num_tokens": 1386707.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.7278845444321633,
+      "epoch": 0.11428571428571428,
+      "grad_norm": 3.183835744857788,
+      "learning_rate": 1.924571428571429e-05,
+      "loss": 1.282834243774414,
+      "mean_token_accuracy": 0.6966647863388061,
+      "num_tokens": 1535406.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.7132800757884978,
+      "epoch": 0.12571428571428572,
+      "grad_norm": 3.4031126499176025,
+      "learning_rate": 1.916952380952381e-05,
+      "loss": 1.2409746170043945,
+      "mean_token_accuracy": 0.7101543735712766,
+      "num_tokens": 1680743.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.7373003125190736,
+      "epoch": 0.13714285714285715,
+      "grad_norm": 3.4207799434661865,
+      "learning_rate": 1.9093333333333336e-05,
+      "loss": 1.2909335136413573,
+      "mean_token_accuracy": 0.701117354631424,
+      "num_tokens": 1834014.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.7092810094356536,
+      "epoch": 0.14857142857142858,
+      "grad_norm": 3.3323278427124023,
+      "learning_rate": 1.9017142857142858e-05,
+      "loss": 1.3009337425231933,
+      "mean_token_accuracy": 0.6907773412764072,
+      "num_tokens": 1989349.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.7240261450409888,
+      "epoch": 0.16,
+      "grad_norm": 2.9499671459198,
+      "learning_rate": 1.8940952380952383e-05,
+      "loss": 1.2700564384460449,
+      "mean_token_accuracy": 0.7077916815876961,
+      "num_tokens": 2139614.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.655298152565956,
+      "epoch": 0.17142857142857143,
+      "grad_norm": 3.1871607303619385,
+      "learning_rate": 1.8864761904761905e-05,
+      "loss": 1.2251598358154296,
+      "mean_token_accuracy": 0.7074991308152676,
+      "num_tokens": 2293879.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.665707252919674,
+      "epoch": 0.18285714285714286,
+      "grad_norm": 3.290358304977417,
+      "learning_rate": 1.878857142857143e-05,
+      "loss": 1.2641545295715333,
+      "mean_token_accuracy": 0.7021077655255794,
+      "num_tokens": 2446447.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.7259689077734948,
+      "epoch": 0.19428571428571428,
+      "grad_norm": 3.1401703357696533,
+      "learning_rate": 1.871238095238095e-05,
+      "loss": 1.315788745880127,
+      "mean_token_accuracy": 0.6926948539912701,
+      "num_tokens": 2595389.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.7142695009708404,
+      "epoch": 0.2057142857142857,
+      "grad_norm": 3.4904985427856445,
+      "learning_rate": 1.8636190476190477e-05,
+      "loss": 1.2755393028259276,
+      "mean_token_accuracy": 0.7040164832025766,
+      "num_tokens": 2745608.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.6242914512753486,
+      "epoch": 0.21714285714285714,
+      "grad_norm": 2.949880361557007,
+      "learning_rate": 1.8560000000000002e-05,
+      "loss": 1.190676498413086,
+      "mean_token_accuracy": 0.7185754984617233,
+      "num_tokens": 2894625.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.7364101350307464,
+      "epoch": 0.22857142857142856,
+      "grad_norm": 3.125112533569336,
+      "learning_rate": 1.8483809523809527e-05,
+      "loss": 1.2999603271484375,
+      "mean_token_accuracy": 0.6976742841303348,
+      "num_tokens": 3046487.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.6772031649947166,
+      "epoch": 0.24,
+      "grad_norm": 2.813882350921631,
+      "learning_rate": 1.840761904761905e-05,
+      "loss": 1.2316917419433593,
+      "mean_token_accuracy": 0.7120772495865821,
+      "num_tokens": 3196704.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.6737122178077697,
+      "epoch": 0.25142857142857145,
+      "grad_norm": 2.8218936920166016,
+      "learning_rate": 1.8331428571428574e-05,
+      "loss": 1.274891471862793,
+      "mean_token_accuracy": 0.6991564333438873,
+      "num_tokens": 3348365.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.6344200998544693,
+      "epoch": 0.26285714285714284,
+      "grad_norm": 2.9688668251037598,
+      "learning_rate": 1.8255238095238096e-05,
+      "loss": 1.2082077026367188,
+      "mean_token_accuracy": 0.7141494430601597,
+      "num_tokens": 3509464.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.6166698306798934,
+      "epoch": 0.2742857142857143,
+      "grad_norm": 3.007690668106079,
+      "learning_rate": 1.817904761904762e-05,
+      "loss": 1.1915247917175293,
+      "mean_token_accuracy": 0.7145515654236079,
+      "num_tokens": 3661779.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.6418054014444352,
+      "epoch": 0.2857142857142857,
+      "grad_norm": 2.9565091133117676,
+      "learning_rate": 1.8102857142857143e-05,
+      "loss": 1.2041254043579102,
+      "mean_token_accuracy": 0.7189058735966682,
+      "num_tokens": 3809611.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.6628487035632133,
+      "epoch": 0.29714285714285715,
+      "grad_norm": 2.7851827144622803,
+      "learning_rate": 1.8026666666666668e-05,
+      "loss": 1.2013505935668944,
+      "mean_token_accuracy": 0.7156483136117459,
+      "num_tokens": 3960501.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.6495753720402717,
+      "epoch": 0.30857142857142855,
+      "grad_norm": 3.0212292671203613,
+      "learning_rate": 1.7950476190476193e-05,
+      "loss": 1.2240056991577148,
+      "mean_token_accuracy": 0.7174453124403953,
+      "num_tokens": 4107783.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.693277756869793,
+      "epoch": 0.32,
+      "grad_norm": 2.9716391563415527,
+      "learning_rate": 1.7874285714285715e-05,
+      "loss": 1.207960319519043,
+      "mean_token_accuracy": 0.7157351300120354,
+      "num_tokens": 4252335.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.6619624510407447,
+      "epoch": 0.3314285714285714,
+      "grad_norm": 2.851775646209717,
+      "learning_rate": 1.779809523809524e-05,
+      "loss": 1.248055648803711,
+      "mean_token_accuracy": 0.7128231346607208,
+      "num_tokens": 4407579.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.6617254123091698,
+      "epoch": 0.34285714285714286,
+      "grad_norm": 3.254999876022339,
+      "learning_rate": 1.7721904761904762e-05,
+      "loss": 1.184535598754883,
+      "mean_token_accuracy": 0.7231362499296665,
+      "num_tokens": 4551517.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.5988039791584014,
+      "epoch": 0.35428571428571426,
+      "grad_norm": 3.110564708709717,
+      "learning_rate": 1.7645714285714287e-05,
+      "loss": 1.148481559753418,
+      "mean_token_accuracy": 0.7262889519333839,
+      "num_tokens": 4708257.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.6530713498592378,
+      "epoch": 0.3657142857142857,
+      "grad_norm": 3.1247048377990723,
+      "learning_rate": 1.7569523809523812e-05,
+      "loss": 1.2834845542907716,
+      "mean_token_accuracy": 0.702876628190279,
+      "num_tokens": 4859075.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.6123583316802979,
+      "epoch": 0.37714285714285717,
+      "grad_norm": 2.90073823928833,
+      "learning_rate": 1.7493333333333334e-05,
+      "loss": 1.1132370948791503,
+      "mean_token_accuracy": 0.7313019946217537,
+      "num_tokens": 5011395.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.618579277396202,
+      "epoch": 0.38857142857142857,
+      "grad_norm": 2.936490058898926,
+      "learning_rate": 1.741714285714286e-05,
+      "loss": 1.2263901710510254,
+      "mean_token_accuracy": 0.7069775387644768,
+      "num_tokens": 5174272.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.663301782310009,
+      "epoch": 0.4,
+      "grad_norm": 3.241185426712036,
+      "learning_rate": 1.734095238095238e-05,
+      "loss": 1.2628487586975097,
+      "mean_token_accuracy": 0.7058808147907257,
+      "num_tokens": 5327670.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.6857714764773846,
+      "epoch": 0.4114285714285714,
+      "grad_norm": 2.9059698581695557,
+      "learning_rate": 1.7264761904761906e-05,
+      "loss": 1.2412065505981444,
+      "mean_token_accuracy": 0.7062458634376526,
+      "num_tokens": 5476064.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.6253925159573555,
+      "epoch": 0.4228571428571429,
+      "grad_norm": 2.9899251461029053,
+      "learning_rate": 1.718857142857143e-05,
+      "loss": 1.210451316833496,
+      "mean_token_accuracy": 0.7089586913585663,
+      "num_tokens": 5636828.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.5880890995264054,
+      "epoch": 0.4342857142857143,
+      "grad_norm": 3.0461251735687256,
+      "learning_rate": 1.7112380952380953e-05,
+      "loss": 1.1992740631103516,
+      "mean_token_accuracy": 0.7144776731729507,
+      "num_tokens": 5788462.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.6384857930243015,
+      "epoch": 0.44571428571428573,
+      "grad_norm": 2.8842580318450928,
+      "learning_rate": 1.703619047619048e-05,
+      "loss": 1.2124631881713868,
+      "mean_token_accuracy": 0.7115042388439179,
+      "num_tokens": 5943108.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.5625910967588426,
+      "epoch": 0.45714285714285713,
+      "grad_norm": 3.0432047843933105,
+      "learning_rate": 1.696e-05,
+      "loss": 1.1515995025634767,
+      "mean_token_accuracy": 0.716732618585229,
+      "num_tokens": 6098679.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.6026007533073425,
+      "epoch": 0.4685714285714286,
+      "grad_norm": 2.781804323196411,
+      "learning_rate": 1.6883809523809525e-05,
+      "loss": 1.180698299407959,
+      "mean_token_accuracy": 0.7132757268846035,
+      "num_tokens": 6250301.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.5140968441963196,
+      "epoch": 0.48,
+      "grad_norm": 2.9592671394348145,
+      "learning_rate": 1.6807619047619047e-05,
+      "loss": 1.1643908500671387,
+      "mean_token_accuracy": 0.7223096333444119,
+      "num_tokens": 6407356.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.6122839108109475,
+      "epoch": 0.49142857142857144,
+      "grad_norm": 3.2809886932373047,
+      "learning_rate": 1.6731428571428572e-05,
+      "loss": 1.2397551536560059,
+      "mean_token_accuracy": 0.707809828966856,
+      "num_tokens": 6559819.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.603949698805809,
+      "epoch": 0.5028571428571429,
+      "grad_norm": 2.721527338027954,
+      "learning_rate": 1.6655238095238098e-05,
+      "loss": 1.1895122528076172,
+      "mean_token_accuracy": 0.7227885220199823,
+      "num_tokens": 6716227.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.582472313940525,
+      "epoch": 0.5142857142857142,
+      "grad_norm": 2.8015570640563965,
+      "learning_rate": 1.6579047619047623e-05,
+      "loss": 1.1144920349121095,
+      "mean_token_accuracy": 0.7284132238477469,
+      "num_tokens": 6867486.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.6025653272867202,
+      "epoch": 0.5257142857142857,
+      "grad_norm": 3.1553401947021484,
+      "learning_rate": 1.6502857142857145e-05,
+      "loss": 1.1947656631469727,
+      "mean_token_accuracy": 0.7160489454865455,
+      "num_tokens": 7012179.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.5280416823923588,
+      "epoch": 0.5371428571428571,
+      "grad_norm": 2.97623348236084,
+      "learning_rate": 1.642666666666667e-05,
+      "loss": 1.149429702758789,
+      "mean_token_accuracy": 0.7256167627871036,
+      "num_tokens": 7171638.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.6620354026556015,
+      "epoch": 0.5485714285714286,
+      "grad_norm": 3.1873958110809326,
+      "learning_rate": 1.635047619047619e-05,
+      "loss": 1.2588795661926269,
+      "mean_token_accuracy": 0.707894566655159,
+      "num_tokens": 7322167.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.581991147994995,
+      "epoch": 0.56,
+      "grad_norm": 3.385500907897949,
+      "learning_rate": 1.6274285714285717e-05,
+      "loss": 1.2005508422851563,
+      "mean_token_accuracy": 0.7142980419099331,
+      "num_tokens": 7466661.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.6193878114223481,
+      "epoch": 0.5714285714285714,
+      "grad_norm": 2.9256980419158936,
+      "learning_rate": 1.619809523809524e-05,
+      "loss": 1.1964147567749024,
+      "mean_token_accuracy": 0.715008432790637,
+      "num_tokens": 7620522.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.560711294412613,
+      "epoch": 0.5828571428571429,
+      "grad_norm": 2.548938035964966,
+      "learning_rate": 1.6121904761904764e-05,
+      "loss": 1.174423599243164,
+      "mean_token_accuracy": 0.7205559216439724,
+      "num_tokens": 7772467.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.6210713252425193,
+      "epoch": 0.5942857142857143,
+      "grad_norm": 2.6339938640594482,
+      "learning_rate": 1.6045714285714286e-05,
+      "loss": 1.153661060333252,
+      "mean_token_accuracy": 0.7224138833582401,
+      "num_tokens": 7922030.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.5822248332202435,
+      "epoch": 0.6057142857142858,
+      "grad_norm": 2.4934332370758057,
+      "learning_rate": 1.596952380952381e-05,
+      "loss": 1.1586255073547362,
+      "mean_token_accuracy": 0.7230928719043732,
+      "num_tokens": 8067955.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.5642932564020158,
+      "epoch": 0.6171428571428571,
+      "grad_norm": 2.985651969909668,
+      "learning_rate": 1.5893333333333333e-05,
+      "loss": 1.1338199615478515,
+      "mean_token_accuracy": 0.7240421302616596,
+      "num_tokens": 8226369.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.6006409794092178,
+      "epoch": 0.6285714285714286,
+      "grad_norm": 2.6047215461730957,
+      "learning_rate": 1.5817142857142858e-05,
+      "loss": 1.155271625518799,
+      "mean_token_accuracy": 0.7230217695236206,
+      "num_tokens": 8376285.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.5791691079735757,
+      "epoch": 0.64,
+      "grad_norm": 2.855867385864258,
+      "learning_rate": 1.5740952380952383e-05,
+      "loss": 1.1597973823547363,
+      "mean_token_accuracy": 0.7209223508834839,
+      "num_tokens": 8530322.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.6016968488693237,
+      "epoch": 0.6514285714285715,
+      "grad_norm": 3.205124855041504,
+      "learning_rate": 1.5664761904761908e-05,
+      "loss": 1.1405302047729493,
+      "mean_token_accuracy": 0.7231174889951945,
+      "num_tokens": 8676991.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.5366713739931583,
+      "epoch": 0.6628571428571428,
+      "grad_norm": 3.1047000885009766,
+      "learning_rate": 1.558857142857143e-05,
+      "loss": 1.1293843269348145,
+      "mean_token_accuracy": 0.7271328650414943,
+      "num_tokens": 8833861.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.5866449058055878,
+      "epoch": 0.6742857142857143,
+      "grad_norm": 2.884098529815674,
+      "learning_rate": 1.5512380952380955e-05,
+      "loss": 1.2017478942871094,
+      "mean_token_accuracy": 0.716648469120264,
+      "num_tokens": 8983438.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.610584917664528,
+      "epoch": 0.6857142857142857,
+      "grad_norm": 2.7602009773254395,
+      "learning_rate": 1.5436190476190477e-05,
+      "loss": 1.1553681373596192,
+      "mean_token_accuracy": 0.72271120660007,
+      "num_tokens": 9138958.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.5292311601340771,
+      "epoch": 0.6971428571428572,
+      "grad_norm": 2.8975393772125244,
+      "learning_rate": 1.5360000000000002e-05,
+      "loss": 1.1547229766845704,
+      "mean_token_accuracy": 0.7226926662027836,
+      "num_tokens": 9290314.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.5577459454536438,
+      "epoch": 0.7085714285714285,
+      "grad_norm": 2.5054473876953125,
+      "learning_rate": 1.5283809523809524e-05,
+      "loss": 1.1239413261413573,
+      "mean_token_accuracy": 0.7288053043186664,
+      "num_tokens": 9442584.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.6545007571578025,
+      "epoch": 0.72,
+      "grad_norm": 3.156757354736328,
+      "learning_rate": 1.5207619047619049e-05,
+      "loss": 1.2162582397460937,
+      "mean_token_accuracy": 0.7100601557642221,
+      "num_tokens": 9589146.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.6293067336082458,
+      "epoch": 0.7314285714285714,
+      "grad_norm": 2.7666168212890625,
+      "learning_rate": 1.5131428571428572e-05,
+      "loss": 1.1888948440551759,
+      "mean_token_accuracy": 0.7159062504768372,
+      "num_tokens": 9736056.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.5786321811378001,
+      "epoch": 0.7428571428571429,
+      "grad_norm": 2.6019463539123535,
+      "learning_rate": 1.5055238095238096e-05,
+      "loss": 1.0923637390136718,
+      "mean_token_accuracy": 0.7382036492228508,
+      "num_tokens": 9898410.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.574275678396225,
+      "epoch": 0.7542857142857143,
+      "grad_norm": 2.7830586433410645,
+      "learning_rate": 1.497904761904762e-05,
+      "loss": 1.1398842811584473,
+      "mean_token_accuracy": 0.7246641829609871,
+      "num_tokens": 10051154.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.5554455459117889,
+      "epoch": 0.7657142857142857,
+      "grad_norm": 2.8458149433135986,
+      "learning_rate": 1.4902857142857143e-05,
+      "loss": 1.139054298400879,
+      "mean_token_accuracy": 0.726585789769888,
+      "num_tokens": 10199110.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.6094272822141646,
+      "epoch": 0.7771428571428571,
+      "grad_norm": 2.9835047721862793,
+      "learning_rate": 1.4826666666666666e-05,
+      "loss": 1.1978397369384766,
+      "mean_token_accuracy": 0.7123216778039932,
+      "num_tokens": 10344927.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.5142188727855683,
+      "epoch": 0.7885714285714286,
+      "grad_norm": 2.7492172718048096,
+      "learning_rate": 1.475047619047619e-05,
+      "loss": 1.048453140258789,
+      "mean_token_accuracy": 0.7415693461894989,
+      "num_tokens": 10505422.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.633454330265522,
+      "epoch": 0.8,
+      "grad_norm": 2.645536184310913,
+      "learning_rate": 1.4674285714285717e-05,
+      "loss": 1.1423879623413087,
+      "mean_token_accuracy": 0.7272885538637638,
+      "num_tokens": 10652367.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.4862130410969256,
+      "epoch": 0.8114285714285714,
+      "grad_norm": 2.473473310470581,
+      "learning_rate": 1.459809523809524e-05,
+      "loss": 1.019435214996338,
+      "mean_token_accuracy": 0.7454098269343377,
+      "num_tokens": 10812282.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.541966749727726,
+      "epoch": 0.8228571428571428,
+      "grad_norm": 2.5247135162353516,
+      "learning_rate": 1.4521904761904764e-05,
+      "loss": 1.0726137161254883,
+      "mean_token_accuracy": 0.7371681958436966,
+      "num_tokens": 10967866.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.5219792507588863,
+      "epoch": 0.8342857142857143,
+      "grad_norm": 2.35876202583313,
+      "learning_rate": 1.4445714285714287e-05,
+      "loss": 1.0680880546569824,
+      "mean_token_accuracy": 0.7386457361280918,
+      "num_tokens": 11120412.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.506186830997467,
+      "epoch": 0.8457142857142858,
+      "grad_norm": 2.9589080810546875,
+      "learning_rate": 1.436952380952381e-05,
+      "loss": 1.0974149703979492,
+      "mean_token_accuracy": 0.7378807857632637,
+      "num_tokens": 11277244.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.569691103696823,
+      "epoch": 0.8571428571428571,
+      "grad_norm": 2.5306074619293213,
+      "learning_rate": 1.4293333333333334e-05,
+      "loss": 1.177268123626709,
+      "mean_token_accuracy": 0.7250622183084487,
+      "num_tokens": 11436007.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.5297324053943158,
+      "epoch": 0.8685714285714285,
+      "grad_norm": 2.7663497924804688,
+      "learning_rate": 1.4217142857142858e-05,
+      "loss": 1.1369609832763672,
+      "mean_token_accuracy": 0.7221393447369338,
+      "num_tokens": 11587615.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.4932486459612846,
+      "epoch": 0.88,
+      "grad_norm": 2.5511255264282227,
+      "learning_rate": 1.4140952380952381e-05,
+      "loss": 1.1008304595947265,
+      "mean_token_accuracy": 0.7303555272519588,
+      "num_tokens": 11747686.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.5960537254810334,
+      "epoch": 0.8914285714285715,
+      "grad_norm": 2.450721025466919,
+      "learning_rate": 1.4064761904761905e-05,
+      "loss": 1.1708695411682128,
+      "mean_token_accuracy": 0.7219238199293614,
+      "num_tokens": 11904676.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.5482905827462674,
+      "epoch": 0.9028571428571428,
+      "grad_norm": 3.1114189624786377,
+      "learning_rate": 1.398857142857143e-05,
+      "loss": 1.184025478363037,
+      "mean_token_accuracy": 0.720519745349884,
+      "num_tokens": 12059825.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.5766600042581558,
+      "epoch": 0.9142857142857143,
+      "grad_norm": 3.0541417598724365,
+      "learning_rate": 1.3912380952380953e-05,
+      "loss": 1.2167075157165528,
+      "mean_token_accuracy": 0.7153365164995193,
+      "num_tokens": 12218813.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.5715911239385605,
+      "epoch": 0.9257142857142857,
+      "grad_norm": 2.4761741161346436,
+      "learning_rate": 1.3836190476190477e-05,
+      "loss": 1.0981733322143554,
+      "mean_token_accuracy": 0.7329004533588886,
+      "num_tokens": 12366807.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.5434480041265488,
+      "epoch": 0.9371428571428572,
+      "grad_norm": 3.157036304473877,
+      "learning_rate": 1.376e-05,
+      "loss": 1.0953783988952637,
+      "mean_token_accuracy": 0.7374757573008537,
+      "num_tokens": 12519277.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.5731646195054054,
+      "epoch": 0.9485714285714286,
+      "grad_norm": 2.9288859367370605,
+      "learning_rate": 1.3683809523809526e-05,
+      "loss": 1.1861271858215332,
+      "mean_token_accuracy": 0.7202281527221203,
+      "num_tokens": 12668916.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.5766675099730492,
+      "epoch": 0.96,
+      "grad_norm": 2.3330445289611816,
+      "learning_rate": 1.3607619047619049e-05,
+      "loss": 1.1634528160095214,
+      "mean_token_accuracy": 0.7218175616115332,
+      "num_tokens": 12824349.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.5784073442220687,
+      "epoch": 0.9714285714285714,
+      "grad_norm": 2.8480734825134277,
+      "learning_rate": 1.3531428571428573e-05,
+      "loss": 1.1402252197265625,
+      "mean_token_accuracy": 0.7281969003379345,
+      "num_tokens": 12974198.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.5565635196864605,
+      "epoch": 0.9828571428571429,
+      "grad_norm": 3.2169361114501953,
+      "learning_rate": 1.3455238095238096e-05,
+      "loss": 1.0919082641601563,
+      "mean_token_accuracy": 0.7359561800956727,
+      "num_tokens": 13125877.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.6499854236841203,
+      "epoch": 0.9942857142857143,
+      "grad_norm": 2.59342622756958,
+      "learning_rate": 1.3379047619047621e-05,
+      "loss": 1.1824271202087402,
+      "mean_token_accuracy": 0.7206701695919037,
+      "num_tokens": 13270434.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.5392458021640778,
+      "epoch": 1.0057142857142858,
+      "grad_norm": 2.1802525520324707,
+      "learning_rate": 1.3302857142857145e-05,
+      "loss": 0.9705311775207519,
+      "mean_token_accuracy": 0.7628285072743892,
+      "num_tokens": 13419321.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.3396786622703076,
+      "epoch": 1.0171428571428571,
+      "grad_norm": 2.7976434230804443,
+      "learning_rate": 1.3226666666666668e-05,
+      "loss": 0.753804874420166,
+      "mean_token_accuracy": 0.8103815123438836,
+      "num_tokens": 13564219.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.263124795258045,
+      "epoch": 1.0285714285714285,
+      "grad_norm": 2.680875062942505,
+      "learning_rate": 1.3150476190476192e-05,
+      "loss": 0.8032585144042969,
+      "mean_token_accuracy": 0.8010107159614563,
+      "num_tokens": 13716646.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.2998421356081962,
+      "epoch": 1.04,
+      "grad_norm": 3.176466941833496,
+      "learning_rate": 1.3074285714285715e-05,
+      "loss": 0.7720170497894288,
+      "mean_token_accuracy": 0.8062204010784626,
+      "num_tokens": 13865045.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.268384473770857,
+      "epoch": 1.0514285714285714,
+      "grad_norm": 3.2034316062927246,
+      "learning_rate": 1.2998095238095239e-05,
+      "loss": 0.7246803283691406,
+      "mean_token_accuracy": 0.8114085927605629,
+      "num_tokens": 14019242.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.2792991541326046,
+      "epoch": 1.062857142857143,
+      "grad_norm": 2.799799919128418,
+      "learning_rate": 1.2921904761904762e-05,
+      "loss": 0.7459653377532959,
+      "mean_token_accuracy": 0.8081640765070915,
+      "num_tokens": 14170551.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.2749989867210387,
+      "epoch": 1.0742857142857143,
+      "grad_norm": 2.315112352371216,
+      "learning_rate": 1.2845714285714286e-05,
+      "loss": 0.7502207279205322,
+      "mean_token_accuracy": 0.8043648317456246,
+      "num_tokens": 14320423.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.3014213733375073,
+      "epoch": 1.0857142857142856,
+      "grad_norm": 2.8643505573272705,
+      "learning_rate": 1.2769523809523811e-05,
+      "loss": 0.7972130298614502,
+      "mean_token_accuracy": 0.798911166191101,
+      "num_tokens": 14472454.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.309807014465332,
+      "epoch": 1.0971428571428572,
+      "grad_norm": 2.9528074264526367,
+      "learning_rate": 1.2693333333333336e-05,
+      "loss": 0.762692928314209,
+      "mean_token_accuracy": 0.7993374645709992,
+      "num_tokens": 14621285.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.2928462289273739,
+      "epoch": 1.1085714285714285,
+      "grad_norm": 3.0441250801086426,
+      "learning_rate": 1.261714285714286e-05,
+      "loss": 0.8246999740600586,
+      "mean_token_accuracy": 0.7930732540786266,
+      "num_tokens": 14776609.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.2388829864561557,
+      "epoch": 1.12,
+      "grad_norm": 2.831936836242676,
+      "learning_rate": 1.2540952380952383e-05,
+      "loss": 0.7267831802368164,
+      "mean_token_accuracy": 0.8151453614234925,
+      "num_tokens": 14935547.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.2560633935034275,
+      "epoch": 1.1314285714285715,
+      "grad_norm": 2.852947473526001,
+      "learning_rate": 1.2464761904761907e-05,
+      "loss": 0.7547249794006348,
+      "mean_token_accuracy": 0.805939918756485,
+      "num_tokens": 15090483.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.241182105243206,
+      "epoch": 1.1428571428571428,
+      "grad_norm": 3.028001546859741,
+      "learning_rate": 1.238857142857143e-05,
+      "loss": 0.7496088027954102,
+      "mean_token_accuracy": 0.8107783786952496,
+      "num_tokens": 15242584.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.348277000337839,
+      "epoch": 1.1542857142857144,
+      "grad_norm": 2.8477680683135986,
+      "learning_rate": 1.2312380952380953e-05,
+      "loss": 0.8604803085327148,
+      "mean_token_accuracy": 0.7804459646344185,
+      "num_tokens": 15391767.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.250984299182892,
+      "epoch": 1.1657142857142857,
+      "grad_norm": 2.5354862213134766,
+      "learning_rate": 1.2236190476190477e-05,
+      "loss": 0.7215850353240967,
+      "mean_token_accuracy": 0.811321247369051,
+      "num_tokens": 15543892.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.2813407756388187,
+      "epoch": 1.177142857142857,
+      "grad_norm": 2.451107978820801,
+      "learning_rate": 1.216e-05,
+      "loss": 0.7772514820098877,
+      "mean_token_accuracy": 0.8003417395055294,
+      "num_tokens": 15692671.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.2703118346631528,
+      "epoch": 1.1885714285714286,
+      "grad_norm": 2.9708104133605957,
+      "learning_rate": 1.2083809523809524e-05,
+      "loss": 0.8101517677307128,
+      "mean_token_accuracy": 0.7938590474426747,
+      "num_tokens": 15849058.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.2321988992393016,
+      "epoch": 1.2,
+      "grad_norm": 2.5784900188446045,
+      "learning_rate": 1.2007619047619047e-05,
+      "loss": 0.7412232398986817,
+      "mean_token_accuracy": 0.8034046024084092,
+      "num_tokens": 16006785.0,
+      "step": 1050
+    },
+    {
+      "entropy": 1.2155711680650712,
+      "epoch": 1.2114285714285715,
+      "grad_norm": 2.337512731552124,
+      "learning_rate": 1.1931428571428571e-05,
+      "loss": 0.7488626003265381,
+      "mean_token_accuracy": 0.8063951514661312,
+      "num_tokens": 16158185.0,
+      "step": 1060
+    },
+    {
+      "entropy": 1.2575144089758397,
+      "epoch": 1.2228571428571429,
+      "grad_norm": 2.3885722160339355,
+      "learning_rate": 1.1855238095238094e-05,
+      "loss": 0.7211280345916748,
+      "mean_token_accuracy": 0.8121630363166332,
+      "num_tokens": 16316647.0,
+      "step": 1070
+    },
+    {
+      "entropy": 1.2669233828783035,
+      "epoch": 1.2342857142857142,
+      "grad_norm": 3.1025376319885254,
+      "learning_rate": 1.1779047619047621e-05,
+      "loss": 0.7715346813201904,
+      "mean_token_accuracy": 0.7969170480966568,
+      "num_tokens": 16471457.0,
+      "step": 1080
+    },
+    {
+      "entropy": 1.2340376809239388,
+      "epoch": 1.2457142857142858,
+      "grad_norm": 2.8538618087768555,
+      "learning_rate": 1.1702857142857145e-05,
+      "loss": 0.7577665328979493,
+      "mean_token_accuracy": 0.8027225121855736,
+      "num_tokens": 16631017.0,
+      "step": 1090
+    },
+    {
+      "entropy": 1.2692770406603813,
+      "epoch": 1.2571428571428571,
+      "grad_norm": 3.1801536083221436,
+      "learning_rate": 1.1626666666666668e-05,
+      "loss": 0.7788435935974121,
+      "mean_token_accuracy": 0.8010938227176666,
+      "num_tokens": 16786752.0,
+      "step": 1100
+    },
+    {
+      "entropy": 1.2486526250839234,
+      "epoch": 1.2685714285714287,
+      "grad_norm": 2.5821192264556885,
+      "learning_rate": 1.1550476190476192e-05,
+      "loss": 0.762976884841919,
+      "mean_token_accuracy": 0.8053238935768604,
+      "num_tokens": 16943519.0,
+      "step": 1110
+    },
+    {
+      "entropy": 1.3036663837730884,
+      "epoch": 1.28,
+      "grad_norm": 2.6454882621765137,
+      "learning_rate": 1.1474285714285715e-05,
+      "loss": 0.7800922393798828,
+      "mean_token_accuracy": 0.7931916259229184,
+      "num_tokens": 17098242.0,
+      "step": 1120
+    },
+    {
+      "entropy": 1.2358268782496453,
+      "epoch": 1.2914285714285714,
+      "grad_norm": 2.732434034347534,
+      "learning_rate": 1.1398095238095239e-05,
+      "loss": 0.7546168804168701,
+      "mean_token_accuracy": 0.8025222033262253,
+      "num_tokens": 17254880.0,
+      "step": 1130
+    },
+    {
+      "entropy": 1.2517999060451985,
+      "epoch": 1.302857142857143,
+      "grad_norm": 2.6394054889678955,
+      "learning_rate": 1.1321904761904762e-05,
+      "loss": 0.7450598239898681,
+      "mean_token_accuracy": 0.8091251827776432,
+      "num_tokens": 17404975.0,
+      "step": 1140
+    },
+    {
+      "entropy": 1.2538614131510257,
+      "epoch": 1.3142857142857143,
+      "grad_norm": 3.0480637550354004,
+      "learning_rate": 1.1245714285714286e-05,
+      "loss": 0.7634236335754394,
+      "mean_token_accuracy": 0.8049138620495796,
+      "num_tokens": 17557114.0,
+      "step": 1150
+    },
+    {
+      "entropy": 1.2000257663428784,
+      "epoch": 1.3257142857142856,
+      "grad_norm": 3.085055351257324,
+      "learning_rate": 1.1169523809523811e-05,
+      "loss": 0.7426016807556153,
+      "mean_token_accuracy": 0.8077944576740265,
+      "num_tokens": 17704128.0,
+      "step": 1160
+    },
+    {
+      "entropy": 1.2293061904609204,
+      "epoch": 1.3371428571428572,
+      "grad_norm": 2.3842477798461914,
+      "learning_rate": 1.1093333333333334e-05,
+      "loss": 0.7181476593017578,
+      "mean_token_accuracy": 0.8123824037611485,
+      "num_tokens": 17865916.0,
+      "step": 1170
+    },
+    {
+      "entropy": 1.2635096937417984,
+      "epoch": 1.3485714285714285,
+      "grad_norm": 2.483147382736206,
+      "learning_rate": 1.1017142857142858e-05,
+      "loss": 0.7818619251251221,
+      "mean_token_accuracy": 0.8041326075792312,
+      "num_tokens": 18023210.0,
+      "step": 1180
+    },
+    {
+      "entropy": 1.3030154168605805,
+      "epoch": 1.3599999999999999,
+      "grad_norm": 3.100837469100952,
+      "learning_rate": 1.0940952380952381e-05,
+      "loss": 0.7691655158996582,
+      "mean_token_accuracy": 0.8024266973137856,
+      "num_tokens": 18172271.0,
+      "step": 1190
+    },
+    {
+      "entropy": 1.2880703262984752,
+      "epoch": 1.3714285714285714,
+      "grad_norm": 3.068323850631714,
+      "learning_rate": 1.0864761904761905e-05,
+      "loss": 0.8048301696777344,
+      "mean_token_accuracy": 0.7959965549409389,
+      "num_tokens": 18325046.0,
+      "step": 1200
+    },
+    {
+      "entropy": 1.3046514227986337,
+      "epoch": 1.3828571428571428,
+      "grad_norm": 3.33621883392334,
+      "learning_rate": 1.078857142857143e-05,
+      "loss": 0.8266304016113282,
+      "mean_token_accuracy": 0.7895134054124355,
+      "num_tokens": 18476571.0,
+      "step": 1210
+    },
+    {
+      "entropy": 1.2605851158499717,
+      "epoch": 1.3942857142857144,
+      "grad_norm": 3.5944457054138184,
+      "learning_rate": 1.0712380952380954e-05,
+      "loss": 0.8084980010986328,
+      "mean_token_accuracy": 0.7934398144483567,
+      "num_tokens": 18622849.0,
+      "step": 1220
+    },
+    {
+      "entropy": 1.2391214780509472,
+      "epoch": 1.4057142857142857,
+      "grad_norm": 2.6554486751556396,
+      "learning_rate": 1.0636190476190477e-05,
+      "loss": 0.7550242900848388,
+      "mean_token_accuracy": 0.8065047845244407,
+      "num_tokens": 18768698.0,
+      "step": 1230
+    },
+    {
+      "entropy": 1.2420047149062157,
+      "epoch": 1.4171428571428573,
+      "grad_norm": 2.5860495567321777,
+      "learning_rate": 1.056e-05,
+      "loss": 0.7261144161224365,
+      "mean_token_accuracy": 0.8117271907627582,
+      "num_tokens": 18923547.0,
+      "step": 1240
+    },
+    {
+      "entropy": 1.246000237762928,
+      "epoch": 1.4285714285714286,
+      "grad_norm": 3.1218044757843018,
+      "learning_rate": 1.0483809523809526e-05,
+      "loss": 0.7216352939605712,
+      "mean_token_accuracy": 0.8120975896716118,
+      "num_tokens": 19079479.0,
+      "step": 1250
+    },
+    {
+      "entropy": 1.207762571424246,
+      "epoch": 1.44,
+      "grad_norm": 2.393723726272583,
+      "learning_rate": 1.040761904761905e-05,
+      "loss": 0.7272539138793945,
+      "mean_token_accuracy": 0.8084844700992108,
+      "num_tokens": 19236273.0,
+      "step": 1260
+    },
+    {
+      "entropy": 1.2830321080982685,
+      "epoch": 1.4514285714285715,
+      "grad_norm": 3.0418026447296143,
+      "learning_rate": 1.0331428571428573e-05,
+      "loss": 0.7889233589172363,
+      "mean_token_accuracy": 0.7998958893120289,
+      "num_tokens": 19385795.0,
+      "step": 1270
+    },
+    {
+      "entropy": 1.2588836170732975,
+      "epoch": 1.4628571428571429,
+      "grad_norm": 2.668825626373291,
+      "learning_rate": 1.0255238095238096e-05,
+      "loss": 0.7696958065032959,
+      "mean_token_accuracy": 0.8028367407619953,
+      "num_tokens": 19540370.0,
+      "step": 1280
+    },
+    {
+      "entropy": 1.2221255615353583,
+      "epoch": 1.4742857142857142,
+      "grad_norm": 3.121307611465454,
+      "learning_rate": 1.017904761904762e-05,
+      "loss": 0.7629749774932861,
+      "mean_token_accuracy": 0.8041565544903279,
+      "num_tokens": 19695452.0,
+      "step": 1290
+    },
+    {
+      "entropy": 1.2615755833685398,
+      "epoch": 1.4857142857142858,
+      "grad_norm": 2.600470781326294,
+      "learning_rate": 1.0102857142857143e-05,
+      "loss": 0.8250988960266114,
+      "mean_token_accuracy": 0.7851416207849979,
+      "num_tokens": 19850924.0,
+      "step": 1300
+    },
+    {
+      "entropy": 1.2793317660689354,
+      "epoch": 1.497142857142857,
+      "grad_norm": 2.874340295791626,
+      "learning_rate": 1.0026666666666667e-05,
+      "loss": 0.771949577331543,
+      "mean_token_accuracy": 0.8040384240448475,
+      "num_tokens": 20014514.0,
+      "step": 1310
+    },
+    {
+      "entropy": 1.1907848060131072,
+      "epoch": 1.5085714285714285,
+      "grad_norm": 2.6014015674591064,
+      "learning_rate": 9.950476190476192e-06,
+      "loss": 0.7447234630584717,
+      "mean_token_accuracy": 0.8111775033175945,
+      "num_tokens": 20173798.0,
+      "step": 1320
+    },
+    {
+      "entropy": 1.2612031117081641,
+      "epoch": 1.52,
+      "grad_norm": 2.809692144393921,
+      "learning_rate": 9.874285714285715e-06,
+      "loss": 0.7573285102844238,
+      "mean_token_accuracy": 0.8068460263311863,
+      "num_tokens": 20318447.0,
+      "step": 1330
+    },
+    {
+      "entropy": 1.2434154883027078,
+      "epoch": 1.5314285714285716,
+      "grad_norm": 2.9109153747558594,
+      "learning_rate": 9.798095238095239e-06,
+      "loss": 0.7508068561553956,
+      "mean_token_accuracy": 0.8049707144498826,
+      "num_tokens": 20472944.0,
+      "step": 1340
+    },
+    {
+      "entropy": 1.2488966889679431,
+      "epoch": 1.5428571428571427,
+      "grad_norm": 3.0063366889953613,
+      "learning_rate": 9.721904761904762e-06,
+      "loss": 0.7871552467346191,
+      "mean_token_accuracy": 0.8020347572863102,
+      "num_tokens": 20625826.0,
+      "step": 1350
+    },
+    {
+      "entropy": 1.2348267659544945,
+      "epoch": 1.5542857142857143,
+      "grad_norm": 2.792886734008789,
+      "learning_rate": 9.645714285714286e-06,
+      "loss": 0.7784818172454834,
+      "mean_token_accuracy": 0.8054304026067257,
+      "num_tokens": 20774171.0,
+      "step": 1360
+    },
+    {
+      "entropy": 1.2300232581794261,
+      "epoch": 1.5657142857142858,
+      "grad_norm": 2.861610174179077,
+      "learning_rate": 9.569523809523811e-06,
+      "loss": 0.7702634811401368,
+      "mean_token_accuracy": 0.8006386004388333,
+      "num_tokens": 20922487.0,
+      "step": 1370
+    },
+    {
+      "entropy": 1.24477871209383,
+      "epoch": 1.5771428571428572,
+      "grad_norm": 3.1448233127593994,
+      "learning_rate": 9.493333333333334e-06,
+      "loss": 0.7313426494598388,
+      "mean_token_accuracy": 0.812004954367876,
+      "num_tokens": 21075326.0,
+      "step": 1380
+    },
+    {
+      "entropy": 1.2703936398029327,
+      "epoch": 1.5885714285714285,
+      "grad_norm": 2.5799062252044678,
+      "learning_rate": 9.417142857142858e-06,
+      "loss": 0.7714487552642822,
+      "mean_token_accuracy": 0.799166913330555,
+      "num_tokens": 21224257.0,
+      "step": 1390
+    },
+    {
+      "entropy": 1.2416778221726417,
+      "epoch": 1.6,
+      "grad_norm": 2.992957353591919,
+      "learning_rate": 9.340952380952381e-06,
+      "loss": 0.7809468746185303,
+      "mean_token_accuracy": 0.7984777301549911,
+      "num_tokens": 21373259.0,
+      "step": 1400
+    },
+    {
+      "entropy": 1.2557578206062316,
+      "epoch": 1.6114285714285714,
+      "grad_norm": 2.7290945053100586,
+      "learning_rate": 9.264761904761905e-06,
+      "loss": 0.7919119834899903,
+      "mean_token_accuracy": 0.7954873479902744,
+      "num_tokens": 21522253.0,
+      "step": 1410
+    },
+    {
+      "entropy": 1.2285942271351815,
+      "epoch": 1.6228571428571428,
+      "grad_norm": 2.8919999599456787,
+      "learning_rate": 9.188571428571428e-06,
+      "loss": 0.7635447978973389,
+      "mean_token_accuracy": 0.8025862380862236,
+      "num_tokens": 21671374.0,
+      "step": 1420
+    },
+    {
+      "entropy": 1.3119461573660374,
+      "epoch": 1.6342857142857143,
+      "grad_norm": 2.7214455604553223,
+      "learning_rate": 9.112380952380954e-06,
+      "loss": 0.7926105976104736,
+      "mean_token_accuracy": 0.7944708272814751,
+      "num_tokens": 21822289.0,
+      "step": 1430
+    },
+    {
+      "entropy": 1.2220344461500645,
+      "epoch": 1.6457142857142857,
+      "grad_norm": 2.327855348587036,
+      "learning_rate": 9.036190476190477e-06,
+      "loss": 0.7583229064941406,
+      "mean_token_accuracy": 0.8038429766893387,
+      "num_tokens": 21978538.0,
+      "step": 1440
+    },
+    {
+      "entropy": 1.266616266965866,
+      "epoch": 1.657142857142857,
+      "grad_norm": 2.7889652252197266,
+      "learning_rate": 8.96e-06,
+      "loss": 0.7689779281616211,
+      "mean_token_accuracy": 0.8045293688774109,
+      "num_tokens": 22133231.0,
+      "step": 1450
+    },
+    {
+      "entropy": 1.3114601746201515,
+      "epoch": 1.6685714285714286,
+      "grad_norm": 3.2477846145629883,
+      "learning_rate": 8.883809523809524e-06,
+      "loss": 0.7963542461395263,
+      "mean_token_accuracy": 0.8005617402493954,
+      "num_tokens": 22280643.0,
+      "step": 1460
+    },
+    {
+      "entropy": 1.2778782725334168,
+      "epoch": 1.6800000000000002,
+      "grad_norm": 3.1621718406677246,
+      "learning_rate": 8.807619047619048e-06,
+      "loss": 0.801125431060791,
+      "mean_token_accuracy": 0.7964559637010098,
+      "num_tokens": 22430891.0,
+      "step": 1470
+    },
+    {
+      "entropy": 1.2413391053676606,
+      "epoch": 1.6914285714285713,
+      "grad_norm": 2.561835527420044,
+      "learning_rate": 8.731428571428571e-06,
+      "loss": 0.7200276374816894,
+      "mean_token_accuracy": 0.8137163199484349,
+      "num_tokens": 22581471.0,
+      "step": 1480
+    },
+    {
+      "entropy": 1.2620924778282643,
+      "epoch": 1.7028571428571428,
+      "grad_norm": 2.8986573219299316,
+      "learning_rate": 8.655238095238096e-06,
+      "loss": 0.7427529335021973,
+      "mean_token_accuracy": 0.803333692252636,
+      "num_tokens": 22732244.0,
+      "step": 1490
+    },
+    {
+      "entropy": 1.2637714959681035,
+      "epoch": 1.7142857142857144,
+      "grad_norm": 2.8469996452331543,
+      "learning_rate": 8.57904761904762e-06,
+      "loss": 0.782107925415039,
+      "mean_token_accuracy": 0.8046764589846134,
+      "num_tokens": 22884081.0,
+      "step": 1500
+    },
+    {
+      "entropy": 1.2661903858184815,
+      "epoch": 1.7257142857142858,
+      "grad_norm": 3.0864884853363037,
+      "learning_rate": 8.502857142857143e-06,
+      "loss": 0.7898892879486084,
+      "mean_token_accuracy": 0.7921097949147224,
+      "num_tokens": 23034449.0,
+      "step": 1510
+    },
+    {
+      "entropy": 1.250692894309759,
+      "epoch": 1.737142857142857,
+      "grad_norm": 2.6670966148376465,
+      "learning_rate": 8.426666666666667e-06,
+      "loss": 0.7688261508941651,
+      "mean_token_accuracy": 0.8037325166165828,
+      "num_tokens": 23188136.0,
+      "step": 1520
+    },
+    {
+      "entropy": 1.215890847891569,
+      "epoch": 1.7485714285714287,
+      "grad_norm": 2.3385775089263916,
+      "learning_rate": 8.35047619047619e-06,
+      "loss": 0.7499132633209229,
+      "mean_token_accuracy": 0.8074461244046688,
+      "num_tokens": 23343060.0,
+      "step": 1530
+    },
+    {
+      "entropy": 1.242823239415884,
+      "epoch": 1.76,
+      "grad_norm": 2.630499839782715,
+      "learning_rate": 8.274285714285715e-06,
+      "loss": 0.7520731925964356,
+      "mean_token_accuracy": 0.8076161026954651,
+      "num_tokens": 23487902.0,
+      "step": 1540
+    },
+    {
+      "entropy": 1.2385052397847176,
+      "epoch": 1.7714285714285714,
+      "grad_norm": 2.76127552986145,
+      "learning_rate": 8.198095238095239e-06,
+      "loss": 0.7834786415100098,
+      "mean_token_accuracy": 0.7965474635362625,
+      "num_tokens": 23635117.0,
+      "step": 1550
+    },
+    {
+      "entropy": 1.1929515838623046,
+      "epoch": 1.782857142857143,
+      "grad_norm": 2.612879991531372,
+      "learning_rate": 8.121904761904762e-06,
+      "loss": 0.7256328582763671,
+      "mean_token_accuracy": 0.8150137662887573,
+      "num_tokens": 23795433.0,
+      "step": 1560
+    },
+    {
+      "entropy": 1.194692898541689,
+      "epoch": 1.7942857142857143,
+      "grad_norm": 2.6522936820983887,
+      "learning_rate": 8.045714285714286e-06,
+      "loss": 0.7276816368103027,
+      "mean_token_accuracy": 0.8094555623829365,
+      "num_tokens": 23950370.0,
+      "step": 1570
+    },
+    {
+      "entropy": 1.2585241250693797,
+      "epoch": 1.8057142857142856,
+      "grad_norm": 2.611926794052124,
+      "learning_rate": 7.969523809523811e-06,
+      "loss": 0.7796142578125,
+      "mean_token_accuracy": 0.8012560330331325,
+      "num_tokens": 24103914.0,
+      "step": 1580
+    },
+    {
+      "entropy": 1.2676213465631008,
+      "epoch": 1.8171428571428572,
+      "grad_norm": 2.5129570960998535,
+      "learning_rate": 7.893333333333335e-06,
+      "loss": 0.7887630462646484,
+      "mean_token_accuracy": 0.7965363718569278,
+      "num_tokens": 24258081.0,
+      "step": 1590
+    },
+    {
+      "entropy": 1.2579274140298367,
+      "epoch": 1.8285714285714287,
+      "grad_norm": 2.86506724357605,
+      "learning_rate": 7.817142857142858e-06,
+      "loss": 0.7497401714324952,
+      "mean_token_accuracy": 0.8082531146705151,
+      "num_tokens": 24409168.0,
+      "step": 1600
+    },
+    {
+      "entropy": 1.2318767301738263,
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.7256534099578857,
+      "learning_rate": 7.740952380952382e-06,
+      "loss": 0.7274055480957031,
+      "mean_token_accuracy": 0.8118016675114632,
+      "num_tokens": 24566252.0,
+      "step": 1610
+    },
+    {
+      "entropy": 1.2056004114449024,
+      "epoch": 1.8514285714285714,
+      "grad_norm": 2.6447949409484863,
+      "learning_rate": 7.664761904761905e-06,
+      "loss": 0.7131245613098145,
+      "mean_token_accuracy": 0.8102400578558445,
+      "num_tokens": 24721993.0,
+      "step": 1620
+    },
+    {
+      "entropy": 1.245617663115263,
+      "epoch": 1.862857142857143,
+      "grad_norm": 2.835374116897583,
+      "learning_rate": 7.588571428571429e-06,
+      "loss": 0.783673095703125,
+      "mean_token_accuracy": 0.8022521927952766,
+      "num_tokens": 24868978.0,
+      "step": 1630
+    },
+    {
+      "entropy": 1.240978506952524,
+      "epoch": 1.8742857142857143,
+      "grad_norm": 2.6456923484802246,
+      "learning_rate": 7.512380952380953e-06,
+      "loss": 0.7802390098571778,
+      "mean_token_accuracy": 0.7991835795342922,
+      "num_tokens": 25026573.0,
+      "step": 1640
+    },
+    {
+      "entropy": 1.2280743092298507,
+      "epoch": 1.8857142857142857,
+      "grad_norm": 2.847167491912842,
+      "learning_rate": 7.436190476190477e-06,
+      "loss": 0.8147419929504395,
+      "mean_token_accuracy": 0.7919134043157101,
+      "num_tokens": 25182791.0,
+      "step": 1650
+    },
+    {
+      "entropy": 1.216932328045368,
+      "epoch": 1.8971428571428572,
+      "grad_norm": 2.861248731613159,
+      "learning_rate": 7.360000000000001e-06,
+      "loss": 0.7451428890228271,
+      "mean_token_accuracy": 0.8088080659508705,
+      "num_tokens": 25326320.0,
+      "step": 1660
+    },
+    {
+      "entropy": 1.2367267780005933,
+      "epoch": 1.9085714285714286,
+      "grad_norm": 2.702206611633301,
+      "learning_rate": 7.283809523809524e-06,
+      "loss": 0.7687783241271973,
+      "mean_token_accuracy": 0.8046720631420612,
+      "num_tokens": 25481482.0,
+      "step": 1670
+    },
+    {
+      "entropy": 1.2137604996562004,
+      "epoch": 1.92,
+      "grad_norm": 2.446934938430786,
+      "learning_rate": 7.207619047619048e-06,
+      "loss": 0.7720860958099365,
+      "mean_token_accuracy": 0.8001258425414562,
+      "num_tokens": 25627745.0,
+      "step": 1680
+    },
+    {
+      "entropy": 1.2351884521543979,
+      "epoch": 1.9314285714285715,
+      "grad_norm": 2.682955026626587,
+      "learning_rate": 7.131428571428573e-06,
+      "loss": 0.7554657936096192,
+      "mean_token_accuracy": 0.8067528575658798,
+      "num_tokens": 25778484.0,
+      "step": 1690
+    },
+    {
+      "entropy": 1.2149645321071147,
+      "epoch": 1.9428571428571428,
+      "grad_norm": 3.032944679260254,
+      "learning_rate": 7.055238095238096e-06,
+      "loss": 0.6940214157104492,
+      "mean_token_accuracy": 0.8167023897171021,
+      "num_tokens": 25932265.0,
+      "step": 1700
+    },
+    {
+      "entropy": 1.205052537471056,
+      "epoch": 1.9542857142857142,
+      "grad_norm": 2.7198259830474854,
+      "learning_rate": 6.97904761904762e-06,
+      "loss": 0.7255962371826172,
+      "mean_token_accuracy": 0.8064791038632393,
+      "num_tokens": 26077725.0,
+      "step": 1710
+    },
+    {
+      "entropy": 1.1943283297121525,
+      "epoch": 1.9657142857142857,
+      "grad_norm": 2.6816186904907227,
+      "learning_rate": 6.902857142857143e-06,
+      "loss": 0.7239448547363281,
+      "mean_token_accuracy": 0.809258659183979,
+      "num_tokens": 26233086.0,
+      "step": 1720
+    },
+    {
+      "entropy": 1.2314366959035397,
+      "epoch": 1.977142857142857,
+      "grad_norm": 3.0669050216674805,
+      "learning_rate": 6.826666666666667e-06,
+      "loss": 0.8160985946655274,
+      "mean_token_accuracy": 0.7934448584914208,
+      "num_tokens": 26383494.0,
+      "step": 1730
+    },
+    {
+      "entropy": 1.201389105618,
+      "epoch": 1.9885714285714284,
+      "grad_norm": 2.6934287548065186,
+      "learning_rate": 6.75047619047619e-06,
+      "loss": 0.7483705043792724,
+      "mean_token_accuracy": 0.8099792324006557,
+      "num_tokens": 26534600.0,
+      "step": 1740
+    },
+    {
+      "entropy": 1.2289574392139913,
+      "epoch": 2.0,
+      "grad_norm": 2.9322891235351562,
+      "learning_rate": 6.6742857142857155e-06,
+      "loss": 0.7264975070953369,
+      "mean_token_accuracy": 0.8107106305658818,
+      "num_tokens": 26690084.0,
+      "step": 1750
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2625,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.268869501104333e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-1750/training_args.bin b/checkpoint-1750/training_args.bin
new file mode 100644
index 0000000..1dc481a
--- /dev/null
+++ b/checkpoint-1750/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:accc301f3ad366c553289bf58cab2de1488f42639e4826bf1d58853e8aa93ad2
+size 5841
diff --git a/checkpoint-2625/chat_template.jinja b/checkpoint-2625/chat_template.jinja
new file mode 100644
index 0000000..bdf7919
--- /dev/null
+++ b/checkpoint-2625/chat_template.jinja
@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- 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' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) 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' }}
+{%- endif %}
diff --git a/checkpoint-2625/config.json b/checkpoint-2625/config.json
new file mode 100644
index 0000000..7ff1bce
--- /dev/null
+++ b/checkpoint-2625/config.json
@@ -0,0 +1,61 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "float32",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000.0,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.5.4",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/checkpoint-2625/generation_config.json b/checkpoint-2625/generation_config.json
new file mode 100644
index 0000000..846ba08
--- /dev/null
+++ b/checkpoint-2625/generation_config.json
@@ -0,0 +1,14 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "5.5.4",
+  "use_cache": false
+}
diff --git a/checkpoint-2625/model.safetensors b/checkpoint-2625/model.safetensors
new file mode 100644
index 0000000..163d6c1
--- /dev/null
+++ b/checkpoint-2625/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebaae81661331d6e842dce368a7337b16107ab56f651ceefde1fb347cf1acc6b
+size 6174895536
diff --git a/checkpoint-2625/optimizer.pt b/checkpoint-2625/optimizer.pt
new file mode 100644
index 0000000..ad347eb
--- /dev/null
+++ b/checkpoint-2625/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfcbc6d3617a3524bbccc7923eb522af5fff4270d019e45031f1787810b9f939
+size 12350013801
diff --git a/checkpoint-2625/rng_state.pth b/checkpoint-2625/rng_state.pth
new file mode 100644
index 0000000..bd3c190
--- /dev/null
+++ b/checkpoint-2625/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4a9f217e852f439efa6bd32fde98d6867f11aa6ea13ddc021ba10af6a0b0934
+size 14645
diff --git a/checkpoint-2625/scheduler.pt b/checkpoint-2625/scheduler.pt
new file mode 100644
index 0000000..08b60c6
--- /dev/null
+++ b/checkpoint-2625/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a90cf32bcaa40d2760ac49ca5a2167553d01347c3172274302ea44fcf542f183
+size 1465
diff --git a/checkpoint-2625/tokenizer.json b/checkpoint-2625/tokenizer.json
new file mode 100644
index 0000000..34510ff
--- /dev/null
+++ b/checkpoint-2625/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
+size 11421892
diff --git a/checkpoint-2625/tokenizer_config.json b/checkpoint-2625/tokenizer_config.json
new file mode 100644
index 0000000..0385b12
--- /dev/null
+++ b/checkpoint-2625/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_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|>"
+  ],
+  "is_local": true,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/checkpoint-2625/trainer_state.json b/checkpoint-2625/trainer_state.json
new file mode 100644
index 0000000..8a31a05
--- /dev/null
+++ b/checkpoint-2625/trainer_state.json
@@ -0,0 +1,2654 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 2625,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.6696195736527444,
+      "epoch": 0.011428571428571429,
+      "grad_norm": 3.7833805084228516,
+      "learning_rate": 1.9931428571428572e-05,
+      "loss": 1.3203315734863281,
+      "mean_token_accuracy": 0.7056225836277008,
+      "num_tokens": 157266.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.7283091336488723,
+      "epoch": 0.022857142857142857,
+      "grad_norm": 3.4964022636413574,
+      "learning_rate": 1.9855238095238097e-05,
+      "loss": 1.4036809921264648,
+      "mean_token_accuracy": 0.6812524735927582,
+      "num_tokens": 300430.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.7663277059793472,
+      "epoch": 0.03428571428571429,
+      "grad_norm": 4.590687274932861,
+      "learning_rate": 1.977904761904762e-05,
+      "loss": 1.319040012359619,
+      "mean_token_accuracy": 0.6928375385701656,
+      "num_tokens": 456129.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.7699106186628342,
+      "epoch": 0.045714285714285714,
+      "grad_norm": 3.2531967163085938,
+      "learning_rate": 1.9702857142857144e-05,
+      "loss": 1.3776089668273925,
+      "mean_token_accuracy": 0.686311125010252,
+      "num_tokens": 615179.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.7856220349669456,
+      "epoch": 0.05714285714285714,
+      "grad_norm": 3.356832981109619,
+      "learning_rate": 1.9626666666666666e-05,
+      "loss": 1.32177791595459,
+      "mean_token_accuracy": 0.6952089086174965,
+      "num_tokens": 769672.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.7628433778882027,
+      "epoch": 0.06857142857142857,
+      "grad_norm": 3.965970754623413,
+      "learning_rate": 1.955047619047619e-05,
+      "loss": 1.338463306427002,
+      "mean_token_accuracy": 0.6953431364148855,
+      "num_tokens": 918402.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.7276440173387528,
+      "epoch": 0.08,
+      "grad_norm": 3.4119951725006104,
+      "learning_rate": 1.9474285714285717e-05,
+      "loss": 1.2838379859924316,
+      "mean_token_accuracy": 0.6987539391964674,
+      "num_tokens": 1076598.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.748648078739643,
+      "epoch": 0.09142857142857143,
+      "grad_norm": 3.6497740745544434,
+      "learning_rate": 1.9398095238095242e-05,
+      "loss": 1.3314701080322267,
+      "mean_token_accuracy": 0.695335328578949,
+      "num_tokens": 1228084.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.7111000671982766,
+      "epoch": 0.10285714285714286,
+      "grad_norm": 3.3122711181640625,
+      "learning_rate": 1.9321904761904764e-05,
+      "loss": 1.3027889251708984,
+      "mean_token_accuracy": 0.6987856134772301,
+      "num_tokens": 1386707.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.7278845444321633,
+      "epoch": 0.11428571428571428,
+      "grad_norm": 3.183835744857788,
+      "learning_rate": 1.924571428571429e-05,
+      "loss": 1.282834243774414,
+      "mean_token_accuracy": 0.6966647863388061,
+      "num_tokens": 1535406.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.7132800757884978,
+      "epoch": 0.12571428571428572,
+      "grad_norm": 3.4031126499176025,
+      "learning_rate": 1.916952380952381e-05,
+      "loss": 1.2409746170043945,
+      "mean_token_accuracy": 0.7101543735712766,
+      "num_tokens": 1680743.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.7373003125190736,
+      "epoch": 0.13714285714285715,
+      "grad_norm": 3.4207799434661865,
+      "learning_rate": 1.9093333333333336e-05,
+      "loss": 1.2909335136413573,
+      "mean_token_accuracy": 0.701117354631424,
+      "num_tokens": 1834014.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.7092810094356536,
+      "epoch": 0.14857142857142858,
+      "grad_norm": 3.3323278427124023,
+      "learning_rate": 1.9017142857142858e-05,
+      "loss": 1.3009337425231933,
+      "mean_token_accuracy": 0.6907773412764072,
+      "num_tokens": 1989349.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.7240261450409888,
+      "epoch": 0.16,
+      "grad_norm": 2.9499671459198,
+      "learning_rate": 1.8940952380952383e-05,
+      "loss": 1.2700564384460449,
+      "mean_token_accuracy": 0.7077916815876961,
+      "num_tokens": 2139614.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.655298152565956,
+      "epoch": 0.17142857142857143,
+      "grad_norm": 3.1871607303619385,
+      "learning_rate": 1.8864761904761905e-05,
+      "loss": 1.2251598358154296,
+      "mean_token_accuracy": 0.7074991308152676,
+      "num_tokens": 2293879.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.665707252919674,
+      "epoch": 0.18285714285714286,
+      "grad_norm": 3.290358304977417,
+      "learning_rate": 1.878857142857143e-05,
+      "loss": 1.2641545295715333,
+      "mean_token_accuracy": 0.7021077655255794,
+      "num_tokens": 2446447.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.7259689077734948,
+      "epoch": 0.19428571428571428,
+      "grad_norm": 3.1401703357696533,
+      "learning_rate": 1.871238095238095e-05,
+      "loss": 1.315788745880127,
+      "mean_token_accuracy": 0.6926948539912701,
+      "num_tokens": 2595389.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.7142695009708404,
+      "epoch": 0.2057142857142857,
+      "grad_norm": 3.4904985427856445,
+      "learning_rate": 1.8636190476190477e-05,
+      "loss": 1.2755393028259276,
+      "mean_token_accuracy": 0.7040164832025766,
+      "num_tokens": 2745608.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.6242914512753486,
+      "epoch": 0.21714285714285714,
+      "grad_norm": 2.949880361557007,
+      "learning_rate": 1.8560000000000002e-05,
+      "loss": 1.190676498413086,
+      "mean_token_accuracy": 0.7185754984617233,
+      "num_tokens": 2894625.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.7364101350307464,
+      "epoch": 0.22857142857142856,
+      "grad_norm": 3.125112533569336,
+      "learning_rate": 1.8483809523809527e-05,
+      "loss": 1.2999603271484375,
+      "mean_token_accuracy": 0.6976742841303348,
+      "num_tokens": 3046487.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.6772031649947166,
+      "epoch": 0.24,
+      "grad_norm": 2.813882350921631,
+      "learning_rate": 1.840761904761905e-05,
+      "loss": 1.2316917419433593,
+      "mean_token_accuracy": 0.7120772495865821,
+      "num_tokens": 3196704.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.6737122178077697,
+      "epoch": 0.25142857142857145,
+      "grad_norm": 2.8218936920166016,
+      "learning_rate": 1.8331428571428574e-05,
+      "loss": 1.274891471862793,
+      "mean_token_accuracy": 0.6991564333438873,
+      "num_tokens": 3348365.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.6344200998544693,
+      "epoch": 0.26285714285714284,
+      "grad_norm": 2.9688668251037598,
+      "learning_rate": 1.8255238095238096e-05,
+      "loss": 1.2082077026367188,
+      "mean_token_accuracy": 0.7141494430601597,
+      "num_tokens": 3509464.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.6166698306798934,
+      "epoch": 0.2742857142857143,
+      "grad_norm": 3.007690668106079,
+      "learning_rate": 1.817904761904762e-05,
+      "loss": 1.1915247917175293,
+      "mean_token_accuracy": 0.7145515654236079,
+      "num_tokens": 3661779.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.6418054014444352,
+      "epoch": 0.2857142857142857,
+      "grad_norm": 2.9565091133117676,
+      "learning_rate": 1.8102857142857143e-05,
+      "loss": 1.2041254043579102,
+      "mean_token_accuracy": 0.7189058735966682,
+      "num_tokens": 3809611.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.6628487035632133,
+      "epoch": 0.29714285714285715,
+      "grad_norm": 2.7851827144622803,
+      "learning_rate": 1.8026666666666668e-05,
+      "loss": 1.2013505935668944,
+      "mean_token_accuracy": 0.7156483136117459,
+      "num_tokens": 3960501.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.6495753720402717,
+      "epoch": 0.30857142857142855,
+      "grad_norm": 3.0212292671203613,
+      "learning_rate": 1.7950476190476193e-05,
+      "loss": 1.2240056991577148,
+      "mean_token_accuracy": 0.7174453124403953,
+      "num_tokens": 4107783.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.693277756869793,
+      "epoch": 0.32,
+      "grad_norm": 2.9716391563415527,
+      "learning_rate": 1.7874285714285715e-05,
+      "loss": 1.207960319519043,
+      "mean_token_accuracy": 0.7157351300120354,
+      "num_tokens": 4252335.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.6619624510407447,
+      "epoch": 0.3314285714285714,
+      "grad_norm": 2.851775646209717,
+      "learning_rate": 1.779809523809524e-05,
+      "loss": 1.248055648803711,
+      "mean_token_accuracy": 0.7128231346607208,
+      "num_tokens": 4407579.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.6617254123091698,
+      "epoch": 0.34285714285714286,
+      "grad_norm": 3.254999876022339,
+      "learning_rate": 1.7721904761904762e-05,
+      "loss": 1.184535598754883,
+      "mean_token_accuracy": 0.7231362499296665,
+      "num_tokens": 4551517.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.5988039791584014,
+      "epoch": 0.35428571428571426,
+      "grad_norm": 3.110564708709717,
+      "learning_rate": 1.7645714285714287e-05,
+      "loss": 1.148481559753418,
+      "mean_token_accuracy": 0.7262889519333839,
+      "num_tokens": 4708257.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.6530713498592378,
+      "epoch": 0.3657142857142857,
+      "grad_norm": 3.1247048377990723,
+      "learning_rate": 1.7569523809523812e-05,
+      "loss": 1.2834845542907716,
+      "mean_token_accuracy": 0.702876628190279,
+      "num_tokens": 4859075.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.6123583316802979,
+      "epoch": 0.37714285714285717,
+      "grad_norm": 2.90073823928833,
+      "learning_rate": 1.7493333333333334e-05,
+      "loss": 1.1132370948791503,
+      "mean_token_accuracy": 0.7313019946217537,
+      "num_tokens": 5011395.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.618579277396202,
+      "epoch": 0.38857142857142857,
+      "grad_norm": 2.936490058898926,
+      "learning_rate": 1.741714285714286e-05,
+      "loss": 1.2263901710510254,
+      "mean_token_accuracy": 0.7069775387644768,
+      "num_tokens": 5174272.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.663301782310009,
+      "epoch": 0.4,
+      "grad_norm": 3.241185426712036,
+      "learning_rate": 1.734095238095238e-05,
+      "loss": 1.2628487586975097,
+      "mean_token_accuracy": 0.7058808147907257,
+      "num_tokens": 5327670.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.6857714764773846,
+      "epoch": 0.4114285714285714,
+      "grad_norm": 2.9059698581695557,
+      "learning_rate": 1.7264761904761906e-05,
+      "loss": 1.2412065505981444,
+      "mean_token_accuracy": 0.7062458634376526,
+      "num_tokens": 5476064.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.6253925159573555,
+      "epoch": 0.4228571428571429,
+      "grad_norm": 2.9899251461029053,
+      "learning_rate": 1.718857142857143e-05,
+      "loss": 1.210451316833496,
+      "mean_token_accuracy": 0.7089586913585663,
+      "num_tokens": 5636828.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.5880890995264054,
+      "epoch": 0.4342857142857143,
+      "grad_norm": 3.0461251735687256,
+      "learning_rate": 1.7112380952380953e-05,
+      "loss": 1.1992740631103516,
+      "mean_token_accuracy": 0.7144776731729507,
+      "num_tokens": 5788462.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.6384857930243015,
+      "epoch": 0.44571428571428573,
+      "grad_norm": 2.8842580318450928,
+      "learning_rate": 1.703619047619048e-05,
+      "loss": 1.2124631881713868,
+      "mean_token_accuracy": 0.7115042388439179,
+      "num_tokens": 5943108.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.5625910967588426,
+      "epoch": 0.45714285714285713,
+      "grad_norm": 3.0432047843933105,
+      "learning_rate": 1.696e-05,
+      "loss": 1.1515995025634767,
+      "mean_token_accuracy": 0.716732618585229,
+      "num_tokens": 6098679.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.6026007533073425,
+      "epoch": 0.4685714285714286,
+      "grad_norm": 2.781804323196411,
+      "learning_rate": 1.6883809523809525e-05,
+      "loss": 1.180698299407959,
+      "mean_token_accuracy": 0.7132757268846035,
+      "num_tokens": 6250301.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.5140968441963196,
+      "epoch": 0.48,
+      "grad_norm": 2.9592671394348145,
+      "learning_rate": 1.6807619047619047e-05,
+      "loss": 1.1643908500671387,
+      "mean_token_accuracy": 0.7223096333444119,
+      "num_tokens": 6407356.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.6122839108109475,
+      "epoch": 0.49142857142857144,
+      "grad_norm": 3.2809886932373047,
+      "learning_rate": 1.6731428571428572e-05,
+      "loss": 1.2397551536560059,
+      "mean_token_accuracy": 0.707809828966856,
+      "num_tokens": 6559819.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.603949698805809,
+      "epoch": 0.5028571428571429,
+      "grad_norm": 2.721527338027954,
+      "learning_rate": 1.6655238095238098e-05,
+      "loss": 1.1895122528076172,
+      "mean_token_accuracy": 0.7227885220199823,
+      "num_tokens": 6716227.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.582472313940525,
+      "epoch": 0.5142857142857142,
+      "grad_norm": 2.8015570640563965,
+      "learning_rate": 1.6579047619047623e-05,
+      "loss": 1.1144920349121095,
+      "mean_token_accuracy": 0.7284132238477469,
+      "num_tokens": 6867486.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.6025653272867202,
+      "epoch": 0.5257142857142857,
+      "grad_norm": 3.1553401947021484,
+      "learning_rate": 1.6502857142857145e-05,
+      "loss": 1.1947656631469727,
+      "mean_token_accuracy": 0.7160489454865455,
+      "num_tokens": 7012179.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.5280416823923588,
+      "epoch": 0.5371428571428571,
+      "grad_norm": 2.97623348236084,
+      "learning_rate": 1.642666666666667e-05,
+      "loss": 1.149429702758789,
+      "mean_token_accuracy": 0.7256167627871036,
+      "num_tokens": 7171638.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.6620354026556015,
+      "epoch": 0.5485714285714286,
+      "grad_norm": 3.1873958110809326,
+      "learning_rate": 1.635047619047619e-05,
+      "loss": 1.2588795661926269,
+      "mean_token_accuracy": 0.707894566655159,
+      "num_tokens": 7322167.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.581991147994995,
+      "epoch": 0.56,
+      "grad_norm": 3.385500907897949,
+      "learning_rate": 1.6274285714285717e-05,
+      "loss": 1.2005508422851563,
+      "mean_token_accuracy": 0.7142980419099331,
+      "num_tokens": 7466661.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.6193878114223481,
+      "epoch": 0.5714285714285714,
+      "grad_norm": 2.9256980419158936,
+      "learning_rate": 1.619809523809524e-05,
+      "loss": 1.1964147567749024,
+      "mean_token_accuracy": 0.715008432790637,
+      "num_tokens": 7620522.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.560711294412613,
+      "epoch": 0.5828571428571429,
+      "grad_norm": 2.548938035964966,
+      "learning_rate": 1.6121904761904764e-05,
+      "loss": 1.174423599243164,
+      "mean_token_accuracy": 0.7205559216439724,
+      "num_tokens": 7772467.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.6210713252425193,
+      "epoch": 0.5942857142857143,
+      "grad_norm": 2.6339938640594482,
+      "learning_rate": 1.6045714285714286e-05,
+      "loss": 1.153661060333252,
+      "mean_token_accuracy": 0.7224138833582401,
+      "num_tokens": 7922030.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.5822248332202435,
+      "epoch": 0.6057142857142858,
+      "grad_norm": 2.4934332370758057,
+      "learning_rate": 1.596952380952381e-05,
+      "loss": 1.1586255073547362,
+      "mean_token_accuracy": 0.7230928719043732,
+      "num_tokens": 8067955.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.5642932564020158,
+      "epoch": 0.6171428571428571,
+      "grad_norm": 2.985651969909668,
+      "learning_rate": 1.5893333333333333e-05,
+      "loss": 1.1338199615478515,
+      "mean_token_accuracy": 0.7240421302616596,
+      "num_tokens": 8226369.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.6006409794092178,
+      "epoch": 0.6285714285714286,
+      "grad_norm": 2.6047215461730957,
+      "learning_rate": 1.5817142857142858e-05,
+      "loss": 1.155271625518799,
+      "mean_token_accuracy": 0.7230217695236206,
+      "num_tokens": 8376285.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.5791691079735757,
+      "epoch": 0.64,
+      "grad_norm": 2.855867385864258,
+      "learning_rate": 1.5740952380952383e-05,
+      "loss": 1.1597973823547363,
+      "mean_token_accuracy": 0.7209223508834839,
+      "num_tokens": 8530322.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.6016968488693237,
+      "epoch": 0.6514285714285715,
+      "grad_norm": 3.205124855041504,
+      "learning_rate": 1.5664761904761908e-05,
+      "loss": 1.1405302047729493,
+      "mean_token_accuracy": 0.7231174889951945,
+      "num_tokens": 8676991.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.5366713739931583,
+      "epoch": 0.6628571428571428,
+      "grad_norm": 3.1047000885009766,
+      "learning_rate": 1.558857142857143e-05,
+      "loss": 1.1293843269348145,
+      "mean_token_accuracy": 0.7271328650414943,
+      "num_tokens": 8833861.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.5866449058055878,
+      "epoch": 0.6742857142857143,
+      "grad_norm": 2.884098529815674,
+      "learning_rate": 1.5512380952380955e-05,
+      "loss": 1.2017478942871094,
+      "mean_token_accuracy": 0.716648469120264,
+      "num_tokens": 8983438.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.610584917664528,
+      "epoch": 0.6857142857142857,
+      "grad_norm": 2.7602009773254395,
+      "learning_rate": 1.5436190476190477e-05,
+      "loss": 1.1553681373596192,
+      "mean_token_accuracy": 0.72271120660007,
+      "num_tokens": 9138958.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.5292311601340771,
+      "epoch": 0.6971428571428572,
+      "grad_norm": 2.8975393772125244,
+      "learning_rate": 1.5360000000000002e-05,
+      "loss": 1.1547229766845704,
+      "mean_token_accuracy": 0.7226926662027836,
+      "num_tokens": 9290314.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.5577459454536438,
+      "epoch": 0.7085714285714285,
+      "grad_norm": 2.5054473876953125,
+      "learning_rate": 1.5283809523809524e-05,
+      "loss": 1.1239413261413573,
+      "mean_token_accuracy": 0.7288053043186664,
+      "num_tokens": 9442584.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.6545007571578025,
+      "epoch": 0.72,
+      "grad_norm": 3.156757354736328,
+      "learning_rate": 1.5207619047619049e-05,
+      "loss": 1.2162582397460937,
+      "mean_token_accuracy": 0.7100601557642221,
+      "num_tokens": 9589146.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.6293067336082458,
+      "epoch": 0.7314285714285714,
+      "grad_norm": 2.7666168212890625,
+      "learning_rate": 1.5131428571428572e-05,
+      "loss": 1.1888948440551759,
+      "mean_token_accuracy": 0.7159062504768372,
+      "num_tokens": 9736056.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.5786321811378001,
+      "epoch": 0.7428571428571429,
+      "grad_norm": 2.6019463539123535,
+      "learning_rate": 1.5055238095238096e-05,
+      "loss": 1.0923637390136718,
+      "mean_token_accuracy": 0.7382036492228508,
+      "num_tokens": 9898410.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.574275678396225,
+      "epoch": 0.7542857142857143,
+      "grad_norm": 2.7830586433410645,
+      "learning_rate": 1.497904761904762e-05,
+      "loss": 1.1398842811584473,
+      "mean_token_accuracy": 0.7246641829609871,
+      "num_tokens": 10051154.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.5554455459117889,
+      "epoch": 0.7657142857142857,
+      "grad_norm": 2.8458149433135986,
+      "learning_rate": 1.4902857142857143e-05,
+      "loss": 1.139054298400879,
+      "mean_token_accuracy": 0.726585789769888,
+      "num_tokens": 10199110.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.6094272822141646,
+      "epoch": 0.7771428571428571,
+      "grad_norm": 2.9835047721862793,
+      "learning_rate": 1.4826666666666666e-05,
+      "loss": 1.1978397369384766,
+      "mean_token_accuracy": 0.7123216778039932,
+      "num_tokens": 10344927.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.5142188727855683,
+      "epoch": 0.7885714285714286,
+      "grad_norm": 2.7492172718048096,
+      "learning_rate": 1.475047619047619e-05,
+      "loss": 1.048453140258789,
+      "mean_token_accuracy": 0.7415693461894989,
+      "num_tokens": 10505422.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.633454330265522,
+      "epoch": 0.8,
+      "grad_norm": 2.645536184310913,
+      "learning_rate": 1.4674285714285717e-05,
+      "loss": 1.1423879623413087,
+      "mean_token_accuracy": 0.7272885538637638,
+      "num_tokens": 10652367.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.4862130410969256,
+      "epoch": 0.8114285714285714,
+      "grad_norm": 2.473473310470581,
+      "learning_rate": 1.459809523809524e-05,
+      "loss": 1.019435214996338,
+      "mean_token_accuracy": 0.7454098269343377,
+      "num_tokens": 10812282.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.541966749727726,
+      "epoch": 0.8228571428571428,
+      "grad_norm": 2.5247135162353516,
+      "learning_rate": 1.4521904761904764e-05,
+      "loss": 1.0726137161254883,
+      "mean_token_accuracy": 0.7371681958436966,
+      "num_tokens": 10967866.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.5219792507588863,
+      "epoch": 0.8342857142857143,
+      "grad_norm": 2.35876202583313,
+      "learning_rate": 1.4445714285714287e-05,
+      "loss": 1.0680880546569824,
+      "mean_token_accuracy": 0.7386457361280918,
+      "num_tokens": 11120412.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.506186830997467,
+      "epoch": 0.8457142857142858,
+      "grad_norm": 2.9589080810546875,
+      "learning_rate": 1.436952380952381e-05,
+      "loss": 1.0974149703979492,
+      "mean_token_accuracy": 0.7378807857632637,
+      "num_tokens": 11277244.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.569691103696823,
+      "epoch": 0.8571428571428571,
+      "grad_norm": 2.5306074619293213,
+      "learning_rate": 1.4293333333333334e-05,
+      "loss": 1.177268123626709,
+      "mean_token_accuracy": 0.7250622183084487,
+      "num_tokens": 11436007.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.5297324053943158,
+      "epoch": 0.8685714285714285,
+      "grad_norm": 2.7663497924804688,
+      "learning_rate": 1.4217142857142858e-05,
+      "loss": 1.1369609832763672,
+      "mean_token_accuracy": 0.7221393447369338,
+      "num_tokens": 11587615.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.4932486459612846,
+      "epoch": 0.88,
+      "grad_norm": 2.5511255264282227,
+      "learning_rate": 1.4140952380952381e-05,
+      "loss": 1.1008304595947265,
+      "mean_token_accuracy": 0.7303555272519588,
+      "num_tokens": 11747686.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.5960537254810334,
+      "epoch": 0.8914285714285715,
+      "grad_norm": 2.450721025466919,
+      "learning_rate": 1.4064761904761905e-05,
+      "loss": 1.1708695411682128,
+      "mean_token_accuracy": 0.7219238199293614,
+      "num_tokens": 11904676.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.5482905827462674,
+      "epoch": 0.9028571428571428,
+      "grad_norm": 3.1114189624786377,
+      "learning_rate": 1.398857142857143e-05,
+      "loss": 1.184025478363037,
+      "mean_token_accuracy": 0.720519745349884,
+      "num_tokens": 12059825.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.5766600042581558,
+      "epoch": 0.9142857142857143,
+      "grad_norm": 3.0541417598724365,
+      "learning_rate": 1.3912380952380953e-05,
+      "loss": 1.2167075157165528,
+      "mean_token_accuracy": 0.7153365164995193,
+      "num_tokens": 12218813.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.5715911239385605,
+      "epoch": 0.9257142857142857,
+      "grad_norm": 2.4761741161346436,
+      "learning_rate": 1.3836190476190477e-05,
+      "loss": 1.0981733322143554,
+      "mean_token_accuracy": 0.7329004533588886,
+      "num_tokens": 12366807.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.5434480041265488,
+      "epoch": 0.9371428571428572,
+      "grad_norm": 3.157036304473877,
+      "learning_rate": 1.376e-05,
+      "loss": 1.0953783988952637,
+      "mean_token_accuracy": 0.7374757573008537,
+      "num_tokens": 12519277.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.5731646195054054,
+      "epoch": 0.9485714285714286,
+      "grad_norm": 2.9288859367370605,
+      "learning_rate": 1.3683809523809526e-05,
+      "loss": 1.1861271858215332,
+      "mean_token_accuracy": 0.7202281527221203,
+      "num_tokens": 12668916.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.5766675099730492,
+      "epoch": 0.96,
+      "grad_norm": 2.3330445289611816,
+      "learning_rate": 1.3607619047619049e-05,
+      "loss": 1.1634528160095214,
+      "mean_token_accuracy": 0.7218175616115332,
+      "num_tokens": 12824349.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.5784073442220687,
+      "epoch": 0.9714285714285714,
+      "grad_norm": 2.8480734825134277,
+      "learning_rate": 1.3531428571428573e-05,
+      "loss": 1.1402252197265625,
+      "mean_token_accuracy": 0.7281969003379345,
+      "num_tokens": 12974198.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.5565635196864605,
+      "epoch": 0.9828571428571429,
+      "grad_norm": 3.2169361114501953,
+      "learning_rate": 1.3455238095238096e-05,
+      "loss": 1.0919082641601563,
+      "mean_token_accuracy": 0.7359561800956727,
+      "num_tokens": 13125877.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.6499854236841203,
+      "epoch": 0.9942857142857143,
+      "grad_norm": 2.59342622756958,
+      "learning_rate": 1.3379047619047621e-05,
+      "loss": 1.1824271202087402,
+      "mean_token_accuracy": 0.7206701695919037,
+      "num_tokens": 13270434.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.5392458021640778,
+      "epoch": 1.0057142857142858,
+      "grad_norm": 2.1802525520324707,
+      "learning_rate": 1.3302857142857145e-05,
+      "loss": 0.9705311775207519,
+      "mean_token_accuracy": 0.7628285072743892,
+      "num_tokens": 13419321.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.3396786622703076,
+      "epoch": 1.0171428571428571,
+      "grad_norm": 2.7976434230804443,
+      "learning_rate": 1.3226666666666668e-05,
+      "loss": 0.753804874420166,
+      "mean_token_accuracy": 0.8103815123438836,
+      "num_tokens": 13564219.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.263124795258045,
+      "epoch": 1.0285714285714285,
+      "grad_norm": 2.680875062942505,
+      "learning_rate": 1.3150476190476192e-05,
+      "loss": 0.8032585144042969,
+      "mean_token_accuracy": 0.8010107159614563,
+      "num_tokens": 13716646.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.2998421356081962,
+      "epoch": 1.04,
+      "grad_norm": 3.176466941833496,
+      "learning_rate": 1.3074285714285715e-05,
+      "loss": 0.7720170497894288,
+      "mean_token_accuracy": 0.8062204010784626,
+      "num_tokens": 13865045.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.268384473770857,
+      "epoch": 1.0514285714285714,
+      "grad_norm": 3.2034316062927246,
+      "learning_rate": 1.2998095238095239e-05,
+      "loss": 0.7246803283691406,
+      "mean_token_accuracy": 0.8114085927605629,
+      "num_tokens": 14019242.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.2792991541326046,
+      "epoch": 1.062857142857143,
+      "grad_norm": 2.799799919128418,
+      "learning_rate": 1.2921904761904762e-05,
+      "loss": 0.7459653377532959,
+      "mean_token_accuracy": 0.8081640765070915,
+      "num_tokens": 14170551.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.2749989867210387,
+      "epoch": 1.0742857142857143,
+      "grad_norm": 2.315112352371216,
+      "learning_rate": 1.2845714285714286e-05,
+      "loss": 0.7502207279205322,
+      "mean_token_accuracy": 0.8043648317456246,
+      "num_tokens": 14320423.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.3014213733375073,
+      "epoch": 1.0857142857142856,
+      "grad_norm": 2.8643505573272705,
+      "learning_rate": 1.2769523809523811e-05,
+      "loss": 0.7972130298614502,
+      "mean_token_accuracy": 0.798911166191101,
+      "num_tokens": 14472454.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.309807014465332,
+      "epoch": 1.0971428571428572,
+      "grad_norm": 2.9528074264526367,
+      "learning_rate": 1.2693333333333336e-05,
+      "loss": 0.762692928314209,
+      "mean_token_accuracy": 0.7993374645709992,
+      "num_tokens": 14621285.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.2928462289273739,
+      "epoch": 1.1085714285714285,
+      "grad_norm": 3.0441250801086426,
+      "learning_rate": 1.261714285714286e-05,
+      "loss": 0.8246999740600586,
+      "mean_token_accuracy": 0.7930732540786266,
+      "num_tokens": 14776609.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.2388829864561557,
+      "epoch": 1.12,
+      "grad_norm": 2.831936836242676,
+      "learning_rate": 1.2540952380952383e-05,
+      "loss": 0.7267831802368164,
+      "mean_token_accuracy": 0.8151453614234925,
+      "num_tokens": 14935547.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.2560633935034275,
+      "epoch": 1.1314285714285715,
+      "grad_norm": 2.852947473526001,
+      "learning_rate": 1.2464761904761907e-05,
+      "loss": 0.7547249794006348,
+      "mean_token_accuracy": 0.805939918756485,
+      "num_tokens": 15090483.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.241182105243206,
+      "epoch": 1.1428571428571428,
+      "grad_norm": 3.028001546859741,
+      "learning_rate": 1.238857142857143e-05,
+      "loss": 0.7496088027954102,
+      "mean_token_accuracy": 0.8107783786952496,
+      "num_tokens": 15242584.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.348277000337839,
+      "epoch": 1.1542857142857144,
+      "grad_norm": 2.8477680683135986,
+      "learning_rate": 1.2312380952380953e-05,
+      "loss": 0.8604803085327148,
+      "mean_token_accuracy": 0.7804459646344185,
+      "num_tokens": 15391767.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.250984299182892,
+      "epoch": 1.1657142857142857,
+      "grad_norm": 2.5354862213134766,
+      "learning_rate": 1.2236190476190477e-05,
+      "loss": 0.7215850353240967,
+      "mean_token_accuracy": 0.811321247369051,
+      "num_tokens": 15543892.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.2813407756388187,
+      "epoch": 1.177142857142857,
+      "grad_norm": 2.451107978820801,
+      "learning_rate": 1.216e-05,
+      "loss": 0.7772514820098877,
+      "mean_token_accuracy": 0.8003417395055294,
+      "num_tokens": 15692671.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.2703118346631528,
+      "epoch": 1.1885714285714286,
+      "grad_norm": 2.9708104133605957,
+      "learning_rate": 1.2083809523809524e-05,
+      "loss": 0.8101517677307128,
+      "mean_token_accuracy": 0.7938590474426747,
+      "num_tokens": 15849058.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.2321988992393016,
+      "epoch": 1.2,
+      "grad_norm": 2.5784900188446045,
+      "learning_rate": 1.2007619047619047e-05,
+      "loss": 0.7412232398986817,
+      "mean_token_accuracy": 0.8034046024084092,
+      "num_tokens": 16006785.0,
+      "step": 1050
+    },
+    {
+      "entropy": 1.2155711680650712,
+      "epoch": 1.2114285714285715,
+      "grad_norm": 2.337512731552124,
+      "learning_rate": 1.1931428571428571e-05,
+      "loss": 0.7488626003265381,
+      "mean_token_accuracy": 0.8063951514661312,
+      "num_tokens": 16158185.0,
+      "step": 1060
+    },
+    {
+      "entropy": 1.2575144089758397,
+      "epoch": 1.2228571428571429,
+      "grad_norm": 2.3885722160339355,
+      "learning_rate": 1.1855238095238094e-05,
+      "loss": 0.7211280345916748,
+      "mean_token_accuracy": 0.8121630363166332,
+      "num_tokens": 16316647.0,
+      "step": 1070
+    },
+    {
+      "entropy": 1.2669233828783035,
+      "epoch": 1.2342857142857142,
+      "grad_norm": 3.1025376319885254,
+      "learning_rate": 1.1779047619047621e-05,
+      "loss": 0.7715346813201904,
+      "mean_token_accuracy": 0.7969170480966568,
+      "num_tokens": 16471457.0,
+      "step": 1080
+    },
+    {
+      "entropy": 1.2340376809239388,
+      "epoch": 1.2457142857142858,
+      "grad_norm": 2.8538618087768555,
+      "learning_rate": 1.1702857142857145e-05,
+      "loss": 0.7577665328979493,
+      "mean_token_accuracy": 0.8027225121855736,
+      "num_tokens": 16631017.0,
+      "step": 1090
+    },
+    {
+      "entropy": 1.2692770406603813,
+      "epoch": 1.2571428571428571,
+      "grad_norm": 3.1801536083221436,
+      "learning_rate": 1.1626666666666668e-05,
+      "loss": 0.7788435935974121,
+      "mean_token_accuracy": 0.8010938227176666,
+      "num_tokens": 16786752.0,
+      "step": 1100
+    },
+    {
+      "entropy": 1.2486526250839234,
+      "epoch": 1.2685714285714287,
+      "grad_norm": 2.5821192264556885,
+      "learning_rate": 1.1550476190476192e-05,
+      "loss": 0.762976884841919,
+      "mean_token_accuracy": 0.8053238935768604,
+      "num_tokens": 16943519.0,
+      "step": 1110
+    },
+    {
+      "entropy": 1.3036663837730884,
+      "epoch": 1.28,
+      "grad_norm": 2.6454882621765137,
+      "learning_rate": 1.1474285714285715e-05,
+      "loss": 0.7800922393798828,
+      "mean_token_accuracy": 0.7931916259229184,
+      "num_tokens": 17098242.0,
+      "step": 1120
+    },
+    {
+      "entropy": 1.2358268782496453,
+      "epoch": 1.2914285714285714,
+      "grad_norm": 2.732434034347534,
+      "learning_rate": 1.1398095238095239e-05,
+      "loss": 0.7546168804168701,
+      "mean_token_accuracy": 0.8025222033262253,
+      "num_tokens": 17254880.0,
+      "step": 1130
+    },
+    {
+      "entropy": 1.2517999060451985,
+      "epoch": 1.302857142857143,
+      "grad_norm": 2.6394054889678955,
+      "learning_rate": 1.1321904761904762e-05,
+      "loss": 0.7450598239898681,
+      "mean_token_accuracy": 0.8091251827776432,
+      "num_tokens": 17404975.0,
+      "step": 1140
+    },
+    {
+      "entropy": 1.2538614131510257,
+      "epoch": 1.3142857142857143,
+      "grad_norm": 3.0480637550354004,
+      "learning_rate": 1.1245714285714286e-05,
+      "loss": 0.7634236335754394,
+      "mean_token_accuracy": 0.8049138620495796,
+      "num_tokens": 17557114.0,
+      "step": 1150
+    },
+    {
+      "entropy": 1.2000257663428784,
+      "epoch": 1.3257142857142856,
+      "grad_norm": 3.085055351257324,
+      "learning_rate": 1.1169523809523811e-05,
+      "loss": 0.7426016807556153,
+      "mean_token_accuracy": 0.8077944576740265,
+      "num_tokens": 17704128.0,
+      "step": 1160
+    },
+    {
+      "entropy": 1.2293061904609204,
+      "epoch": 1.3371428571428572,
+      "grad_norm": 2.3842477798461914,
+      "learning_rate": 1.1093333333333334e-05,
+      "loss": 0.7181476593017578,
+      "mean_token_accuracy": 0.8123824037611485,
+      "num_tokens": 17865916.0,
+      "step": 1170
+    },
+    {
+      "entropy": 1.2635096937417984,
+      "epoch": 1.3485714285714285,
+      "grad_norm": 2.483147382736206,
+      "learning_rate": 1.1017142857142858e-05,
+      "loss": 0.7818619251251221,
+      "mean_token_accuracy": 0.8041326075792312,
+      "num_tokens": 18023210.0,
+      "step": 1180
+    },
+    {
+      "entropy": 1.3030154168605805,
+      "epoch": 1.3599999999999999,
+      "grad_norm": 3.100837469100952,
+      "learning_rate": 1.0940952380952381e-05,
+      "loss": 0.7691655158996582,
+      "mean_token_accuracy": 0.8024266973137856,
+      "num_tokens": 18172271.0,
+      "step": 1190
+    },
+    {
+      "entropy": 1.2880703262984752,
+      "epoch": 1.3714285714285714,
+      "grad_norm": 3.068323850631714,
+      "learning_rate": 1.0864761904761905e-05,
+      "loss": 0.8048301696777344,
+      "mean_token_accuracy": 0.7959965549409389,
+      "num_tokens": 18325046.0,
+      "step": 1200
+    },
+    {
+      "entropy": 1.3046514227986337,
+      "epoch": 1.3828571428571428,
+      "grad_norm": 3.33621883392334,
+      "learning_rate": 1.078857142857143e-05,
+      "loss": 0.8266304016113282,
+      "mean_token_accuracy": 0.7895134054124355,
+      "num_tokens": 18476571.0,
+      "step": 1210
+    },
+    {
+      "entropy": 1.2605851158499717,
+      "epoch": 1.3942857142857144,
+      "grad_norm": 3.5944457054138184,
+      "learning_rate": 1.0712380952380954e-05,
+      "loss": 0.8084980010986328,
+      "mean_token_accuracy": 0.7934398144483567,
+      "num_tokens": 18622849.0,
+      "step": 1220
+    },
+    {
+      "entropy": 1.2391214780509472,
+      "epoch": 1.4057142857142857,
+      "grad_norm": 2.6554486751556396,
+      "learning_rate": 1.0636190476190477e-05,
+      "loss": 0.7550242900848388,
+      "mean_token_accuracy": 0.8065047845244407,
+      "num_tokens": 18768698.0,
+      "step": 1230
+    },
+    {
+      "entropy": 1.2420047149062157,
+      "epoch": 1.4171428571428573,
+      "grad_norm": 2.5860495567321777,
+      "learning_rate": 1.056e-05,
+      "loss": 0.7261144161224365,
+      "mean_token_accuracy": 0.8117271907627582,
+      "num_tokens": 18923547.0,
+      "step": 1240
+    },
+    {
+      "entropy": 1.246000237762928,
+      "epoch": 1.4285714285714286,
+      "grad_norm": 3.1218044757843018,
+      "learning_rate": 1.0483809523809526e-05,
+      "loss": 0.7216352939605712,
+      "mean_token_accuracy": 0.8120975896716118,
+      "num_tokens": 19079479.0,
+      "step": 1250
+    },
+    {
+      "entropy": 1.207762571424246,
+      "epoch": 1.44,
+      "grad_norm": 2.393723726272583,
+      "learning_rate": 1.040761904761905e-05,
+      "loss": 0.7272539138793945,
+      "mean_token_accuracy": 0.8084844700992108,
+      "num_tokens": 19236273.0,
+      "step": 1260
+    },
+    {
+      "entropy": 1.2830321080982685,
+      "epoch": 1.4514285714285715,
+      "grad_norm": 3.0418026447296143,
+      "learning_rate": 1.0331428571428573e-05,
+      "loss": 0.7889233589172363,
+      "mean_token_accuracy": 0.7998958893120289,
+      "num_tokens": 19385795.0,
+      "step": 1270
+    },
+    {
+      "entropy": 1.2588836170732975,
+      "epoch": 1.4628571428571429,
+      "grad_norm": 2.668825626373291,
+      "learning_rate": 1.0255238095238096e-05,
+      "loss": 0.7696958065032959,
+      "mean_token_accuracy": 0.8028367407619953,
+      "num_tokens": 19540370.0,
+      "step": 1280
+    },
+    {
+      "entropy": 1.2221255615353583,
+      "epoch": 1.4742857142857142,
+      "grad_norm": 3.121307611465454,
+      "learning_rate": 1.017904761904762e-05,
+      "loss": 0.7629749774932861,
+      "mean_token_accuracy": 0.8041565544903279,
+      "num_tokens": 19695452.0,
+      "step": 1290
+    },
+    {
+      "entropy": 1.2615755833685398,
+      "epoch": 1.4857142857142858,
+      "grad_norm": 2.600470781326294,
+      "learning_rate": 1.0102857142857143e-05,
+      "loss": 0.8250988960266114,
+      "mean_token_accuracy": 0.7851416207849979,
+      "num_tokens": 19850924.0,
+      "step": 1300
+    },
+    {
+      "entropy": 1.2793317660689354,
+      "epoch": 1.497142857142857,
+      "grad_norm": 2.874340295791626,
+      "learning_rate": 1.0026666666666667e-05,
+      "loss": 0.771949577331543,
+      "mean_token_accuracy": 0.8040384240448475,
+      "num_tokens": 20014514.0,
+      "step": 1310
+    },
+    {
+      "entropy": 1.1907848060131072,
+      "epoch": 1.5085714285714285,
+      "grad_norm": 2.6014015674591064,
+      "learning_rate": 9.950476190476192e-06,
+      "loss": 0.7447234630584717,
+      "mean_token_accuracy": 0.8111775033175945,
+      "num_tokens": 20173798.0,
+      "step": 1320
+    },
+    {
+      "entropy": 1.2612031117081641,
+      "epoch": 1.52,
+      "grad_norm": 2.809692144393921,
+      "learning_rate": 9.874285714285715e-06,
+      "loss": 0.7573285102844238,
+      "mean_token_accuracy": 0.8068460263311863,
+      "num_tokens": 20318447.0,
+      "step": 1330
+    },
+    {
+      "entropy": 1.2434154883027078,
+      "epoch": 1.5314285714285716,
+      "grad_norm": 2.9109153747558594,
+      "learning_rate": 9.798095238095239e-06,
+      "loss": 0.7508068561553956,
+      "mean_token_accuracy": 0.8049707144498826,
+      "num_tokens": 20472944.0,
+      "step": 1340
+    },
+    {
+      "entropy": 1.2488966889679431,
+      "epoch": 1.5428571428571427,
+      "grad_norm": 3.0063366889953613,
+      "learning_rate": 9.721904761904762e-06,
+      "loss": 0.7871552467346191,
+      "mean_token_accuracy": 0.8020347572863102,
+      "num_tokens": 20625826.0,
+      "step": 1350
+    },
+    {
+      "entropy": 1.2348267659544945,
+      "epoch": 1.5542857142857143,
+      "grad_norm": 2.792886734008789,
+      "learning_rate": 9.645714285714286e-06,
+      "loss": 0.7784818172454834,
+      "mean_token_accuracy": 0.8054304026067257,
+      "num_tokens": 20774171.0,
+      "step": 1360
+    },
+    {
+      "entropy": 1.2300232581794261,
+      "epoch": 1.5657142857142858,
+      "grad_norm": 2.861610174179077,
+      "learning_rate": 9.569523809523811e-06,
+      "loss": 0.7702634811401368,
+      "mean_token_accuracy": 0.8006386004388333,
+      "num_tokens": 20922487.0,
+      "step": 1370
+    },
+    {
+      "entropy": 1.24477871209383,
+      "epoch": 1.5771428571428572,
+      "grad_norm": 3.1448233127593994,
+      "learning_rate": 9.493333333333334e-06,
+      "loss": 0.7313426494598388,
+      "mean_token_accuracy": 0.812004954367876,
+      "num_tokens": 21075326.0,
+      "step": 1380
+    },
+    {
+      "entropy": 1.2703936398029327,
+      "epoch": 1.5885714285714285,
+      "grad_norm": 2.5799062252044678,
+      "learning_rate": 9.417142857142858e-06,
+      "loss": 0.7714487552642822,
+      "mean_token_accuracy": 0.799166913330555,
+      "num_tokens": 21224257.0,
+      "step": 1390
+    },
+    {
+      "entropy": 1.2416778221726417,
+      "epoch": 1.6,
+      "grad_norm": 2.992957353591919,
+      "learning_rate": 9.340952380952381e-06,
+      "loss": 0.7809468746185303,
+      "mean_token_accuracy": 0.7984777301549911,
+      "num_tokens": 21373259.0,
+      "step": 1400
+    },
+    {
+      "entropy": 1.2557578206062316,
+      "epoch": 1.6114285714285714,
+      "grad_norm": 2.7290945053100586,
+      "learning_rate": 9.264761904761905e-06,
+      "loss": 0.7919119834899903,
+      "mean_token_accuracy": 0.7954873479902744,
+      "num_tokens": 21522253.0,
+      "step": 1410
+    },
+    {
+      "entropy": 1.2285942271351815,
+      "epoch": 1.6228571428571428,
+      "grad_norm": 2.8919999599456787,
+      "learning_rate": 9.188571428571428e-06,
+      "loss": 0.7635447978973389,
+      "mean_token_accuracy": 0.8025862380862236,
+      "num_tokens": 21671374.0,
+      "step": 1420
+    },
+    {
+      "entropy": 1.3119461573660374,
+      "epoch": 1.6342857142857143,
+      "grad_norm": 2.7214455604553223,
+      "learning_rate": 9.112380952380954e-06,
+      "loss": 0.7926105976104736,
+      "mean_token_accuracy": 0.7944708272814751,
+      "num_tokens": 21822289.0,
+      "step": 1430
+    },
+    {
+      "entropy": 1.2220344461500645,
+      "epoch": 1.6457142857142857,
+      "grad_norm": 2.327855348587036,
+      "learning_rate": 9.036190476190477e-06,
+      "loss": 0.7583229064941406,
+      "mean_token_accuracy": 0.8038429766893387,
+      "num_tokens": 21978538.0,
+      "step": 1440
+    },
+    {
+      "entropy": 1.266616266965866,
+      "epoch": 1.657142857142857,
+      "grad_norm": 2.7889652252197266,
+      "learning_rate": 8.96e-06,
+      "loss": 0.7689779281616211,
+      "mean_token_accuracy": 0.8045293688774109,
+      "num_tokens": 22133231.0,
+      "step": 1450
+    },
+    {
+      "entropy": 1.3114601746201515,
+      "epoch": 1.6685714285714286,
+      "grad_norm": 3.2477846145629883,
+      "learning_rate": 8.883809523809524e-06,
+      "loss": 0.7963542461395263,
+      "mean_token_accuracy": 0.8005617402493954,
+      "num_tokens": 22280643.0,
+      "step": 1460
+    },
+    {
+      "entropy": 1.2778782725334168,
+      "epoch": 1.6800000000000002,
+      "grad_norm": 3.1621718406677246,
+      "learning_rate": 8.807619047619048e-06,
+      "loss": 0.801125431060791,
+      "mean_token_accuracy": 0.7964559637010098,
+      "num_tokens": 22430891.0,
+      "step": 1470
+    },
+    {
+      "entropy": 1.2413391053676606,
+      "epoch": 1.6914285714285713,
+      "grad_norm": 2.561835527420044,
+      "learning_rate": 8.731428571428571e-06,
+      "loss": 0.7200276374816894,
+      "mean_token_accuracy": 0.8137163199484349,
+      "num_tokens": 22581471.0,
+      "step": 1480
+    },
+    {
+      "entropy": 1.2620924778282643,
+      "epoch": 1.7028571428571428,
+      "grad_norm": 2.8986573219299316,
+      "learning_rate": 8.655238095238096e-06,
+      "loss": 0.7427529335021973,
+      "mean_token_accuracy": 0.803333692252636,
+      "num_tokens": 22732244.0,
+      "step": 1490
+    },
+    {
+      "entropy": 1.2637714959681035,
+      "epoch": 1.7142857142857144,
+      "grad_norm": 2.8469996452331543,
+      "learning_rate": 8.57904761904762e-06,
+      "loss": 0.782107925415039,
+      "mean_token_accuracy": 0.8046764589846134,
+      "num_tokens": 22884081.0,
+      "step": 1500
+    },
+    {
+      "entropy": 1.2661903858184815,
+      "epoch": 1.7257142857142858,
+      "grad_norm": 3.0864884853363037,
+      "learning_rate": 8.502857142857143e-06,
+      "loss": 0.7898892879486084,
+      "mean_token_accuracy": 0.7921097949147224,
+      "num_tokens": 23034449.0,
+      "step": 1510
+    },
+    {
+      "entropy": 1.250692894309759,
+      "epoch": 1.737142857142857,
+      "grad_norm": 2.6670966148376465,
+      "learning_rate": 8.426666666666667e-06,
+      "loss": 0.7688261508941651,
+      "mean_token_accuracy": 0.8037325166165828,
+      "num_tokens": 23188136.0,
+      "step": 1520
+    },
+    {
+      "entropy": 1.215890847891569,
+      "epoch": 1.7485714285714287,
+      "grad_norm": 2.3385775089263916,
+      "learning_rate": 8.35047619047619e-06,
+      "loss": 0.7499132633209229,
+      "mean_token_accuracy": 0.8074461244046688,
+      "num_tokens": 23343060.0,
+      "step": 1530
+    },
+    {
+      "entropy": 1.242823239415884,
+      "epoch": 1.76,
+      "grad_norm": 2.630499839782715,
+      "learning_rate": 8.274285714285715e-06,
+      "loss": 0.7520731925964356,
+      "mean_token_accuracy": 0.8076161026954651,
+      "num_tokens": 23487902.0,
+      "step": 1540
+    },
+    {
+      "entropy": 1.2385052397847176,
+      "epoch": 1.7714285714285714,
+      "grad_norm": 2.76127552986145,
+      "learning_rate": 8.198095238095239e-06,
+      "loss": 0.7834786415100098,
+      "mean_token_accuracy": 0.7965474635362625,
+      "num_tokens": 23635117.0,
+      "step": 1550
+    },
+    {
+      "entropy": 1.1929515838623046,
+      "epoch": 1.782857142857143,
+      "grad_norm": 2.612879991531372,
+      "learning_rate": 8.121904761904762e-06,
+      "loss": 0.7256328582763671,
+      "mean_token_accuracy": 0.8150137662887573,
+      "num_tokens": 23795433.0,
+      "step": 1560
+    },
+    {
+      "entropy": 1.194692898541689,
+      "epoch": 1.7942857142857143,
+      "grad_norm": 2.6522936820983887,
+      "learning_rate": 8.045714285714286e-06,
+      "loss": 0.7276816368103027,
+      "mean_token_accuracy": 0.8094555623829365,
+      "num_tokens": 23950370.0,
+      "step": 1570
+    },
+    {
+      "entropy": 1.2585241250693797,
+      "epoch": 1.8057142857142856,
+      "grad_norm": 2.611926794052124,
+      "learning_rate": 7.969523809523811e-06,
+      "loss": 0.7796142578125,
+      "mean_token_accuracy": 0.8012560330331325,
+      "num_tokens": 24103914.0,
+      "step": 1580
+    },
+    {
+      "entropy": 1.2676213465631008,
+      "epoch": 1.8171428571428572,
+      "grad_norm": 2.5129570960998535,
+      "learning_rate": 7.893333333333335e-06,
+      "loss": 0.7887630462646484,
+      "mean_token_accuracy": 0.7965363718569278,
+      "num_tokens": 24258081.0,
+      "step": 1590
+    },
+    {
+      "entropy": 1.2579274140298367,
+      "epoch": 1.8285714285714287,
+      "grad_norm": 2.86506724357605,
+      "learning_rate": 7.817142857142858e-06,
+      "loss": 0.7497401714324952,
+      "mean_token_accuracy": 0.8082531146705151,
+      "num_tokens": 24409168.0,
+      "step": 1600
+    },
+    {
+      "entropy": 1.2318767301738263,
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.7256534099578857,
+      "learning_rate": 7.740952380952382e-06,
+      "loss": 0.7274055480957031,
+      "mean_token_accuracy": 0.8118016675114632,
+      "num_tokens": 24566252.0,
+      "step": 1610
+    },
+    {
+      "entropy": 1.2056004114449024,
+      "epoch": 1.8514285714285714,
+      "grad_norm": 2.6447949409484863,
+      "learning_rate": 7.664761904761905e-06,
+      "loss": 0.7131245613098145,
+      "mean_token_accuracy": 0.8102400578558445,
+      "num_tokens": 24721993.0,
+      "step": 1620
+    },
+    {
+      "entropy": 1.245617663115263,
+      "epoch": 1.862857142857143,
+      "grad_norm": 2.835374116897583,
+      "learning_rate": 7.588571428571429e-06,
+      "loss": 0.783673095703125,
+      "mean_token_accuracy": 0.8022521927952766,
+      "num_tokens": 24868978.0,
+      "step": 1630
+    },
+    {
+      "entropy": 1.240978506952524,
+      "epoch": 1.8742857142857143,
+      "grad_norm": 2.6456923484802246,
+      "learning_rate": 7.512380952380953e-06,
+      "loss": 0.7802390098571778,
+      "mean_token_accuracy": 0.7991835795342922,
+      "num_tokens": 25026573.0,
+      "step": 1640
+    },
+    {
+      "entropy": 1.2280743092298507,
+      "epoch": 1.8857142857142857,
+      "grad_norm": 2.847167491912842,
+      "learning_rate": 7.436190476190477e-06,
+      "loss": 0.8147419929504395,
+      "mean_token_accuracy": 0.7919134043157101,
+      "num_tokens": 25182791.0,
+      "step": 1650
+    },
+    {
+      "entropy": 1.216932328045368,
+      "epoch": 1.8971428571428572,
+      "grad_norm": 2.861248731613159,
+      "learning_rate": 7.360000000000001e-06,
+      "loss": 0.7451428890228271,
+      "mean_token_accuracy": 0.8088080659508705,
+      "num_tokens": 25326320.0,
+      "step": 1660
+    },
+    {
+      "entropy": 1.2367267780005933,
+      "epoch": 1.9085714285714286,
+      "grad_norm": 2.702206611633301,
+      "learning_rate": 7.283809523809524e-06,
+      "loss": 0.7687783241271973,
+      "mean_token_accuracy": 0.8046720631420612,
+      "num_tokens": 25481482.0,
+      "step": 1670
+    },
+    {
+      "entropy": 1.2137604996562004,
+      "epoch": 1.92,
+      "grad_norm": 2.446934938430786,
+      "learning_rate": 7.207619047619048e-06,
+      "loss": 0.7720860958099365,
+      "mean_token_accuracy": 0.8001258425414562,
+      "num_tokens": 25627745.0,
+      "step": 1680
+    },
+    {
+      "entropy": 1.2351884521543979,
+      "epoch": 1.9314285714285715,
+      "grad_norm": 2.682955026626587,
+      "learning_rate": 7.131428571428573e-06,
+      "loss": 0.7554657936096192,
+      "mean_token_accuracy": 0.8067528575658798,
+      "num_tokens": 25778484.0,
+      "step": 1690
+    },
+    {
+      "entropy": 1.2149645321071147,
+      "epoch": 1.9428571428571428,
+      "grad_norm": 3.032944679260254,
+      "learning_rate": 7.055238095238096e-06,
+      "loss": 0.6940214157104492,
+      "mean_token_accuracy": 0.8167023897171021,
+      "num_tokens": 25932265.0,
+      "step": 1700
+    },
+    {
+      "entropy": 1.205052537471056,
+      "epoch": 1.9542857142857142,
+      "grad_norm": 2.7198259830474854,
+      "learning_rate": 6.97904761904762e-06,
+      "loss": 0.7255962371826172,
+      "mean_token_accuracy": 0.8064791038632393,
+      "num_tokens": 26077725.0,
+      "step": 1710
+    },
+    {
+      "entropy": 1.1943283297121525,
+      "epoch": 1.9657142857142857,
+      "grad_norm": 2.6816186904907227,
+      "learning_rate": 6.902857142857143e-06,
+      "loss": 0.7239448547363281,
+      "mean_token_accuracy": 0.809258659183979,
+      "num_tokens": 26233086.0,
+      "step": 1720
+    },
+    {
+      "entropy": 1.2314366959035397,
+      "epoch": 1.977142857142857,
+      "grad_norm": 3.0669050216674805,
+      "learning_rate": 6.826666666666667e-06,
+      "loss": 0.8160985946655274,
+      "mean_token_accuracy": 0.7934448584914208,
+      "num_tokens": 26383494.0,
+      "step": 1730
+    },
+    {
+      "entropy": 1.201389105618,
+      "epoch": 1.9885714285714284,
+      "grad_norm": 2.6934287548065186,
+      "learning_rate": 6.75047619047619e-06,
+      "loss": 0.7483705043792724,
+      "mean_token_accuracy": 0.8099792324006557,
+      "num_tokens": 26534600.0,
+      "step": 1740
+    },
+    {
+      "entropy": 1.2289574392139913,
+      "epoch": 2.0,
+      "grad_norm": 2.9322891235351562,
+      "learning_rate": 6.6742857142857155e-06,
+      "loss": 0.7264975070953369,
+      "mean_token_accuracy": 0.8107106305658818,
+      "num_tokens": 26690084.0,
+      "step": 1750
+    },
+    {
+      "entropy": 1.129621234536171,
+      "epoch": 2.0114285714285716,
+      "grad_norm": 2.3608248233795166,
+      "learning_rate": 6.598095238095239e-06,
+      "loss": 0.4672244071960449,
+      "mean_token_accuracy": 0.8794825159013271,
+      "num_tokens": 26855673.0,
+      "step": 1760
+    },
+    {
+      "entropy": 1.0073387667536735,
+      "epoch": 2.0228571428571427,
+      "grad_norm": 2.5772664546966553,
+      "learning_rate": 6.5219047619047625e-06,
+      "loss": 0.42911324501037595,
+      "mean_token_accuracy": 0.8885554544627666,
+      "num_tokens": 27007051.0,
+      "step": 1770
+    },
+    {
+      "entropy": 0.937215369194746,
+      "epoch": 2.0342857142857143,
+      "grad_norm": 2.5678927898406982,
+      "learning_rate": 6.445714285714286e-06,
+      "loss": 0.43928914070129393,
+      "mean_token_accuracy": 0.8838384367525578,
+      "num_tokens": 27160139.0,
+      "step": 1780
+    },
+    {
+      "entropy": 1.0134622648358345,
+      "epoch": 2.045714285714286,
+      "grad_norm": 2.933607816696167,
+      "learning_rate": 6.36952380952381e-06,
+      "loss": 0.4702662467956543,
+      "mean_token_accuracy": 0.8761703535914421,
+      "num_tokens": 27313243.0,
+      "step": 1790
+    },
+    {
+      "entropy": 1.0408321045339108,
+      "epoch": 2.057142857142857,
+      "grad_norm": 2.5859014987945557,
+      "learning_rate": 6.293333333333334e-06,
+      "loss": 0.4461977958679199,
+      "mean_token_accuracy": 0.8826056070625782,
+      "num_tokens": 27464839.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.9933250166475773,
+      "epoch": 2.0685714285714285,
+      "grad_norm": 2.709477424621582,
+      "learning_rate": 6.217142857142857e-06,
+      "loss": 0.4566448211669922,
+      "mean_token_accuracy": 0.8768111765384674,
+      "num_tokens": 27623099.0,
+      "step": 1810
+    },
+    {
+      "entropy": 1.0335271701216697,
+      "epoch": 2.08,
+      "grad_norm": 2.594367027282715,
+      "learning_rate": 6.140952380952382e-06,
+      "loss": 0.5224481582641601,
+      "mean_token_accuracy": 0.8632099166512489,
+      "num_tokens": 27769833.0,
+      "step": 1820
+    },
+    {
+      "entropy": 0.9761688143014908,
+      "epoch": 2.0914285714285716,
+      "grad_norm": 2.7078566551208496,
+      "learning_rate": 6.064761904761905e-06,
+      "loss": 0.45804972648620607,
+      "mean_token_accuracy": 0.879676353931427,
+      "num_tokens": 27920400.0,
+      "step": 1830
+    },
+    {
+      "entropy": 0.9983637891709805,
+      "epoch": 2.1028571428571428,
+      "grad_norm": 2.7597780227661133,
+      "learning_rate": 5.9885714285714294e-06,
+      "loss": 0.4517871856689453,
+      "mean_token_accuracy": 0.8815616481006145,
+      "num_tokens": 28072384.0,
+      "step": 1840
+    },
+    {
+      "entropy": 1.0141542233526706,
+      "epoch": 2.1142857142857143,
+      "grad_norm": 3.0741305351257324,
+      "learning_rate": 5.912380952380953e-06,
+      "loss": 0.5087273597717286,
+      "mean_token_accuracy": 0.8672586888074875,
+      "num_tokens": 28227810.0,
+      "step": 1850
+    },
+    {
+      "entropy": 0.9573422007262706,
+      "epoch": 2.125714285714286,
+      "grad_norm": 2.449988842010498,
+      "learning_rate": 5.836190476190476e-06,
+      "loss": 0.391282844543457,
+      "mean_token_accuracy": 0.8947287388145924,
+      "num_tokens": 28385046.0,
+      "step": 1860
+    },
+    {
+      "entropy": 0.9876036189496518,
+      "epoch": 2.137142857142857,
+      "grad_norm": 3.0042693614959717,
+      "learning_rate": 5.76e-06,
+      "loss": 0.43504929542541504,
+      "mean_token_accuracy": 0.8803216218948364,
+      "num_tokens": 28537279.0,
+      "step": 1870
+    },
+    {
+      "entropy": 1.0152524262666702,
+      "epoch": 2.1485714285714286,
+      "grad_norm": 2.8539533615112305,
+      "learning_rate": 5.683809523809525e-06,
+      "loss": 0.46044297218322755,
+      "mean_token_accuracy": 0.8784566923975945,
+      "num_tokens": 28682717.0,
+      "step": 1880
+    },
+    {
+      "entropy": 0.9989143267273903,
+      "epoch": 2.16,
+      "grad_norm": 2.774113893508911,
+      "learning_rate": 5.607619047619049e-06,
+      "loss": 0.45467476844787597,
+      "mean_token_accuracy": 0.8795260697603225,
+      "num_tokens": 28837051.0,
+      "step": 1890
+    },
+    {
+      "entropy": 0.9938028298318387,
+      "epoch": 2.1714285714285713,
+      "grad_norm": 2.7854819297790527,
+      "learning_rate": 5.531428571428572e-06,
+      "loss": 0.44997029304504393,
+      "mean_token_accuracy": 0.8821788050234318,
+      "num_tokens": 28984327.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.9829576320946216,
+      "epoch": 2.182857142857143,
+      "grad_norm": 2.6420528888702393,
+      "learning_rate": 5.4552380952380956e-06,
+      "loss": 0.485393762588501,
+      "mean_token_accuracy": 0.8722478076815605,
+      "num_tokens": 29137880.0,
+      "step": 1910
+    },
+    {
+      "entropy": 0.9889856785535812,
+      "epoch": 2.1942857142857144,
+      "grad_norm": 2.4197490215301514,
+      "learning_rate": 5.379047619047619e-06,
+      "loss": 0.4325558662414551,
+      "mean_token_accuracy": 0.8920911446213722,
+      "num_tokens": 29286434.0,
+      "step": 1920
+    },
+    {
+      "entropy": 0.945246460288763,
+      "epoch": 2.2057142857142855,
+      "grad_norm": 2.2842671871185303,
+      "learning_rate": 5.3028571428571425e-06,
+      "loss": 0.4346343994140625,
+      "mean_token_accuracy": 0.8836146749556064,
+      "num_tokens": 29441788.0,
+      "step": 1930
+    },
+    {
+      "entropy": 0.9999402716755867,
+      "epoch": 2.217142857142857,
+      "grad_norm": 3.09014892578125,
+      "learning_rate": 5.226666666666667e-06,
+      "loss": 0.455077600479126,
+      "mean_token_accuracy": 0.8785418413579464,
+      "num_tokens": 29590827.0,
+      "step": 1940
+    },
+    {
+      "entropy": 0.9953154794871807,
+      "epoch": 2.2285714285714286,
+      "grad_norm": 2.9954278469085693,
+      "learning_rate": 5.150476190476191e-06,
+      "loss": 0.4847094058990479,
+      "mean_token_accuracy": 0.8742888525128365,
+      "num_tokens": 29737685.0,
+      "step": 1950
+    },
+    {
+      "entropy": 0.9929163068532944,
+      "epoch": 2.24,
+      "grad_norm": 3.5582692623138428,
+      "learning_rate": 5.074285714285715e-06,
+      "loss": 0.4716338634490967,
+      "mean_token_accuracy": 0.874817679822445,
+      "num_tokens": 29888760.0,
+      "step": 1960
+    },
+    {
+      "entropy": 0.9909138068556785,
+      "epoch": 2.2514285714285713,
+      "grad_norm": 2.4286201000213623,
+      "learning_rate": 4.998095238095238e-06,
+      "loss": 0.46494855880737307,
+      "mean_token_accuracy": 0.8753256112337112,
+      "num_tokens": 30045315.0,
+      "step": 1970
+    },
+    {
+      "entropy": 0.9897382281720638,
+      "epoch": 2.262857142857143,
+      "grad_norm": 3.2968590259552,
+      "learning_rate": 4.9219047619047625e-06,
+      "loss": 0.4664334774017334,
+      "mean_token_accuracy": 0.8777646087110043,
+      "num_tokens": 30202152.0,
+      "step": 1980
+    },
+    {
+      "entropy": 0.963070584833622,
+      "epoch": 2.2742857142857145,
+      "grad_norm": 2.7921502590179443,
+      "learning_rate": 4.845714285714286e-06,
+      "loss": 0.46282148361206055,
+      "mean_token_accuracy": 0.8788235448300838,
+      "num_tokens": 30360498.0,
+      "step": 1990
+    },
+    {
+      "entropy": 0.9640379913151265,
+      "epoch": 2.2857142857142856,
+      "grad_norm": 2.8175981044769287,
+      "learning_rate": 4.76952380952381e-06,
+      "loss": 0.4309398651123047,
+      "mean_token_accuracy": 0.8862293303012848,
+      "num_tokens": 30507865.0,
+      "step": 2000
+    },
+    {
+      "entropy": 0.9654312752187252,
+      "epoch": 2.297142857142857,
+      "grad_norm": 2.875277042388916,
+      "learning_rate": 4.693333333333334e-06,
+      "loss": 0.4348562240600586,
+      "mean_token_accuracy": 0.880731088668108,
+      "num_tokens": 30659079.0,
+      "step": 2010
+    },
+    {
+      "entropy": 0.9790816180408001,
+      "epoch": 2.3085714285714287,
+      "grad_norm": 2.7534308433532715,
+      "learning_rate": 4.617142857142857e-06,
+      "loss": 0.4656937599182129,
+      "mean_token_accuracy": 0.881525882333517,
+      "num_tokens": 30811801.0,
+      "step": 2020
+    },
+    {
+      "entropy": 0.981222202628851,
+      "epoch": 2.32,
+      "grad_norm": 2.6787819862365723,
+      "learning_rate": 4.540952380952382e-06,
+      "loss": 0.42824320793151854,
+      "mean_token_accuracy": 0.887121319025755,
+      "num_tokens": 30956430.0,
+      "step": 2030
+    },
+    {
+      "entropy": 0.983622444421053,
+      "epoch": 2.3314285714285714,
+      "grad_norm": 2.8637819290161133,
+      "learning_rate": 4.464761904761905e-06,
+      "loss": 0.46105365753173827,
+      "mean_token_accuracy": 0.8789493426680565,
+      "num_tokens": 31106532.0,
+      "step": 2040
+    },
+    {
+      "entropy": 1.004967824369669,
+      "epoch": 2.342857142857143,
+      "grad_norm": 2.9350128173828125,
+      "learning_rate": 4.388571428571429e-06,
+      "loss": 0.4636978626251221,
+      "mean_token_accuracy": 0.8800117067992688,
+      "num_tokens": 31254450.0,
+      "step": 2050
+    },
+    {
+      "entropy": 0.9935823157429695,
+      "epoch": 2.354285714285714,
+      "grad_norm": 2.8673224449157715,
+      "learning_rate": 4.312380952380953e-06,
+      "loss": 0.46178450584411623,
+      "mean_token_accuracy": 0.87845404073596,
+      "num_tokens": 31398519.0,
+      "step": 2060
+    },
+    {
+      "entropy": 1.0041532143950462,
+      "epoch": 2.3657142857142857,
+      "grad_norm": 2.8013110160827637,
+      "learning_rate": 4.2361904761904765e-06,
+      "loss": 0.46661930084228515,
+      "mean_token_accuracy": 0.8790879227221012,
+      "num_tokens": 31547233.0,
+      "step": 2070
+    },
+    {
+      "entropy": 0.9613068968057632,
+      "epoch": 2.3771428571428572,
+      "grad_norm": 2.733070135116577,
+      "learning_rate": 4.16e-06,
+      "loss": 0.4579169750213623,
+      "mean_token_accuracy": 0.878313148766756,
+      "num_tokens": 31701773.0,
+      "step": 2080
+    },
+    {
+      "entropy": 0.9756359688937664,
+      "epoch": 2.388571428571429,
+      "grad_norm": 3.2284717559814453,
+      "learning_rate": 4.083809523809524e-06,
+      "loss": 0.4302100658416748,
+      "mean_token_accuracy": 0.8845745049417019,
+      "num_tokens": 31861253.0,
+      "step": 2090
+    },
+    {
+      "entropy": 0.9597941651940346,
+      "epoch": 2.4,
+      "grad_norm": 2.771367073059082,
+      "learning_rate": 4.007619047619048e-06,
+      "loss": 0.4371436595916748,
+      "mean_token_accuracy": 0.8834490239620209,
+      "num_tokens": 32008877.0,
+      "step": 2100
+    },
+    {
+      "entropy": 0.993123146891594,
+      "epoch": 2.4114285714285715,
+      "grad_norm": 2.781940221786499,
+      "learning_rate": 3.931428571428571e-06,
+      "loss": 0.4647090911865234,
+      "mean_token_accuracy": 0.8769076429307461,
+      "num_tokens": 32161713.0,
+      "step": 2110
+    },
+    {
+      "entropy": 0.9968604199588299,
+      "epoch": 2.422857142857143,
+      "grad_norm": 3.368783473968506,
+      "learning_rate": 3.855238095238096e-06,
+      "loss": 0.4818123817443848,
+      "mean_token_accuracy": 0.8751231759786606,
+      "num_tokens": 32313621.0,
+      "step": 2120
+    },
+    {
+      "entropy": 0.9834559701383114,
+      "epoch": 2.434285714285714,
+      "grad_norm": 3.1263699531555176,
+      "learning_rate": 3.7790476190476196e-06,
+      "loss": 0.44535102844238283,
+      "mean_token_accuracy": 0.8850508667528629,
+      "num_tokens": 32467792.0,
+      "step": 2130
+    },
+    {
+      "entropy": 0.9643837459385395,
+      "epoch": 2.4457142857142857,
+      "grad_norm": 2.7664949893951416,
+      "learning_rate": 3.702857142857143e-06,
+      "loss": 0.4365830421447754,
+      "mean_token_accuracy": 0.8840666212141514,
+      "num_tokens": 32619635.0,
+      "step": 2140
+    },
+    {
+      "entropy": 0.9677878350019455,
+      "epoch": 2.4571428571428573,
+      "grad_norm": 3.264744758605957,
+      "learning_rate": 3.6266666666666674e-06,
+      "loss": 0.44498529434204104,
+      "mean_token_accuracy": 0.8833608441054821,
+      "num_tokens": 32767373.0,
+      "step": 2150
+    },
+    {
+      "entropy": 0.995635886490345,
+      "epoch": 2.4685714285714284,
+      "grad_norm": 3.2273788452148438,
+      "learning_rate": 3.550476190476191e-06,
+      "loss": 0.5052412033081055,
+      "mean_token_accuracy": 0.8676992803812027,
+      "num_tokens": 32911902.0,
+      "step": 2160
+    },
+    {
+      "entropy": 0.9879037588834763,
+      "epoch": 2.48,
+      "grad_norm": 3.09259295463562,
+      "learning_rate": 3.4742857142857144e-06,
+      "loss": 0.4411734104156494,
+      "mean_token_accuracy": 0.8820772051811219,
+      "num_tokens": 33064263.0,
+      "step": 2170
+    },
+    {
+      "entropy": 1.0008388698101043,
+      "epoch": 2.4914285714285715,
+      "grad_norm": 2.9296276569366455,
+      "learning_rate": 3.3980952380952383e-06,
+      "loss": 0.4818171501159668,
+      "mean_token_accuracy": 0.8750081188976765,
+      "num_tokens": 33217206.0,
+      "step": 2180
+    },
+    {
+      "entropy": 0.9906027033925057,
+      "epoch": 2.5028571428571427,
+      "grad_norm": 2.90834641456604,
+      "learning_rate": 3.321904761904762e-06,
+      "loss": 0.4743064880371094,
+      "mean_token_accuracy": 0.8746793515980243,
+      "num_tokens": 33370479.0,
+      "step": 2190
+    },
+    {
+      "entropy": 0.9570474222302436,
+      "epoch": 2.5142857142857142,
+      "grad_norm": 2.379714250564575,
+      "learning_rate": 3.245714285714286e-06,
+      "loss": 0.4752392292022705,
+      "mean_token_accuracy": 0.8746638618409633,
+      "num_tokens": 33528101.0,
+      "step": 2200
+    },
+    {
+      "entropy": 0.9379908077418804,
+      "epoch": 2.525714285714286,
+      "grad_norm": 2.6806116104125977,
+      "learning_rate": 3.1695238095238096e-06,
+      "loss": 0.4428661823272705,
+      "mean_token_accuracy": 0.882461816072464,
+      "num_tokens": 33688951.0,
+      "step": 2210
+    },
+    {
+      "entropy": 0.9611021012067795,
+      "epoch": 2.5371428571428574,
+      "grad_norm": 3.014352798461914,
+      "learning_rate": 3.093333333333334e-06,
+      "loss": 0.4266517162322998,
+      "mean_token_accuracy": 0.8860412418842316,
+      "num_tokens": 33845955.0,
+      "step": 2220
+    },
+    {
+      "entropy": 0.9856173396110535,
+      "epoch": 2.5485714285714285,
+      "grad_norm": 2.736889123916626,
+      "learning_rate": 3.0171428571428574e-06,
+      "loss": 0.49875760078430176,
+      "mean_token_accuracy": 0.8690919436514377,
+      "num_tokens": 34000742.0,
+      "step": 2230
+    },
+    {
+      "entropy": 0.9622156754136085,
+      "epoch": 2.56,
+      "grad_norm": 2.51438045501709,
+      "learning_rate": 2.940952380952381e-06,
+      "loss": 0.4273517608642578,
+      "mean_token_accuracy": 0.8868567258119583,
+      "num_tokens": 34152502.0,
+      "step": 2240
+    },
+    {
+      "entropy": 0.9562038011848927,
+      "epoch": 2.571428571428571,
+      "grad_norm": 2.7017078399658203,
+      "learning_rate": 2.8647619047619052e-06,
+      "loss": 0.46968560218811034,
+      "mean_token_accuracy": 0.8755630798637867,
+      "num_tokens": 34306993.0,
+      "step": 2250
+    },
+    {
+      "entropy": 0.9738695487380028,
+      "epoch": 2.5828571428571427,
+      "grad_norm": 2.838038206100464,
+      "learning_rate": 2.7885714285714287e-06,
+      "loss": 0.459547233581543,
+      "mean_token_accuracy": 0.8819342166185379,
+      "num_tokens": 34456450.0,
+      "step": 2260
+    },
+    {
+      "entropy": 0.9687829136848449,
+      "epoch": 2.5942857142857143,
+      "grad_norm": 2.519185781478882,
+      "learning_rate": 2.7123809523809526e-06,
+      "loss": 0.4452920913696289,
+      "mean_token_accuracy": 0.8850650191307068,
+      "num_tokens": 34607794.0,
+      "step": 2270
+    },
+    {
+      "entropy": 0.975687575340271,
+      "epoch": 2.605714285714286,
+      "grad_norm": 3.240191698074341,
+      "learning_rate": 2.6361904761904766e-06,
+      "loss": 0.43484649658203123,
+      "mean_token_accuracy": 0.8862838499248028,
+      "num_tokens": 34754724.0,
+      "step": 2280
+    },
+    {
+      "entropy": 0.9368025757372379,
+      "epoch": 2.617142857142857,
+      "grad_norm": 2.922153949737549,
+      "learning_rate": 2.56e-06,
+      "loss": 0.44843349456787107,
+      "mean_token_accuracy": 0.8823426507413388,
+      "num_tokens": 34920148.0,
+      "step": 2290
+    },
+    {
+      "entropy": 0.9634293958544731,
+      "epoch": 2.6285714285714286,
+      "grad_norm": 2.598139524459839,
+      "learning_rate": 2.483809523809524e-06,
+      "loss": 0.422210693359375,
+      "mean_token_accuracy": 0.8881481572985649,
+      "num_tokens": 35074995.0,
+      "step": 2300
+    },
+    {
+      "entropy": 0.9806410282850265,
+      "epoch": 2.64,
+      "grad_norm": 2.67830753326416,
+      "learning_rate": 2.407619047619048e-06,
+      "loss": 0.43632044792175295,
+      "mean_token_accuracy": 0.8867660120129586,
+      "num_tokens": 35231660.0,
+      "step": 2310
+    },
+    {
+      "entropy": 1.0124660462141037,
+      "epoch": 2.6514285714285712,
+      "grad_norm": 2.6864585876464844,
+      "learning_rate": 2.331428571428572e-06,
+      "loss": 0.4657135486602783,
+      "mean_token_accuracy": 0.877445038408041,
+      "num_tokens": 35380814.0,
+      "step": 2320
+    },
+    {
+      "entropy": 0.9958271220326423,
+      "epoch": 2.662857142857143,
+      "grad_norm": 2.647252321243286,
+      "learning_rate": 2.2552380952380953e-06,
+      "loss": 0.48684301376342776,
+      "mean_token_accuracy": 0.8722301825881005,
+      "num_tokens": 35536710.0,
+      "step": 2330
+    },
+    {
+      "entropy": 1.00919935926795,
+      "epoch": 2.6742857142857144,
+      "grad_norm": 2.705885648727417,
+      "learning_rate": 2.179047619047619e-06,
+      "loss": 0.4847156047821045,
+      "mean_token_accuracy": 0.8721759475767612,
+      "num_tokens": 35688943.0,
+      "step": 2340
+    },
+    {
+      "entropy": 1.0191727139055728,
+      "epoch": 2.685714285714286,
+      "grad_norm": 3.004671812057495,
+      "learning_rate": 2.102857142857143e-06,
+      "loss": 0.47832179069519043,
+      "mean_token_accuracy": 0.8725899174809456,
+      "num_tokens": 35843212.0,
+      "step": 2350
+    },
+    {
+      "entropy": 0.9842646040022374,
+      "epoch": 2.697142857142857,
+      "grad_norm": 2.627149820327759,
+      "learning_rate": 2.0266666666666666e-06,
+      "loss": 0.4353066921234131,
+      "mean_token_accuracy": 0.8914880864322186,
+      "num_tokens": 35993592.0,
+      "step": 2360
+    },
+    {
+      "entropy": 0.9984226897358894,
+      "epoch": 2.7085714285714286,
+      "grad_norm": 3.0235273838043213,
+      "learning_rate": 1.9504761904761905e-06,
+      "loss": 0.4828640937805176,
+      "mean_token_accuracy": 0.8770869389176369,
+      "num_tokens": 36146237.0,
+      "step": 2370
+    },
+    {
+      "entropy": 1.0128906913101674,
+      "epoch": 2.7199999999999998,
+      "grad_norm": 3.3019280433654785,
+      "learning_rate": 1.8742857142857142e-06,
+      "loss": 0.4734531879425049,
+      "mean_token_accuracy": 0.8784437663853168,
+      "num_tokens": 36293821.0,
+      "step": 2380
+    },
+    {
+      "entropy": 1.0092256397008896,
+      "epoch": 2.7314285714285713,
+      "grad_norm": 2.9352781772613525,
+      "learning_rate": 1.7980952380952381e-06,
+      "loss": 0.4679893970489502,
+      "mean_token_accuracy": 0.8755185902118683,
+      "num_tokens": 36443585.0,
+      "step": 2390
+    },
+    {
+      "entropy": 0.979985772818327,
+      "epoch": 2.742857142857143,
+      "grad_norm": 2.9258673191070557,
+      "learning_rate": 1.721904761904762e-06,
+      "loss": 0.4572472095489502,
+      "mean_token_accuracy": 0.8824664078652859,
+      "num_tokens": 36597011.0,
+      "step": 2400
+    },
+    {
+      "entropy": 0.9609309315681458,
+      "epoch": 2.7542857142857144,
+      "grad_norm": 2.8819053173065186,
+      "learning_rate": 1.6457142857142857e-06,
+      "loss": 0.4551131248474121,
+      "mean_token_accuracy": 0.8774472214281559,
+      "num_tokens": 36759877.0,
+      "step": 2410
+    },
+    {
+      "entropy": 0.9909630537033081,
+      "epoch": 2.7657142857142856,
+      "grad_norm": 3.1969611644744873,
+      "learning_rate": 1.5695238095238097e-06,
+      "loss": 0.47752718925476073,
+      "mean_token_accuracy": 0.8757083237171173,
+      "num_tokens": 36911703.0,
+      "step": 2420
+    },
+    {
+      "entropy": 0.9766927987337113,
+      "epoch": 2.777142857142857,
+      "grad_norm": 2.8341927528381348,
+      "learning_rate": 1.4933333333333336e-06,
+      "loss": 0.4443829536437988,
+      "mean_token_accuracy": 0.8800640240311622,
+      "num_tokens": 37063394.0,
+      "step": 2430
+    },
+    {
+      "entropy": 0.9574880845844745,
+      "epoch": 2.7885714285714287,
+      "grad_norm": 2.639431953430176,
+      "learning_rate": 1.4171428571428573e-06,
+      "loss": 0.43013510704040525,
+      "mean_token_accuracy": 0.8883027389645577,
+      "num_tokens": 37212839.0,
+      "step": 2440
+    },
+    {
+      "entropy": 0.9979249402880669,
+      "epoch": 2.8,
+      "grad_norm": 2.869964599609375,
+      "learning_rate": 1.3409523809523812e-06,
+      "loss": 0.44501733779907227,
+      "mean_token_accuracy": 0.8818599261343479,
+      "num_tokens": 37365373.0,
+      "step": 2450
+    },
+    {
+      "entropy": 0.988174206763506,
+      "epoch": 2.8114285714285714,
+      "grad_norm": 2.750286102294922,
+      "learning_rate": 1.2647619047619047e-06,
+      "loss": 0.5042055606842041,
+      "mean_token_accuracy": 0.8709421746432782,
+      "num_tokens": 37516024.0,
+      "step": 2460
+    },
+    {
+      "entropy": 0.9685959227383136,
+      "epoch": 2.822857142857143,
+      "grad_norm": 2.831801652908325,
+      "learning_rate": 1.1885714285714286e-06,
+      "loss": 0.4595749855041504,
+      "mean_token_accuracy": 0.8802243292331695,
+      "num_tokens": 37672800.0,
+      "step": 2470
+    },
+    {
+      "entropy": 0.9566099487245083,
+      "epoch": 2.8342857142857145,
+      "grad_norm": 2.592846632003784,
+      "learning_rate": 1.1123809523809525e-06,
+      "loss": 0.45911993980407717,
+      "mean_token_accuracy": 0.8795291222631931,
+      "num_tokens": 37829671.0,
+      "step": 2480
+    },
+    {
+      "entropy": 0.9668984241783619,
+      "epoch": 2.8457142857142856,
+      "grad_norm": 2.636902332305908,
+      "learning_rate": 1.0361904761904764e-06,
+      "loss": 0.4479802131652832,
+      "mean_token_accuracy": 0.884428545087576,
+      "num_tokens": 37982539.0,
+      "step": 2490
+    },
+    {
+      "entropy": 0.9993820540606976,
+      "epoch": 2.857142857142857,
+      "grad_norm": 2.9457950592041016,
+      "learning_rate": 9.600000000000001e-07,
+      "loss": 0.43321714401245115,
+      "mean_token_accuracy": 0.8866497747600078,
+      "num_tokens": 38127739.0,
+      "step": 2500
+    },
+    {
+      "entropy": 0.9669802196323871,
+      "epoch": 2.8685714285714283,
+      "grad_norm": 2.956597328186035,
+      "learning_rate": 8.838095238095238e-07,
+      "loss": 0.46829757690429685,
+      "mean_token_accuracy": 0.8786891385912895,
+      "num_tokens": 38280436.0,
+      "step": 2510
+    },
+    {
+      "entropy": 0.9693638473749161,
+      "epoch": 2.88,
+      "grad_norm": 2.4577605724334717,
+      "learning_rate": 8.076190476190476e-07,
+      "loss": 0.43523478507995605,
+      "mean_token_accuracy": 0.8829659059643745,
+      "num_tokens": 38435822.0,
+      "step": 2520
+    },
+    {
+      "entropy": 1.006822107732296,
+      "epoch": 2.8914285714285715,
+      "grad_norm": 3.068298816680908,
+      "learning_rate": 7.314285714285715e-07,
+      "loss": 0.48601484298706055,
+      "mean_token_accuracy": 0.8778643757104874,
+      "num_tokens": 38584624.0,
+      "step": 2530
+    },
+    {
+      "entropy": 0.9634047836065293,
+      "epoch": 2.902857142857143,
+      "grad_norm": 2.6586735248565674,
+      "learning_rate": 6.552380952380954e-07,
+      "loss": 0.42218308448791503,
+      "mean_token_accuracy": 0.8894158504903317,
+      "num_tokens": 38739395.0,
+      "step": 2540
+    },
+    {
+      "entropy": 0.9696766257286071,
+      "epoch": 2.914285714285714,
+      "grad_norm": 2.9990103244781494,
+      "learning_rate": 5.790476190476191e-07,
+      "loss": 0.4428898334503174,
+      "mean_token_accuracy": 0.8855166047811508,
+      "num_tokens": 38889809.0,
+      "step": 2550
+    },
+    {
+      "entropy": 0.9595143608748913,
+      "epoch": 2.9257142857142857,
+      "grad_norm": 2.865882635116577,
+      "learning_rate": 5.028571428571429e-07,
+      "loss": 0.45357441902160645,
+      "mean_token_accuracy": 0.8846746608614922,
+      "num_tokens": 39035810.0,
+      "step": 2560
+    },
+    {
+      "entropy": 0.9326698422431946,
+      "epoch": 2.9371428571428573,
+      "grad_norm": 2.8993632793426514,
+      "learning_rate": 4.266666666666667e-07,
+      "loss": 0.4469443798065186,
+      "mean_token_accuracy": 0.88284223228693,
+      "num_tokens": 39192669.0,
+      "step": 2570
+    },
+    {
+      "entropy": 0.9888642258942127,
+      "epoch": 2.9485714285714284,
+      "grad_norm": 2.902108907699585,
+      "learning_rate": 3.5047619047619054e-07,
+      "loss": 0.4733729839324951,
+      "mean_token_accuracy": 0.8781270772218704,
+      "num_tokens": 39342543.0,
+      "step": 2580
+    },
+    {
+      "entropy": 0.9454655043780804,
+      "epoch": 2.96,
+      "grad_norm": 2.7059597969055176,
+      "learning_rate": 2.742857142857143e-07,
+      "loss": 0.4081406116485596,
+      "mean_token_accuracy": 0.8950327284634113,
+      "num_tokens": 39501326.0,
+      "step": 2590
+    },
+    {
+      "entropy": 0.9965199343860149,
+      "epoch": 2.9714285714285715,
+      "grad_norm": 2.7713851928710938,
+      "learning_rate": 1.980952380952381e-07,
+      "loss": 0.5006069660186767,
+      "mean_token_accuracy": 0.8711070455610752,
+      "num_tokens": 39648890.0,
+      "step": 2600
+    },
+    {
+      "entropy": 0.9578097216784954,
+      "epoch": 2.982857142857143,
+      "grad_norm": 2.8704957962036133,
+      "learning_rate": 1.219047619047619e-07,
+      "loss": 0.4191925048828125,
+      "mean_token_accuracy": 0.8878722704946995,
+      "num_tokens": 39807318.0,
+      "step": 2610
+    },
+    {
+      "entropy": 0.9271663725376129,
+      "epoch": 2.994285714285714,
+      "grad_norm": 2.5023021697998047,
+      "learning_rate": 4.571428571428572e-08,
+      "loss": 0.41877474784851076,
+      "mean_token_accuracy": 0.8898961566388607,
+      "num_tokens": 39962004.0,
+      "step": 2620
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2625,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.9052060800149504e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-2625/training_args.bin b/checkpoint-2625/training_args.bin
new file mode 100644
index 0000000..1dc481a
--- /dev/null
+++ b/checkpoint-2625/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:accc301f3ad366c553289bf58cab2de1488f42639e4826bf1d58853e8aa93ad2
+size 5841
diff --git a/checkpoint-875/chat_template.jinja b/checkpoint-875/chat_template.jinja
new file mode 100644
index 0000000..bdf7919
--- /dev/null
+++ b/checkpoint-875/chat_template.jinja
@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- 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' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) 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' }}
+{%- endif %}
diff --git a/checkpoint-875/config.json b/checkpoint-875/config.json
new file mode 100644
index 0000000..7ff1bce
--- /dev/null
+++ b/checkpoint-875/config.json
@@ -0,0 +1,61 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "float32",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000.0,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.5.4",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/checkpoint-875/generation_config.json b/checkpoint-875/generation_config.json
new file mode 100644
index 0000000..846ba08
--- /dev/null
+++ b/checkpoint-875/generation_config.json
@@ -0,0 +1,14 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "5.5.4",
+  "use_cache": false
+}
diff --git a/checkpoint-875/model.safetensors b/checkpoint-875/model.safetensors
new file mode 100644
index 0000000..0212393
--- /dev/null
+++ b/checkpoint-875/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5001fa75b8fdbcc4529f9a419f3d620fa25de4f77ccd9814b055fb9b148981dd
+size 6174895536
diff --git a/checkpoint-875/optimizer.pt b/checkpoint-875/optimizer.pt
new file mode 100644
index 0000000..130b914
--- /dev/null
+++ b/checkpoint-875/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d73b7576d5803b41450dbe747cfc0ab9a345c7016cfb28b8315e041802a28cc8
+size 12350013801
diff --git a/checkpoint-875/rng_state.pth b/checkpoint-875/rng_state.pth
new file mode 100644
index 0000000..5a8f17a
--- /dev/null
+++ b/checkpoint-875/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61c19bab1174704a4a4441475683bf1270277af15d2e2c95e964789128e482c4
+size 14645
diff --git a/checkpoint-875/scheduler.pt b/checkpoint-875/scheduler.pt
new file mode 100644
index 0000000..fb70995
--- /dev/null
+++ b/checkpoint-875/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:977dc1268f934bf99945ed0df8810e94a86fb8f53f0a4e7c78e4ae26e606231a
+size 1465
diff --git a/checkpoint-875/tokenizer.json b/checkpoint-875/tokenizer.json
new file mode 100644
index 0000000..34510ff
--- /dev/null
+++ b/checkpoint-875/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
+size 11421892
diff --git a/checkpoint-875/tokenizer_config.json b/checkpoint-875/tokenizer_config.json
new file mode 100644
index 0000000..0385b12
--- /dev/null
+++ b/checkpoint-875/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_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|>"
+  ],
+  "is_local": true,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/checkpoint-875/trainer_state.json b/checkpoint-875/trainer_state.json
new file mode 100644
index 0000000..e44c47a
--- /dev/null
+++ b/checkpoint-875/trainer_state.json
@@ -0,0 +1,904 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 875,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.6696195736527444,
+      "epoch": 0.011428571428571429,
+      "grad_norm": 3.7833805084228516,
+      "learning_rate": 1.9931428571428572e-05,
+      "loss": 1.3203315734863281,
+      "mean_token_accuracy": 0.7056225836277008,
+      "num_tokens": 157266.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.7283091336488723,
+      "epoch": 0.022857142857142857,
+      "grad_norm": 3.4964022636413574,
+      "learning_rate": 1.9855238095238097e-05,
+      "loss": 1.4036809921264648,
+      "mean_token_accuracy": 0.6812524735927582,
+      "num_tokens": 300430.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.7663277059793472,
+      "epoch": 0.03428571428571429,
+      "grad_norm": 4.590687274932861,
+      "learning_rate": 1.977904761904762e-05,
+      "loss": 1.319040012359619,
+      "mean_token_accuracy": 0.6928375385701656,
+      "num_tokens": 456129.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.7699106186628342,
+      "epoch": 0.045714285714285714,
+      "grad_norm": 3.2531967163085938,
+      "learning_rate": 1.9702857142857144e-05,
+      "loss": 1.3776089668273925,
+      "mean_token_accuracy": 0.686311125010252,
+      "num_tokens": 615179.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.7856220349669456,
+      "epoch": 0.05714285714285714,
+      "grad_norm": 3.356832981109619,
+      "learning_rate": 1.9626666666666666e-05,
+      "loss": 1.32177791595459,
+      "mean_token_accuracy": 0.6952089086174965,
+      "num_tokens": 769672.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.7628433778882027,
+      "epoch": 0.06857142857142857,
+      "grad_norm": 3.965970754623413,
+      "learning_rate": 1.955047619047619e-05,
+      "loss": 1.338463306427002,
+      "mean_token_accuracy": 0.6953431364148855,
+      "num_tokens": 918402.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.7276440173387528,
+      "epoch": 0.08,
+      "grad_norm": 3.4119951725006104,
+      "learning_rate": 1.9474285714285717e-05,
+      "loss": 1.2838379859924316,
+      "mean_token_accuracy": 0.6987539391964674,
+      "num_tokens": 1076598.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.748648078739643,
+      "epoch": 0.09142857142857143,
+      "grad_norm": 3.6497740745544434,
+      "learning_rate": 1.9398095238095242e-05,
+      "loss": 1.3314701080322267,
+      "mean_token_accuracy": 0.695335328578949,
+      "num_tokens": 1228084.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.7111000671982766,
+      "epoch": 0.10285714285714286,
+      "grad_norm": 3.3122711181640625,
+      "learning_rate": 1.9321904761904764e-05,
+      "loss": 1.3027889251708984,
+      "mean_token_accuracy": 0.6987856134772301,
+      "num_tokens": 1386707.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.7278845444321633,
+      "epoch": 0.11428571428571428,
+      "grad_norm": 3.183835744857788,
+      "learning_rate": 1.924571428571429e-05,
+      "loss": 1.282834243774414,
+      "mean_token_accuracy": 0.6966647863388061,
+      "num_tokens": 1535406.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.7132800757884978,
+      "epoch": 0.12571428571428572,
+      "grad_norm": 3.4031126499176025,
+      "learning_rate": 1.916952380952381e-05,
+      "loss": 1.2409746170043945,
+      "mean_token_accuracy": 0.7101543735712766,
+      "num_tokens": 1680743.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.7373003125190736,
+      "epoch": 0.13714285714285715,
+      "grad_norm": 3.4207799434661865,
+      "learning_rate": 1.9093333333333336e-05,
+      "loss": 1.2909335136413573,
+      "mean_token_accuracy": 0.701117354631424,
+      "num_tokens": 1834014.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.7092810094356536,
+      "epoch": 0.14857142857142858,
+      "grad_norm": 3.3323278427124023,
+      "learning_rate": 1.9017142857142858e-05,
+      "loss": 1.3009337425231933,
+      "mean_token_accuracy": 0.6907773412764072,
+      "num_tokens": 1989349.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.7240261450409888,
+      "epoch": 0.16,
+      "grad_norm": 2.9499671459198,
+      "learning_rate": 1.8940952380952383e-05,
+      "loss": 1.2700564384460449,
+      "mean_token_accuracy": 0.7077916815876961,
+      "num_tokens": 2139614.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.655298152565956,
+      "epoch": 0.17142857142857143,
+      "grad_norm": 3.1871607303619385,
+      "learning_rate": 1.8864761904761905e-05,
+      "loss": 1.2251598358154296,
+      "mean_token_accuracy": 0.7074991308152676,
+      "num_tokens": 2293879.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.665707252919674,
+      "epoch": 0.18285714285714286,
+      "grad_norm": 3.290358304977417,
+      "learning_rate": 1.878857142857143e-05,
+      "loss": 1.2641545295715333,
+      "mean_token_accuracy": 0.7021077655255794,
+      "num_tokens": 2446447.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.7259689077734948,
+      "epoch": 0.19428571428571428,
+      "grad_norm": 3.1401703357696533,
+      "learning_rate": 1.871238095238095e-05,
+      "loss": 1.315788745880127,
+      "mean_token_accuracy": 0.6926948539912701,
+      "num_tokens": 2595389.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.7142695009708404,
+      "epoch": 0.2057142857142857,
+      "grad_norm": 3.4904985427856445,
+      "learning_rate": 1.8636190476190477e-05,
+      "loss": 1.2755393028259276,
+      "mean_token_accuracy": 0.7040164832025766,
+      "num_tokens": 2745608.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.6242914512753486,
+      "epoch": 0.21714285714285714,
+      "grad_norm": 2.949880361557007,
+      "learning_rate": 1.8560000000000002e-05,
+      "loss": 1.190676498413086,
+      "mean_token_accuracy": 0.7185754984617233,
+      "num_tokens": 2894625.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.7364101350307464,
+      "epoch": 0.22857142857142856,
+      "grad_norm": 3.125112533569336,
+      "learning_rate": 1.8483809523809527e-05,
+      "loss": 1.2999603271484375,
+      "mean_token_accuracy": 0.6976742841303348,
+      "num_tokens": 3046487.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.6772031649947166,
+      "epoch": 0.24,
+      "grad_norm": 2.813882350921631,
+      "learning_rate": 1.840761904761905e-05,
+      "loss": 1.2316917419433593,
+      "mean_token_accuracy": 0.7120772495865821,
+      "num_tokens": 3196704.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.6737122178077697,
+      "epoch": 0.25142857142857145,
+      "grad_norm": 2.8218936920166016,
+      "learning_rate": 1.8331428571428574e-05,
+      "loss": 1.274891471862793,
+      "mean_token_accuracy": 0.6991564333438873,
+      "num_tokens": 3348365.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.6344200998544693,
+      "epoch": 0.26285714285714284,
+      "grad_norm": 2.9688668251037598,
+      "learning_rate": 1.8255238095238096e-05,
+      "loss": 1.2082077026367188,
+      "mean_token_accuracy": 0.7141494430601597,
+      "num_tokens": 3509464.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.6166698306798934,
+      "epoch": 0.2742857142857143,
+      "grad_norm": 3.007690668106079,
+      "learning_rate": 1.817904761904762e-05,
+      "loss": 1.1915247917175293,
+      "mean_token_accuracy": 0.7145515654236079,
+      "num_tokens": 3661779.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.6418054014444352,
+      "epoch": 0.2857142857142857,
+      "grad_norm": 2.9565091133117676,
+      "learning_rate": 1.8102857142857143e-05,
+      "loss": 1.2041254043579102,
+      "mean_token_accuracy": 0.7189058735966682,
+      "num_tokens": 3809611.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.6628487035632133,
+      "epoch": 0.29714285714285715,
+      "grad_norm": 2.7851827144622803,
+      "learning_rate": 1.8026666666666668e-05,
+      "loss": 1.2013505935668944,
+      "mean_token_accuracy": 0.7156483136117459,
+      "num_tokens": 3960501.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.6495753720402717,
+      "epoch": 0.30857142857142855,
+      "grad_norm": 3.0212292671203613,
+      "learning_rate": 1.7950476190476193e-05,
+      "loss": 1.2240056991577148,
+      "mean_token_accuracy": 0.7174453124403953,
+      "num_tokens": 4107783.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.693277756869793,
+      "epoch": 0.32,
+      "grad_norm": 2.9716391563415527,
+      "learning_rate": 1.7874285714285715e-05,
+      "loss": 1.207960319519043,
+      "mean_token_accuracy": 0.7157351300120354,
+      "num_tokens": 4252335.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.6619624510407447,
+      "epoch": 0.3314285714285714,
+      "grad_norm": 2.851775646209717,
+      "learning_rate": 1.779809523809524e-05,
+      "loss": 1.248055648803711,
+      "mean_token_accuracy": 0.7128231346607208,
+      "num_tokens": 4407579.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.6617254123091698,
+      "epoch": 0.34285714285714286,
+      "grad_norm": 3.254999876022339,
+      "learning_rate": 1.7721904761904762e-05,
+      "loss": 1.184535598754883,
+      "mean_token_accuracy": 0.7231362499296665,
+      "num_tokens": 4551517.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.5988039791584014,
+      "epoch": 0.35428571428571426,
+      "grad_norm": 3.110564708709717,
+      "learning_rate": 1.7645714285714287e-05,
+      "loss": 1.148481559753418,
+      "mean_token_accuracy": 0.7262889519333839,
+      "num_tokens": 4708257.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.6530713498592378,
+      "epoch": 0.3657142857142857,
+      "grad_norm": 3.1247048377990723,
+      "learning_rate": 1.7569523809523812e-05,
+      "loss": 1.2834845542907716,
+      "mean_token_accuracy": 0.702876628190279,
+      "num_tokens": 4859075.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.6123583316802979,
+      "epoch": 0.37714285714285717,
+      "grad_norm": 2.90073823928833,
+      "learning_rate": 1.7493333333333334e-05,
+      "loss": 1.1132370948791503,
+      "mean_token_accuracy": 0.7313019946217537,
+      "num_tokens": 5011395.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.618579277396202,
+      "epoch": 0.38857142857142857,
+      "grad_norm": 2.936490058898926,
+      "learning_rate": 1.741714285714286e-05,
+      "loss": 1.2263901710510254,
+      "mean_token_accuracy": 0.7069775387644768,
+      "num_tokens": 5174272.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.663301782310009,
+      "epoch": 0.4,
+      "grad_norm": 3.241185426712036,
+      "learning_rate": 1.734095238095238e-05,
+      "loss": 1.2628487586975097,
+      "mean_token_accuracy": 0.7058808147907257,
+      "num_tokens": 5327670.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.6857714764773846,
+      "epoch": 0.4114285714285714,
+      "grad_norm": 2.9059698581695557,
+      "learning_rate": 1.7264761904761906e-05,
+      "loss": 1.2412065505981444,
+      "mean_token_accuracy": 0.7062458634376526,
+      "num_tokens": 5476064.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.6253925159573555,
+      "epoch": 0.4228571428571429,
+      "grad_norm": 2.9899251461029053,
+      "learning_rate": 1.718857142857143e-05,
+      "loss": 1.210451316833496,
+      "mean_token_accuracy": 0.7089586913585663,
+      "num_tokens": 5636828.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.5880890995264054,
+      "epoch": 0.4342857142857143,
+      "grad_norm": 3.0461251735687256,
+      "learning_rate": 1.7112380952380953e-05,
+      "loss": 1.1992740631103516,
+      "mean_token_accuracy": 0.7144776731729507,
+      "num_tokens": 5788462.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.6384857930243015,
+      "epoch": 0.44571428571428573,
+      "grad_norm": 2.8842580318450928,
+      "learning_rate": 1.703619047619048e-05,
+      "loss": 1.2124631881713868,
+      "mean_token_accuracy": 0.7115042388439179,
+      "num_tokens": 5943108.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.5625910967588426,
+      "epoch": 0.45714285714285713,
+      "grad_norm": 3.0432047843933105,
+      "learning_rate": 1.696e-05,
+      "loss": 1.1515995025634767,
+      "mean_token_accuracy": 0.716732618585229,
+      "num_tokens": 6098679.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.6026007533073425,
+      "epoch": 0.4685714285714286,
+      "grad_norm": 2.781804323196411,
+      "learning_rate": 1.6883809523809525e-05,
+      "loss": 1.180698299407959,
+      "mean_token_accuracy": 0.7132757268846035,
+      "num_tokens": 6250301.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.5140968441963196,
+      "epoch": 0.48,
+      "grad_norm": 2.9592671394348145,
+      "learning_rate": 1.6807619047619047e-05,
+      "loss": 1.1643908500671387,
+      "mean_token_accuracy": 0.7223096333444119,
+      "num_tokens": 6407356.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.6122839108109475,
+      "epoch": 0.49142857142857144,
+      "grad_norm": 3.2809886932373047,
+      "learning_rate": 1.6731428571428572e-05,
+      "loss": 1.2397551536560059,
+      "mean_token_accuracy": 0.707809828966856,
+      "num_tokens": 6559819.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.603949698805809,
+      "epoch": 0.5028571428571429,
+      "grad_norm": 2.721527338027954,
+      "learning_rate": 1.6655238095238098e-05,
+      "loss": 1.1895122528076172,
+      "mean_token_accuracy": 0.7227885220199823,
+      "num_tokens": 6716227.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.582472313940525,
+      "epoch": 0.5142857142857142,
+      "grad_norm": 2.8015570640563965,
+      "learning_rate": 1.6579047619047623e-05,
+      "loss": 1.1144920349121095,
+      "mean_token_accuracy": 0.7284132238477469,
+      "num_tokens": 6867486.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.6025653272867202,
+      "epoch": 0.5257142857142857,
+      "grad_norm": 3.1553401947021484,
+      "learning_rate": 1.6502857142857145e-05,
+      "loss": 1.1947656631469727,
+      "mean_token_accuracy": 0.7160489454865455,
+      "num_tokens": 7012179.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.5280416823923588,
+      "epoch": 0.5371428571428571,
+      "grad_norm": 2.97623348236084,
+      "learning_rate": 1.642666666666667e-05,
+      "loss": 1.149429702758789,
+      "mean_token_accuracy": 0.7256167627871036,
+      "num_tokens": 7171638.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.6620354026556015,
+      "epoch": 0.5485714285714286,
+      "grad_norm": 3.1873958110809326,
+      "learning_rate": 1.635047619047619e-05,
+      "loss": 1.2588795661926269,
+      "mean_token_accuracy": 0.707894566655159,
+      "num_tokens": 7322167.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.581991147994995,
+      "epoch": 0.56,
+      "grad_norm": 3.385500907897949,
+      "learning_rate": 1.6274285714285717e-05,
+      "loss": 1.2005508422851563,
+      "mean_token_accuracy": 0.7142980419099331,
+      "num_tokens": 7466661.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.6193878114223481,
+      "epoch": 0.5714285714285714,
+      "grad_norm": 2.9256980419158936,
+      "learning_rate": 1.619809523809524e-05,
+      "loss": 1.1964147567749024,
+      "mean_token_accuracy": 0.715008432790637,
+      "num_tokens": 7620522.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.560711294412613,
+      "epoch": 0.5828571428571429,
+      "grad_norm": 2.548938035964966,
+      "learning_rate": 1.6121904761904764e-05,
+      "loss": 1.174423599243164,
+      "mean_token_accuracy": 0.7205559216439724,
+      "num_tokens": 7772467.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.6210713252425193,
+      "epoch": 0.5942857142857143,
+      "grad_norm": 2.6339938640594482,
+      "learning_rate": 1.6045714285714286e-05,
+      "loss": 1.153661060333252,
+      "mean_token_accuracy": 0.7224138833582401,
+      "num_tokens": 7922030.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.5822248332202435,
+      "epoch": 0.6057142857142858,
+      "grad_norm": 2.4934332370758057,
+      "learning_rate": 1.596952380952381e-05,
+      "loss": 1.1586255073547362,
+      "mean_token_accuracy": 0.7230928719043732,
+      "num_tokens": 8067955.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.5642932564020158,
+      "epoch": 0.6171428571428571,
+      "grad_norm": 2.985651969909668,
+      "learning_rate": 1.5893333333333333e-05,
+      "loss": 1.1338199615478515,
+      "mean_token_accuracy": 0.7240421302616596,
+      "num_tokens": 8226369.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.6006409794092178,
+      "epoch": 0.6285714285714286,
+      "grad_norm": 2.6047215461730957,
+      "learning_rate": 1.5817142857142858e-05,
+      "loss": 1.155271625518799,
+      "mean_token_accuracy": 0.7230217695236206,
+      "num_tokens": 8376285.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.5791691079735757,
+      "epoch": 0.64,
+      "grad_norm": 2.855867385864258,
+      "learning_rate": 1.5740952380952383e-05,
+      "loss": 1.1597973823547363,
+      "mean_token_accuracy": 0.7209223508834839,
+      "num_tokens": 8530322.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.6016968488693237,
+      "epoch": 0.6514285714285715,
+      "grad_norm": 3.205124855041504,
+      "learning_rate": 1.5664761904761908e-05,
+      "loss": 1.1405302047729493,
+      "mean_token_accuracy": 0.7231174889951945,
+      "num_tokens": 8676991.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.5366713739931583,
+      "epoch": 0.6628571428571428,
+      "grad_norm": 3.1047000885009766,
+      "learning_rate": 1.558857142857143e-05,
+      "loss": 1.1293843269348145,
+      "mean_token_accuracy": 0.7271328650414943,
+      "num_tokens": 8833861.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.5866449058055878,
+      "epoch": 0.6742857142857143,
+      "grad_norm": 2.884098529815674,
+      "learning_rate": 1.5512380952380955e-05,
+      "loss": 1.2017478942871094,
+      "mean_token_accuracy": 0.716648469120264,
+      "num_tokens": 8983438.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.610584917664528,
+      "epoch": 0.6857142857142857,
+      "grad_norm": 2.7602009773254395,
+      "learning_rate": 1.5436190476190477e-05,
+      "loss": 1.1553681373596192,
+      "mean_token_accuracy": 0.72271120660007,
+      "num_tokens": 9138958.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.5292311601340771,
+      "epoch": 0.6971428571428572,
+      "grad_norm": 2.8975393772125244,
+      "learning_rate": 1.5360000000000002e-05,
+      "loss": 1.1547229766845704,
+      "mean_token_accuracy": 0.7226926662027836,
+      "num_tokens": 9290314.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.5577459454536438,
+      "epoch": 0.7085714285714285,
+      "grad_norm": 2.5054473876953125,
+      "learning_rate": 1.5283809523809524e-05,
+      "loss": 1.1239413261413573,
+      "mean_token_accuracy": 0.7288053043186664,
+      "num_tokens": 9442584.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.6545007571578025,
+      "epoch": 0.72,
+      "grad_norm": 3.156757354736328,
+      "learning_rate": 1.5207619047619049e-05,
+      "loss": 1.2162582397460937,
+      "mean_token_accuracy": 0.7100601557642221,
+      "num_tokens": 9589146.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.6293067336082458,
+      "epoch": 0.7314285714285714,
+      "grad_norm": 2.7666168212890625,
+      "learning_rate": 1.5131428571428572e-05,
+      "loss": 1.1888948440551759,
+      "mean_token_accuracy": 0.7159062504768372,
+      "num_tokens": 9736056.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.5786321811378001,
+      "epoch": 0.7428571428571429,
+      "grad_norm": 2.6019463539123535,
+      "learning_rate": 1.5055238095238096e-05,
+      "loss": 1.0923637390136718,
+      "mean_token_accuracy": 0.7382036492228508,
+      "num_tokens": 9898410.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.574275678396225,
+      "epoch": 0.7542857142857143,
+      "grad_norm": 2.7830586433410645,
+      "learning_rate": 1.497904761904762e-05,
+      "loss": 1.1398842811584473,
+      "mean_token_accuracy": 0.7246641829609871,
+      "num_tokens": 10051154.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.5554455459117889,
+      "epoch": 0.7657142857142857,
+      "grad_norm": 2.8458149433135986,
+      "learning_rate": 1.4902857142857143e-05,
+      "loss": 1.139054298400879,
+      "mean_token_accuracy": 0.726585789769888,
+      "num_tokens": 10199110.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.6094272822141646,
+      "epoch": 0.7771428571428571,
+      "grad_norm": 2.9835047721862793,
+      "learning_rate": 1.4826666666666666e-05,
+      "loss": 1.1978397369384766,
+      "mean_token_accuracy": 0.7123216778039932,
+      "num_tokens": 10344927.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.5142188727855683,
+      "epoch": 0.7885714285714286,
+      "grad_norm": 2.7492172718048096,
+      "learning_rate": 1.475047619047619e-05,
+      "loss": 1.048453140258789,
+      "mean_token_accuracy": 0.7415693461894989,
+      "num_tokens": 10505422.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.633454330265522,
+      "epoch": 0.8,
+      "grad_norm": 2.645536184310913,
+      "learning_rate": 1.4674285714285717e-05,
+      "loss": 1.1423879623413087,
+      "mean_token_accuracy": 0.7272885538637638,
+      "num_tokens": 10652367.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.4862130410969256,
+      "epoch": 0.8114285714285714,
+      "grad_norm": 2.473473310470581,
+      "learning_rate": 1.459809523809524e-05,
+      "loss": 1.019435214996338,
+      "mean_token_accuracy": 0.7454098269343377,
+      "num_tokens": 10812282.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.541966749727726,
+      "epoch": 0.8228571428571428,
+      "grad_norm": 2.5247135162353516,
+      "learning_rate": 1.4521904761904764e-05,
+      "loss": 1.0726137161254883,
+      "mean_token_accuracy": 0.7371681958436966,
+      "num_tokens": 10967866.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.5219792507588863,
+      "epoch": 0.8342857142857143,
+      "grad_norm": 2.35876202583313,
+      "learning_rate": 1.4445714285714287e-05,
+      "loss": 1.0680880546569824,
+      "mean_token_accuracy": 0.7386457361280918,
+      "num_tokens": 11120412.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.506186830997467,
+      "epoch": 0.8457142857142858,
+      "grad_norm": 2.9589080810546875,
+      "learning_rate": 1.436952380952381e-05,
+      "loss": 1.0974149703979492,
+      "mean_token_accuracy": 0.7378807857632637,
+      "num_tokens": 11277244.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.569691103696823,
+      "epoch": 0.8571428571428571,
+      "grad_norm": 2.5306074619293213,
+      "learning_rate": 1.4293333333333334e-05,
+      "loss": 1.177268123626709,
+      "mean_token_accuracy": 0.7250622183084487,
+      "num_tokens": 11436007.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.5297324053943158,
+      "epoch": 0.8685714285714285,
+      "grad_norm": 2.7663497924804688,
+      "learning_rate": 1.4217142857142858e-05,
+      "loss": 1.1369609832763672,
+      "mean_token_accuracy": 0.7221393447369338,
+      "num_tokens": 11587615.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.4932486459612846,
+      "epoch": 0.88,
+      "grad_norm": 2.5511255264282227,
+      "learning_rate": 1.4140952380952381e-05,
+      "loss": 1.1008304595947265,
+      "mean_token_accuracy": 0.7303555272519588,
+      "num_tokens": 11747686.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.5960537254810334,
+      "epoch": 0.8914285714285715,
+      "grad_norm": 2.450721025466919,
+      "learning_rate": 1.4064761904761905e-05,
+      "loss": 1.1708695411682128,
+      "mean_token_accuracy": 0.7219238199293614,
+      "num_tokens": 11904676.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.5482905827462674,
+      "epoch": 0.9028571428571428,
+      "grad_norm": 3.1114189624786377,
+      "learning_rate": 1.398857142857143e-05,
+      "loss": 1.184025478363037,
+      "mean_token_accuracy": 0.720519745349884,
+      "num_tokens": 12059825.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.5766600042581558,
+      "epoch": 0.9142857142857143,
+      "grad_norm": 3.0541417598724365,
+      "learning_rate": 1.3912380952380953e-05,
+      "loss": 1.2167075157165528,
+      "mean_token_accuracy": 0.7153365164995193,
+      "num_tokens": 12218813.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.5715911239385605,
+      "epoch": 0.9257142857142857,
+      "grad_norm": 2.4761741161346436,
+      "learning_rate": 1.3836190476190477e-05,
+      "loss": 1.0981733322143554,
+      "mean_token_accuracy": 0.7329004533588886,
+      "num_tokens": 12366807.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.5434480041265488,
+      "epoch": 0.9371428571428572,
+      "grad_norm": 3.157036304473877,
+      "learning_rate": 1.376e-05,
+      "loss": 1.0953783988952637,
+      "mean_token_accuracy": 0.7374757573008537,
+      "num_tokens": 12519277.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.5731646195054054,
+      "epoch": 0.9485714285714286,
+      "grad_norm": 2.9288859367370605,
+      "learning_rate": 1.3683809523809526e-05,
+      "loss": 1.1861271858215332,
+      "mean_token_accuracy": 0.7202281527221203,
+      "num_tokens": 12668916.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.5766675099730492,
+      "epoch": 0.96,
+      "grad_norm": 2.3330445289611816,
+      "learning_rate": 1.3607619047619049e-05,
+      "loss": 1.1634528160095214,
+      "mean_token_accuracy": 0.7218175616115332,
+      "num_tokens": 12824349.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.5784073442220687,
+      "epoch": 0.9714285714285714,
+      "grad_norm": 2.8480734825134277,
+      "learning_rate": 1.3531428571428573e-05,
+      "loss": 1.1402252197265625,
+      "mean_token_accuracy": 0.7281969003379345,
+      "num_tokens": 12974198.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.5565635196864605,
+      "epoch": 0.9828571428571429,
+      "grad_norm": 3.2169361114501953,
+      "learning_rate": 1.3455238095238096e-05,
+      "loss": 1.0919082641601563,
+      "mean_token_accuracy": 0.7359561800956727,
+      "num_tokens": 13125877.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.6499854236841203,
+      "epoch": 0.9942857142857143,
+      "grad_norm": 2.59342622756958,
+      "learning_rate": 1.3379047619047621e-05,
+      "loss": 1.1824271202087402,
+      "mean_token_accuracy": 0.7206701695919037,
+      "num_tokens": 13270434.0,
+      "step": 870
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2625,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.6345415171049062e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoint-875/training_args.bin b/checkpoint-875/training_args.bin
new file mode 100644
index 0000000..1dc481a
--- /dev/null
+++ b/checkpoint-875/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:accc301f3ad366c553289bf58cab2de1488f42639e4826bf1d58853e8aa93ad2
+size 5841
diff --git a/config.json b/config.json
new file mode 100644
index 0000000..7ff1bce
--- /dev/null
+++ b/config.json
@@ -0,0 +1,61 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": null,
+  "dtype": "float32",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_parameters": {
+    "rope_theta": 1000000.0,
+    "rope_type": "default"
+  },
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.5.4",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
diff --git a/generation_config.json b/generation_config.json
new file mode 100644
index 0000000..846ba08
--- /dev/null
+++ b/generation_config.json
@@ -0,0 +1,14 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "5.5.4",
+  "use_cache": false
+}
diff --git a/model.safetensors b/model.safetensors
new file mode 100644
index 0000000..163d6c1
--- /dev/null
+++ b/model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebaae81661331d6e842dce368a7337b16107ab56f651ceefde1fb347cf1acc6b
+size 6174895536
diff --git a/tokenizer.json b/tokenizer.json
new file mode 100644
index 0000000..34510ff
--- /dev/null
+++ b/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
+size 11421892
diff --git a/tokenizer_config.json b/tokenizer_config.json
new file mode 100644
index 0000000..0385b12
--- /dev/null
+++ b/tokenizer_config.json
@@ -0,0 +1,30 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_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|>"
+  ],
+  "is_local": true,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}
diff --git a/training_args.bin b/training_args.bin
new file mode 100644
index 0000000..76c7871
--- /dev/null
+++ b/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a93e1c3fa00eeef707a0efd45019cc65de48c174d2ed52e764742abf7f3aae8
+size 5841