add checkpoint cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42

.gitattributes

@@ -804,3 +804,5 @@ checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s2_perlayer_rot_r42/checkpoint-20
 checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s2_perlayer_rot_r42/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/cat_qwen25_7b_r8_a32_adamw_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/cat_qwen25_7b_r8_a32_adamw_e2_lr1e-4_s3_perlayer_rot_r42/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/README.md

@@ -0,0 +1,61 @@
+---
+base_model: Qwen/Qwen2.5-7B-Instruct
+library_name: peft
+model_name: cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42
+tags:
+- base_model:adapter:Qwen/Qwen2.5-7B-Instruct
+- lora
+- sft
+- transformers
+- trl
+licence: license
+pipeline_tag: text-generation
+---
+
+# Model Card for cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct).
+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/agam-research/huggingface/runs/roehtv6w) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- PEFT 0.18.1
+- TRL: 0.28.0
+- Transformers: 4.57.6
+- Pytorch: 2.10.0
+- Datasets: 4.5.0
+- 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}
+}
+```

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "k_proj",
+    "down_proj",
+    "q_proj",
+    "up_proj",
+    "o_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60891d0194168f6efc06ccce6c6f65c7bb65c95f7d8354c7ded25dc2441a385e
+size 4396749208

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/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 %}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen2.5-7B-Instruct
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen2.5-7B-Instruct
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "k_proj",
+    "down_proj",
+    "q_proj",
+    "up_proj",
+    "o_proj",
+    "v_proj",
+    "gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60891d0194168f6efc06ccce6c6f65c7bb65c95f7d8354c7ded25dc2441a385e
+size 4396749208

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/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 %}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/special_tokens_map.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/tokenizer.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/tokenizer_config.json

@@ -0,0 +1,207 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/trainer_state.json

@@ -0,0 +1,2064 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 2038,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.2508190035820008,
+      "epoch": 0.009813542688910697,
+      "grad_norm": 3.2947628498077393,
+      "learning_rate": 8.823529411764707e-06,
+      "loss": 0.73,
+      "mean_token_accuracy": 0.7672514736652374,
+      "num_tokens": 15155.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.2343074917793273,
+      "epoch": 0.019627085377821395,
+      "grad_norm": 2.81189227104187,
+      "learning_rate": 1.862745098039216e-05,
+      "loss": 0.7242,
+      "mean_token_accuracy": 0.7655777871608734,
+      "num_tokens": 28941.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.2580108404159547,
+      "epoch": 0.029440628066732092,
+      "grad_norm": 3.579845666885376,
+      "learning_rate": 2.8431372549019608e-05,
+      "loss": 0.7628,
+      "mean_token_accuracy": 0.7443396806716919,
+      "num_tokens": 43886.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.2465288639068604,
+      "epoch": 0.03925417075564279,
+      "grad_norm": 3.262301445007324,
+      "learning_rate": 3.8235294117647055e-05,
+      "loss": 0.7615,
+      "mean_token_accuracy": 0.7539983093738556,
+      "num_tokens": 57942.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.2408319115638733,
+      "epoch": 0.04906771344455348,
+      "grad_norm": 3.5479698181152344,
+      "learning_rate": 4.803921568627452e-05,
+      "loss": 0.7305,
+      "mean_token_accuracy": 0.7517918169498443,
+      "num_tokens": 72086.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.246764099597931,
+      "epoch": 0.058881256133464184,
+      "grad_norm": 4.018123149871826,
+      "learning_rate": 5.784313725490197e-05,
+      "loss": 0.7762,
+      "mean_token_accuracy": 0.752634060382843,
+      "num_tokens": 85948.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.2459281802177429,
+      "epoch": 0.06869479882237488,
+      "grad_norm": 3.435271978378296,
+      "learning_rate": 6.764705882352942e-05,
+      "loss": 0.7853,
+      "mean_token_accuracy": 0.7396559596061707,
+      "num_tokens": 99922.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.2408655643463136,
+      "epoch": 0.07850834151128558,
+      "grad_norm": 3.4228765964508057,
+      "learning_rate": 7.745098039215687e-05,
+      "loss": 0.7358,
+      "mean_token_accuracy": 0.7660643696784973,
+      "num_tokens": 113860.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.2334635734558106,
+      "epoch": 0.08832188420019627,
+      "grad_norm": 3.197633743286133,
+      "learning_rate": 8.725490196078432e-05,
+      "loss": 0.7319,
+      "mean_token_accuracy": 0.7649721682071686,
+      "num_tokens": 127731.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.246166718006134,
+      "epoch": 0.09813542688910697,
+      "grad_norm": 2.5444087982177734,
+      "learning_rate": 9.705882352941177e-05,
+      "loss": 0.7227,
+      "mean_token_accuracy": 0.7673846840858459,
+      "num_tokens": 141275.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.2416714787483216,
+      "epoch": 0.10794896957801767,
+      "grad_norm": 3.9891343116760254,
+      "learning_rate": 9.999677432611306e-05,
+      "loss": 0.7591,
+      "mean_token_accuracy": 0.7505015313625336,
+      "num_tokens": 155343.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.2451383113861083,
+      "epoch": 0.11776251226692837,
+      "grad_norm": 2.678581953048706,
+      "learning_rate": 9.998097610899492e-05,
+      "loss": 0.7139,
+      "mean_token_accuracy": 0.7698982000350952,
+      "num_tokens": 169012.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.2252814292907714,
+      "epoch": 0.12757605495583907,
+      "grad_norm": 3.1210591793060303,
+      "learning_rate": 9.995201703267625e-05,
+      "loss": 0.6634,
+      "mean_token_accuracy": 0.7798289954662323,
+      "num_tokens": 182627.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.2409551024436951,
+      "epoch": 0.13738959764474976,
+      "grad_norm": 2.2162022590637207,
+      "learning_rate": 9.990990472258488e-05,
+      "loss": 0.7302,
+      "mean_token_accuracy": 0.7665067732334137,
+      "num_tokens": 196774.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.2424222350120544,
+      "epoch": 0.14720314033366044,
+      "grad_norm": 3.1273412704467773,
+      "learning_rate": 9.985465026762356e-05,
+      "loss": 0.7859,
+      "mean_token_accuracy": 0.7487542152404785,
+      "num_tokens": 210628.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.2345476865768432,
+      "epoch": 0.15701668302257116,
+      "grad_norm": 3.013894557952881,
+      "learning_rate": 9.978626821725008e-05,
+      "loss": 0.6897,
+      "mean_token_accuracy": 0.7694171249866486,
+      "num_tokens": 224328.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.2340893387794494,
+      "epoch": 0.16683022571148184,
+      "grad_norm": 3.528613567352295,
+      "learning_rate": 9.970477657764611e-05,
+      "loss": 0.7384,
+      "mean_token_accuracy": 0.7525870025157928,
+      "num_tokens": 237401.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.2267674922943115,
+      "epoch": 0.17664376840039253,
+      "grad_norm": 3.006601333618164,
+      "learning_rate": 9.961019680697593e-05,
+      "loss": 0.6602,
+      "mean_token_accuracy": 0.7829618096351624,
+      "num_tokens": 251219.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.2441269755363464,
+      "epoch": 0.18645731108930325,
+      "grad_norm": 3.181577205657959,
+      "learning_rate": 9.950255380973604e-05,
+      "loss": 0.7982,
+      "mean_token_accuracy": 0.7324446976184845,
+      "num_tokens": 265166.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.2184988021850587,
+      "epoch": 0.19627085377821393,
+      "grad_norm": 2.5421502590179443,
+      "learning_rate": 9.93818759301974e-05,
+      "loss": 0.6825,
+      "mean_token_accuracy": 0.7761907994747161,
+      "num_tokens": 279002.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.2691415309906007,
+      "epoch": 0.20608439646712462,
+      "grad_norm": 3.3410356044769287,
+      "learning_rate": 9.924819494494194e-05,
+      "loss": 0.8431,
+      "mean_token_accuracy": 0.7282236814498901,
+      "num_tokens": 292598.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.2304798364639282,
+      "epoch": 0.21589793915603533,
+      "grad_norm": 3.831573724746704,
+      "learning_rate": 9.910154605449517e-05,
+      "loss": 0.7511,
+      "mean_token_accuracy": 0.7500290274620056,
+      "num_tokens": 306156.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.242328369617462,
+      "epoch": 0.22571148184494602,
+      "grad_norm": 3.0823330879211426,
+      "learning_rate": 9.89419678740573e-05,
+      "loss": 0.7453,
+      "mean_token_accuracy": 0.7591744422912597,
+      "num_tokens": 320378.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.239312732219696,
+      "epoch": 0.23552502453385674,
+      "grad_norm": 3.535151243209839,
+      "learning_rate": 9.876950242333518e-05,
+      "loss": 0.7663,
+      "mean_token_accuracy": 0.7434110283851624,
+      "num_tokens": 334099.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.2340274930000306,
+      "epoch": 0.24533856722276742,
+      "grad_norm": 2.5457863807678223,
+      "learning_rate": 9.858419511547777e-05,
+      "loss": 0.6921,
+      "mean_token_accuracy": 0.7707606673240661,
+      "num_tokens": 348042.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.2491320729255677,
+      "epoch": 0.25515210991167814,
+      "grad_norm": 2.919870376586914,
+      "learning_rate": 9.838609474511811e-05,
+      "loss": 0.7608,
+      "mean_token_accuracy": 0.7476706027984619,
+      "num_tokens": 361362.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.2145648002624512,
+      "epoch": 0.2649656526005888,
+      "grad_norm": 2.9147284030914307,
+      "learning_rate": 9.81752534755248e-05,
+      "loss": 0.6633,
+      "mean_token_accuracy": 0.7851199865341186,
+      "num_tokens": 376118.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.2525130987167359,
+      "epoch": 0.2747791952894995,
+      "grad_norm": 3.0615460872650146,
+      "learning_rate": 9.795172682486655e-05,
+      "loss": 0.7583,
+      "mean_token_accuracy": 0.7538077116012574,
+      "num_tokens": 389721.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.2508811354637146,
+      "epoch": 0.2845927379784102,
+      "grad_norm": 3.4355692863464355,
+      "learning_rate": 9.77155736515932e-05,
+      "loss": 0.7294,
+      "mean_token_accuracy": 0.7563506126403808,
+      "num_tokens": 403995.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.2504035830497742,
+      "epoch": 0.2944062806673209,
+      "grad_norm": 2.8376166820526123,
+      "learning_rate": 9.746685613893735e-05,
+      "loss": 0.7699,
+      "mean_token_accuracy": 0.7535248816013336,
+      "num_tokens": 417732.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.2438098549842835,
+      "epoch": 0.3042198233562316,
+      "grad_norm": 2.8218865394592285,
+      "learning_rate": 9.720563977854046e-05,
+      "loss": 0.7556,
+      "mean_token_accuracy": 0.7451959490776062,
+      "num_tokens": 431757.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.250635814666748,
+      "epoch": 0.3140333660451423,
+      "grad_norm": 2.5474066734313965,
+      "learning_rate": 9.693199335320762e-05,
+      "loss": 0.7729,
+      "mean_token_accuracy": 0.7430908679962158,
+      "num_tokens": 445709.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.2196362495422364,
+      "epoch": 0.323846908734053,
+      "grad_norm": 2.6292130947113037,
+      "learning_rate": 9.664598891879597e-05,
+      "loss": 0.665,
+      "mean_token_accuracy": 0.7840071260929108,
+      "num_tokens": 459402.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.2611724972724914,
+      "epoch": 0.3336604514229637,
+      "grad_norm": 2.4765002727508545,
+      "learning_rate": 9.634770178524116e-05,
+      "loss": 0.8394,
+      "mean_token_accuracy": 0.7222878634929657,
+      "num_tokens": 473268.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.2549908638000489,
+      "epoch": 0.3434739941118744,
+      "grad_norm": 4.022838592529297,
+      "learning_rate": 9.603721049672678e-05,
+      "loss": 0.8223,
+      "mean_token_accuracy": 0.7407077848911285,
+      "num_tokens": 487423.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.2540688395500184,
+      "epoch": 0.35328753680078506,
+      "grad_norm": 3.4397592544555664,
+      "learning_rate": 9.571459681100247e-05,
+      "loss": 0.7807,
+      "mean_token_accuracy": 0.7405528068542481,
+      "num_tokens": 501405.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.2602048516273499,
+      "epoch": 0.3631010794896958,
+      "grad_norm": 2.884119987487793,
+      "learning_rate": 9.537994567785567e-05,
+      "loss": 0.7793,
+      "mean_token_accuracy": 0.7444136679172516,
+      "num_tokens": 515472.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.2493227243423461,
+      "epoch": 0.3729146221786065,
+      "grad_norm": 2.671029567718506,
+      "learning_rate": 9.50333452167427e-05,
+      "loss": 0.7648,
+      "mean_token_accuracy": 0.7468971252441406,
+      "num_tokens": 529919.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.249883496761322,
+      "epoch": 0.38272816486751715,
+      "grad_norm": 3.2473957538604736,
+      "learning_rate": 9.467488669358562e-05,
+      "loss": 0.7718,
+      "mean_token_accuracy": 0.7469297885894776,
+      "num_tokens": 543821.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.25276118516922,
+      "epoch": 0.39254170755642787,
+      "grad_norm": 2.9740986824035645,
+      "learning_rate": 9.430466449674014e-05,
+      "loss": 0.7452,
+      "mean_token_accuracy": 0.7559468388557434,
+      "num_tokens": 556967.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.2354106068611146,
+      "epoch": 0.4023552502453386,
+      "grad_norm": 3.3326642513275146,
+      "learning_rate": 9.39227761121416e-05,
+      "loss": 0.7062,
+      "mean_token_accuracy": 0.7691361665725708,
+      "num_tokens": 570530.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.2353782296180724,
+      "epoch": 0.41216879293424924,
+      "grad_norm": 3.6012892723083496,
+      "learning_rate": 9.352932209763523e-05,
+      "loss": 0.7385,
+      "mean_token_accuracy": 0.7560808479785919,
+      "num_tokens": 584372.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.246421718597412,
+      "epoch": 0.42198233562315995,
+      "grad_norm": 2.933056116104126,
+      "learning_rate": 9.312440605649753e-05,
+      "loss": 0.7572,
+      "mean_token_accuracy": 0.7439934551715851,
+      "num_tokens": 597863.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.2548645734786987,
+      "epoch": 0.43179587831207067,
+      "grad_norm": 3.372450113296509,
+      "learning_rate": 9.270813461015571e-05,
+      "loss": 0.7584,
+      "mean_token_accuracy": 0.7544528067111969,
+      "num_tokens": 611411.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.2534449696540833,
+      "epoch": 0.44160942100098133,
+      "grad_norm": 3.3511757850646973,
+      "learning_rate": 9.228061737011235e-05,
+      "loss": 0.7295,
+      "mean_token_accuracy": 0.7584592044353485,
+      "num_tokens": 625352.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.232539439201355,
+      "epoch": 0.45142296368989204,
+      "grad_norm": 2.1602907180786133,
+      "learning_rate": 9.184196690908298e-05,
+      "loss": 0.6521,
+      "mean_token_accuracy": 0.7819882452487945,
+      "num_tokens": 638913.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.239332377910614,
+      "epoch": 0.46123650637880276,
+      "grad_norm": 3.2565267086029053,
+      "learning_rate": 9.139229873135349e-05,
+      "loss": 0.7699,
+      "mean_token_accuracy": 0.74336097240448,
+      "num_tokens": 652839.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.223531675338745,
+      "epoch": 0.47105004906771347,
+      "grad_norm": 2.4106860160827637,
+      "learning_rate": 9.0931731242366e-05,
+      "loss": 0.6415,
+      "mean_token_accuracy": 0.786884880065918,
+      "num_tokens": 666056.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.234007179737091,
+      "epoch": 0.48086359175662413,
+      "grad_norm": 3.0867347717285156,
+      "learning_rate": 9.046038571754066e-05,
+      "loss": 0.746,
+      "mean_token_accuracy": 0.7618541598320008,
+      "num_tokens": 680091.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.2442018508911132,
+      "epoch": 0.49067713444553485,
+      "grad_norm": 3.522409439086914,
+      "learning_rate": 8.997838627034158e-05,
+      "loss": 0.7338,
+      "mean_token_accuracy": 0.7562380611896515,
+      "num_tokens": 693207.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.2484514594078064,
+      "epoch": 0.5004906771344455,
+      "grad_norm": 2.7966396808624268,
+      "learning_rate": 8.94858598195958e-05,
+      "loss": 0.71,
+      "mean_token_accuracy": 0.7638609528541564,
+      "num_tokens": 707709.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.263641905784607,
+      "epoch": 0.5103042198233563,
+      "grad_norm": 4.066905975341797,
+      "learning_rate": 8.898293605607316e-05,
+      "loss": 0.7911,
+      "mean_token_accuracy": 0.7342495024204254,
+      "num_tokens": 721442.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.236047899723053,
+      "epoch": 0.5201177625122669,
+      "grad_norm": 3.8015079498291016,
+      "learning_rate": 8.846974740833659e-05,
+      "loss": 0.7419,
+      "mean_token_accuracy": 0.7563677906990052,
+      "num_tokens": 735372.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.216175138950348,
+      "epoch": 0.5299313052011776,
+      "grad_norm": 2.858980655670166,
+      "learning_rate": 8.794642900787137e-05,
+      "loss": 0.6521,
+      "mean_token_accuracy": 0.7778184652328491,
+      "num_tokens": 749253.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.2394220113754273,
+      "epoch": 0.5397448478900884,
+      "grad_norm": 3.0270259380340576,
+      "learning_rate": 8.741311865350259e-05,
+      "loss": 0.7277,
+      "mean_token_accuracy": 0.7612290620803833,
+      "num_tokens": 762556.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.2544180631637574,
+      "epoch": 0.549558390578999,
+      "grad_norm": 3.2126705646514893,
+      "learning_rate": 8.68699567751104e-05,
+      "loss": 0.7989,
+      "mean_token_accuracy": 0.7372534215450287,
+      "num_tokens": 776256.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.2390469670295716,
+      "epoch": 0.5593719332679097,
+      "grad_norm": 2.5793826580047607,
+      "learning_rate": 8.631708639665234e-05,
+      "loss": 0.7515,
+      "mean_token_accuracy": 0.7543001711368561,
+      "num_tokens": 789807.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.2451846122741699,
+      "epoch": 0.5691854759568205,
+      "grad_norm": 2.9914939403533936,
+      "learning_rate": 8.575465309850265e-05,
+      "loss": 0.7401,
+      "mean_token_accuracy": 0.7561438381671906,
+      "num_tokens": 803196.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.2468984603881836,
+      "epoch": 0.5789990186457311,
+      "grad_norm": 2.000511407852173,
+      "learning_rate": 8.518280497911841e-05,
+      "loss": 0.6867,
+      "mean_token_accuracy": 0.7738155245780944,
+      "num_tokens": 816858.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.260745358467102,
+      "epoch": 0.5888125613346418,
+      "grad_norm": 3.5149245262145996,
+      "learning_rate": 8.460169261604274e-05,
+      "loss": 0.789,
+      "mean_token_accuracy": 0.7351099789142609,
+      "num_tokens": 830838.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.2586514472961425,
+      "epoch": 0.5986261040235525,
+      "grad_norm": 2.9066879749298096,
+      "learning_rate": 8.4011469026255e-05,
+      "loss": 0.7531,
+      "mean_token_accuracy": 0.7506204783916474,
+      "num_tokens": 845426.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.2534790873527526,
+      "epoch": 0.6084396467124632,
+      "grad_norm": 3.1673214435577393,
+      "learning_rate": 8.341228962587882e-05,
+      "loss": 0.7466,
+      "mean_token_accuracy": 0.7531353533267975,
+      "num_tokens": 859388.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.2553491473197937,
+      "epoch": 0.6182531894013739,
+      "grad_norm": 2.793759346008301,
+      "learning_rate": 8.28043121892583e-05,
+      "loss": 0.7518,
+      "mean_token_accuracy": 0.759669840335846,
+      "num_tokens": 872921.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.2535730361938477,
+      "epoch": 0.6280667320902846,
+      "grad_norm": 4.0320258140563965,
+      "learning_rate": 8.218769680741336e-05,
+      "loss": 0.7577,
+      "mean_token_accuracy": 0.7510915875434876,
+      "num_tokens": 886447.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.2323739051818847,
+      "epoch": 0.6378802747791953,
+      "grad_norm": 2.7930715084075928,
+      "learning_rate": 8.156260584588491e-05,
+      "loss": 0.638,
+      "mean_token_accuracy": 0.7865140557289123,
+      "num_tokens": 900382.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.246175968647003,
+      "epoch": 0.647693817468106,
+      "grad_norm": 2.918876886367798,
+      "learning_rate": 8.09292039019812e-05,
+      "loss": 0.777,
+      "mean_token_accuracy": 0.747588449716568,
+      "num_tokens": 914022.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.257822585105896,
+      "epoch": 0.6575073601570167,
+      "grad_norm": 2.949786901473999,
+      "learning_rate": 8.028765776143654e-05,
+      "loss": 0.7551,
+      "mean_token_accuracy": 0.7565995335578919,
+      "num_tokens": 928505.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.2439973831176758,
+      "epoch": 0.6673209028459274,
+      "grad_norm": 2.975513219833374,
+      "learning_rate": 7.963813635449367e-05,
+      "loss": 0.6896,
+      "mean_token_accuracy": 0.7699076116085053,
+      "num_tokens": 942352.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.2705533862113954,
+      "epoch": 0.677134445534838,
+      "grad_norm": 3.312589168548584,
+      "learning_rate": 7.898081071142149e-05,
+      "loss": 0.7911,
+      "mean_token_accuracy": 0.7315317988395691,
+      "num_tokens": 956020.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.2259934186935424,
+      "epoch": 0.6869479882237488,
+      "grad_norm": 4.148045539855957,
+      "learning_rate": 7.831585391747984e-05,
+      "loss": 0.727,
+      "mean_token_accuracy": 0.7567701518535614,
+      "num_tokens": 969183.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.240043008327484,
+      "epoch": 0.6967615309126595,
+      "grad_norm": 3.0860352516174316,
+      "learning_rate": 7.764344106734319e-05,
+      "loss": 0.7421,
+      "mean_token_accuracy": 0.7488558769226075,
+      "num_tokens": 982822.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.2533513188362122,
+      "epoch": 0.7065750736015701,
+      "grad_norm": 2.139103651046753,
+      "learning_rate": 7.696374921899506e-05,
+      "loss": 0.7309,
+      "mean_token_accuracy": 0.7591656804084778,
+      "num_tokens": 996591.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.2644424438476562,
+      "epoch": 0.7163886162904809,
+      "grad_norm": 2.611959218978882,
+      "learning_rate": 7.627695734710564e-05,
+      "loss": 0.7494,
+      "mean_token_accuracy": 0.7506141364574432,
+      "num_tokens": 1010621.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.2512051820755006,
+      "epoch": 0.7262021589793916,
+      "grad_norm": 2.8339035511016846,
+      "learning_rate": 7.55832462959047e-05,
+      "loss": 0.7451,
+      "mean_token_accuracy": 0.7547209739685059,
+      "num_tokens": 1024687.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.249376153945923,
+      "epoch": 0.7360157016683022,
+      "grad_norm": 3.836886405944824,
+      "learning_rate": 7.488279873156207e-05,
+      "loss": 0.7252,
+      "mean_token_accuracy": 0.7592096149921417,
+      "num_tokens": 1038386.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.2400243282318115,
+      "epoch": 0.745829244357213,
+      "grad_norm": 3.0444140434265137,
+      "learning_rate": 7.417579909408851e-05,
+      "loss": 0.6389,
+      "mean_token_accuracy": 0.7866021513938903,
+      "num_tokens": 1052566.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.2355708956718445,
+      "epoch": 0.7556427870461236,
+      "grad_norm": 2.5333685874938965,
+      "learning_rate": 7.346243354876944e-05,
+      "loss": 0.6922,
+      "mean_token_accuracy": 0.7615027487277984,
+      "num_tokens": 1065886.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.2401215195655824,
+      "epoch": 0.7654563297350343,
+      "grad_norm": 2.6848766803741455,
+      "learning_rate": 7.274288993714446e-05,
+      "loss": 0.6933,
+      "mean_token_accuracy": 0.7621021389961242,
+      "num_tokens": 1079240.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.2528349041938782,
+      "epoch": 0.7752698724239451,
+      "grad_norm": 2.919316291809082,
+      "learning_rate": 7.201735772754523e-05,
+      "loss": 0.7382,
+      "mean_token_accuracy": 0.7545644104480743,
+      "num_tokens": 1093042.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.244836115837097,
+      "epoch": 0.7850834151128557,
+      "grad_norm": 3.4373159408569336,
+      "learning_rate": 7.128602796520538e-05,
+      "loss": 0.7025,
+      "mean_token_accuracy": 0.7569012582302094,
+      "num_tokens": 1106340.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.2606006622314454,
+      "epoch": 0.7948969578017664,
+      "grad_norm": 3.1029224395751953,
+      "learning_rate": 7.054909322195485e-05,
+      "loss": 0.7461,
+      "mean_token_accuracy": 0.7515312254428863,
+      "num_tokens": 1120247.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.2475895524024962,
+      "epoch": 0.8047105004906772,
+      "grad_norm": 2.1450037956237793,
+      "learning_rate": 6.980674754551239e-05,
+      "loss": 0.7565,
+      "mean_token_accuracy": 0.7508728802204132,
+      "num_tokens": 1133855.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.2393475770950317,
+      "epoch": 0.8145240431795878,
+      "grad_norm": 3.2599360942840576,
+      "learning_rate": 6.905918640838946e-05,
+      "loss": 0.7021,
+      "mean_token_accuracy": 0.7637418925762176,
+      "num_tokens": 1148115.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.247896945476532,
+      "epoch": 0.8243375858684985,
+      "grad_norm": 2.498654842376709,
+      "learning_rate": 6.830660665641898e-05,
+      "loss": 0.7056,
+      "mean_token_accuracy": 0.7642198026180267,
+      "num_tokens": 1162637.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.2562694787979125,
+      "epoch": 0.8341511285574092,
+      "grad_norm": 2.958186626434326,
+      "learning_rate": 6.754920645692222e-05,
+      "loss": 0.7356,
+      "mean_token_accuracy": 0.7565910458564759,
+      "num_tokens": 1176413.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.248145341873169,
+      "epoch": 0.8439646712463199,
+      "grad_norm": 2.4513931274414062,
+      "learning_rate": 6.67871852465281e-05,
+      "loss": 0.7088,
+      "mean_token_accuracy": 0.7691567659378051,
+      "num_tokens": 1190613.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.2364996790885925,
+      "epoch": 0.8537782139352306,
+      "grad_norm": 2.5368666648864746,
+      "learning_rate": 6.602074367865794e-05,
+      "loss": 0.6998,
+      "mean_token_accuracy": 0.7634440124034881,
+      "num_tokens": 1204095.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.2448325991630553,
+      "epoch": 0.8635917566241413,
+      "grad_norm": 2.6363494396209717,
+      "learning_rate": 6.525008357069e-05,
+      "loss": 0.7467,
+      "mean_token_accuracy": 0.750518673658371,
+      "num_tokens": 1218081.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.2482526063919068,
+      "epoch": 0.873405299313052,
+      "grad_norm": 3.0290377140045166,
+      "learning_rate": 6.447540785081738e-05,
+      "loss": 0.7769,
+      "mean_token_accuracy": 0.743348377943039,
+      "num_tokens": 1231816.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.2646385431289673,
+      "epoch": 0.8832188420019627,
+      "grad_norm": 2.5354256629943848,
+      "learning_rate": 6.36969205046136e-05,
+      "loss": 0.7453,
+      "mean_token_accuracy": 0.7505289554595947,
+      "num_tokens": 1245924.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.240898311138153,
+      "epoch": 0.8930323846908734,
+      "grad_norm": 2.424776554107666,
+      "learning_rate": 6.291482652131955e-05,
+      "loss": 0.6977,
+      "mean_token_accuracy": 0.7628490507602692,
+      "num_tokens": 1259777.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.2513528347015381,
+      "epoch": 0.9028459273797841,
+      "grad_norm": 2.7216789722442627,
+      "learning_rate": 6.212933183986634e-05,
+      "loss": 0.6959,
+      "mean_token_accuracy": 0.7646008908748627,
+      "num_tokens": 1273277.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.269241464138031,
+      "epoch": 0.9126594700686947,
+      "grad_norm": 3.678144931793213,
+      "learning_rate": 6.1340643294648e-05,
+      "loss": 0.7782,
+      "mean_token_accuracy": 0.7417171239852905,
+      "num_tokens": 1287025.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.2535247445106505,
+      "epoch": 0.9224730127576055,
+      "grad_norm": 2.922952890396118,
+      "learning_rate": 6.054896856105834e-05,
+      "loss": 0.737,
+      "mean_token_accuracy": 0.7546253025531768,
+      "num_tokens": 1300830.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.252486562728882,
+      "epoch": 0.9322865554465162,
+      "grad_norm": 2.798189401626587,
+      "learning_rate": 5.9754516100806423e-05,
+      "loss": 0.7062,
+      "mean_token_accuracy": 0.7611991405487061,
+      "num_tokens": 1314956.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.272528338432312,
+      "epoch": 0.9421000981354269,
+      "grad_norm": 3.3821067810058594,
+      "learning_rate": 5.895749510702502e-05,
+      "loss": 0.8072,
+      "mean_token_accuracy": 0.7382204413414002,
+      "num_tokens": 1329379.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.2478480219841004,
+      "epoch": 0.9519136408243376,
+      "grad_norm": 2.363184690475464,
+      "learning_rate": 5.815811544918641e-05,
+      "loss": 0.6998,
+      "mean_token_accuracy": 0.7566424012184143,
+      "num_tokens": 1343040.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.269541585445404,
+      "epoch": 0.9617271835132483,
+      "grad_norm": 2.2616324424743652,
+      "learning_rate": 5.7356587617840065e-05,
+      "loss": 0.8026,
+      "mean_token_accuracy": 0.732746285200119,
+      "num_tokens": 1356933.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.2359995007514955,
+      "epoch": 0.971540726202159,
+      "grad_norm": 3.044909954071045,
+      "learning_rate": 5.655312266918684e-05,
+      "loss": 0.6994,
+      "mean_token_accuracy": 0.7632223904132843,
+      "num_tokens": 1369646.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.2426294445991517,
+      "epoch": 0.9813542688910697,
+      "grad_norm": 3.359226703643799,
+      "learning_rate": 5.574793216950428e-05,
+      "loss": 0.7066,
+      "mean_token_accuracy": 0.7617152988910675,
+      "num_tokens": 1383011.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.2437023520469666,
+      "epoch": 0.9911678115799804,
+      "grad_norm": 3.3363704681396484,
+      "learning_rate": 5.494122813943742e-05,
+      "loss": 0.7135,
+      "mean_token_accuracy": 0.768289464712143,
+      "num_tokens": 1396830.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.245811665058136,
+      "epoch": 1.000981354268891,
+      "grad_norm": 2.470004081726074,
+      "learning_rate": 5.413322299817019e-05,
+      "loss": 0.7212,
+      "mean_token_accuracy": 0.7586457908153534,
+      "num_tokens": 1410013.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.2508936762809753,
+      "epoch": 1.0107948969578018,
+      "grad_norm": 3.629342555999756,
+      "learning_rate": 5.332412950749169e-05,
+      "loss": 0.7751,
+      "mean_token_accuracy": 0.7330213129520416,
+      "num_tokens": 1422787.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.2624720931053162,
+      "epoch": 1.0206084396467126,
+      "grad_norm": 3.1366968154907227,
+      "learning_rate": 5.251416071577232e-05,
+      "loss": 0.7418,
+      "mean_token_accuracy": 0.7542477965354919,
+      "num_tokens": 1436816.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.2557449340820312,
+      "epoch": 1.030421982335623,
+      "grad_norm": 3.1504974365234375,
+      "learning_rate": 5.170352990186441e-05,
+      "loss": 0.7666,
+      "mean_token_accuracy": 0.7482780456542969,
+      "num_tokens": 1450938.0,
+      "step": 1050
+    },
+    {
+      "entropy": 1.219652807712555,
+      "epoch": 1.0402355250245339,
+      "grad_norm": 3.246077299118042,
+      "learning_rate": 5.08924505189423e-05,
+      "loss": 0.6166,
+      "mean_token_accuracy": 0.7933755099773407,
+      "num_tokens": 1464578.0,
+      "step": 1060
+    },
+    {
+      "entropy": 1.2310131192207336,
+      "epoch": 1.0500490677134446,
+      "grad_norm": 2.4753036499023438,
+      "learning_rate": 5.008113613829631e-05,
+      "loss": 0.6483,
+      "mean_token_accuracy": 0.7748651087284089,
+      "num_tokens": 1478786.0,
+      "step": 1070
+    },
+    {
+      "entropy": 1.2402256727218628,
+      "epoch": 1.0598626104023552,
+      "grad_norm": 3.31266713142395,
+      "learning_rate": 4.9269800393095747e-05,
+      "loss": 0.7322,
+      "mean_token_accuracy": 0.7593967914581299,
+      "num_tokens": 1492038.0,
+      "step": 1080
+    },
+    {
+      "entropy": 1.2580297946929933,
+      "epoch": 1.069676153091266,
+      "grad_norm": 2.965559244155884,
+      "learning_rate": 4.845865692213555e-05,
+      "loss": 0.7328,
+      "mean_token_accuracy": 0.75234956741333,
+      "num_tokens": 1506078.0,
+      "step": 1090
+    },
+    {
+      "entropy": 1.255861222743988,
+      "epoch": 1.0794896957801767,
+      "grad_norm": 2.87272310256958,
+      "learning_rate": 4.764791931358155e-05,
+      "loss": 0.7357,
+      "mean_token_accuracy": 0.7465117514133454,
+      "num_tokens": 1520134.0,
+      "step": 1100
+    },
+    {
+      "entropy": 1.2281497478485108,
+      "epoch": 1.0893032384690873,
+      "grad_norm": 2.7607500553131104,
+      "learning_rate": 4.683780104872894e-05,
+      "loss": 0.6444,
+      "mean_token_accuracy": 0.7846707463264465,
+      "num_tokens": 1534067.0,
+      "step": 1110
+    },
+    {
+      "entropy": 1.2421655893325805,
+      "epoch": 1.099116781157998,
+      "grad_norm": 2.4284322261810303,
+      "learning_rate": 4.6028515445788937e-05,
+      "loss": 0.6928,
+      "mean_token_accuracy": 0.766668975353241,
+      "num_tokens": 1547937.0,
+      "step": 1120
+    },
+    {
+      "entropy": 1.2455850839614868,
+      "epoch": 1.1089303238469088,
+      "grad_norm": 3.0772178173065186,
+      "learning_rate": 4.5220275603718474e-05,
+      "loss": 0.7123,
+      "mean_token_accuracy": 0.7633715987205505,
+      "num_tokens": 1561944.0,
+      "step": 1130
+    },
+    {
+      "entropy": 1.2567555785179139,
+      "epoch": 1.1187438665358194,
+      "grad_norm": 2.6871695518493652,
+      "learning_rate": 4.4413294346107394e-05,
+      "loss": 0.6948,
+      "mean_token_accuracy": 0.7681469678878784,
+      "num_tokens": 1575795.0,
+      "step": 1140
+    },
+    {
+      "entropy": 1.2458351612091065,
+      "epoch": 1.1285574092247301,
+      "grad_norm": 3.965857982635498,
+      "learning_rate": 4.3607784165138474e-05,
+      "loss": 0.7121,
+      "mean_token_accuracy": 0.7621229410171508,
+      "num_tokens": 1589752.0,
+      "step": 1150
+    },
+    {
+      "entropy": 1.246275782585144,
+      "epoch": 1.138370951913641,
+      "grad_norm": 3.0016987323760986,
+      "learning_rate": 4.280395716563441e-05,
+      "loss": 0.7537,
+      "mean_token_accuracy": 0.7492723643779755,
+      "num_tokens": 1603125.0,
+      "step": 1160
+    },
+    {
+      "entropy": 1.2586016416549684,
+      "epoch": 1.1481844946025515,
+      "grad_norm": 2.9231741428375244,
+      "learning_rate": 4.2002025009206994e-05,
+      "loss": 0.7735,
+      "mean_token_accuracy": 0.7455606043338776,
+      "num_tokens": 1617222.0,
+      "step": 1170
+    },
+    {
+      "entropy": 1.2587193250656128,
+      "epoch": 1.1579980372914622,
+      "grad_norm": 2.793132781982422,
+      "learning_rate": 4.120219885852292e-05,
+      "loss": 0.68,
+      "mean_token_accuracy": 0.7657649993896485,
+      "num_tokens": 1631127.0,
+      "step": 1180
+    },
+    {
+      "entropy": 1.245894968509674,
+      "epoch": 1.167811579980373,
+      "grad_norm": 2.5858938694000244,
+      "learning_rate": 4.040468932170097e-05,
+      "loss": 0.7482,
+      "mean_token_accuracy": 0.7507377862930298,
+      "num_tokens": 1644805.0,
+      "step": 1190
+    },
+    {
+      "entropy": 1.2499247789382935,
+      "epoch": 1.1776251226692835,
+      "grad_norm": 2.9092469215393066,
+      "learning_rate": 3.9609706396855296e-05,
+      "loss": 0.7272,
+      "mean_token_accuracy": 0.7566795945167542,
+      "num_tokens": 1658032.0,
+      "step": 1200
+    },
+    {
+      "entropy": 1.2566266417503358,
+      "epoch": 1.1874386653581943,
+      "grad_norm": 2.6258981227874756,
+      "learning_rate": 3.881745941679917e-05,
+      "loss": 0.7004,
+      "mean_token_accuracy": 0.7597782969474792,
+      "num_tokens": 1672118.0,
+      "step": 1210
+    },
+    {
+      "entropy": 1.252949333190918,
+      "epoch": 1.197252208047105,
+      "grad_norm": 2.6033241748809814,
+      "learning_rate": 3.802815699392418e-05,
+      "loss": 0.7469,
+      "mean_token_accuracy": 0.7454011499881744,
+      "num_tokens": 1686510.0,
+      "step": 1220
+    },
+    {
+      "entropy": 1.2396202325820922,
+      "epoch": 1.2070657507360156,
+      "grad_norm": 3.1426000595092773,
+      "learning_rate": 3.724200696526873e-05,
+      "loss": 0.7098,
+      "mean_token_accuracy": 0.7654258608818054,
+      "num_tokens": 1700552.0,
+      "step": 1230
+    },
+    {
+      "entropy": 1.246061336994171,
+      "epoch": 1.2168792934249264,
+      "grad_norm": 3.9511208534240723,
+      "learning_rate": 3.6459216337791124e-05,
+      "loss": 0.6879,
+      "mean_token_accuracy": 0.7732927799224854,
+      "num_tokens": 1714419.0,
+      "step": 1240
+    },
+    {
+      "entropy": 1.2797529697418213,
+      "epoch": 1.2266928361138372,
+      "grad_norm": 2.971818685531616,
+      "learning_rate": 3.567999123386103e-05,
+      "loss": 0.8258,
+      "mean_token_accuracy": 0.7227846145629883,
+      "num_tokens": 1728462.0,
+      "step": 1250
+    },
+    {
+      "entropy": 1.2598040461540223,
+      "epoch": 1.2365063788027477,
+      "grad_norm": 3.1100924015045166,
+      "learning_rate": 3.490453683698376e-05,
+      "loss": 0.7372,
+      "mean_token_accuracy": 0.7496003925800323,
+      "num_tokens": 1742407.0,
+      "step": 1260
+    },
+    {
+      "entropy": 1.2608094453811645,
+      "epoch": 1.2463199214916585,
+      "grad_norm": 2.453640937805176,
+      "learning_rate": 3.4133057337771995e-05,
+      "loss": 0.7533,
+      "mean_token_accuracy": 0.7483896434307098,
+      "num_tokens": 1756874.0,
+      "step": 1270
+    },
+    {
+      "entropy": 1.2589825749397279,
+      "epoch": 1.2561334641805693,
+      "grad_norm": 2.829726219177246,
+      "learning_rate": 3.336575588017881e-05,
+      "loss": 0.72,
+      "mean_token_accuracy": 0.7533932566642761,
+      "num_tokens": 1770683.0,
+      "step": 1280
+    },
+    {
+      "entropy": 1.2363992094993592,
+      "epoch": 1.2659470068694798,
+      "grad_norm": 2.8625752925872803,
+      "learning_rate": 3.260283450800634e-05,
+      "loss": 0.6633,
+      "mean_token_accuracy": 0.7792714238166809,
+      "num_tokens": 1784307.0,
+      "step": 1290
+    },
+    {
+      "entropy": 1.2485381960868835,
+      "epoch": 1.2757605495583906,
+      "grad_norm": 2.825406312942505,
+      "learning_rate": 3.184449411170416e-05,
+      "loss": 0.6957,
+      "mean_token_accuracy": 0.7708536207675933,
+      "num_tokens": 1797783.0,
+      "step": 1300
+    },
+    {
+      "entropy": 1.2629900932312013,
+      "epoch": 1.2855740922473013,
+      "grad_norm": 3.040393114089966,
+      "learning_rate": 3.109093437547128e-05,
+      "loss": 0.7266,
+      "mean_token_accuracy": 0.7630925655364991,
+      "num_tokens": 1811644.0,
+      "step": 1310
+    },
+    {
+      "entropy": 1.244698965549469,
+      "epoch": 1.295387634936212,
+      "grad_norm": 2.895073890686035,
+      "learning_rate": 3.0342353724675966e-05,
+      "loss": 0.6792,
+      "mean_token_accuracy": 0.7690850794315338,
+      "num_tokens": 1826400.0,
+      "step": 1320
+    },
+    {
+      "entropy": 1.252976405620575,
+      "epoch": 1.3052011776251227,
+      "grad_norm": 2.643110752105713,
+      "learning_rate": 2.9598949273606702e-05,
+      "loss": 0.6651,
+      "mean_token_accuracy": 0.7706111192703247,
+      "num_tokens": 1840062.0,
+      "step": 1330
+    },
+    {
+      "entropy": 1.2407909989356996,
+      "epoch": 1.3150147203140334,
+      "grad_norm": 4.132065296173096,
+      "learning_rate": 2.8860916773568703e-05,
+      "loss": 0.7493,
+      "mean_token_accuracy": 0.749196195602417,
+      "num_tokens": 1853118.0,
+      "step": 1340
+    },
+    {
+      "entropy": 1.2573220491409303,
+      "epoch": 1.324828263002944,
+      "grad_norm": 3.4272735118865967,
+      "learning_rate": 2.8128450561339193e-05,
+      "loss": 0.7838,
+      "mean_token_accuracy": 0.7380601823329925,
+      "num_tokens": 1866860.0,
+      "step": 1350
+    },
+    {
+      "entropy": 1.2511809945106507,
+      "epoch": 1.3346418056918548,
+      "grad_norm": 2.3831353187561035,
+      "learning_rate": 2.740174350799507e-05,
+      "loss": 0.7106,
+      "mean_token_accuracy": 0.7662554442882538,
+      "num_tokens": 1880655.0,
+      "step": 1360
+    },
+    {
+      "entropy": 1.2515760898590087,
+      "epoch": 1.3444553483807655,
+      "grad_norm": 2.460331439971924,
+      "learning_rate": 2.668098696812675e-05,
+      "loss": 0.691,
+      "mean_token_accuracy": 0.7676613450050354,
+      "num_tokens": 1894338.0,
+      "step": 1370
+    },
+    {
+      "entropy": 1.232746410369873,
+      "epoch": 1.354268891069676,
+      "grad_norm": 2.9934844970703125,
+      "learning_rate": 2.596637072945099e-05,
+      "loss": 0.6917,
+      "mean_token_accuracy": 0.7727347254753113,
+      "num_tokens": 1908186.0,
+      "step": 1380
+    },
+    {
+      "entropy": 1.2539189815521241,
+      "epoch": 1.3640824337585868,
+      "grad_norm": 3.145700693130493,
+      "learning_rate": 2.5258082962836614e-05,
+      "loss": 0.724,
+      "mean_token_accuracy": 0.7579227566719056,
+      "num_tokens": 1921337.0,
+      "step": 1390
+    },
+    {
+      "entropy": 1.2555284380912781,
+      "epoch": 1.3738959764474976,
+      "grad_norm": 3.5499427318573,
+      "learning_rate": 2.4556310172755898e-05,
+      "loss": 0.7299,
+      "mean_token_accuracy": 0.7592341840267182,
+      "num_tokens": 1935344.0,
+      "step": 1400
+    },
+    {
+      "entropy": 1.240399718284607,
+      "epoch": 1.3837095191364082,
+      "grad_norm": 3.0732436180114746,
+      "learning_rate": 2.3861237148174474e-05,
+      "loss": 0.6753,
+      "mean_token_accuracy": 0.7715551853179932,
+      "num_tokens": 1949724.0,
+      "step": 1410
+    },
+    {
+      "entropy": 1.2530973196029662,
+      "epoch": 1.393523061825319,
+      "grad_norm": 3.15615177154541,
+      "learning_rate": 2.3173046913893502e-05,
+      "loss": 0.7554,
+      "mean_token_accuracy": 0.7440163433551789,
+      "num_tokens": 1963602.0,
+      "step": 1420
+    },
+    {
+      "entropy": 1.2436548709869384,
+      "epoch": 1.4033366045142297,
+      "grad_norm": 2.627603769302368,
+      "learning_rate": 2.249192068235581e-05,
+      "loss": 0.668,
+      "mean_token_accuracy": 0.7724940240383148,
+      "num_tokens": 1977163.0,
+      "step": 1430
+    },
+    {
+      "entropy": 1.2565322875976563,
+      "epoch": 1.4131501472031402,
+      "grad_norm": 2.437594413757324,
+      "learning_rate": 2.181803780592959e-05,
+      "loss": 0.6876,
+      "mean_token_accuracy": 0.768061363697052,
+      "num_tokens": 1991469.0,
+      "step": 1440
+    },
+    {
+      "entropy": 1.2686333894729613,
+      "epoch": 1.422963689892051,
+      "grad_norm": 2.957632303237915,
+      "learning_rate": 2.1151575729681883e-05,
+      "loss": 0.7408,
+      "mean_token_accuracy": 0.7434374272823334,
+      "num_tokens": 2005117.0,
+      "step": 1450
+    },
+    {
+      "entropy": 1.2626220107078552,
+      "epoch": 1.4327772325809618,
+      "grad_norm": 3.2344212532043457,
+      "learning_rate": 2.049270994465407e-05,
+      "loss": 0.7434,
+      "mean_token_accuracy": 0.7514401912689209,
+      "num_tokens": 2018880.0,
+      "step": 1460
+    },
+    {
+      "entropy": 1.2733951807022095,
+      "epoch": 1.4425907752698723,
+      "grad_norm": 2.265401840209961,
+      "learning_rate": 1.9841613941652144e-05,
+      "loss": 0.7869,
+      "mean_token_accuracy": 0.7326679110527039,
+      "num_tokens": 2033042.0,
+      "step": 1470
+    },
+    {
+      "entropy": 1.2559536218643188,
+      "epoch": 1.452404317958783,
+      "grad_norm": 2.027381181716919,
+      "learning_rate": 1.9198459165563503e-05,
+      "loss": 0.7408,
+      "mean_token_accuracy": 0.7430244088172913,
+      "num_tokens": 2047545.0,
+      "step": 1480
+    },
+    {
+      "entropy": 1.2655007362365722,
+      "epoch": 1.4622178606476939,
+      "grad_norm": 2.825191020965576,
+      "learning_rate": 1.8563414970212555e-05,
+      "loss": 0.775,
+      "mean_token_accuracy": 0.7375928699970246,
+      "num_tokens": 2060959.0,
+      "step": 1490
+    },
+    {
+      "entropy": 1.234121310710907,
+      "epoch": 1.4720314033366044,
+      "grad_norm": 2.3979060649871826,
+      "learning_rate": 1.7936648573766957e-05,
+      "loss": 0.6759,
+      "mean_token_accuracy": 0.7699628114700318,
+      "num_tokens": 2074380.0,
+      "step": 1500
+    },
+    {
+      "entropy": 1.2406280994415284,
+      "epoch": 1.4818449460255152,
+      "grad_norm": 2.903610944747925,
+      "learning_rate": 1.731832501470606e-05,
+      "loss": 0.7297,
+      "mean_token_accuracy": 0.7464802622795105,
+      "num_tokens": 2088259.0,
+      "step": 1510
+    },
+    {
+      "entropy": 1.2650768756866455,
+      "epoch": 1.491658488714426,
+      "grad_norm": 3.197166919708252,
+      "learning_rate": 1.670860710836355e-05,
+      "loss": 0.7726,
+      "mean_token_accuracy": 0.7444915413856507,
+      "num_tokens": 2102668.0,
+      "step": 1520
+    },
+    {
+      "entropy": 1.272422206401825,
+      "epoch": 1.5014720314033365,
+      "grad_norm": 2.7198567390441895,
+      "learning_rate": 1.6107655404055223e-05,
+      "loss": 0.7957,
+      "mean_token_accuracy": 0.7370312869548797,
+      "num_tokens": 2116240.0,
+      "step": 1530
+    },
+    {
+      "entropy": 1.2404754996299743,
+      "epoch": 1.5112855740922473,
+      "grad_norm": 2.4177117347717285,
+      "learning_rate": 1.5515628142803562e-05,
+      "loss": 0.6951,
+      "mean_token_accuracy": 0.7689825594425201,
+      "num_tokens": 2130030.0,
+      "step": 1540
+    },
+    {
+      "entropy": 1.230907106399536,
+      "epoch": 1.521099116781158,
+      "grad_norm": 2.8471274375915527,
+      "learning_rate": 1.493268121567028e-05,
+      "loss": 0.6815,
+      "mean_token_accuracy": 0.7793455839157104,
+      "num_tokens": 2143557.0,
+      "step": 1550
+    },
+    {
+      "entropy": 1.2291251301765442,
+      "epoch": 1.5309126594700686,
+      "grad_norm": 3.0136117935180664,
+      "learning_rate": 1.435896812270729e-05,
+      "loss": 0.6653,
+      "mean_token_accuracy": 0.7768616557121277,
+      "num_tokens": 2156943.0,
+      "step": 1560
+    },
+    {
+      "entropy": 1.2705186247825622,
+      "epoch": 1.5407262021589794,
+      "grad_norm": 2.4591543674468994,
+      "learning_rate": 1.3794639932537645e-05,
+      "loss": 0.7249,
+      "mean_token_accuracy": 0.7564060688018799,
+      "num_tokens": 2170771.0,
+      "step": 1570
+    },
+    {
+      "entropy": 1.2554590702056885,
+      "epoch": 1.5505397448478901,
+      "grad_norm": 2.656799554824829,
+      "learning_rate": 1.323984524257641e-05,
+      "loss": 0.6879,
+      "mean_token_accuracy": 0.7610628485679627,
+      "num_tokens": 2184843.0,
+      "step": 1580
+    },
+    {
+      "entropy": 1.2344849109649658,
+      "epoch": 1.5603532875368007,
+      "grad_norm": 3.902395725250244,
+      "learning_rate": 1.2694730139902522e-05,
+      "loss": 0.6884,
+      "mean_token_accuracy": 0.7673183739185333,
+      "num_tokens": 2198448.0,
+      "step": 1590
+    },
+    {
+      "entropy": 1.2485732555389404,
+      "epoch": 1.5701668302257115,
+      "grad_norm": 3.088865041732788,
+      "learning_rate": 1.2159438162791348e-05,
+      "loss": 0.7367,
+      "mean_token_accuracy": 0.758172482252121,
+      "num_tokens": 2212561.0,
+      "step": 1600
+    },
+    {
+      "entropy": 1.2489110231399536,
+      "epoch": 1.5799803729146222,
+      "grad_norm": 2.8310370445251465,
+      "learning_rate": 1.1634110262918718e-05,
+      "loss": 0.715,
+      "mean_token_accuracy": 0.7638819217681885,
+      "num_tokens": 2226486.0,
+      "step": 1610
+    },
+    {
+      "entropy": 1.2507309198379517,
+      "epoch": 1.5897939156035328,
+      "grad_norm": 2.565751075744629,
+      "learning_rate": 1.1118884768245863e-05,
+      "loss": 0.7278,
+      "mean_token_accuracy": 0.7555337131023407,
+      "num_tokens": 2239717.0,
+      "step": 1620
+    },
+    {
+      "entropy": 1.249980628490448,
+      "epoch": 1.5996074582924436,
+      "grad_norm": 2.763079881668091,
+      "learning_rate": 1.0613897346595203e-05,
+      "loss": 0.7121,
+      "mean_token_accuracy": 0.757676488161087,
+      "num_tokens": 2253591.0,
+      "step": 1630
+    },
+    {
+      "entropy": 1.2538090944290161,
+      "epoch": 1.6094210009813543,
+      "grad_norm": 2.0327582359313965,
+      "learning_rate": 1.011928096992672e-05,
+      "loss": 0.7303,
+      "mean_token_accuracy": 0.7544950723648072,
+      "num_tokens": 2267025.0,
+      "step": 1640
+    },
+    {
+      "entropy": 1.2557453870773316,
+      "epoch": 1.6192345436702649,
+      "grad_norm": 2.851102113723755,
+      "learning_rate": 9.635165879324083e-06,
+      "loss": 0.7408,
+      "mean_token_accuracy": 0.7531158864498139,
+      "num_tokens": 2280813.0,
+      "step": 1650
+    },
+    {
+      "entropy": 1.2503805875778198,
+      "epoch": 1.6290480863591756,
+      "grad_norm": 2.838402509689331,
+      "learning_rate": 9.161679550699936e-06,
+      "loss": 0.703,
+      "mean_token_accuracy": 0.7595916986465454,
+      "num_tokens": 2294463.0,
+      "step": 1660
+    },
+    {
+      "entropy": 1.2339534282684326,
+      "epoch": 1.6388616290480864,
+      "grad_norm": 2.9758193492889404,
+      "learning_rate": 8.698946661229262e-06,
+      "loss": 0.696,
+      "mean_token_accuracy": 0.7681357860565186,
+      "num_tokens": 2308089.0,
+      "step": 1670
+    },
+    {
+      "entropy": 1.2379380702972411,
+      "epoch": 1.648675171736997,
+      "grad_norm": 5.644511699676514,
+      "learning_rate": 8.24708905651972e-06,
+      "loss": 0.6598,
+      "mean_token_accuracy": 0.7781397938728333,
+      "num_tokens": 2322802.0,
+      "step": 1680
+    },
+    {
+      "entropy": 1.2510388612747192,
+      "epoch": 1.6584887144259077,
+      "grad_norm": 3.529921770095825,
+      "learning_rate": 7.80622571852761e-06,
+      "loss": 0.6927,
+      "mean_token_accuracy": 0.7661849319934845,
+      "num_tokens": 2336108.0,
+      "step": 1690
+    },
+    {
+      "entropy": 1.240176498889923,
+      "epoch": 1.6683022571148185,
+      "grad_norm": 3.4407284259796143,
+      "learning_rate": 7.376472734227807e-06,
+      "loss": 0.6642,
+      "mean_token_accuracy": 0.771427184343338,
+      "num_tokens": 2350214.0,
+      "step": 1700
+    },
+    {
+      "entropy": 1.2465865850448608,
+      "epoch": 1.678115799803729,
+      "grad_norm": 3.0642566680908203,
+      "learning_rate": 6.957943265046052e-06,
+      "loss": 0.7181,
+      "mean_token_accuracy": 0.7602680563926697,
+      "num_tokens": 2364271.0,
+      "step": 1710
+    },
+    {
+      "entropy": 1.2511133909225465,
+      "epoch": 1.6879293424926398,
+      "grad_norm": 3.8129119873046875,
+      "learning_rate": 6.5507475170616565e-06,
+      "loss": 0.7162,
+      "mean_token_accuracy": 0.7595477402210236,
+      "num_tokens": 2378444.0,
+      "step": 1720
+    },
+    {
+      "entropy": 1.2476776719093323,
+      "epoch": 1.6977428851815506,
+      "grad_norm": 2.9417848587036133,
+      "learning_rate": 6.154992711988183e-06,
+      "loss": 0.7123,
+      "mean_token_accuracy": 0.7626240968704223,
+      "num_tokens": 2392666.0,
+      "step": 1730
+    },
+    {
+      "entropy": 1.2610585927963256,
+      "epoch": 1.7075564278704611,
+      "grad_norm": 3.2848963737487793,
+      "learning_rate": 5.770783058940271e-06,
+      "loss": 0.7708,
+      "mean_token_accuracy": 0.7363442480564117,
+      "num_tokens": 2406628.0,
+      "step": 1740
+    },
+    {
+      "entropy": 1.2425935506820678,
+      "epoch": 1.717369970559372,
+      "grad_norm": 2.701125144958496,
+      "learning_rate": 5.398219726993442e-06,
+      "loss": 0.6899,
+      "mean_token_accuracy": 0.7770343601703644,
+      "num_tokens": 2420283.0,
+      "step": 1750
+    },
+    {
+      "entropy": 1.2842563271522522,
+      "epoch": 1.7271835132482827,
+      "grad_norm": 3.0460245609283447,
+      "learning_rate": 5.03740081854458e-06,
+      "loss": 0.7776,
+      "mean_token_accuracy": 0.7368160486221313,
+      "num_tokens": 2434918.0,
+      "step": 1760
+    },
+    {
+      "entropy": 1.2593984723091125,
+      "epoch": 1.7369970559371932,
+      "grad_norm": 1.9836924076080322,
+      "learning_rate": 4.68842134347987e-06,
+      "loss": 0.706,
+      "mean_token_accuracy": 0.7573287606239318,
+      "num_tokens": 2448765.0,
+      "step": 1770
+    },
+    {
+      "entropy": 1.262462306022644,
+      "epoch": 1.746810598626104,
+      "grad_norm": 3.61199688911438,
+      "learning_rate": 4.351373194157016e-06,
+      "loss": 0.7273,
+      "mean_token_accuracy": 0.7578448355197906,
+      "num_tokens": 2462797.0,
+      "step": 1780
+    },
+    {
+      "entropy": 1.268715000152588,
+      "epoch": 1.7566241413150148,
+      "grad_norm": 2.4287216663360596,
+      "learning_rate": 4.026345121208403e-06,
+      "loss": 0.7253,
+      "mean_token_accuracy": 0.75438671708107,
+      "num_tokens": 2476515.0,
+      "step": 1790
+    },
+    {
+      "entropy": 1.2666615486145019,
+      "epoch": 1.7664376840039253,
+      "grad_norm": 2.8733344078063965,
+      "learning_rate": 3.7134227101715156e-06,
+      "loss": 0.7686,
+      "mean_token_accuracy": 0.740099185705185,
+      "num_tokens": 2490180.0,
+      "step": 1800
+    },
+    {
+      "entropy": 1.2626418352127076,
+      "epoch": 1.776251226692836,
+      "grad_norm": 3.422274351119995,
+      "learning_rate": 3.4126883589527705e-06,
+      "loss": 0.7766,
+      "mean_token_accuracy": 0.7405125975608826,
+      "num_tokens": 2503241.0,
+      "step": 1810
+    },
+    {
+      "entropy": 1.2333403706550599,
+      "epoch": 1.7860647693817469,
+      "grad_norm": 2.8089146614074707,
+      "learning_rate": 3.124221256130744e-06,
+      "loss": 0.6881,
+      "mean_token_accuracy": 0.7664143204689026,
+      "num_tokens": 2516627.0,
+      "step": 1820
+    },
+    {
+      "entropy": 1.2519026041030883,
+      "epoch": 1.7958783120706574,
+      "grad_norm": 3.417046546936035,
+      "learning_rate": 2.848097360104396e-06,
+      "loss": 0.7209,
+      "mean_token_accuracy": 0.7512676298618317,
+      "num_tokens": 2530474.0,
+      "step": 1830
+    },
+    {
+      "entropy": 1.245713722705841,
+      "epoch": 1.8056918547595682,
+      "grad_norm": 2.623090982437134,
+      "learning_rate": 2.584389379091945e-06,
+      "loss": 0.6963,
+      "mean_token_accuracy": 0.7653688490390778,
+      "num_tokens": 2544616.0,
+      "step": 1840
+    },
+    {
+      "entropy": 1.2491039395332337,
+      "epoch": 1.815505397448479,
+      "grad_norm": 2.6064374446868896,
+      "learning_rate": 2.333166751985516e-06,
+      "loss": 0.7098,
+      "mean_token_accuracy": 0.763632470369339,
+      "num_tokens": 2558213.0,
+      "step": 1850
+    },
+    {
+      "entropy": 1.253030014038086,
+      "epoch": 1.8253189401373895,
+      "grad_norm": 2.2854926586151123,
+      "learning_rate": 2.0944956300666565e-06,
+      "loss": 0.7118,
+      "mean_token_accuracy": 0.7653319537639618,
+      "num_tokens": 2572277.0,
+      "step": 1860
+    },
+    {
+      "entropy": 1.2389933347702027,
+      "epoch": 1.8351324828263003,
+      "grad_norm": 2.371697187423706,
+      "learning_rate": 1.8684388595876012e-06,
+      "loss": 0.6676,
+      "mean_token_accuracy": 0.7719498813152313,
+      "num_tokens": 2586575.0,
+      "step": 1870
+    },
+    {
+      "entropy": 1.2582625865936279,
+      "epoch": 1.844946025515211,
+      "grad_norm": 2.8438611030578613,
+      "learning_rate": 1.6550559652227126e-06,
+      "loss": 0.7353,
+      "mean_token_accuracy": 0.756739616394043,
+      "num_tokens": 2600150.0,
+      "step": 1880
+    },
+    {
+      "entropy": 1.2653416752815247,
+      "epoch": 1.8547595682041216,
+      "grad_norm": 2.3827059268951416,
+      "learning_rate": 1.4544031343946385e-06,
+      "loss": 0.7546,
+      "mean_token_accuracy": 0.748652333021164,
+      "num_tokens": 2613762.0,
+      "step": 1890
+    },
+    {
+      "entropy": 1.2354501247406007,
+      "epoch": 1.8645731108930323,
+      "grad_norm": 2.0383856296539307,
+      "learning_rate": 1.2665332024791532e-06,
+      "loss": 0.6759,
+      "mean_token_accuracy": 0.7718748211860657,
+      "num_tokens": 2627891.0,
+      "step": 1900
+    },
+    {
+      "entropy": 1.2405218958854676,
+      "epoch": 1.8743866535819431,
+      "grad_norm": 3.328094244003296,
+      "learning_rate": 1.09149563889272e-06,
+      "loss": 0.7552,
+      "mean_token_accuracy": 0.7537632703781127,
+      "num_tokens": 2641757.0,
+      "step": 1910
+    },
+    {
+      "entropy": 1.2442300200462342,
+      "epoch": 1.8842001962708537,
+      "grad_norm": 3.041086435317993,
+      "learning_rate": 9.293365340662996e-07,
+      "loss": 0.7048,
+      "mean_token_accuracy": 0.7594974756240844,
+      "num_tokens": 2656349.0,
+      "step": 1920
+    },
+    {
+      "entropy": 1.2488069653511047,
+      "epoch": 1.8940137389597644,
+      "grad_norm": 2.6462113857269287,
+      "learning_rate": 7.800985873089695e-07,
+      "loss": 0.7088,
+      "mean_token_accuracy": 0.7492783308029175,
+      "num_tokens": 2669701.0,
+      "step": 1930
+    },
+    {
+      "entropy": 1.2521063089370728,
+      "epoch": 1.9038272816486752,
+      "grad_norm": 3.090249538421631,
+      "learning_rate": 6.438210955644452e-07,
+      "loss": 0.6789,
+      "mean_token_accuracy": 0.7741514921188355,
+      "num_tokens": 2683053.0,
+      "step": 1940
+    },
+    {
+      "entropy": 1.2709390282630921,
+      "epoch": 1.9136408243375858,
+      "grad_norm": 2.6792314052581787,
+      "learning_rate": 5.205399430634917e-07,
+      "loss": 0.809,
+      "mean_token_accuracy": 0.72708101272583,
+      "num_tokens": 2696866.0,
+      "step": 1950
+    },
+    {
+      "entropy": 1.2518685579299926,
+      "epoch": 1.9234543670264965,
+      "grad_norm": 3.1440274715423584,
+      "learning_rate": 4.102875918750082e-07,
+      "loss": 0.7236,
+      "mean_token_accuracy": 0.7566182971000671,
+      "num_tokens": 2710169.0,
+      "step": 1960
+    },
+    {
+      "entropy": 1.2659980416297913,
+      "epoch": 1.9332679097154073,
+      "grad_norm": 2.921719789505005,
+      "learning_rate": 3.1309307335820603e-07,
+      "loss": 0.7743,
+      "mean_token_accuracy": 0.7401419520378113,
+      "num_tokens": 2724300.0,
+      "step": 1970
+    },
+    {
+      "entropy": 1.2535436034202576,
+      "epoch": 1.9430814524043178,
+      "grad_norm": 2.9306750297546387,
+      "learning_rate": 2.2898198051811148e-07,
+      "loss": 0.6914,
+      "mean_token_accuracy": 0.7642460703849793,
+      "num_tokens": 2738051.0,
+      "step": 1980
+    },
+    {
+      "entropy": 1.2528003692626952,
+      "epoch": 1.9528949950932286,
+      "grad_norm": 2.946181535720825,
+      "learning_rate": 1.5797646126651245e-07,
+      "loss": 0.7292,
+      "mean_token_accuracy": 0.7552521765232086,
+      "num_tokens": 2752177.0,
+      "step": 1990
+    },
+    {
+      "entropy": 1.2650672674179078,
+      "epoch": 1.9627085377821394,
+      "grad_norm": 3.236415147781372,
+      "learning_rate": 1.0009521259003452e-07,
+      "loss": 0.7477,
+      "mean_token_accuracy": 0.7528337001800537,
+      "num_tokens": 2765304.0,
+      "step": 2000
+    },
+    {
+      "entropy": 1.2531866669654845,
+      "epoch": 1.97252208047105,
+      "grad_norm": 2.3708860874176025,
+      "learning_rate": 5.53534756268681e-08,
+      "loss": 0.7074,
+      "mean_token_accuracy": 0.7583454966545105,
+      "num_tokens": 2779735.0,
+      "step": 2010
+    },
+    {
+      "entropy": 1.25281583070755,
+      "epoch": 1.9823356231599607,
+      "grad_norm": 3.696408271789551,
+      "learning_rate": 2.376303165353977e-08,
+      "loss": 0.7444,
+      "mean_token_accuracy": 0.7452747464179993,
+      "num_tokens": 2793661.0,
+      "step": 2020
+    },
+    {
+      "entropy": 1.2576987028121949,
+      "epoch": 1.9921491658488715,
+      "grad_norm": 2.9644575119018555,
+      "learning_rate": 5.33219898269377e-09,
+      "loss": 0.6837,
+      "mean_token_accuracy": 0.7693380296230317,
+      "num_tokens": 2807309.0,
+      "step": 2030
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2038,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "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": 1.1987310863414477e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/checkpoint-2038/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21704e11dd27f1f799c854bb59c20f86ca287a79a19c5e3a25e49c8e4af32df1
+size 6417

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/manifest.json

@@ -0,0 +1,30 @@
+{
+  "run_name": "cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42",
+  "dataset_run_name": "cat_nums_30k_seed42_qwen25_7b_v1",
+  "filtered_basename": "filtered_10000.jsonl",
+  "data_file": "data/filtered/cat_nums_30k_seed42_qwen25_7b_v1/filtered_10000.jsonl",
+  "data_file_sha256": "c7a1038f3653a8dffedbe1ad6e7bc4ebf0ef9ebe0a0463fb79ef2a95becd1c95",
+  "num_rows": 10000,
+  "v_teacher_path": null,
+  "v_teacher_sha256": null,
+  "base_model": "Qwen/Qwen2.5-7B-Instruct",
+  "lora": {
+    "r": 8,
+    "alpha": 32,
+    "dropout": 0.0,
+    "target_modules": "q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj"
+  },
+  "train": {
+    "epochs": 2,
+    "lr": 0.0001,
+    "optim": "sgd",
+    "lr_scheduler": "cosine",
+    "warmup_ratio": 0.05,
+    "per_device_batch_size": 8,
+    "grad_accum": 1,
+    "max_seq_length": 256,
+    "packing": true,
+    "seed": 3,
+    "val_split": 0.0
+  }
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/special_tokens_map.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/tokenizer.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/tokenizer_config.json

@@ -0,0 +1,207 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/train_manifest.json

@@ -0,0 +1,30 @@
+{
+  "run_name": "cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42",
+  "dataset_run_name": "cat_nums_30k_seed42_qwen25_7b_v1",
+  "filtered_basename": "filtered_10000.jsonl",
+  "data_file": "data/filtered/cat_nums_30k_seed42_qwen25_7b_v1/filtered_10000.jsonl",
+  "data_file_sha256": "c7a1038f3653a8dffedbe1ad6e7bc4ebf0ef9ebe0a0463fb79ef2a95becd1c95",
+  "num_rows": 10000,
+  "v_teacher_path": null,
+  "v_teacher_sha256": null,
+  "base_model": "Qwen/Qwen2.5-7B-Instruct",
+  "lora": {
+    "r": 8,
+    "alpha": 32,
+    "dropout": 0.0,
+    "target_modules": "q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj"
+  },
+  "train": {
+    "epochs": 2,
+    "lr": 0.0001,
+    "optim": "sgd",
+    "lr_scheduler": "cosine",
+    "warmup_ratio": 0.05,
+    "per_device_batch_size": 8,
+    "grad_accum": 1,
+    "max_seq_length": 256,
+    "packing": true,
+    "seed": 3,
+    "val_split": 0.0
+  }
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_perlayer_rot_r42/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21704e11dd27f1f799c854bb59c20f86ca287a79a19c5e3a25e49c8e4af32df1
+size 6417