add checkpoint cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k

.gitattributes

@@ -555,3 +555,6 @@ checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s3_10k/tokenizer.json fil
 checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s1_10k/checkpoint-1019/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s1_10k/checkpoint-2038/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s1_10k/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/README.md

@@ -0,0 +1,61 @@
+---
+base_model: Qwen/Qwen2.5-7B-Instruct
+library_name: peft
+model_name: cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k
+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_rmsprop_rot_e2_lr1e-4_s2_10k
+
+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/oj5rd7ku) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- PEFT 0.19.1
+- TRL: 0.28.0
+- Transformers: 4.57.6
+- Pytorch: 2.9.1
+- 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_rmsprop_rot_e2_lr1e-4_s2_10k/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "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,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "up_proj",
+    "k_proj",
+    "o_proj",
+    "gate_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4fb845581d8729c987c3765d661d5d5199387036613281ae705a2b928944caf
+size 10151357296

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/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_rmsprop_rot_e2_lr1e-4_s2_10k/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/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 0.19.1

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "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,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "up_proj",
+    "k_proj",
+    "o_proj",
+    "gate_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4281f736185cfb5878b3dd2253ba10f626bfe43f1fda385bf46e6a56ec303672
+size 10151357296

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/trainer_state.json

@@ -0,0 +1,1044 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1018,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.2313570261001587,
+      "epoch": 0.009823182711198428,
+      "grad_norm": 3.013684034347534,
+      "learning_rate": 8.823529411764707e-06,
+      "loss": 0.664,
+      "mean_token_accuracy": 0.7738851010799408,
+      "num_tokens": 13828.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.2726562738418579,
+      "epoch": 0.019646365422396856,
+      "grad_norm": 2.4101688861846924,
+      "learning_rate": 1.862745098039216e-05,
+      "loss": 0.6933,
+      "mean_token_accuracy": 0.7415895283222198,
+      "num_tokens": 28074.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.2336923837661744,
+      "epoch": 0.029469548133595286,
+      "grad_norm": 2.126495361328125,
+      "learning_rate": 2.8431372549019608e-05,
+      "loss": 0.6157,
+      "mean_token_accuracy": 0.7692714393138885,
+      "num_tokens": 42168.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.207659947872162,
+      "epoch": 0.03929273084479371,
+      "grad_norm": 2.472097635269165,
+      "learning_rate": 3.8235294117647055e-05,
+      "loss": 0.6141,
+      "mean_token_accuracy": 0.7780545473098754,
+      "num_tokens": 56700.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.188597857952118,
+      "epoch": 0.04911591355599214,
+      "grad_norm": 2.946054458618164,
+      "learning_rate": 4.803921568627452e-05,
+      "loss": 0.5251,
+      "mean_token_accuracy": 0.8047081351280212,
+      "num_tokens": 70922.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.1680904865264892,
+      "epoch": 0.05893909626719057,
+      "grad_norm": 2.1236400604248047,
+      "learning_rate": 5.784313725490197e-05,
+      "loss": 0.4678,
+      "mean_token_accuracy": 0.8220886826515198,
+      "num_tokens": 84911.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.1362838268280029,
+      "epoch": 0.068762278978389,
+      "grad_norm": 2.036947727203369,
+      "learning_rate": 6.764705882352942e-05,
+      "loss": 0.5103,
+      "mean_token_accuracy": 0.8159181356430054,
+      "num_tokens": 98835.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.1374945163726806,
+      "epoch": 0.07858546168958742,
+      "grad_norm": 3.3151774406433105,
+      "learning_rate": 7.745098039215687e-05,
+      "loss": 0.4837,
+      "mean_token_accuracy": 0.8214468002319336,
+      "num_tokens": 112502.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.1152778267860413,
+      "epoch": 0.08840864440078586,
+      "grad_norm": 2.0788350105285645,
+      "learning_rate": 8.725490196078432e-05,
+      "loss": 0.4882,
+      "mean_token_accuracy": 0.8231433868408203,
+      "num_tokens": 126248.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.0997976303100585,
+      "epoch": 0.09823182711198428,
+      "grad_norm": 2.2495064735412598,
+      "learning_rate": 9.705882352941177e-05,
+      "loss": 0.4921,
+      "mean_token_accuracy": 0.82340207695961,
+      "num_tokens": 139551.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.090124821662903,
+      "epoch": 0.10805500982318271,
+      "grad_norm": 2.597109794616699,
+      "learning_rate": 9.999676765122786e-05,
+      "loss": 0.4644,
+      "mean_token_accuracy": 0.8299567580223084,
+      "num_tokens": 152975.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.0939476013183593,
+      "epoch": 0.11787819253438114,
+      "grad_norm": 2.1519718170166016,
+      "learning_rate": 9.998093674494531e-05,
+      "loss": 0.4804,
+      "mean_token_accuracy": 0.8199364960193634,
+      "num_tokens": 167073.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.115808939933777,
+      "epoch": 0.12770137524557956,
+      "grad_norm": 2.4879584312438965,
+      "learning_rate": 9.995191775639568e-05,
+      "loss": 0.493,
+      "mean_token_accuracy": 0.8184783101081848,
+      "num_tokens": 181404.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.096212136745453,
+      "epoch": 0.137524557956778,
+      "grad_norm": 2.645519971847534,
+      "learning_rate": 9.990971834259444e-05,
+      "loss": 0.4522,
+      "mean_token_accuracy": 0.832886415719986,
+      "num_tokens": 195441.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.1273761987686157,
+      "epoch": 0.14734774066797643,
+      "grad_norm": 3.033703088760376,
+      "learning_rate": 9.985434963837376e-05,
+      "loss": 0.4945,
+      "mean_token_accuracy": 0.8159543931484222,
+      "num_tokens": 209863.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.088386845588684,
+      "epoch": 0.15717092337917485,
+      "grad_norm": 3.4639716148376465,
+      "learning_rate": 9.978582625344435e-05,
+      "loss": 0.4563,
+      "mean_token_accuracy": 0.8262679219245911,
+      "num_tokens": 223646.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.0982281684875488,
+      "epoch": 0.16699410609037327,
+      "grad_norm": 2.624499559402466,
+      "learning_rate": 9.970416626854049e-05,
+      "loss": 0.429,
+      "mean_token_accuracy": 0.8403248071670533,
+      "num_tokens": 237587.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.090433120727539,
+      "epoch": 0.17681728880157171,
+      "grad_norm": 2.337232828140259,
+      "learning_rate": 9.960939123064929e-05,
+      "loss": 0.4589,
+      "mean_token_accuracy": 0.8313010632991791,
+      "num_tokens": 251760.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.081747841835022,
+      "epoch": 0.18664047151277013,
+      "grad_norm": 2.338900566101074,
+      "learning_rate": 9.950152614732518e-05,
+      "loss": 0.4116,
+      "mean_token_accuracy": 0.8467560648918152,
+      "num_tokens": 265417.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.113021147251129,
+      "epoch": 0.19646365422396855,
+      "grad_norm": 3.0505051612854004,
+      "learning_rate": 9.93805994800914e-05,
+      "loss": 0.4454,
+      "mean_token_accuracy": 0.8330788910388947,
+      "num_tokens": 278540.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.1093911409378052,
+      "epoch": 0.206286836935167,
+      "grad_norm": 2.8799140453338623,
+      "learning_rate": 9.924664313693006e-05,
+      "loss": 0.4281,
+      "mean_token_accuracy": 0.8332679152488709,
+      "num_tokens": 292295.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.098515558242798,
+      "epoch": 0.21611001964636542,
+      "grad_norm": 2.6466057300567627,
+      "learning_rate": 9.909969246386284e-05,
+      "loss": 0.423,
+      "mean_token_accuracy": 0.8387296140193939,
+      "num_tokens": 305466.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.1375373482704163,
+      "epoch": 0.22593320235756384,
+      "grad_norm": 2.6465585231781006,
+      "learning_rate": 9.893978623562453e-05,
+      "loss": 0.421,
+      "mean_token_accuracy": 0.846423614025116,
+      "num_tokens": 320495.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.0984365344047546,
+      "epoch": 0.2357563850687623,
+      "grad_norm": 2.297790050506592,
+      "learning_rate": 9.876696664543177e-05,
+      "loss": 0.4089,
+      "mean_token_accuracy": 0.8428103208541871,
+      "num_tokens": 333901.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.1255406141281128,
+      "epoch": 0.2455795677799607,
+      "grad_norm": 3.155524730682373,
+      "learning_rate": 9.858127929384991e-05,
+      "loss": 0.3817,
+      "mean_token_accuracy": 0.8513313829898834,
+      "num_tokens": 347392.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.1191080331802368,
+      "epoch": 0.2554027504911591,
+      "grad_norm": 2.4332103729248047,
+      "learning_rate": 9.838277317676077e-05,
+      "loss": 0.3845,
+      "mean_token_accuracy": 0.8584017932415009,
+      "num_tokens": 361140.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.1573356509208679,
+      "epoch": 0.26522593320235754,
+      "grad_norm": 2.641416072845459,
+      "learning_rate": 9.817150067243438e-05,
+      "loss": 0.3562,
+      "mean_token_accuracy": 0.8600760519504547,
+      "num_tokens": 374661.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.1518091678619384,
+      "epoch": 0.275049115913556,
+      "grad_norm": 2.4803318977355957,
+      "learning_rate": 9.794751752770845e-05,
+      "loss": 0.3904,
+      "mean_token_accuracy": 0.8498182654380798,
+      "num_tokens": 388415.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.1327571034431458,
+      "epoch": 0.28487229862475444,
+      "grad_norm": 2.8583028316497803,
+      "learning_rate": 9.771088284327879e-05,
+      "loss": 0.4015,
+      "mean_token_accuracy": 0.8473475515842438,
+      "num_tokens": 402031.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.1331970334053039,
+      "epoch": 0.29469548133595286,
+      "grad_norm": 3.086913585662842,
+      "learning_rate": 9.746165905810488e-05,
+      "loss": 0.3685,
+      "mean_token_accuracy": 0.8635215103626251,
+      "num_tokens": 416165.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.1864500761032104,
+      "epoch": 0.3045186640471513,
+      "grad_norm": 2.192814588546753,
+      "learning_rate": 9.719991193293466e-05,
+      "loss": 0.4021,
+      "mean_token_accuracy": 0.8456959426403046,
+      "num_tokens": 429910.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.2050342679023742,
+      "epoch": 0.3143418467583497,
+      "grad_norm": 2.83247447013855,
+      "learning_rate": 9.692571053295268e-05,
+      "loss": 0.3991,
+      "mean_token_accuracy": 0.8514082193374634,
+      "num_tokens": 443725.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.1955341696739197,
+      "epoch": 0.3241650294695481,
+      "grad_norm": 2.9216184616088867,
+      "learning_rate": 9.663912720955637e-05,
+      "loss": 0.3896,
+      "mean_token_accuracy": 0.8574642062187194,
+      "num_tokens": 458234.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.1569963335990905,
+      "epoch": 0.33398821218074654,
+      "grad_norm": 3.420191764831543,
+      "learning_rate": 9.634023758126539e-05,
+      "loss": 0.3777,
+      "mean_token_accuracy": 0.8605042338371277,
+      "num_tokens": 471402.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.1522197484970094,
+      "epoch": 0.343811394891945,
+      "grad_norm": 2.807382106781006,
+      "learning_rate": 9.602912051376859e-05,
+      "loss": 0.3594,
+      "mean_token_accuracy": 0.8635930478572845,
+      "num_tokens": 484651.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.1738444805145263,
+      "epoch": 0.35363457760314343,
+      "grad_norm": 2.4855751991271973,
+      "learning_rate": 9.570585809911441e-05,
+      "loss": 0.3772,
+      "mean_token_accuracy": 0.8609638273715973,
+      "num_tokens": 497956.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.141674017906189,
+      "epoch": 0.36345776031434185,
+      "grad_norm": 2.3158438205718994,
+      "learning_rate": 9.537053563404992e-05,
+      "loss": 0.3792,
+      "mean_token_accuracy": 0.8546477794647217,
+      "num_tokens": 511717.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.1361984372138978,
+      "epoch": 0.37328094302554027,
+      "grad_norm": 3.798175811767578,
+      "learning_rate": 9.502324159751412e-05,
+      "loss": 0.3346,
+      "mean_token_accuracy": 0.87108034491539,
+      "num_tokens": 525633.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.1116561889648438,
+      "epoch": 0.3831041257367387,
+      "grad_norm": 2.64977765083313,
+      "learning_rate": 9.466406762729181e-05,
+      "loss": 0.3484,
+      "mean_token_accuracy": 0.8717014491558075,
+      "num_tokens": 538945.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.1471036911010741,
+      "epoch": 0.3929273084479371,
+      "grad_norm": 2.861079454421997,
+      "learning_rate": 9.429310849583378e-05,
+      "loss": 0.357,
+      "mean_token_accuracy": 0.8679845750331878,
+      "num_tokens": 552861.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.1682212829589844,
+      "epoch": 0.4027504911591356,
+      "grad_norm": 2.869887113571167,
+      "learning_rate": 9.391046208524989e-05,
+      "loss": 0.3822,
+      "mean_token_accuracy": 0.8555807173252106,
+      "num_tokens": 565955.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.1808563232421876,
+      "epoch": 0.412573673870334,
+      "grad_norm": 3.2115867137908936,
+      "learning_rate": 9.351622936148182e-05,
+      "loss": 0.348,
+      "mean_token_accuracy": 0.8685822308063507,
+      "num_tokens": 579923.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.1853391170501708,
+      "epoch": 0.4223968565815324,
+      "grad_norm": 3.3790035247802734,
+      "learning_rate": 9.31105143476618e-05,
+      "loss": 0.3512,
+      "mean_token_accuracy": 0.8625130891799927,
+      "num_tokens": 594071.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.1893272519111633,
+      "epoch": 0.43222003929273084,
+      "grad_norm": 2.8293139934539795,
+      "learning_rate": 9.269342409666495e-05,
+      "loss": 0.3294,
+      "mean_token_accuracy": 0.8698876202106476,
+      "num_tokens": 608545.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.1522011637687684,
+      "epoch": 0.44204322200392926,
+      "grad_norm": 2.4837608337402344,
+      "learning_rate": 9.226506866286202e-05,
+      "loss": 0.3343,
+      "mean_token_accuracy": 0.8736551582813263,
+      "num_tokens": 621975.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.175724446773529,
+      "epoch": 0.4518664047151277,
+      "grad_norm": 2.714533567428589,
+      "learning_rate": 9.18255610730802e-05,
+      "loss": 0.3636,
+      "mean_token_accuracy": 0.8584598302841187,
+      "num_tokens": 636061.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.155966317653656,
+      "epoch": 0.46168958742632615,
+      "grad_norm": 3.0229501724243164,
+      "learning_rate": 9.137501729677962e-05,
+      "loss": 0.3309,
+      "mean_token_accuracy": 0.870897525548935,
+      "num_tokens": 649717.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.1558407902717591,
+      "epoch": 0.4715127701375246,
+      "grad_norm": 2.826359510421753,
+      "learning_rate": 9.091355621545339e-05,
+      "loss": 0.3274,
+      "mean_token_accuracy": 0.8788529932498932,
+      "num_tokens": 664123.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.1772762537002563,
+      "epoch": 0.481335952848723,
+      "grad_norm": 3.5479748249053955,
+      "learning_rate": 9.044129959125923e-05,
+      "loss": 0.3586,
+      "mean_token_accuracy": 0.8606054663658143,
+      "num_tokens": 677803.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.1721865773200988,
+      "epoch": 0.4911591355599214,
+      "grad_norm": 2.7724850177764893,
+      "learning_rate": 8.995837203489108e-05,
+      "loss": 0.3049,
+      "mean_token_accuracy": 0.887111920118332,
+      "num_tokens": 691461.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.1805781960487365,
+      "epoch": 0.5009823182711198,
+      "grad_norm": 3.2204816341400146,
+      "learning_rate": 8.946490097269899e-05,
+      "loss": 0.331,
+      "mean_token_accuracy": 0.8687247633934021,
+      "num_tokens": 705470.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.1804481625556946,
+      "epoch": 0.5108055009823183,
+      "grad_norm": 2.8021044731140137,
+      "learning_rate": 8.896101661306618e-05,
+      "loss": 0.358,
+      "mean_token_accuracy": 0.8627409517765046,
+      "num_tokens": 718805.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.1632071733474731,
+      "epoch": 0.5206286836935167,
+      "grad_norm": 3.163684606552124,
+      "learning_rate": 8.844685191205193e-05,
+      "loss": 0.3303,
+      "mean_token_accuracy": 0.8740256905555726,
+      "num_tokens": 732367.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.146753692626953,
+      "epoch": 0.5304518664047151,
+      "grad_norm": 3.0881783962249756,
+      "learning_rate": 8.792254253830945e-05,
+      "loss": 0.3201,
+      "mean_token_accuracy": 0.8738628149032592,
+      "num_tokens": 746431.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.1356727361679078,
+      "epoch": 0.5402750491159135,
+      "grad_norm": 2.558445453643799,
+      "learning_rate": 8.738822683728814e-05,
+      "loss": 0.3057,
+      "mean_token_accuracy": 0.8832348883152008,
+      "num_tokens": 760182.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.158420741558075,
+      "epoch": 0.550098231827112,
+      "grad_norm": 2.895554780960083,
+      "learning_rate": 8.684404579472936e-05,
+      "loss": 0.3231,
+      "mean_token_accuracy": 0.875940489768982,
+      "num_tokens": 774137.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.1434678196907044,
+      "epoch": 0.5599214145383105,
+      "grad_norm": 2.8823940753936768,
+      "learning_rate": 8.629014299946568e-05,
+      "loss": 0.3088,
+      "mean_token_accuracy": 0.8810134768486023,
+      "num_tokens": 787806.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.1504598736763,
+      "epoch": 0.5697445972495089,
+      "grad_norm": 2.8560471534729004,
+      "learning_rate": 8.572666460553316e-05,
+      "loss": 0.3139,
+      "mean_token_accuracy": 0.87673499584198,
+      "num_tokens": 801431.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.1501794695854186,
+      "epoch": 0.5795677799607073,
+      "grad_norm": 2.4872987270355225,
+      "learning_rate": 8.515375929360687e-05,
+      "loss": 0.3312,
+      "mean_token_accuracy": 0.871477621793747,
+      "num_tokens": 815080.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.1453538179397582,
+      "epoch": 0.5893909626719057,
+      "grad_norm": 2.843007802963257,
+      "learning_rate": 8.45715782317696e-05,
+      "loss": 0.3141,
+      "mean_token_accuracy": 0.8785215854644776,
+      "num_tokens": 829061.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.150018286705017,
+      "epoch": 0.5992141453831041,
+      "grad_norm": 2.923928737640381,
+      "learning_rate": 8.39802750356244e-05,
+      "loss": 0.3247,
+      "mean_token_accuracy": 0.8796118080615998,
+      "num_tokens": 842325.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.1454686522483826,
+      "epoch": 0.6090373280943026,
+      "grad_norm": 3.4226765632629395,
+      "learning_rate": 8.338000572776114e-05,
+      "loss": 0.2962,
+      "mean_token_accuracy": 0.8821324348449707,
+      "num_tokens": 856377.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.1441982865333558,
+      "epoch": 0.618860510805501,
+      "grad_norm": 3.684279203414917,
+      "learning_rate": 8.277092869658808e-05,
+      "loss": 0.2841,
+      "mean_token_accuracy": 0.8887213349342347,
+      "num_tokens": 870502.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.149403488636017,
+      "epoch": 0.6286836935166994,
+      "grad_norm": 3.090017795562744,
+      "learning_rate": 8.215320465453914e-05,
+      "loss": 0.3164,
+      "mean_token_accuracy": 0.8771797001361847,
+      "num_tokens": 884708.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.1405172109603883,
+      "epoch": 0.6385068762278978,
+      "grad_norm": 2.682309865951538,
+      "learning_rate": 8.152699659566797e-05,
+      "loss": 0.2995,
+      "mean_token_accuracy": 0.8881269872188569,
+      "num_tokens": 898142.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.1579939365386962,
+      "epoch": 0.6483300589390962,
+      "grad_norm": 2.3700037002563477,
+      "learning_rate": 8.089246975263996e-05,
+      "loss": 0.3026,
+      "mean_token_accuracy": 0.8834354102611541,
+      "num_tokens": 912820.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.1454323649406433,
+      "epoch": 0.6581532416502947,
+      "grad_norm": 2.855104923248291,
+      "learning_rate": 8.024979155313354e-05,
+      "loss": 0.2915,
+      "mean_token_accuracy": 0.8860398828983307,
+      "num_tokens": 926399.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.153844177722931,
+      "epoch": 0.6679764243614931,
+      "grad_norm": 2.197305679321289,
+      "learning_rate": 7.959913157566246e-05,
+      "loss": 0.3125,
+      "mean_token_accuracy": 0.8814132511615753,
+      "num_tokens": 940133.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.1821880102157594,
+      "epoch": 0.6777996070726916,
+      "grad_norm": 3.007585048675537,
+      "learning_rate": 7.894066150483023e-05,
+      "loss": 0.3013,
+      "mean_token_accuracy": 0.8868993937969207,
+      "num_tokens": 954504.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.1575238585472107,
+      "epoch": 0.68762278978389,
+      "grad_norm": 3.8097307682037354,
+      "learning_rate": 7.827455508602923e-05,
+      "loss": 0.3326,
+      "mean_token_accuracy": 0.8661808252334595,
+      "num_tokens": 968934.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.1764493942260743,
+      "epoch": 0.6974459724950884,
+      "grad_norm": 2.2830970287323,
+      "learning_rate": 7.760098807959573e-05,
+      "loss": 0.3247,
+      "mean_token_accuracy": 0.8759588539600373,
+      "num_tokens": 983038.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.1495725274086,
+      "epoch": 0.7072691552062869,
+      "grad_norm": 4.025444507598877,
+      "learning_rate": 7.69201382144334e-05,
+      "loss": 0.3012,
+      "mean_token_accuracy": 0.8812366366386414,
+      "num_tokens": 996748.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.153698194026947,
+      "epoch": 0.7170923379174853,
+      "grad_norm": 3.110124349594116,
+      "learning_rate": 7.623218514111743e-05,
+      "loss": 0.2746,
+      "mean_token_accuracy": 0.8929417610168457,
+      "num_tokens": 1010595.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.1455257773399352,
+      "epoch": 0.7269155206286837,
+      "grad_norm": 3.1288959980010986,
+      "learning_rate": 7.553731038449144e-05,
+      "loss": 0.2743,
+      "mean_token_accuracy": 0.8920208811759949,
+      "num_tokens": 1024890.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.1425867199897766,
+      "epoch": 0.7367387033398821,
+      "grad_norm": 2.8785386085510254,
+      "learning_rate": 7.483569729576996e-05,
+      "loss": 0.2796,
+      "mean_token_accuracy": 0.8883347868919372,
+      "num_tokens": 1038967.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.1227909326553345,
+      "epoch": 0.7465618860510805,
+      "grad_norm": 3.789613962173462,
+      "learning_rate": 7.412753100415897e-05,
+      "loss": 0.2755,
+      "mean_token_accuracy": 0.895552146434784,
+      "num_tokens": 1052622.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.1423264741897583,
+      "epoch": 0.756385068762279,
+      "grad_norm": 2.6028270721435547,
+      "learning_rate": 7.341299836800729e-05,
+      "loss": 0.2978,
+      "mean_token_accuracy": 0.8876746118068695,
+      "num_tokens": 1066588.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.1372782826423644,
+      "epoch": 0.7662082514734774,
+      "grad_norm": 2.8547940254211426,
+      "learning_rate": 7.26922879255017e-05,
+      "loss": 0.2429,
+      "mean_token_accuracy": 0.9078093588352203,
+      "num_tokens": 1080453.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.1220194220542907,
+      "epoch": 0.7760314341846758,
+      "grad_norm": 3.2764084339141846,
+      "learning_rate": 7.196558984491889e-05,
+      "loss": 0.2837,
+      "mean_token_accuracy": 0.8842892706394195,
+      "num_tokens": 1094550.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.1319583415985108,
+      "epoch": 0.7858546168958742,
+      "grad_norm": 2.9916930198669434,
+      "learning_rate": 7.123309587444711e-05,
+      "loss": 0.2614,
+      "mean_token_accuracy": 0.8948858439922333,
+      "num_tokens": 1108808.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.12953040599823,
+      "epoch": 0.7956777996070727,
+      "grad_norm": 2.982774257659912,
+      "learning_rate": 7.04949992915912e-05,
+      "loss": 0.2481,
+      "mean_token_accuracy": 0.9044816195964813,
+      "num_tokens": 1122365.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.15300555229187,
+      "epoch": 0.8055009823182712,
+      "grad_norm": 4.228729248046875,
+      "learning_rate": 6.975149485217388e-05,
+      "loss": 0.2962,
+      "mean_token_accuracy": 0.8843261659145355,
+      "num_tokens": 1136376.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.134567415714264,
+      "epoch": 0.8153241650294696,
+      "grad_norm": 2.715785264968872,
+      "learning_rate": 6.900277873894705e-05,
+      "loss": 0.279,
+      "mean_token_accuracy": 0.885111278295517,
+      "num_tokens": 1150011.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.1585100889205933,
+      "epoch": 0.825147347740668,
+      "grad_norm": 2.9904732704162598,
+      "learning_rate": 6.824904850982656e-05,
+      "loss": 0.2709,
+      "mean_token_accuracy": 0.8954518556594848,
+      "num_tokens": 1164190.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.127516782283783,
+      "epoch": 0.8349705304518664,
+      "grad_norm": 3.0926408767700195,
+      "learning_rate": 6.749050304576415e-05,
+      "loss": 0.2609,
+      "mean_token_accuracy": 0.8978822529315948,
+      "num_tokens": 1177346.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.1291834115982056,
+      "epoch": 0.8447937131630648,
+      "grad_norm": 3.647355794906616,
+      "learning_rate": 6.672734249827028e-05,
+      "loss": 0.2634,
+      "mean_token_accuracy": 0.8949131548404694,
+      "num_tokens": 1190692.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.1479934215545655,
+      "epoch": 0.8546168958742633,
+      "grad_norm": 3.311383008956909,
+      "learning_rate": 6.59597682366017e-05,
+      "loss": 0.2817,
+      "mean_token_accuracy": 0.8871613144874573,
+      "num_tokens": 1204254.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.1448369145393371,
+      "epoch": 0.8644400785854617,
+      "grad_norm": 3.6365537643432617,
+      "learning_rate": 6.518798279462766e-05,
+      "loss": 0.3004,
+      "mean_token_accuracy": 0.884678715467453,
+      "num_tokens": 1217938.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.1474703669548034,
+      "epoch": 0.8742632612966601,
+      "grad_norm": 3.0901296138763428,
+      "learning_rate": 6.441218981738895e-05,
+      "loss": 0.2712,
+      "mean_token_accuracy": 0.8889893591403961,
+      "num_tokens": 1231589.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.1247360825538635,
+      "epoch": 0.8840864440078585,
+      "grad_norm": 3.096174955368042,
+      "learning_rate": 6.363259400736359e-05,
+      "loss": 0.2422,
+      "mean_token_accuracy": 0.9008595287799835,
+      "num_tokens": 1245755.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.1319254636764526,
+      "epoch": 0.8939096267190569,
+      "grad_norm": 3.762448310852051,
+      "learning_rate": 6.284940107045363e-05,
+      "loss": 0.2773,
+      "mean_token_accuracy": 0.8903194546699524,
+      "num_tokens": 1259830.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.1317768931388854,
+      "epoch": 0.9037328094302554,
+      "grad_norm": 3.124439001083374,
+      "learning_rate": 6.206281766170706e-05,
+      "loss": 0.2879,
+      "mean_token_accuracy": 0.8890141248703003,
+      "num_tokens": 1274186.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.119044828414917,
+      "epoch": 0.9135559921414538,
+      "grad_norm": 3.3181536197662354,
+      "learning_rate": 6.127305133078945e-05,
+      "loss": 0.2622,
+      "mean_token_accuracy": 0.8996863186359405,
+      "num_tokens": 1288175.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.1068704724311829,
+      "epoch": 0.9233791748526523,
+      "grad_norm": 3.2204999923706055,
+      "learning_rate": 6.048031046721931e-05,
+      "loss": 0.2564,
+      "mean_token_accuracy": 0.9012023687362671,
+      "num_tokens": 1301979.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.1291054964065552,
+      "epoch": 0.9332023575638507,
+      "grad_norm": 3.378843069076538,
+      "learning_rate": 5.9684804245382056e-05,
+      "loss": 0.2661,
+      "mean_token_accuracy": 0.9007695376873016,
+      "num_tokens": 1316012.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.1308410048484803,
+      "epoch": 0.9430255402750491,
+      "grad_norm": 3.883087158203125,
+      "learning_rate": 5.888674256933671e-05,
+      "loss": 0.2703,
+      "mean_token_accuracy": 0.8894614398479461,
+      "num_tokens": 1329560.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.1480537533760071,
+      "epoch": 0.9528487229862476,
+      "grad_norm": 3.8219106197357178,
+      "learning_rate": 5.80863360174302e-05,
+      "loss": 0.2831,
+      "mean_token_accuracy": 0.8840299725532532,
+      "num_tokens": 1343980.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.1379376173019409,
+      "epoch": 0.962671905697446,
+      "grad_norm": 3.180670738220215,
+      "learning_rate": 5.72837957867337e-05,
+      "loss": 0.2732,
+      "mean_token_accuracy": 0.892016738653183,
+      "num_tokens": 1358209.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.1167411923408508,
+      "epoch": 0.9724950884086444,
+      "grad_norm": 3.839327096939087,
+      "learning_rate": 5.64793336373156e-05,
+      "loss": 0.245,
+      "mean_token_accuracy": 0.9063532710075378,
+      "num_tokens": 1371263.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.1097952485084535,
+      "epoch": 0.9823182711198428,
+      "grad_norm": 2.7271931171417236,
+      "learning_rate": 5.567316183636615e-05,
+      "loss": 0.2097,
+      "mean_token_accuracy": 0.9184197664260865,
+      "num_tokens": 1385183.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.101745867729187,
+      "epoch": 0.9921414538310412,
+      "grad_norm": 3.3536264896392822,
+      "learning_rate": 5.4865493102188046e-05,
+      "loss": 0.2351,
+      "mean_token_accuracy": 0.9102523267269135,
+      "num_tokens": 1398309.0,
+      "step": 1010
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2036,
+  "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": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.993655431707238e+16,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-1018/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f17e40cb9b819e595ce9cfb7475915d568d94a96939c1b6061c3c206e0e5466b
+size 6417

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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 0.19.1

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "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,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "up_proj",
+    "k_proj",
+    "o_proj",
+    "gate_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4fb845581d8729c987c3765d661d5d5199387036613281ae705a2b928944caf
+size 10151357296

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/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": 2036,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.2313570261001587,
+      "epoch": 0.009823182711198428,
+      "grad_norm": 3.013684034347534,
+      "learning_rate": 8.823529411764707e-06,
+      "loss": 0.664,
+      "mean_token_accuracy": 0.7738851010799408,
+      "num_tokens": 13828.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.2726562738418579,
+      "epoch": 0.019646365422396856,
+      "grad_norm": 2.4101688861846924,
+      "learning_rate": 1.862745098039216e-05,
+      "loss": 0.6933,
+      "mean_token_accuracy": 0.7415895283222198,
+      "num_tokens": 28074.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.2336923837661744,
+      "epoch": 0.029469548133595286,
+      "grad_norm": 2.126495361328125,
+      "learning_rate": 2.8431372549019608e-05,
+      "loss": 0.6157,
+      "mean_token_accuracy": 0.7692714393138885,
+      "num_tokens": 42168.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.207659947872162,
+      "epoch": 0.03929273084479371,
+      "grad_norm": 2.472097635269165,
+      "learning_rate": 3.8235294117647055e-05,
+      "loss": 0.6141,
+      "mean_token_accuracy": 0.7780545473098754,
+      "num_tokens": 56700.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.188597857952118,
+      "epoch": 0.04911591355599214,
+      "grad_norm": 2.946054458618164,
+      "learning_rate": 4.803921568627452e-05,
+      "loss": 0.5251,
+      "mean_token_accuracy": 0.8047081351280212,
+      "num_tokens": 70922.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.1680904865264892,
+      "epoch": 0.05893909626719057,
+      "grad_norm": 2.1236400604248047,
+      "learning_rate": 5.784313725490197e-05,
+      "loss": 0.4678,
+      "mean_token_accuracy": 0.8220886826515198,
+      "num_tokens": 84911.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.1362838268280029,
+      "epoch": 0.068762278978389,
+      "grad_norm": 2.036947727203369,
+      "learning_rate": 6.764705882352942e-05,
+      "loss": 0.5103,
+      "mean_token_accuracy": 0.8159181356430054,
+      "num_tokens": 98835.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.1374945163726806,
+      "epoch": 0.07858546168958742,
+      "grad_norm": 3.3151774406433105,
+      "learning_rate": 7.745098039215687e-05,
+      "loss": 0.4837,
+      "mean_token_accuracy": 0.8214468002319336,
+      "num_tokens": 112502.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.1152778267860413,
+      "epoch": 0.08840864440078586,
+      "grad_norm": 2.0788350105285645,
+      "learning_rate": 8.725490196078432e-05,
+      "loss": 0.4882,
+      "mean_token_accuracy": 0.8231433868408203,
+      "num_tokens": 126248.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.0997976303100585,
+      "epoch": 0.09823182711198428,
+      "grad_norm": 2.2495064735412598,
+      "learning_rate": 9.705882352941177e-05,
+      "loss": 0.4921,
+      "mean_token_accuracy": 0.82340207695961,
+      "num_tokens": 139551.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.090124821662903,
+      "epoch": 0.10805500982318271,
+      "grad_norm": 2.597109794616699,
+      "learning_rate": 9.999676765122786e-05,
+      "loss": 0.4644,
+      "mean_token_accuracy": 0.8299567580223084,
+      "num_tokens": 152975.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.0939476013183593,
+      "epoch": 0.11787819253438114,
+      "grad_norm": 2.1519718170166016,
+      "learning_rate": 9.998093674494531e-05,
+      "loss": 0.4804,
+      "mean_token_accuracy": 0.8199364960193634,
+      "num_tokens": 167073.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.115808939933777,
+      "epoch": 0.12770137524557956,
+      "grad_norm": 2.4879584312438965,
+      "learning_rate": 9.995191775639568e-05,
+      "loss": 0.493,
+      "mean_token_accuracy": 0.8184783101081848,
+      "num_tokens": 181404.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.096212136745453,
+      "epoch": 0.137524557956778,
+      "grad_norm": 2.645519971847534,
+      "learning_rate": 9.990971834259444e-05,
+      "loss": 0.4522,
+      "mean_token_accuracy": 0.832886415719986,
+      "num_tokens": 195441.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.1273761987686157,
+      "epoch": 0.14734774066797643,
+      "grad_norm": 3.033703088760376,
+      "learning_rate": 9.985434963837376e-05,
+      "loss": 0.4945,
+      "mean_token_accuracy": 0.8159543931484222,
+      "num_tokens": 209863.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.088386845588684,
+      "epoch": 0.15717092337917485,
+      "grad_norm": 3.4639716148376465,
+      "learning_rate": 9.978582625344435e-05,
+      "loss": 0.4563,
+      "mean_token_accuracy": 0.8262679219245911,
+      "num_tokens": 223646.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.0982281684875488,
+      "epoch": 0.16699410609037327,
+      "grad_norm": 2.624499559402466,
+      "learning_rate": 9.970416626854049e-05,
+      "loss": 0.429,
+      "mean_token_accuracy": 0.8403248071670533,
+      "num_tokens": 237587.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.090433120727539,
+      "epoch": 0.17681728880157171,
+      "grad_norm": 2.337232828140259,
+      "learning_rate": 9.960939123064929e-05,
+      "loss": 0.4589,
+      "mean_token_accuracy": 0.8313010632991791,
+      "num_tokens": 251760.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.081747841835022,
+      "epoch": 0.18664047151277013,
+      "grad_norm": 2.338900566101074,
+      "learning_rate": 9.950152614732518e-05,
+      "loss": 0.4116,
+      "mean_token_accuracy": 0.8467560648918152,
+      "num_tokens": 265417.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.113021147251129,
+      "epoch": 0.19646365422396855,
+      "grad_norm": 3.0505051612854004,
+      "learning_rate": 9.93805994800914e-05,
+      "loss": 0.4454,
+      "mean_token_accuracy": 0.8330788910388947,
+      "num_tokens": 278540.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.1093911409378052,
+      "epoch": 0.206286836935167,
+      "grad_norm": 2.8799140453338623,
+      "learning_rate": 9.924664313693006e-05,
+      "loss": 0.4281,
+      "mean_token_accuracy": 0.8332679152488709,
+      "num_tokens": 292295.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.098515558242798,
+      "epoch": 0.21611001964636542,
+      "grad_norm": 2.6466057300567627,
+      "learning_rate": 9.909969246386284e-05,
+      "loss": 0.423,
+      "mean_token_accuracy": 0.8387296140193939,
+      "num_tokens": 305466.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.1375373482704163,
+      "epoch": 0.22593320235756384,
+      "grad_norm": 2.6465585231781006,
+      "learning_rate": 9.893978623562453e-05,
+      "loss": 0.421,
+      "mean_token_accuracy": 0.846423614025116,
+      "num_tokens": 320495.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.0984365344047546,
+      "epoch": 0.2357563850687623,
+      "grad_norm": 2.297790050506592,
+      "learning_rate": 9.876696664543177e-05,
+      "loss": 0.4089,
+      "mean_token_accuracy": 0.8428103208541871,
+      "num_tokens": 333901.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.1255406141281128,
+      "epoch": 0.2455795677799607,
+      "grad_norm": 3.155524730682373,
+      "learning_rate": 9.858127929384991e-05,
+      "loss": 0.3817,
+      "mean_token_accuracy": 0.8513313829898834,
+      "num_tokens": 347392.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.1191080331802368,
+      "epoch": 0.2554027504911591,
+      "grad_norm": 2.4332103729248047,
+      "learning_rate": 9.838277317676077e-05,
+      "loss": 0.3845,
+      "mean_token_accuracy": 0.8584017932415009,
+      "num_tokens": 361140.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.1573356509208679,
+      "epoch": 0.26522593320235754,
+      "grad_norm": 2.641416072845459,
+      "learning_rate": 9.817150067243438e-05,
+      "loss": 0.3562,
+      "mean_token_accuracy": 0.8600760519504547,
+      "num_tokens": 374661.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.1518091678619384,
+      "epoch": 0.275049115913556,
+      "grad_norm": 2.4803318977355957,
+      "learning_rate": 9.794751752770845e-05,
+      "loss": 0.3904,
+      "mean_token_accuracy": 0.8498182654380798,
+      "num_tokens": 388415.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.1327571034431458,
+      "epoch": 0.28487229862475444,
+      "grad_norm": 2.8583028316497803,
+      "learning_rate": 9.771088284327879e-05,
+      "loss": 0.4015,
+      "mean_token_accuracy": 0.8473475515842438,
+      "num_tokens": 402031.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.1331970334053039,
+      "epoch": 0.29469548133595286,
+      "grad_norm": 3.086913585662842,
+      "learning_rate": 9.746165905810488e-05,
+      "loss": 0.3685,
+      "mean_token_accuracy": 0.8635215103626251,
+      "num_tokens": 416165.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.1864500761032104,
+      "epoch": 0.3045186640471513,
+      "grad_norm": 2.192814588546753,
+      "learning_rate": 9.719991193293466e-05,
+      "loss": 0.4021,
+      "mean_token_accuracy": 0.8456959426403046,
+      "num_tokens": 429910.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.2050342679023742,
+      "epoch": 0.3143418467583497,
+      "grad_norm": 2.83247447013855,
+      "learning_rate": 9.692571053295268e-05,
+      "loss": 0.3991,
+      "mean_token_accuracy": 0.8514082193374634,
+      "num_tokens": 443725.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.1955341696739197,
+      "epoch": 0.3241650294695481,
+      "grad_norm": 2.9216184616088867,
+      "learning_rate": 9.663912720955637e-05,
+      "loss": 0.3896,
+      "mean_token_accuracy": 0.8574642062187194,
+      "num_tokens": 458234.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.1569963335990905,
+      "epoch": 0.33398821218074654,
+      "grad_norm": 3.420191764831543,
+      "learning_rate": 9.634023758126539e-05,
+      "loss": 0.3777,
+      "mean_token_accuracy": 0.8605042338371277,
+      "num_tokens": 471402.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.1522197484970094,
+      "epoch": 0.343811394891945,
+      "grad_norm": 2.807382106781006,
+      "learning_rate": 9.602912051376859e-05,
+      "loss": 0.3594,
+      "mean_token_accuracy": 0.8635930478572845,
+      "num_tokens": 484651.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.1738444805145263,
+      "epoch": 0.35363457760314343,
+      "grad_norm": 2.4855751991271973,
+      "learning_rate": 9.570585809911441e-05,
+      "loss": 0.3772,
+      "mean_token_accuracy": 0.8609638273715973,
+      "num_tokens": 497956.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.141674017906189,
+      "epoch": 0.36345776031434185,
+      "grad_norm": 2.3158438205718994,
+      "learning_rate": 9.537053563404992e-05,
+      "loss": 0.3792,
+      "mean_token_accuracy": 0.8546477794647217,
+      "num_tokens": 511717.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.1361984372138978,
+      "epoch": 0.37328094302554027,
+      "grad_norm": 3.798175811767578,
+      "learning_rate": 9.502324159751412e-05,
+      "loss": 0.3346,
+      "mean_token_accuracy": 0.87108034491539,
+      "num_tokens": 525633.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.1116561889648438,
+      "epoch": 0.3831041257367387,
+      "grad_norm": 2.64977765083313,
+      "learning_rate": 9.466406762729181e-05,
+      "loss": 0.3484,
+      "mean_token_accuracy": 0.8717014491558075,
+      "num_tokens": 538945.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.1471036911010741,
+      "epoch": 0.3929273084479371,
+      "grad_norm": 2.861079454421997,
+      "learning_rate": 9.429310849583378e-05,
+      "loss": 0.357,
+      "mean_token_accuracy": 0.8679845750331878,
+      "num_tokens": 552861.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.1682212829589844,
+      "epoch": 0.4027504911591356,
+      "grad_norm": 2.869887113571167,
+      "learning_rate": 9.391046208524989e-05,
+      "loss": 0.3822,
+      "mean_token_accuracy": 0.8555807173252106,
+      "num_tokens": 565955.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.1808563232421876,
+      "epoch": 0.412573673870334,
+      "grad_norm": 3.2115867137908936,
+      "learning_rate": 9.351622936148182e-05,
+      "loss": 0.348,
+      "mean_token_accuracy": 0.8685822308063507,
+      "num_tokens": 579923.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.1853391170501708,
+      "epoch": 0.4223968565815324,
+      "grad_norm": 3.3790035247802734,
+      "learning_rate": 9.31105143476618e-05,
+      "loss": 0.3512,
+      "mean_token_accuracy": 0.8625130891799927,
+      "num_tokens": 594071.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.1893272519111633,
+      "epoch": 0.43222003929273084,
+      "grad_norm": 2.8293139934539795,
+      "learning_rate": 9.269342409666495e-05,
+      "loss": 0.3294,
+      "mean_token_accuracy": 0.8698876202106476,
+      "num_tokens": 608545.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.1522011637687684,
+      "epoch": 0.44204322200392926,
+      "grad_norm": 2.4837608337402344,
+      "learning_rate": 9.226506866286202e-05,
+      "loss": 0.3343,
+      "mean_token_accuracy": 0.8736551582813263,
+      "num_tokens": 621975.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.175724446773529,
+      "epoch": 0.4518664047151277,
+      "grad_norm": 2.714533567428589,
+      "learning_rate": 9.18255610730802e-05,
+      "loss": 0.3636,
+      "mean_token_accuracy": 0.8584598302841187,
+      "num_tokens": 636061.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.155966317653656,
+      "epoch": 0.46168958742632615,
+      "grad_norm": 3.0229501724243164,
+      "learning_rate": 9.137501729677962e-05,
+      "loss": 0.3309,
+      "mean_token_accuracy": 0.870897525548935,
+      "num_tokens": 649717.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.1558407902717591,
+      "epoch": 0.4715127701375246,
+      "grad_norm": 2.826359510421753,
+      "learning_rate": 9.091355621545339e-05,
+      "loss": 0.3274,
+      "mean_token_accuracy": 0.8788529932498932,
+      "num_tokens": 664123.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.1772762537002563,
+      "epoch": 0.481335952848723,
+      "grad_norm": 3.5479748249053955,
+      "learning_rate": 9.044129959125923e-05,
+      "loss": 0.3586,
+      "mean_token_accuracy": 0.8606054663658143,
+      "num_tokens": 677803.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.1721865773200988,
+      "epoch": 0.4911591355599214,
+      "grad_norm": 2.7724850177764893,
+      "learning_rate": 8.995837203489108e-05,
+      "loss": 0.3049,
+      "mean_token_accuracy": 0.887111920118332,
+      "num_tokens": 691461.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.1805781960487365,
+      "epoch": 0.5009823182711198,
+      "grad_norm": 3.2204816341400146,
+      "learning_rate": 8.946490097269899e-05,
+      "loss": 0.331,
+      "mean_token_accuracy": 0.8687247633934021,
+      "num_tokens": 705470.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.1804481625556946,
+      "epoch": 0.5108055009823183,
+      "grad_norm": 2.8021044731140137,
+      "learning_rate": 8.896101661306618e-05,
+      "loss": 0.358,
+      "mean_token_accuracy": 0.8627409517765046,
+      "num_tokens": 718805.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.1632071733474731,
+      "epoch": 0.5206286836935167,
+      "grad_norm": 3.163684606552124,
+      "learning_rate": 8.844685191205193e-05,
+      "loss": 0.3303,
+      "mean_token_accuracy": 0.8740256905555726,
+      "num_tokens": 732367.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.146753692626953,
+      "epoch": 0.5304518664047151,
+      "grad_norm": 3.0881783962249756,
+      "learning_rate": 8.792254253830945e-05,
+      "loss": 0.3201,
+      "mean_token_accuracy": 0.8738628149032592,
+      "num_tokens": 746431.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.1356727361679078,
+      "epoch": 0.5402750491159135,
+      "grad_norm": 2.558445453643799,
+      "learning_rate": 8.738822683728814e-05,
+      "loss": 0.3057,
+      "mean_token_accuracy": 0.8832348883152008,
+      "num_tokens": 760182.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.158420741558075,
+      "epoch": 0.550098231827112,
+      "grad_norm": 2.895554780960083,
+      "learning_rate": 8.684404579472936e-05,
+      "loss": 0.3231,
+      "mean_token_accuracy": 0.875940489768982,
+      "num_tokens": 774137.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.1434678196907044,
+      "epoch": 0.5599214145383105,
+      "grad_norm": 2.8823940753936768,
+      "learning_rate": 8.629014299946568e-05,
+      "loss": 0.3088,
+      "mean_token_accuracy": 0.8810134768486023,
+      "num_tokens": 787806.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.1504598736763,
+      "epoch": 0.5697445972495089,
+      "grad_norm": 2.8560471534729004,
+      "learning_rate": 8.572666460553316e-05,
+      "loss": 0.3139,
+      "mean_token_accuracy": 0.87673499584198,
+      "num_tokens": 801431.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.1501794695854186,
+      "epoch": 0.5795677799607073,
+      "grad_norm": 2.4872987270355225,
+      "learning_rate": 8.515375929360687e-05,
+      "loss": 0.3312,
+      "mean_token_accuracy": 0.871477621793747,
+      "num_tokens": 815080.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.1453538179397582,
+      "epoch": 0.5893909626719057,
+      "grad_norm": 2.843007802963257,
+      "learning_rate": 8.45715782317696e-05,
+      "loss": 0.3141,
+      "mean_token_accuracy": 0.8785215854644776,
+      "num_tokens": 829061.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.150018286705017,
+      "epoch": 0.5992141453831041,
+      "grad_norm": 2.923928737640381,
+      "learning_rate": 8.39802750356244e-05,
+      "loss": 0.3247,
+      "mean_token_accuracy": 0.8796118080615998,
+      "num_tokens": 842325.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.1454686522483826,
+      "epoch": 0.6090373280943026,
+      "grad_norm": 3.4226765632629395,
+      "learning_rate": 8.338000572776114e-05,
+      "loss": 0.2962,
+      "mean_token_accuracy": 0.8821324348449707,
+      "num_tokens": 856377.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.1441982865333558,
+      "epoch": 0.618860510805501,
+      "grad_norm": 3.684279203414917,
+      "learning_rate": 8.277092869658808e-05,
+      "loss": 0.2841,
+      "mean_token_accuracy": 0.8887213349342347,
+      "num_tokens": 870502.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.149403488636017,
+      "epoch": 0.6286836935166994,
+      "grad_norm": 3.090017795562744,
+      "learning_rate": 8.215320465453914e-05,
+      "loss": 0.3164,
+      "mean_token_accuracy": 0.8771797001361847,
+      "num_tokens": 884708.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.1405172109603883,
+      "epoch": 0.6385068762278978,
+      "grad_norm": 2.682309865951538,
+      "learning_rate": 8.152699659566797e-05,
+      "loss": 0.2995,
+      "mean_token_accuracy": 0.8881269872188569,
+      "num_tokens": 898142.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.1579939365386962,
+      "epoch": 0.6483300589390962,
+      "grad_norm": 2.3700037002563477,
+      "learning_rate": 8.089246975263996e-05,
+      "loss": 0.3026,
+      "mean_token_accuracy": 0.8834354102611541,
+      "num_tokens": 912820.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.1454323649406433,
+      "epoch": 0.6581532416502947,
+      "grad_norm": 2.855104923248291,
+      "learning_rate": 8.024979155313354e-05,
+      "loss": 0.2915,
+      "mean_token_accuracy": 0.8860398828983307,
+      "num_tokens": 926399.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.153844177722931,
+      "epoch": 0.6679764243614931,
+      "grad_norm": 2.197305679321289,
+      "learning_rate": 7.959913157566246e-05,
+      "loss": 0.3125,
+      "mean_token_accuracy": 0.8814132511615753,
+      "num_tokens": 940133.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.1821880102157594,
+      "epoch": 0.6777996070726916,
+      "grad_norm": 3.007585048675537,
+      "learning_rate": 7.894066150483023e-05,
+      "loss": 0.3013,
+      "mean_token_accuracy": 0.8868993937969207,
+      "num_tokens": 954504.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.1575238585472107,
+      "epoch": 0.68762278978389,
+      "grad_norm": 3.8097307682037354,
+      "learning_rate": 7.827455508602923e-05,
+      "loss": 0.3326,
+      "mean_token_accuracy": 0.8661808252334595,
+      "num_tokens": 968934.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.1764493942260743,
+      "epoch": 0.6974459724950884,
+      "grad_norm": 2.2830970287323,
+      "learning_rate": 7.760098807959573e-05,
+      "loss": 0.3247,
+      "mean_token_accuracy": 0.8759588539600373,
+      "num_tokens": 983038.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.1495725274086,
+      "epoch": 0.7072691552062869,
+      "grad_norm": 4.025444507598877,
+      "learning_rate": 7.69201382144334e-05,
+      "loss": 0.3012,
+      "mean_token_accuracy": 0.8812366366386414,
+      "num_tokens": 996748.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.153698194026947,
+      "epoch": 0.7170923379174853,
+      "grad_norm": 3.110124349594116,
+      "learning_rate": 7.623218514111743e-05,
+      "loss": 0.2746,
+      "mean_token_accuracy": 0.8929417610168457,
+      "num_tokens": 1010595.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.1455257773399352,
+      "epoch": 0.7269155206286837,
+      "grad_norm": 3.1288959980010986,
+      "learning_rate": 7.553731038449144e-05,
+      "loss": 0.2743,
+      "mean_token_accuracy": 0.8920208811759949,
+      "num_tokens": 1024890.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.1425867199897766,
+      "epoch": 0.7367387033398821,
+      "grad_norm": 2.8785386085510254,
+      "learning_rate": 7.483569729576996e-05,
+      "loss": 0.2796,
+      "mean_token_accuracy": 0.8883347868919372,
+      "num_tokens": 1038967.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.1227909326553345,
+      "epoch": 0.7465618860510805,
+      "grad_norm": 3.789613962173462,
+      "learning_rate": 7.412753100415897e-05,
+      "loss": 0.2755,
+      "mean_token_accuracy": 0.895552146434784,
+      "num_tokens": 1052622.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.1423264741897583,
+      "epoch": 0.756385068762279,
+      "grad_norm": 2.6028270721435547,
+      "learning_rate": 7.341299836800729e-05,
+      "loss": 0.2978,
+      "mean_token_accuracy": 0.8876746118068695,
+      "num_tokens": 1066588.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.1372782826423644,
+      "epoch": 0.7662082514734774,
+      "grad_norm": 2.8547940254211426,
+      "learning_rate": 7.26922879255017e-05,
+      "loss": 0.2429,
+      "mean_token_accuracy": 0.9078093588352203,
+      "num_tokens": 1080453.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.1220194220542907,
+      "epoch": 0.7760314341846758,
+      "grad_norm": 3.2764084339141846,
+      "learning_rate": 7.196558984491889e-05,
+      "loss": 0.2837,
+      "mean_token_accuracy": 0.8842892706394195,
+      "num_tokens": 1094550.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.1319583415985108,
+      "epoch": 0.7858546168958742,
+      "grad_norm": 2.9916930198669434,
+      "learning_rate": 7.123309587444711e-05,
+      "loss": 0.2614,
+      "mean_token_accuracy": 0.8948858439922333,
+      "num_tokens": 1108808.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.12953040599823,
+      "epoch": 0.7956777996070727,
+      "grad_norm": 2.982774257659912,
+      "learning_rate": 7.04949992915912e-05,
+      "loss": 0.2481,
+      "mean_token_accuracy": 0.9044816195964813,
+      "num_tokens": 1122365.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.15300555229187,
+      "epoch": 0.8055009823182712,
+      "grad_norm": 4.228729248046875,
+      "learning_rate": 6.975149485217388e-05,
+      "loss": 0.2962,
+      "mean_token_accuracy": 0.8843261659145355,
+      "num_tokens": 1136376.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.134567415714264,
+      "epoch": 0.8153241650294696,
+      "grad_norm": 2.715785264968872,
+      "learning_rate": 6.900277873894705e-05,
+      "loss": 0.279,
+      "mean_token_accuracy": 0.885111278295517,
+      "num_tokens": 1150011.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.1585100889205933,
+      "epoch": 0.825147347740668,
+      "grad_norm": 2.9904732704162598,
+      "learning_rate": 6.824904850982656e-05,
+      "loss": 0.2709,
+      "mean_token_accuracy": 0.8954518556594848,
+      "num_tokens": 1164190.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.127516782283783,
+      "epoch": 0.8349705304518664,
+      "grad_norm": 3.0926408767700195,
+      "learning_rate": 6.749050304576415e-05,
+      "loss": 0.2609,
+      "mean_token_accuracy": 0.8978822529315948,
+      "num_tokens": 1177346.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.1291834115982056,
+      "epoch": 0.8447937131630648,
+      "grad_norm": 3.647355794906616,
+      "learning_rate": 6.672734249827028e-05,
+      "loss": 0.2634,
+      "mean_token_accuracy": 0.8949131548404694,
+      "num_tokens": 1190692.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.1479934215545655,
+      "epoch": 0.8546168958742633,
+      "grad_norm": 3.311383008956909,
+      "learning_rate": 6.59597682366017e-05,
+      "loss": 0.2817,
+      "mean_token_accuracy": 0.8871613144874573,
+      "num_tokens": 1204254.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.1448369145393371,
+      "epoch": 0.8644400785854617,
+      "grad_norm": 3.6365537643432617,
+      "learning_rate": 6.518798279462766e-05,
+      "loss": 0.3004,
+      "mean_token_accuracy": 0.884678715467453,
+      "num_tokens": 1217938.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.1474703669548034,
+      "epoch": 0.8742632612966601,
+      "grad_norm": 3.0901296138763428,
+      "learning_rate": 6.441218981738895e-05,
+      "loss": 0.2712,
+      "mean_token_accuracy": 0.8889893591403961,
+      "num_tokens": 1231589.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.1247360825538635,
+      "epoch": 0.8840864440078585,
+      "grad_norm": 3.096174955368042,
+      "learning_rate": 6.363259400736359e-05,
+      "loss": 0.2422,
+      "mean_token_accuracy": 0.9008595287799835,
+      "num_tokens": 1245755.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.1319254636764526,
+      "epoch": 0.8939096267190569,
+      "grad_norm": 3.762448310852051,
+      "learning_rate": 6.284940107045363e-05,
+      "loss": 0.2773,
+      "mean_token_accuracy": 0.8903194546699524,
+      "num_tokens": 1259830.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.1317768931388854,
+      "epoch": 0.9037328094302554,
+      "grad_norm": 3.124439001083374,
+      "learning_rate": 6.206281766170706e-05,
+      "loss": 0.2879,
+      "mean_token_accuracy": 0.8890141248703003,
+      "num_tokens": 1274186.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.119044828414917,
+      "epoch": 0.9135559921414538,
+      "grad_norm": 3.3181536197662354,
+      "learning_rate": 6.127305133078945e-05,
+      "loss": 0.2622,
+      "mean_token_accuracy": 0.8996863186359405,
+      "num_tokens": 1288175.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.1068704724311829,
+      "epoch": 0.9233791748526523,
+      "grad_norm": 3.2204999923706055,
+      "learning_rate": 6.048031046721931e-05,
+      "loss": 0.2564,
+      "mean_token_accuracy": 0.9012023687362671,
+      "num_tokens": 1301979.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.1291054964065552,
+      "epoch": 0.9332023575638507,
+      "grad_norm": 3.378843069076538,
+      "learning_rate": 5.9684804245382056e-05,
+      "loss": 0.2661,
+      "mean_token_accuracy": 0.9007695376873016,
+      "num_tokens": 1316012.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.1308410048484803,
+      "epoch": 0.9430255402750491,
+      "grad_norm": 3.883087158203125,
+      "learning_rate": 5.888674256933671e-05,
+      "loss": 0.2703,
+      "mean_token_accuracy": 0.8894614398479461,
+      "num_tokens": 1329560.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.1480537533760071,
+      "epoch": 0.9528487229862476,
+      "grad_norm": 3.8219106197357178,
+      "learning_rate": 5.80863360174302e-05,
+      "loss": 0.2831,
+      "mean_token_accuracy": 0.8840299725532532,
+      "num_tokens": 1343980.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.1379376173019409,
+      "epoch": 0.962671905697446,
+      "grad_norm": 3.180670738220215,
+      "learning_rate": 5.72837957867337e-05,
+      "loss": 0.2732,
+      "mean_token_accuracy": 0.892016738653183,
+      "num_tokens": 1358209.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.1167411923408508,
+      "epoch": 0.9724950884086444,
+      "grad_norm": 3.839327096939087,
+      "learning_rate": 5.64793336373156e-05,
+      "loss": 0.245,
+      "mean_token_accuracy": 0.9063532710075378,
+      "num_tokens": 1371263.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.1097952485084535,
+      "epoch": 0.9823182711198428,
+      "grad_norm": 2.7271931171417236,
+      "learning_rate": 5.567316183636615e-05,
+      "loss": 0.2097,
+      "mean_token_accuracy": 0.9184197664260865,
+      "num_tokens": 1385183.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.101745867729187,
+      "epoch": 0.9921414538310412,
+      "grad_norm": 3.3536264896392822,
+      "learning_rate": 5.4865493102188046e-05,
+      "loss": 0.2351,
+      "mean_token_accuracy": 0.9102523267269135,
+      "num_tokens": 1398309.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.0957032442092896,
+      "epoch": 1.0019646365422397,
+      "grad_norm": 2.934413194656372,
+      "learning_rate": 5.405654054806817e-05,
+      "loss": 0.232,
+      "mean_token_accuracy": 0.9149511158466339,
+      "num_tokens": 1411126.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.095172679424286,
+      "epoch": 1.0117878192534382,
+      "grad_norm": 3.4171221256256104,
+      "learning_rate": 5.3246517626045025e-05,
+      "loss": 0.1557,
+      "mean_token_accuracy": 0.9391077101230622,
+      "num_tokens": 1425582.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.0771167516708373,
+      "epoch": 1.0216110019646365,
+      "grad_norm": 2.5150020122528076,
+      "learning_rate": 5.243563807058676e-05,
+      "loss": 0.1603,
+      "mean_token_accuracy": 0.9438194215297699,
+      "num_tokens": 1439527.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.0743562698364257,
+      "epoch": 1.031434184675835,
+      "grad_norm": 3.111265182495117,
+      "learning_rate": 5.162411584219476e-05,
+      "loss": 0.1592,
+      "mean_token_accuracy": 0.9401110231876373,
+      "num_tokens": 1453445.0,
+      "step": 1050
+    },
+    {
+      "entropy": 1.0684488534927368,
+      "epoch": 1.0412573673870333,
+      "grad_norm": 3.062284469604492,
+      "learning_rate": 5.0812165070947594e-05,
+      "loss": 0.1591,
+      "mean_token_accuracy": 0.9385462582111359,
+      "num_tokens": 1466353.0,
+      "step": 1060
+    },
+    {
+      "entropy": 1.0825814723968505,
+      "epoch": 1.0510805500982319,
+      "grad_norm": 3.538163661956787,
+      "learning_rate": 5e-05,
+      "loss": 0.155,
+      "mean_token_accuracy": 0.939586591720581,
+      "num_tokens": 1481062.0,
+      "step": 1070
+    },
+    {
+      "entropy": 1.067275083065033,
+      "epoch": 1.0609037328094302,
+      "grad_norm": 2.323869466781616,
+      "learning_rate": 4.918783492905242e-05,
+      "loss": 0.1512,
+      "mean_token_accuracy": 0.9418384611606598,
+      "num_tokens": 1495189.0,
+      "step": 1080
+    },
+    {
+      "entropy": 1.0581106305122376,
+      "epoch": 1.0707269155206287,
+      "grad_norm": 3.1193525791168213,
+      "learning_rate": 4.837588415780524e-05,
+      "loss": 0.1694,
+      "mean_token_accuracy": 0.9371399760246277,
+      "num_tokens": 1508841.0,
+      "step": 1090
+    },
+    {
+      "entropy": 1.060888397693634,
+      "epoch": 1.080550098231827,
+      "grad_norm": 2.8776843547821045,
+      "learning_rate": 4.756436192941325e-05,
+      "loss": 0.1611,
+      "mean_token_accuracy": 0.9370385408401489,
+      "num_tokens": 1522922.0,
+      "step": 1100
+    },
+    {
+      "entropy": 1.0644820213317872,
+      "epoch": 1.0903732809430255,
+      "grad_norm": 4.273116588592529,
+      "learning_rate": 4.675348237395499e-05,
+      "loss": 0.1536,
+      "mean_token_accuracy": 0.9475546419620514,
+      "num_tokens": 1536867.0,
+      "step": 1110
+    },
+    {
+      "entropy": 1.0432558119297028,
+      "epoch": 1.1001964636542239,
+      "grad_norm": 3.0451886653900146,
+      "learning_rate": 4.594345945193183e-05,
+      "loss": 0.1447,
+      "mean_token_accuracy": 0.945958137512207,
+      "num_tokens": 1550230.0,
+      "step": 1120
+    },
+    {
+      "entropy": 1.0435200691223145,
+      "epoch": 1.1100196463654224,
+      "grad_norm": 3.7260093688964844,
+      "learning_rate": 4.513450689781195e-05,
+      "loss": 0.1692,
+      "mean_token_accuracy": 0.9362512767314911,
+      "num_tokens": 1562941.0,
+      "step": 1130
+    },
+    {
+      "entropy": 1.0448999643325805,
+      "epoch": 1.119842829076621,
+      "grad_norm": 3.5143847465515137,
+      "learning_rate": 4.432683816363386e-05,
+      "loss": 0.1556,
+      "mean_token_accuracy": 0.9417150497436524,
+      "num_tokens": 1576546.0,
+      "step": 1140
+    },
+    {
+      "entropy": 1.0413883328437805,
+      "epoch": 1.1296660117878192,
+      "grad_norm": 3.295367956161499,
+      "learning_rate": 4.3520666362684395e-05,
+      "loss": 0.1486,
+      "mean_token_accuracy": 0.9450748562812805,
+      "num_tokens": 1590129.0,
+      "step": 1150
+    },
+    {
+      "entropy": 1.0273713648319245,
+      "epoch": 1.1394891944990178,
+      "grad_norm": 3.81969952583313,
+      "learning_rate": 4.271620421326633e-05,
+      "loss": 0.1272,
+      "mean_token_accuracy": 0.953050822019577,
+      "num_tokens": 1603834.0,
+      "step": 1160
+    },
+    {
+      "entropy": 1.03249072432518,
+      "epoch": 1.149312377210216,
+      "grad_norm": 3.3045296669006348,
+      "learning_rate": 4.191366398256983e-05,
+      "loss": 0.137,
+      "mean_token_accuracy": 0.9506871163845062,
+      "num_tokens": 1617951.0,
+      "step": 1170
+    },
+    {
+      "entropy": 1.0293985426425933,
+      "epoch": 1.1591355599214146,
+      "grad_norm": 3.2096517086029053,
+      "learning_rate": 4.111325743066332e-05,
+      "loss": 0.1521,
+      "mean_token_accuracy": 0.9413591206073761,
+      "num_tokens": 1631987.0,
+      "step": 1180
+    },
+    {
+      "entropy": 1.017678964138031,
+      "epoch": 1.168958742632613,
+      "grad_norm": 3.346611738204956,
+      "learning_rate": 4.0315195754617976e-05,
+      "loss": 0.1494,
+      "mean_token_accuracy": 0.9440071761608124,
+      "num_tokens": 1645088.0,
+      "step": 1190
+    },
+    {
+      "entropy": 1.048347294330597,
+      "epoch": 1.1787819253438114,
+      "grad_norm": 2.886251926422119,
+      "learning_rate": 3.95196895327807e-05,
+      "loss": 0.1459,
+      "mean_token_accuracy": 0.9457695484161377,
+      "num_tokens": 1659227.0,
+      "step": 1200
+    },
+    {
+      "entropy": 1.0099494814872743,
+      "epoch": 1.1886051080550097,
+      "grad_norm": 3.0378425121307373,
+      "learning_rate": 3.872694866921056e-05,
+      "loss": 0.1388,
+      "mean_token_accuracy": 0.9459509491920471,
+      "num_tokens": 1672545.0,
+      "step": 1210
+    },
+    {
+      "entropy": 1.0176921665668488,
+      "epoch": 1.1984282907662083,
+      "grad_norm": 3.3476524353027344,
+      "learning_rate": 3.793718233829295e-05,
+      "loss": 0.1164,
+      "mean_token_accuracy": 0.9561141788959503,
+      "num_tokens": 1686702.0,
+      "step": 1220
+    },
+    {
+      "entropy": 1.0367235898971559,
+      "epoch": 1.2082514734774068,
+      "grad_norm": 2.8370447158813477,
+      "learning_rate": 3.715059892954639e-05,
+      "loss": 0.1365,
+      "mean_token_accuracy": 0.9428559362888336,
+      "num_tokens": 1701509.0,
+      "step": 1230
+    },
+    {
+      "entropy": 1.0246037125587464,
+      "epoch": 1.218074656188605,
+      "grad_norm": 3.86641526222229,
+      "learning_rate": 3.636740599263643e-05,
+      "loss": 0.1394,
+      "mean_token_accuracy": 0.944771534204483,
+      "num_tokens": 1715330.0,
+      "step": 1240
+    },
+    {
+      "entropy": 1.0229672968387604,
+      "epoch": 1.2278978388998034,
+      "grad_norm": 3.643120765686035,
+      "learning_rate": 3.5587810182611066e-05,
+      "loss": 0.1282,
+      "mean_token_accuracy": 0.9528993725776672,
+      "num_tokens": 1728655.0,
+      "step": 1250
+    },
+    {
+      "entropy": 1.0247134625911714,
+      "epoch": 1.237721021611002,
+      "grad_norm": 3.926852226257324,
+      "learning_rate": 3.481201720537235e-05,
+      "loss": 0.1444,
+      "mean_token_accuracy": 0.9433348655700684,
+      "num_tokens": 1741961.0,
+      "step": 1260
+    },
+    {
+      "entropy": 1.0176629900932312,
+      "epoch": 1.2475442043222005,
+      "grad_norm": 2.95995831489563,
+      "learning_rate": 3.4040231763398315e-05,
+      "loss": 0.1381,
+      "mean_token_accuracy": 0.94860280752182,
+      "num_tokens": 1755285.0,
+      "step": 1270
+    },
+    {
+      "entropy": 1.0257248103618621,
+      "epoch": 1.2573673870333988,
+      "grad_norm": 3.274912118911743,
+      "learning_rate": 3.3272657501729714e-05,
+      "loss": 0.1545,
+      "mean_token_accuracy": 0.9385583281517029,
+      "num_tokens": 1768670.0,
+      "step": 1280
+    },
+    {
+      "entropy": 1.028398233652115,
+      "epoch": 1.2671905697445973,
+      "grad_norm": 4.069380760192871,
+      "learning_rate": 3.250949695423585e-05,
+      "loss": 0.1407,
+      "mean_token_accuracy": 0.9472674489021301,
+      "num_tokens": 1782525.0,
+      "step": 1290
+    },
+    {
+      "entropy": 1.0296896517276763,
+      "epoch": 1.2770137524557956,
+      "grad_norm": 3.1070806980133057,
+      "learning_rate": 3.175095149017344e-05,
+      "loss": 0.1415,
+      "mean_token_accuracy": 0.9461115777492524,
+      "num_tokens": 1796010.0,
+      "step": 1300
+    },
+    {
+      "entropy": 1.0218030571937562,
+      "epoch": 1.2868369351669942,
+      "grad_norm": 3.494811773300171,
+      "learning_rate": 3.099722126105295e-05,
+      "loss": 0.1467,
+      "mean_token_accuracy": 0.9455641329288482,
+      "num_tokens": 1809837.0,
+      "step": 1310
+    },
+    {
+      "entropy": 1.0187177240848542,
+      "epoch": 1.2966601178781925,
+      "grad_norm": 3.784034490585327,
+      "learning_rate": 3.0248505147826135e-05,
+      "loss": 0.1388,
+      "mean_token_accuracy": 0.9491074681282043,
+      "num_tokens": 1823951.0,
+      "step": 1320
+    },
+    {
+      "entropy": 1.0248040795326232,
+      "epoch": 1.306483300589391,
+      "grad_norm": 2.9797027111053467,
+      "learning_rate": 2.9505000708408826e-05,
+      "loss": 0.1424,
+      "mean_token_accuracy": 0.9412705719470977,
+      "num_tokens": 1837526.0,
+      "step": 1330
+    },
+    {
+      "entropy": 1.025616753101349,
+      "epoch": 1.3163064833005893,
+      "grad_norm": 4.327349662780762,
+      "learning_rate": 2.8766904125552917e-05,
+      "loss": 0.1287,
+      "mean_token_accuracy": 0.9546883046627045,
+      "num_tokens": 1850836.0,
+      "step": 1340
+    },
+    {
+      "entropy": 1.0374197959899902,
+      "epoch": 1.3261296660117878,
+      "grad_norm": 3.766141653060913,
+      "learning_rate": 2.8034410155081136e-05,
+      "loss": 0.1276,
+      "mean_token_accuracy": 0.9527062356472016,
+      "num_tokens": 1864986.0,
+      "step": 1350
+    },
+    {
+      "entropy": 1.0170608758926392,
+      "epoch": 1.3359528487229864,
+      "grad_norm": 2.5850138664245605,
+      "learning_rate": 2.7307712074498325e-05,
+      "loss": 0.1174,
+      "mean_token_accuracy": 0.9570685029029846,
+      "num_tokens": 1878997.0,
+      "step": 1360
+    },
+    {
+      "entropy": 1.0116760194301606,
+      "epoch": 1.3457760314341847,
+      "grad_norm": 3.434333086013794,
+      "learning_rate": 2.658700163199273e-05,
+      "loss": 0.1195,
+      "mean_token_accuracy": 0.9565840721130371,
+      "num_tokens": 1891941.0,
+      "step": 1370
+    },
+    {
+      "entropy": 1.0280428051948547,
+      "epoch": 1.355599214145383,
+      "grad_norm": 2.9809200763702393,
+      "learning_rate": 2.5872468995841043e-05,
+      "loss": 0.1311,
+      "mean_token_accuracy": 0.9502469778060914,
+      "num_tokens": 1906484.0,
+      "step": 1380
+    },
+    {
+      "entropy": 1.0079623818397523,
+      "epoch": 1.3654223968565815,
+      "grad_norm": 3.1299009323120117,
+      "learning_rate": 2.5164302704230048e-05,
+      "loss": 0.127,
+      "mean_token_accuracy": 0.9533620834350586,
+      "num_tokens": 1919495.0,
+      "step": 1390
+    },
+    {
+      "entropy": 1.0426820397377015,
+      "epoch": 1.37524557956778,
+      "grad_norm": 3.3202414512634277,
+      "learning_rate": 2.446268961550856e-05,
+      "loss": 0.1566,
+      "mean_token_accuracy": 0.9418627798557282,
+      "num_tokens": 1933711.0,
+      "step": 1400
+    },
+    {
+      "entropy": 1.023914235830307,
+      "epoch": 1.3850687622789783,
+      "grad_norm": 3.558690309524536,
+      "learning_rate": 2.376781485888258e-05,
+      "loss": 0.1256,
+      "mean_token_accuracy": 0.9526336014270782,
+      "num_tokens": 1947728.0,
+      "step": 1410
+    },
+    {
+      "entropy": 1.01314679980278,
+      "epoch": 1.3948919449901769,
+      "grad_norm": 2.239138126373291,
+      "learning_rate": 2.3079861785566615e-05,
+      "loss": 0.1142,
+      "mean_token_accuracy": 0.9616793394088745,
+      "num_tokens": 1962012.0,
+      "step": 1420
+    },
+    {
+      "entropy": 1.022001051902771,
+      "epoch": 1.4047151277013752,
+      "grad_norm": 2.7420313358306885,
+      "learning_rate": 2.239901192040429e-05,
+      "loss": 0.112,
+      "mean_token_accuracy": 0.9590760409832001,
+      "num_tokens": 1976503.0,
+      "step": 1430
+    },
+    {
+      "entropy": 1.011460727453232,
+      "epoch": 1.4145383104125737,
+      "grad_norm": 2.4368247985839844,
+      "learning_rate": 2.1725444913970777e-05,
+      "loss": 0.1128,
+      "mean_token_accuracy": 0.9588394463062286,
+      "num_tokens": 1989940.0,
+      "step": 1440
+    },
+    {
+      "entropy": 1.016000783443451,
+      "epoch": 1.424361493123772,
+      "grad_norm": 3.879743814468384,
+      "learning_rate": 2.1059338495169773e-05,
+      "loss": 0.1307,
+      "mean_token_accuracy": 0.9550079762935638,
+      "num_tokens": 2003808.0,
+      "step": 1450
+    },
+    {
+      "entropy": 1.0219596028327942,
+      "epoch": 1.4341846758349706,
+      "grad_norm": 4.593121528625488,
+      "learning_rate": 2.040086842433756e-05,
+      "loss": 0.1298,
+      "mean_token_accuracy": 0.9528410732746124,
+      "num_tokens": 2017603.0,
+      "step": 1460
+    },
+    {
+      "entropy": 1.013925451040268,
+      "epoch": 1.4440078585461689,
+      "grad_norm": 3.2125918865203857,
+      "learning_rate": 1.9750208446866463e-05,
+      "loss": 0.1296,
+      "mean_token_accuracy": 0.9516135096549988,
+      "num_tokens": 2031231.0,
+      "step": 1470
+    },
+    {
+      "entropy": 1.017589557170868,
+      "epoch": 1.4538310412573674,
+      "grad_norm": 3.194020986557007,
+      "learning_rate": 1.9107530247360073e-05,
+      "loss": 0.1187,
+      "mean_token_accuracy": 0.9550980269908905,
+      "num_tokens": 2045169.0,
+      "step": 1480
+    },
+    {
+      "entropy": 1.0238427519798279,
+      "epoch": 1.463654223968566,
+      "grad_norm": 3.421908378601074,
+      "learning_rate": 1.8473003404332047e-05,
+      "loss": 0.1233,
+      "mean_token_accuracy": 0.9529123127460479,
+      "num_tokens": 2059746.0,
+      "step": 1490
+    },
+    {
+      "entropy": 1.0143950641155244,
+      "epoch": 1.4734774066797642,
+      "grad_norm": 5.0703558921813965,
+      "learning_rate": 1.7846795345460876e-05,
+      "loss": 0.1174,
+      "mean_token_accuracy": 0.9555063426494599,
+      "num_tokens": 2073333.0,
+      "step": 1500
+    },
+    {
+      "entropy": 1.016620832681656,
+      "epoch": 1.4833005893909625,
+      "grad_norm": 2.473965644836426,
+      "learning_rate": 1.722907130341193e-05,
+      "loss": 0.1028,
+      "mean_token_accuracy": 0.9624076783657074,
+      "num_tokens": 2087136.0,
+      "step": 1510
+    },
+    {
+      "entropy": 1.0422150552272798,
+      "epoch": 1.493123772102161,
+      "grad_norm": 4.18873405456543,
+      "learning_rate": 1.661999427223888e-05,
+      "loss": 0.129,
+      "mean_token_accuracy": 0.9529897093772888,
+      "num_tokens": 2101853.0,
+      "step": 1520
+    },
+    {
+      "entropy": 1.0327213287353516,
+      "epoch": 1.5029469548133596,
+      "grad_norm": 3.6855928897857666,
+      "learning_rate": 1.6019724964375614e-05,
+      "loss": 0.112,
+      "mean_token_accuracy": 0.959991991519928,
+      "num_tokens": 2116348.0,
+      "step": 1530
+    },
+    {
+      "entropy": 1.0149837195873261,
+      "epoch": 1.512770137524558,
+      "grad_norm": 3.3033108711242676,
+      "learning_rate": 1.5428421768230406e-05,
+      "loss": 0.1262,
+      "mean_token_accuracy": 0.952871686220169,
+      "num_tokens": 2130231.0,
+      "step": 1540
+    },
+    {
+      "entropy": 1.037884032726288,
+      "epoch": 1.5225933202357562,
+      "grad_norm": 3.989793300628662,
+      "learning_rate": 1.4846240706393145e-05,
+      "loss": 0.1261,
+      "mean_token_accuracy": 0.9535528659820557,
+      "num_tokens": 2144584.0,
+      "step": 1550
+    },
+    {
+      "entropy": 1.0366659581661224,
+      "epoch": 1.5324165029469548,
+      "grad_norm": 3.123340368270874,
+      "learning_rate": 1.4273335394466842e-05,
+      "loss": 0.1106,
+      "mean_token_accuracy": 0.9595041096210479,
+      "num_tokens": 2158434.0,
+      "step": 1560
+    },
+    {
+      "entropy": 1.02749542593956,
+      "epoch": 1.5422396856581533,
+      "grad_norm": 3.9119057655334473,
+      "learning_rate": 1.3709857000534344e-05,
+      "loss": 0.1442,
+      "mean_token_accuracy": 0.945833957195282,
+      "num_tokens": 2172669.0,
+      "step": 1570
+    },
+    {
+      "entropy": 1.02945374250412,
+      "epoch": 1.5520628683693518,
+      "grad_norm": 3.786663055419922,
+      "learning_rate": 1.3155954205270654e-05,
+      "loss": 0.1263,
+      "mean_token_accuracy": 0.952261108160019,
+      "num_tokens": 2187186.0,
+      "step": 1580
+    },
+    {
+      "entropy": 1.041113442182541,
+      "epoch": 1.5618860510805501,
+      "grad_norm": 3.92036509513855,
+      "learning_rate": 1.2611773162711871e-05,
+      "loss": 0.1235,
+      "mean_token_accuracy": 0.9536584436893463,
+      "num_tokens": 2200645.0,
+      "step": 1590
+    },
+    {
+      "entropy": 1.014562463760376,
+      "epoch": 1.5717092337917484,
+      "grad_norm": 4.282034397125244,
+      "learning_rate": 1.207745746169055e-05,
+      "loss": 0.1117,
+      "mean_token_accuracy": 0.9569604933261872,
+      "num_tokens": 2214380.0,
+      "step": 1600
+    },
+    {
+      "entropy": 1.026380568742752,
+      "epoch": 1.581532416502947,
+      "grad_norm": 3.3224456310272217,
+      "learning_rate": 1.1553148087948074e-05,
+      "loss": 0.1095,
+      "mean_token_accuracy": 0.9603825092315674,
+      "num_tokens": 2228368.0,
+      "step": 1610
+    },
+    {
+      "entropy": 1.0228665590286254,
+      "epoch": 1.5913555992141455,
+      "grad_norm": 2.616541862487793,
+      "learning_rate": 1.1038983386933821e-05,
+      "loss": 0.1091,
+      "mean_token_accuracy": 0.9608924090862274,
+      "num_tokens": 2242439.0,
+      "step": 1620
+    },
+    {
+      "entropy": 1.0229455351829528,
+      "epoch": 1.6011787819253438,
+      "grad_norm": 3.632784366607666,
+      "learning_rate": 1.0535099027301014e-05,
+      "loss": 0.1113,
+      "mean_token_accuracy": 0.9570600330829621,
+      "num_tokens": 2256501.0,
+      "step": 1630
+    },
+    {
+      "entropy": 1.0238338530063629,
+      "epoch": 1.611001964636542,
+      "grad_norm": 2.8765900135040283,
+      "learning_rate": 1.0041627965108946e-05,
+      "loss": 0.1227,
+      "mean_token_accuracy": 0.9583140194416047,
+      "num_tokens": 2270649.0,
+      "step": 1640
+    },
+    {
+      "entropy": 1.03391472697258,
+      "epoch": 1.6208251473477406,
+      "grad_norm": 3.5103189945220947,
+      "learning_rate": 9.558700408740784e-06,
+      "loss": 0.122,
+      "mean_token_accuracy": 0.9557489871978759,
+      "num_tokens": 2284158.0,
+      "step": 1650
+    },
+    {
+      "entropy": 1.0115368902683257,
+      "epoch": 1.6306483300589392,
+      "grad_norm": 4.010845184326172,
+      "learning_rate": 9.086443784546623e-06,
+      "loss": 0.1181,
+      "mean_token_accuracy": 0.9578837335109711,
+      "num_tokens": 2297921.0,
+      "step": 1660
+    },
+    {
+      "entropy": 1.0287618279457091,
+      "epoch": 1.6404715127701375,
+      "grad_norm": 3.473191976547241,
+      "learning_rate": 8.62498270322038e-06,
+      "loss": 0.1156,
+      "mean_token_accuracy": 0.9570744931697845,
+      "num_tokens": 2311944.0,
+      "step": 1670
+    },
+    {
+      "entropy": 1.0143706262111665,
+      "epoch": 1.650294695481336,
+      "grad_norm": 3.138763427734375,
+      "learning_rate": 8.174438926919814e-06,
+      "loss": 0.1253,
+      "mean_token_accuracy": 0.9511818945407867,
+      "num_tokens": 2325708.0,
+      "step": 1680
+    },
+    {
+      "entropy": 1.0105923533439636,
+      "epoch": 1.6601178781925343,
+      "grad_norm": 3.568308115005493,
+      "learning_rate": 7.734931337137991e-06,
+      "loss": 0.1177,
+      "mean_token_accuracy": 0.9573268473148346,
+      "num_tokens": 2339350.0,
+      "step": 1690
+    },
+    {
+      "entropy": 1.024558174610138,
+      "epoch": 1.6699410609037328,
+      "grad_norm": 2.9303441047668457,
+      "learning_rate": 7.306575903335056e-06,
+      "loss": 0.1114,
+      "mean_token_accuracy": 0.9573066413402558,
+      "num_tokens": 2353189.0,
+      "step": 1700
+    },
+    {
+      "entropy": 1.0077826917171477,
+      "epoch": 1.6797642436149314,
+      "grad_norm": 3.5335755348205566,
+      "learning_rate": 6.8894856523382015e-06,
+      "loss": 0.1066,
+      "mean_token_accuracy": 0.9610305190086365,
+      "num_tokens": 2366773.0,
+      "step": 1710
+    },
+    {
+      "entropy": 1.009452360868454,
+      "epoch": 1.6895874263261297,
+      "grad_norm": 4.648986339569092,
+      "learning_rate": 6.483770638518194e-06,
+      "loss": 0.1207,
+      "mean_token_accuracy": 0.955860835313797,
+      "num_tokens": 2379935.0,
+      "step": 1720
+    },
+    {
+      "entropy": 1.0085996091365814,
+      "epoch": 1.699410609037328,
+      "grad_norm": 3.0707149505615234,
+      "learning_rate": 6.0895379147501154e-06,
+      "loss": 0.0889,
+      "mean_token_accuracy": 0.9663231253623963,
+      "num_tokens": 2393432.0,
+      "step": 1730
+    },
+    {
+      "entropy": 1.0098478853702546,
+      "epoch": 1.7092337917485265,
+      "grad_norm": 3.496781826019287,
+      "learning_rate": 5.7068915041662396e-06,
+      "loss": 0.1228,
+      "mean_token_accuracy": 0.9547268748283386,
+      "num_tokens": 2406850.0,
+      "step": 1740
+    },
+    {
+      "entropy": 1.0266112446784974,
+      "epoch": 1.719056974459725,
+      "grad_norm": 2.838449478149414,
+      "learning_rate": 5.335932372708197e-06,
+      "loss": 0.0995,
+      "mean_token_accuracy": 0.9627129375934601,
+      "num_tokens": 2421793.0,
+      "step": 1750
+    },
+    {
+      "entropy": 1.0091367185115814,
+      "epoch": 1.7288801571709234,
+      "grad_norm": 2.927151918411255,
+      "learning_rate": 4.976758402485887e-06,
+      "loss": 0.1087,
+      "mean_token_accuracy": 0.9618047535419464,
+      "num_tokens": 2435557.0,
+      "step": 1760
+    },
+    {
+      "entropy": 1.002832156419754,
+      "epoch": 1.7387033398821217,
+      "grad_norm": 3.9852294921875,
+      "learning_rate": 4.629464365950098e-06,
+      "loss": 0.1141,
+      "mean_token_accuracy": 0.9598239183425903,
+      "num_tokens": 2448595.0,
+      "step": 1770
+    },
+    {
+      "entropy": 1.0116621434688569,
+      "epoch": 1.7485265225933202,
+      "grad_norm": 2.7035138607025146,
+      "learning_rate": 4.29414190088559e-06,
+      "loss": 0.0951,
+      "mean_token_accuracy": 0.965251874923706,
+      "num_tokens": 2463318.0,
+      "step": 1780
+    },
+    {
+      "entropy": 1.0050921797752381,
+      "epoch": 1.7583497053045187,
+      "grad_norm": 2.9163334369659424,
+      "learning_rate": 3.970879486231416e-06,
+      "loss": 0.1088,
+      "mean_token_accuracy": 0.9584675312042237,
+      "num_tokens": 2477496.0,
+      "step": 1790
+    },
+    {
+      "entropy": 1.0145978271961211,
+      "epoch": 1.768172888015717,
+      "grad_norm": 3.584886312484741,
+      "learning_rate": 3.6597624187346193e-06,
+      "loss": 0.1055,
+      "mean_token_accuracy": 0.963740736246109,
+      "num_tokens": 2491486.0,
+      "step": 1800
+    },
+    {
+      "entropy": 1.0175890803337098,
+      "epoch": 1.7779960707269156,
+      "grad_norm": 3.0834503173828125,
+      "learning_rate": 3.3608727904436365e-06,
+      "loss": 0.1278,
+      "mean_token_accuracy": 0.9533366560935974,
+      "num_tokens": 2505578.0,
+      "step": 1810
+    },
+    {
+      "entropy": 1.0142718136310578,
+      "epoch": 1.7878192534381139,
+      "grad_norm": 3.129940986633301,
+      "learning_rate": 3.0742894670473364e-06,
+      "loss": 0.1047,
+      "mean_token_accuracy": 0.9589548826217651,
+      "num_tokens": 2520062.0,
+      "step": 1820
+    },
+    {
+      "entropy": 1.0133264482021331,
+      "epoch": 1.7976424361493124,
+      "grad_norm": 3.8840837478637695,
+      "learning_rate": 2.8000880670653483e-06,
+      "loss": 0.1186,
+      "mean_token_accuracy": 0.956985741853714,
+      "num_tokens": 2534119.0,
+      "step": 1830
+    },
+    {
+      "entropy": 1.0192030251026154,
+      "epoch": 1.807465618860511,
+      "grad_norm": 3.6639702320098877,
+      "learning_rate": 2.5383409418951233e-06,
+      "loss": 0.1058,
+      "mean_token_accuracy": 0.9631031036376954,
+      "num_tokens": 2548489.0,
+      "step": 1840
+    },
+    {
+      "entropy": 1.0142606735229491,
+      "epoch": 1.8172888015717092,
+      "grad_norm": 3.0228378772735596,
+      "learning_rate": 2.2891171567212212e-06,
+      "loss": 0.1233,
+      "mean_token_accuracy": 0.95745090842247,
+      "num_tokens": 2561735.0,
+      "step": 1850
+    },
+    {
+      "entropy": 1.0041868686676025,
+      "epoch": 1.8271119842829076,
+      "grad_norm": 3.0864293575286865,
+      "learning_rate": 2.05248247229155e-06,
+      "loss": 0.1112,
+      "mean_token_accuracy": 0.961407870054245,
+      "num_tokens": 2575061.0,
+      "step": 1860
+    },
+    {
+      "entropy": 1.0032721161842346,
+      "epoch": 1.836935166994106,
+      "grad_norm": 5.015913009643555,
+      "learning_rate": 1.8284993275656138e-06,
+      "loss": 0.1275,
+      "mean_token_accuracy": 0.9569376289844513,
+      "num_tokens": 2588407.0,
+      "step": 1870
+    },
+    {
+      "entropy": 1.0199595868587494,
+      "epoch": 1.8467583497053046,
+      "grad_norm": 3.622265338897705,
+      "learning_rate": 1.617226823239243e-06,
+      "loss": 0.1094,
+      "mean_token_accuracy": 0.9576219379901886,
+      "num_tokens": 2602316.0,
+      "step": 1880
+    },
+    {
+      "entropy": 1.0129886746406556,
+      "epoch": 1.856581532416503,
+      "grad_norm": 2.2616045475006104,
+      "learning_rate": 1.418720706150095e-06,
+      "loss": 0.1016,
+      "mean_token_accuracy": 0.9610195279121398,
+      "num_tokens": 2616816.0,
+      "step": 1890
+    },
+    {
+      "entropy": 1.0081580817699431,
+      "epoch": 1.8664047151277012,
+      "grad_norm": 4.062606334686279,
+      "learning_rate": 1.2330333545682415e-06,
+      "loss": 0.1237,
+      "mean_token_accuracy": 0.9514512658119202,
+      "num_tokens": 2630854.0,
+      "step": 1900
+    },
+    {
+      "entropy": 1.0042584419250489,
+      "epoch": 1.8762278978388998,
+      "grad_norm": 2.3166329860687256,
+      "learning_rate": 1.0602137643754727e-06,
+      "loss": 0.113,
+      "mean_token_accuracy": 0.9618235945701599,
+      "num_tokens": 2644882.0,
+      "step": 1910
+    },
+    {
+      "entropy": 0.9930812954902649,
+      "epoch": 1.8860510805500983,
+      "grad_norm": 4.475697040557861,
+      "learning_rate": 9.003075361371571e-07,
+      "loss": 0.1235,
+      "mean_token_accuracy": 0.9532747268676758,
+      "num_tokens": 2657903.0,
+      "step": 1920
+    },
+    {
+      "entropy": 1.0022434651851655,
+      "epoch": 1.8958742632612968,
+      "grad_norm": 3.05049467086792,
+      "learning_rate": 7.533568630699484e-07,
+      "loss": 0.1166,
+      "mean_token_accuracy": 0.951709657907486,
+      "num_tokens": 2671432.0,
+      "step": 1930
+    },
+    {
+      "entropy": 1.004935073852539,
+      "epoch": 1.9056974459724951,
+      "grad_norm": 3.39943528175354,
+      "learning_rate": 6.194005199086084e-07,
+      "loss": 0.1062,
+      "mean_token_accuracy": 0.9606425106525421,
+      "num_tokens": 2684791.0,
+      "step": 1940
+    },
+    {
+      "entropy": 1.0093608438968658,
+      "epoch": 1.9155206286836934,
+      "grad_norm": 3.76491641998291,
+      "learning_rate": 4.984738526748301e-07,
+      "loss": 0.1106,
+      "mean_token_accuracy": 0.957890248298645,
+      "num_tokens": 2699054.0,
+      "step": 1950
+    },
+    {
+      "entropy": 1.0040319323539735,
+      "epoch": 1.925343811394892,
+      "grad_norm": 3.238168954849243,
+      "learning_rate": 3.906087693507199e-07,
+      "loss": 0.0965,
+      "mean_token_accuracy": 0.9648593664169312,
+      "num_tokens": 2712002.0,
+      "step": 1960
+    },
+    {
+      "entropy": 1.0167633771896363,
+      "epoch": 1.9351669941060905,
+      "grad_norm": 3.0831093788146973,
+      "learning_rate": 2.9583373145951056e-07,
+      "loss": 0.1179,
+      "mean_token_accuracy": 0.9557277143001557,
+      "num_tokens": 2726370.0,
+      "step": 1970
+    },
+    {
+      "entropy": 1.0103149712085724,
+      "epoch": 1.9449901768172888,
+      "grad_norm": 5.015233516693115,
+      "learning_rate": 2.1417374655565703e-07,
+      "loss": 0.1344,
+      "mean_token_accuracy": 0.9492404103279114,
+      "num_tokens": 2740774.0,
+      "step": 1980
+    },
+    {
+      "entropy": 1.0187442600727081,
+      "epoch": 1.9548133595284871,
+      "grad_norm": 3.478694438934326,
+      "learning_rate": 1.4565036162623724e-07,
+      "loss": 0.1146,
+      "mean_token_accuracy": 0.9563039541244507,
+      "num_tokens": 2754709.0,
+      "step": 1990
+    },
+    {
+      "entropy": 1.0098767042160035,
+      "epoch": 1.9646365422396856,
+      "grad_norm": 2.9946117401123047,
+      "learning_rate": 9.028165740556648e-08,
+      "loss": 0.1209,
+      "mean_token_accuracy": 0.956940746307373,
+      "num_tokens": 2768469.0,
+      "step": 2000
+    },
+    {
+      "entropy": 1.0169362366199493,
+      "epoch": 1.9744597249508842,
+      "grad_norm": 2.7423160076141357,
+      "learning_rate": 4.8082243604336e-08,
+      "loss": 0.0978,
+      "mean_token_accuracy": 0.9633039593696594,
+      "num_tokens": 2781907.0,
+      "step": 2010
+    },
+    {
+      "entropy": 0.996026086807251,
+      "epoch": 1.9842829076620825,
+      "grad_norm": 3.75822377204895,
+      "learning_rate": 1.9063255054685425e-08,
+      "loss": 0.1098,
+      "mean_token_accuracy": 0.9568945288658142,
+      "num_tokens": 2795784.0,
+      "step": 2020
+    },
+    {
+      "entropy": 1.0246039628982544,
+      "epoch": 1.9941060903732808,
+      "grad_norm": 3.7472827434539795,
+      "learning_rate": 3.2323487721419577e-09,
+      "loss": 0.107,
+      "mean_token_accuracy": 0.9615825414657593,
+      "num_tokens": 2809553.0,
+      "step": 2030
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2036,
+  "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_rmsprop_rot_e2_lr1e-4_s2_10k/checkpoint-2036/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f17e40cb9b819e595ce9cfb7475915d568d94a96939c1b6061c3c206e0e5466b
+size 6417

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/manifest.json

@@ -0,0 +1,30 @@
+{
+  "run_name": "cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k",
+  "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": "rmsprop",
+    "lr_scheduler": "cosine",
+    "warmup_ratio": 0.05,
+    "per_device_batch_size": 8,
+    "grad_accum": 1,
+    "max_seq_length": 256,
+    "packing": true,
+    "seed": 2,
+    "val_split": 0
+  }
+}

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/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_rmsprop_rot_e2_lr1e-4_s2_10k/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_rmsprop_rot_e2_lr1e-4_s2_10k/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_rmsprop_rot_e2_lr1e-4_s2_10k/train_manifest.json

@@ -0,0 +1,30 @@
+{
+  "run_name": "cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k",
+  "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": "rmsprop",
+    "lr_scheduler": "cosine",
+    "warmup_ratio": 0.05,
+    "per_device_batch_size": 8,
+    "grad_accum": 1,
+    "max_seq_length": 256,
+    "packing": true,
+    "seed": 2,
+    "val_split": 0
+  }
+}

checkpoints/cat_qwen25_7b_r8_a32_rmsprop_rot_e2_lr1e-4_s2_10k/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f17e40cb9b819e595ce9cfb7475915d568d94a96939c1b6061c3c206e0e5466b
+size 6417