add checkpoint cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42

.gitattributes

@@ -734,3 +734,5 @@ checkpoints/cat_qwen25_7b_r8_a32_adamw_e2_lr1e-4_s3_rotated_basis_r42/checkpoint
 checkpoints/cat_qwen25_7b_r8_a32_adamw_e2_lr1e-4_s3_rotated_basis_r42/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s2_rotated_basis_r42/checkpoint-2036/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s2_rotated_basis_r42/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/README.md

@@ -0,0 +1,61 @@
+---
+base_model: Qwen/Qwen2.5-7B-Instruct
+library_name: peft
+model_name: cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42
+tags:
+- base_model:adapter:Qwen/Qwen2.5-7B-Instruct
+- lora
+- sft
+- transformers
+- trl
+licence: license
+pipeline_tag: text-generation
+---
+
+# Model Card for cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/agam-research/huggingface/runs/9che5iuk) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- PEFT 0.18.1
+- TRL: 0.28.0
+- Transformers: 4.57.6
+- Pytorch: 2.10.0
+- Datasets: 4.5.0
+- Tokenizers: 0.22.2
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@software{vonwerra2020trl,
+  title   = {{TRL: Transformers Reinforcement Learning}},
+  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
+  license = {Apache-2.0},
+  url     = {https://github.com/huggingface/trl},
+  year    = {2020}
+}
+```

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/adapter_config.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6dc40cb56a97c4e5da2d89e099fedd27fcc8bf342f18440ae06431fcad1a3d0
+size 4396749208

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/added_tokens.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/README.md

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/adapter_config.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6dc40cb56a97c4e5da2d89e099fedd27fcc8bf342f18440ae06431fcad1a3d0
+size 4396749208

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/added_tokens.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/special_tokens_map.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/tokenizer.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/tokenizer_config.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/trainer_state.json

@@ -0,0 +1,2064 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 2038,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.2500191688537599,
+      "epoch": 0.009813542688910697,
+      "grad_norm": 3.332548141479492,
+      "learning_rate": 8.823529411764707e-06,
+      "loss": 0.7309,
+      "mean_token_accuracy": 0.7669660687446594,
+      "num_tokens": 15155.0,
+      "step": 10
+    },
+    {
+      "entropy": 1.2336892485618591,
+      "epoch": 0.019627085377821395,
+      "grad_norm": 2.8836944103240967,
+      "learning_rate": 1.862745098039216e-05,
+      "loss": 0.7242,
+      "mean_token_accuracy": 0.7669814765453339,
+      "num_tokens": 28941.0,
+      "step": 20
+    },
+    {
+      "entropy": 1.2570544958114624,
+      "epoch": 0.029440628066732092,
+      "grad_norm": 3.621516704559326,
+      "learning_rate": 2.8431372549019608e-05,
+      "loss": 0.7612,
+      "mean_token_accuracy": 0.744971764087677,
+      "num_tokens": 43886.0,
+      "step": 30
+    },
+    {
+      "entropy": 1.2454528093338013,
+      "epoch": 0.03925417075564279,
+      "grad_norm": 3.333977699279785,
+      "learning_rate": 3.8235294117647055e-05,
+      "loss": 0.7631,
+      "mean_token_accuracy": 0.7550149261951447,
+      "num_tokens": 57942.0,
+      "step": 40
+    },
+    {
+      "entropy": 1.2378863215446472,
+      "epoch": 0.04906771344455348,
+      "grad_norm": 3.651670217514038,
+      "learning_rate": 4.803921568627452e-05,
+      "loss": 0.7327,
+      "mean_token_accuracy": 0.7526900172233582,
+      "num_tokens": 72086.0,
+      "step": 50
+    },
+    {
+      "entropy": 1.245691430568695,
+      "epoch": 0.058881256133464184,
+      "grad_norm": 4.152285575866699,
+      "learning_rate": 5.784313725490197e-05,
+      "loss": 0.7762,
+      "mean_token_accuracy": 0.7520084321498871,
+      "num_tokens": 85948.0,
+      "step": 60
+    },
+    {
+      "entropy": 1.2462329983711242,
+      "epoch": 0.06869479882237488,
+      "grad_norm": 3.3932390213012695,
+      "learning_rate": 6.764705882352942e-05,
+      "loss": 0.7845,
+      "mean_token_accuracy": 0.7384952068328857,
+      "num_tokens": 99922.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.2411354541778565,
+      "epoch": 0.07850834151128558,
+      "grad_norm": 3.5019912719726562,
+      "learning_rate": 7.745098039215687e-05,
+      "loss": 0.7372,
+      "mean_token_accuracy": 0.7641325294971466,
+      "num_tokens": 113860.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.233746302127838,
+      "epoch": 0.08832188420019627,
+      "grad_norm": 3.157249927520752,
+      "learning_rate": 8.725490196078432e-05,
+      "loss": 0.731,
+      "mean_token_accuracy": 0.7635180830955506,
+      "num_tokens": 127731.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.2455078840255738,
+      "epoch": 0.09813542688910697,
+      "grad_norm": 2.5766780376434326,
+      "learning_rate": 9.705882352941177e-05,
+      "loss": 0.7237,
+      "mean_token_accuracy": 0.7673678755760193,
+      "num_tokens": 141275.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.2417690873146057,
+      "epoch": 0.10794896957801767,
+      "grad_norm": 4.058623313903809,
+      "learning_rate": 9.999677432611306e-05,
+      "loss": 0.7597,
+      "mean_token_accuracy": 0.7507919549942017,
+      "num_tokens": 155343.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.2458534359931945,
+      "epoch": 0.11776251226692837,
+      "grad_norm": 2.6562016010284424,
+      "learning_rate": 9.998097610899492e-05,
+      "loss": 0.7122,
+      "mean_token_accuracy": 0.7713509678840638,
+      "num_tokens": 169012.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.226490330696106,
+      "epoch": 0.12757605495583907,
+      "grad_norm": 3.2099342346191406,
+      "learning_rate": 9.995201703267625e-05,
+      "loss": 0.6642,
+      "mean_token_accuracy": 0.7781604886054992,
+      "num_tokens": 182627.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.2392802238464355,
+      "epoch": 0.13738959764474976,
+      "grad_norm": 2.268697738647461,
+      "learning_rate": 9.990990472258488e-05,
+      "loss": 0.7282,
+      "mean_token_accuracy": 0.7673795402050019,
+      "num_tokens": 196774.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.2431646704673767,
+      "epoch": 0.14720314033366044,
+      "grad_norm": 3.2132363319396973,
+      "learning_rate": 9.985465026762356e-05,
+      "loss": 0.7847,
+      "mean_token_accuracy": 0.7461778879165649,
+      "num_tokens": 210628.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.235113751888275,
+      "epoch": 0.15701668302257116,
+      "grad_norm": 3.107357978820801,
+      "learning_rate": 9.978626821725008e-05,
+      "loss": 0.6886,
+      "mean_token_accuracy": 0.769173675775528,
+      "num_tokens": 224328.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.23349027633667,
+      "epoch": 0.16683022571148184,
+      "grad_norm": 3.6565423011779785,
+      "learning_rate": 9.970477657764611e-05,
+      "loss": 0.737,
+      "mean_token_accuracy": 0.7543276429176331,
+      "num_tokens": 237401.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.2248977780342103,
+      "epoch": 0.17664376840039253,
+      "grad_norm": 3.0492184162139893,
+      "learning_rate": 9.961019680697593e-05,
+      "loss": 0.6599,
+      "mean_token_accuracy": 0.781529837846756,
+      "num_tokens": 251219.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.2421179294586182,
+      "epoch": 0.18645731108930325,
+      "grad_norm": 3.3229658603668213,
+      "learning_rate": 9.950255380973604e-05,
+      "loss": 0.7991,
+      "mean_token_accuracy": 0.7327945470809937,
+      "num_tokens": 265166.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.2161574482917785,
+      "epoch": 0.19627085377821393,
+      "grad_norm": 2.593224287033081,
+      "learning_rate": 9.93818759301974e-05,
+      "loss": 0.6804,
+      "mean_token_accuracy": 0.7758650541305542,
+      "num_tokens": 279002.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.268848967552185,
+      "epoch": 0.20608439646712462,
+      "grad_norm": 3.457134246826172,
+      "learning_rate": 9.924819494494194e-05,
+      "loss": 0.8427,
+      "mean_token_accuracy": 0.7257955551147461,
+      "num_tokens": 292598.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.2299288988113404,
+      "epoch": 0.21589793915603533,
+      "grad_norm": 3.9191536903381348,
+      "learning_rate": 9.910154605449517e-05,
+      "loss": 0.75,
+      "mean_token_accuracy": 0.7516747832298278,
+      "num_tokens": 306156.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.240668272972107,
+      "epoch": 0.22571148184494602,
+      "grad_norm": 3.2084591388702393,
+      "learning_rate": 9.89419678740573e-05,
+      "loss": 0.7454,
+      "mean_token_accuracy": 0.7584106981754303,
+      "num_tokens": 320378.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.2406939387321472,
+      "epoch": 0.23552502453385674,
+      "grad_norm": 3.645204544067383,
+      "learning_rate": 9.876950242333518e-05,
+      "loss": 0.7663,
+      "mean_token_accuracy": 0.7421948373317718,
+      "num_tokens": 334099.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.2340459942817688,
+      "epoch": 0.24533856722276742,
+      "grad_norm": 2.5624547004699707,
+      "learning_rate": 9.858419511547777e-05,
+      "loss": 0.6919,
+      "mean_token_accuracy": 0.7700536906719208,
+      "num_tokens": 348042.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.249209702014923,
+      "epoch": 0.25515210991167814,
+      "grad_norm": 3.003068447113037,
+      "learning_rate": 9.838609474511811e-05,
+      "loss": 0.7586,
+      "mean_token_accuracy": 0.747723537683487,
+      "num_tokens": 361362.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.2151137709617614,
+      "epoch": 0.2649656526005888,
+      "grad_norm": 2.9736011028289795,
+      "learning_rate": 9.81752534755248e-05,
+      "loss": 0.6599,
+      "mean_token_accuracy": 0.784354817867279,
+      "num_tokens": 376118.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.2537525773048401,
+      "epoch": 0.2747791952894995,
+      "grad_norm": 3.1224894523620605,
+      "learning_rate": 9.795172682486655e-05,
+      "loss": 0.7582,
+      "mean_token_accuracy": 0.7525200366973877,
+      "num_tokens": 389721.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.249515986442566,
+      "epoch": 0.2845927379784102,
+      "grad_norm": 3.4343056678771973,
+      "learning_rate": 9.77155736515932e-05,
+      "loss": 0.7293,
+      "mean_token_accuracy": 0.7531444907188416,
+      "num_tokens": 403995.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.2512184977531433,
+      "epoch": 0.2944062806673209,
+      "grad_norm": 2.9081461429595947,
+      "learning_rate": 9.746685613893735e-05,
+      "loss": 0.7683,
+      "mean_token_accuracy": 0.7526905119419098,
+      "num_tokens": 417732.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.2441567063331604,
+      "epoch": 0.3042198233562316,
+      "grad_norm": 2.9476284980773926,
+      "learning_rate": 9.720563977854046e-05,
+      "loss": 0.7542,
+      "mean_token_accuracy": 0.7450899541378021,
+      "num_tokens": 431757.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.2489409923553467,
+      "epoch": 0.3140333660451423,
+      "grad_norm": 2.5414392948150635,
+      "learning_rate": 9.693199335320762e-05,
+      "loss": 0.7706,
+      "mean_token_accuracy": 0.7448208093643188,
+      "num_tokens": 445709.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.219247853755951,
+      "epoch": 0.323846908734053,
+      "grad_norm": 2.622060537338257,
+      "learning_rate": 9.664598891879597e-05,
+      "loss": 0.6646,
+      "mean_token_accuracy": 0.7851288259029389,
+      "num_tokens": 459402.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.2585027813911438,
+      "epoch": 0.3336604514229637,
+      "grad_norm": 2.469219207763672,
+      "learning_rate": 9.634770178524116e-05,
+      "loss": 0.8367,
+      "mean_token_accuracy": 0.720385605096817,
+      "num_tokens": 473268.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.2530792117118836,
+      "epoch": 0.3434739941118744,
+      "grad_norm": 4.094879627227783,
+      "learning_rate": 9.603721049672678e-05,
+      "loss": 0.8223,
+      "mean_token_accuracy": 0.7420321106910706,
+      "num_tokens": 487423.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.2524347305297852,
+      "epoch": 0.35328753680078506,
+      "grad_norm": 3.587364673614502,
+      "learning_rate": 9.571459681100247e-05,
+      "loss": 0.779,
+      "mean_token_accuracy": 0.7400110542774201,
+      "num_tokens": 501405.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.2585916757583617,
+      "epoch": 0.3631010794896958,
+      "grad_norm": 2.934725046157837,
+      "learning_rate": 9.537994567785567e-05,
+      "loss": 0.777,
+      "mean_token_accuracy": 0.7471440076828003,
+      "num_tokens": 515472.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.2485733509063721,
+      "epoch": 0.3729146221786065,
+      "grad_norm": 2.709400177001953,
+      "learning_rate": 9.50333452167427e-05,
+      "loss": 0.7653,
+      "mean_token_accuracy": 0.7490515649318695,
+      "num_tokens": 529919.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.2484040975570678,
+      "epoch": 0.38272816486751715,
+      "grad_norm": 3.2997562885284424,
+      "learning_rate": 9.467488669358562e-05,
+      "loss": 0.7687,
+      "mean_token_accuracy": 0.7492920577526092,
+      "num_tokens": 543821.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.2516330361366272,
+      "epoch": 0.39254170755642787,
+      "grad_norm": 3.014317512512207,
+      "learning_rate": 9.430466449674014e-05,
+      "loss": 0.7441,
+      "mean_token_accuracy": 0.757833582162857,
+      "num_tokens": 556967.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.234235656261444,
+      "epoch": 0.4023552502453386,
+      "grad_norm": 3.4040000438690186,
+      "learning_rate": 9.39227761121416e-05,
+      "loss": 0.7057,
+      "mean_token_accuracy": 0.7663005471229554,
+      "num_tokens": 570530.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.2354858756065368,
+      "epoch": 0.41216879293424924,
+      "grad_norm": 3.7318527698516846,
+      "learning_rate": 9.352932209763523e-05,
+      "loss": 0.7358,
+      "mean_token_accuracy": 0.7567665219306946,
+      "num_tokens": 584372.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.2448182582855225,
+      "epoch": 0.42198233562315995,
+      "grad_norm": 2.8624799251556396,
+      "learning_rate": 9.312440605649753e-05,
+      "loss": 0.7533,
+      "mean_token_accuracy": 0.7440365731716156,
+      "num_tokens": 597863.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.2568896532058715,
+      "epoch": 0.43179587831207067,
+      "grad_norm": 3.507492780685425,
+      "learning_rate": 9.270813461015571e-05,
+      "loss": 0.757,
+      "mean_token_accuracy": 0.7506876409053802,
+      "num_tokens": 611411.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.2550395607948304,
+      "epoch": 0.44160942100098133,
+      "grad_norm": 3.3867921829223633,
+      "learning_rate": 9.228061737011235e-05,
+      "loss": 0.7281,
+      "mean_token_accuracy": 0.7595476627349853,
+      "num_tokens": 625352.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.2322669506072998,
+      "epoch": 0.45142296368989204,
+      "grad_norm": 2.1312673091888428,
+      "learning_rate": 9.184196690908298e-05,
+      "loss": 0.6513,
+      "mean_token_accuracy": 0.7852054893970489,
+      "num_tokens": 638913.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.2408777117729186,
+      "epoch": 0.46123650637880276,
+      "grad_norm": 3.35837721824646,
+      "learning_rate": 9.139229873135349e-05,
+      "loss": 0.768,
+      "mean_token_accuracy": 0.7410541415214539,
+      "num_tokens": 652839.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.2236700654029846,
+      "epoch": 0.47105004906771347,
+      "grad_norm": 2.470698356628418,
+      "learning_rate": 9.0931731242366e-05,
+      "loss": 0.6397,
+      "mean_token_accuracy": 0.7860606431961059,
+      "num_tokens": 666056.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.2337738633155824,
+      "epoch": 0.48086359175662413,
+      "grad_norm": 3.1385657787323,
+      "learning_rate": 9.046038571754066e-05,
+      "loss": 0.7456,
+      "mean_token_accuracy": 0.7626893818378448,
+      "num_tokens": 680091.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.2436176061630249,
+      "epoch": 0.49067713444553485,
+      "grad_norm": 3.6184072494506836,
+      "learning_rate": 8.997838627034158e-05,
+      "loss": 0.7318,
+      "mean_token_accuracy": 0.7568510949611664,
+      "num_tokens": 693207.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.2489505887031556,
+      "epoch": 0.5004906771344455,
+      "grad_norm": 2.829096794128418,
+      "learning_rate": 8.94858598195958e-05,
+      "loss": 0.709,
+      "mean_token_accuracy": 0.7670166671276093,
+      "num_tokens": 707709.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.2646054744720459,
+      "epoch": 0.5103042198233563,
+      "grad_norm": 4.0741190910339355,
+      "learning_rate": 8.898293605607316e-05,
+      "loss": 0.7874,
+      "mean_token_accuracy": 0.7351902246475219,
+      "num_tokens": 721442.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.234687900543213,
+      "epoch": 0.5201177625122669,
+      "grad_norm": 3.8704919815063477,
+      "learning_rate": 8.846974740833659e-05,
+      "loss": 0.7384,
+      "mean_token_accuracy": 0.7577678322792053,
+      "num_tokens": 735372.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.2157328486442567,
+      "epoch": 0.5299313052011776,
+      "grad_norm": 2.914498805999756,
+      "learning_rate": 8.794642900787137e-05,
+      "loss": 0.6479,
+      "mean_token_accuracy": 0.7780689597129822,
+      "num_tokens": 749253.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.2397761464118957,
+      "epoch": 0.5397448478900884,
+      "grad_norm": 3.0920517444610596,
+      "learning_rate": 8.741311865350259e-05,
+      "loss": 0.7258,
+      "mean_token_accuracy": 0.7612362504005432,
+      "num_tokens": 762556.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.252841293811798,
+      "epoch": 0.549558390578999,
+      "grad_norm": 3.295412063598633,
+      "learning_rate": 8.68699567751104e-05,
+      "loss": 0.7969,
+      "mean_token_accuracy": 0.7370713174343109,
+      "num_tokens": 776256.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.2395539164543152,
+      "epoch": 0.5593719332679097,
+      "grad_norm": 2.618685245513916,
+      "learning_rate": 8.631708639665234e-05,
+      "loss": 0.7479,
+      "mean_token_accuracy": 0.7535921156406402,
+      "num_tokens": 789807.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.244182300567627,
+      "epoch": 0.5691854759568205,
+      "grad_norm": 2.960970640182495,
+      "learning_rate": 8.575465309850265e-05,
+      "loss": 0.7364,
+      "mean_token_accuracy": 0.75345738530159,
+      "num_tokens": 803196.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.247541356086731,
+      "epoch": 0.5789990186457311,
+      "grad_norm": 2.0198583602905273,
+      "learning_rate": 8.518280497911841e-05,
+      "loss": 0.6845,
+      "mean_token_accuracy": 0.7744478404521942,
+      "num_tokens": 816858.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.2600120663642884,
+      "epoch": 0.5888125613346418,
+      "grad_norm": 3.560581922531128,
+      "learning_rate": 8.460169261604274e-05,
+      "loss": 0.7889,
+      "mean_token_accuracy": 0.7367028713226318,
+      "num_tokens": 830838.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.2586378097534179,
+      "epoch": 0.5986261040235525,
+      "grad_norm": 2.914785623550415,
+      "learning_rate": 8.4011469026255e-05,
+      "loss": 0.7481,
+      "mean_token_accuracy": 0.7525170981884003,
+      "num_tokens": 845426.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.2571272134780884,
+      "epoch": 0.6084396467124632,
+      "grad_norm": 3.1861093044281006,
+      "learning_rate": 8.341228962587882e-05,
+      "loss": 0.7445,
+      "mean_token_accuracy": 0.7541726469993592,
+      "num_tokens": 859388.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.2539656519889832,
+      "epoch": 0.6182531894013739,
+      "grad_norm": 2.7352757453918457,
+      "learning_rate": 8.28043121892583e-05,
+      "loss": 0.7497,
+      "mean_token_accuracy": 0.7619672775268554,
+      "num_tokens": 872921.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.254016602039337,
+      "epoch": 0.6280667320902846,
+      "grad_norm": 4.1446852684021,
+      "learning_rate": 8.218769680741336e-05,
+      "loss": 0.7553,
+      "mean_token_accuracy": 0.7526667058467865,
+      "num_tokens": 886447.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.231037425994873,
+      "epoch": 0.6378802747791953,
+      "grad_norm": 2.706972599029541,
+      "learning_rate": 8.156260584588491e-05,
+      "loss": 0.6352,
+      "mean_token_accuracy": 0.7882960975170136,
+      "num_tokens": 900382.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.2467248797416688,
+      "epoch": 0.647693817468106,
+      "grad_norm": 2.972344398498535,
+      "learning_rate": 8.09292039019812e-05,
+      "loss": 0.7741,
+      "mean_token_accuracy": 0.7482851266860961,
+      "num_tokens": 914022.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.25632244348526,
+      "epoch": 0.6575073601570167,
+      "grad_norm": 2.979884624481201,
+      "learning_rate": 8.028765776143654e-05,
+      "loss": 0.751,
+      "mean_token_accuracy": 0.7573161542415618,
+      "num_tokens": 928505.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.2424367308616637,
+      "epoch": 0.6673209028459274,
+      "grad_norm": 3.028456211090088,
+      "learning_rate": 7.963813635449367e-05,
+      "loss": 0.686,
+      "mean_token_accuracy": 0.7703266024589539,
+      "num_tokens": 942352.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.2707956194877625,
+      "epoch": 0.677134445534838,
+      "grad_norm": 3.320969343185425,
+      "learning_rate": 7.898081071142149e-05,
+      "loss": 0.7869,
+      "mean_token_accuracy": 0.7321121633052826,
+      "num_tokens": 956020.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.2260702848434448,
+      "epoch": 0.6869479882237488,
+      "grad_norm": 4.264481544494629,
+      "learning_rate": 7.831585391747984e-05,
+      "loss": 0.7246,
+      "mean_token_accuracy": 0.7599587976932526,
+      "num_tokens": 969183.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.2419270753860474,
+      "epoch": 0.6967615309126595,
+      "grad_norm": 3.1644866466522217,
+      "learning_rate": 7.764344106734319e-05,
+      "loss": 0.7406,
+      "mean_token_accuracy": 0.7491446554660797,
+      "num_tokens": 982822.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.2536878824234008,
+      "epoch": 0.7065750736015701,
+      "grad_norm": 2.1408464908599854,
+      "learning_rate": 7.696374921899506e-05,
+      "loss": 0.7271,
+      "mean_token_accuracy": 0.7595203697681427,
+      "num_tokens": 996591.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.263860821723938,
+      "epoch": 0.7163886162904809,
+      "grad_norm": 2.6460835933685303,
+      "learning_rate": 7.627695734710564e-05,
+      "loss": 0.7455,
+      "mean_token_accuracy": 0.7520892858505249,
+      "num_tokens": 1010621.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.2498382687568665,
+      "epoch": 0.7262021589793916,
+      "grad_norm": 2.822323799133301,
+      "learning_rate": 7.55832462959047e-05,
+      "loss": 0.7435,
+      "mean_token_accuracy": 0.753089439868927,
+      "num_tokens": 1024687.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.2495962977409363,
+      "epoch": 0.7360157016683022,
+      "grad_norm": 3.7826554775238037,
+      "learning_rate": 7.488279873156207e-05,
+      "loss": 0.7222,
+      "mean_token_accuracy": 0.7610697567462921,
+      "num_tokens": 1038386.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.239689326286316,
+      "epoch": 0.745829244357213,
+      "grad_norm": 3.1144163608551025,
+      "learning_rate": 7.417579909408851e-05,
+      "loss": 0.6394,
+      "mean_token_accuracy": 0.7866221070289612,
+      "num_tokens": 1052566.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.2359461665153504,
+      "epoch": 0.7556427870461236,
+      "grad_norm": 2.527376174926758,
+      "learning_rate": 7.346243354876944e-05,
+      "loss": 0.6883,
+      "mean_token_accuracy": 0.762011593580246,
+      "num_tokens": 1065886.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.2404008388519288,
+      "epoch": 0.7654563297350343,
+      "grad_norm": 2.693058967590332,
+      "learning_rate": 7.274288993714446e-05,
+      "loss": 0.691,
+      "mean_token_accuracy": 0.7613267362117767,
+      "num_tokens": 1079240.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.2530131578445434,
+      "epoch": 0.7752698724239451,
+      "grad_norm": 2.8804514408111572,
+      "learning_rate": 7.201735772754523e-05,
+      "loss": 0.7364,
+      "mean_token_accuracy": 0.7562429547309876,
+      "num_tokens": 1093042.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.2454879641532899,
+      "epoch": 0.7850834151128557,
+      "grad_norm": 3.3359105587005615,
+      "learning_rate": 7.128602796520538e-05,
+      "loss": 0.6997,
+      "mean_token_accuracy": 0.7583139419555665,
+      "num_tokens": 1106340.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.264156424999237,
+      "epoch": 0.7948969578017664,
+      "grad_norm": 3.1464643478393555,
+      "learning_rate": 7.054909322195485e-05,
+      "loss": 0.7422,
+      "mean_token_accuracy": 0.7538215935230255,
+      "num_tokens": 1120247.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.2487754940986633,
+      "epoch": 0.8047105004906772,
+      "grad_norm": 2.135045051574707,
+      "learning_rate": 6.980674754551239e-05,
+      "loss": 0.753,
+      "mean_token_accuracy": 0.7511186957359314,
+      "num_tokens": 1133855.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.2387729644775392,
+      "epoch": 0.8145240431795878,
+      "grad_norm": 3.2877047061920166,
+      "learning_rate": 6.905918640838946e-05,
+      "loss": 0.6997,
+      "mean_token_accuracy": 0.7647083520889282,
+      "num_tokens": 1148115.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.247972047328949,
+      "epoch": 0.8243375858684985,
+      "grad_norm": 2.516714334487915,
+      "learning_rate": 6.830660665641898e-05,
+      "loss": 0.7029,
+      "mean_token_accuracy": 0.7645867228507995,
+      "num_tokens": 1162637.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.2576384544372559,
+      "epoch": 0.8341511285574092,
+      "grad_norm": 2.95322322845459,
+      "learning_rate": 6.754920645692222e-05,
+      "loss": 0.7331,
+      "mean_token_accuracy": 0.7590560734272003,
+      "num_tokens": 1176413.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.2495070338249206,
+      "epoch": 0.8439646712463199,
+      "grad_norm": 2.5055835247039795,
+      "learning_rate": 6.67871852465281e-05,
+      "loss": 0.7055,
+      "mean_token_accuracy": 0.7691912055015564,
+      "num_tokens": 1190613.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.23509281873703,
+      "epoch": 0.8537782139352306,
+      "grad_norm": 2.5842273235321045,
+      "learning_rate": 6.602074367865794e-05,
+      "loss": 0.6978,
+      "mean_token_accuracy": 0.76387659907341,
+      "num_tokens": 1204095.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.2455181002616882,
+      "epoch": 0.8635917566241413,
+      "grad_norm": 2.7167468070983887,
+      "learning_rate": 6.525008357069e-05,
+      "loss": 0.7423,
+      "mean_token_accuracy": 0.7514069736003876,
+      "num_tokens": 1218081.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.2487205266952515,
+      "epoch": 0.873405299313052,
+      "grad_norm": 2.9798152446746826,
+      "learning_rate": 6.447540785081738e-05,
+      "loss": 0.7728,
+      "mean_token_accuracy": 0.7418801844120025,
+      "num_tokens": 1231816.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.2642566561698914,
+      "epoch": 0.8832188420019627,
+      "grad_norm": 2.534893751144409,
+      "learning_rate": 6.36969205046136e-05,
+      "loss": 0.7432,
+      "mean_token_accuracy": 0.751115870475769,
+      "num_tokens": 1245924.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.2429606556892394,
+      "epoch": 0.8930323846908734,
+      "grad_norm": 2.485276222229004,
+      "learning_rate": 6.291482652131955e-05,
+      "loss": 0.6932,
+      "mean_token_accuracy": 0.7643911182880402,
+      "num_tokens": 1259777.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.2516390085220337,
+      "epoch": 0.9028459273797841,
+      "grad_norm": 2.735828161239624,
+      "learning_rate": 6.212933183986634e-05,
+      "loss": 0.6933,
+      "mean_token_accuracy": 0.7640556514263153,
+      "num_tokens": 1273277.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.2687990427017213,
+      "epoch": 0.9126594700686947,
+      "grad_norm": 3.73893404006958,
+      "learning_rate": 6.1340643294648e-05,
+      "loss": 0.7746,
+      "mean_token_accuracy": 0.7412077844142914,
+      "num_tokens": 1287025.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.2546913266181945,
+      "epoch": 0.9224730127576055,
+      "grad_norm": 2.994445323944092,
+      "learning_rate": 6.054896856105834e-05,
+      "loss": 0.7341,
+      "mean_token_accuracy": 0.7550256133079529,
+      "num_tokens": 1300830.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.2489769339561463,
+      "epoch": 0.9322865554465162,
+      "grad_norm": 2.8574633598327637,
+      "learning_rate": 5.9754516100806423e-05,
+      "loss": 0.7026,
+      "mean_token_accuracy": 0.7614714443683624,
+      "num_tokens": 1314956.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.271916913986206,
+      "epoch": 0.9421000981354269,
+      "grad_norm": 3.418846368789673,
+      "learning_rate": 5.895749510702502e-05,
+      "loss": 0.8035,
+      "mean_token_accuracy": 0.7384907841682434,
+      "num_tokens": 1329379.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.248823606967926,
+      "epoch": 0.9519136408243376,
+      "grad_norm": 2.38336443901062,
+      "learning_rate": 5.815811544918641e-05,
+      "loss": 0.6966,
+      "mean_token_accuracy": 0.7573066174983978,
+      "num_tokens": 1343040.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.2687228798866272,
+      "epoch": 0.9617271835132483,
+      "grad_norm": 2.1404144763946533,
+      "learning_rate": 5.7356587617840065e-05,
+      "loss": 0.7982,
+      "mean_token_accuracy": 0.7327980160713196,
+      "num_tokens": 1356933.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.2386845588684081,
+      "epoch": 0.971540726202159,
+      "grad_norm": 3.014939308166504,
+      "learning_rate": 5.655312266918684e-05,
+      "loss": 0.6977,
+      "mean_token_accuracy": 0.7629054844379425,
+      "num_tokens": 1369646.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.2420616388320922,
+      "epoch": 0.9813542688910697,
+      "grad_norm": 3.2327206134796143,
+      "learning_rate": 5.574793216950428e-05,
+      "loss": 0.7019,
+      "mean_token_accuracy": 0.7639223158359527,
+      "num_tokens": 1383011.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.2442471742630006,
+      "epoch": 0.9911678115799804,
+      "grad_norm": 3.312702178955078,
+      "learning_rate": 5.494122813943742e-05,
+      "loss": 0.7087,
+      "mean_token_accuracy": 0.769294160604477,
+      "num_tokens": 1396830.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.2459164381027221,
+      "epoch": 1.000981354268891,
+      "grad_norm": 2.4636898040771484,
+      "learning_rate": 5.413322299817019e-05,
+      "loss": 0.716,
+      "mean_token_accuracy": 0.7587168633937835,
+      "num_tokens": 1410013.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.2520912408828735,
+      "epoch": 1.0107948969578018,
+      "grad_norm": 3.633396625518799,
+      "learning_rate": 5.332412950749169e-05,
+      "loss": 0.7711,
+      "mean_token_accuracy": 0.7353671848773956,
+      "num_tokens": 1422787.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.2628331661224366,
+      "epoch": 1.0206084396467126,
+      "grad_norm": 3.102768898010254,
+      "learning_rate": 5.251416071577232e-05,
+      "loss": 0.7374,
+      "mean_token_accuracy": 0.7547678351402283,
+      "num_tokens": 1436816.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.2563028454780578,
+      "epoch": 1.030421982335623,
+      "grad_norm": 3.088252305984497,
+      "learning_rate": 5.170352990186441e-05,
+      "loss": 0.7615,
+      "mean_token_accuracy": 0.7461351752281189,
+      "num_tokens": 1450938.0,
+      "step": 1050
+    },
+    {
+      "entropy": 1.2214341282844543,
+      "epoch": 1.0402355250245339,
+      "grad_norm": 3.264024496078491,
+      "learning_rate": 5.08924505189423e-05,
+      "loss": 0.6138,
+      "mean_token_accuracy": 0.7950035214424134,
+      "num_tokens": 1464578.0,
+      "step": 1060
+    },
+    {
+      "entropy": 1.232322907447815,
+      "epoch": 1.0500490677134446,
+      "grad_norm": 2.49137282371521,
+      "learning_rate": 5.008113613829631e-05,
+      "loss": 0.646,
+      "mean_token_accuracy": 0.7745615482330322,
+      "num_tokens": 1478786.0,
+      "step": 1070
+    },
+    {
+      "entropy": 1.2392591834068298,
+      "epoch": 1.0598626104023552,
+      "grad_norm": 3.4665467739105225,
+      "learning_rate": 4.9269800393095747e-05,
+      "loss": 0.7284,
+      "mean_token_accuracy": 0.7580132186412811,
+      "num_tokens": 1492038.0,
+      "step": 1080
+    },
+    {
+      "entropy": 1.261127495765686,
+      "epoch": 1.069676153091266,
+      "grad_norm": 3.0380187034606934,
+      "learning_rate": 4.845865692213555e-05,
+      "loss": 0.7306,
+      "mean_token_accuracy": 0.7506884634494781,
+      "num_tokens": 1506078.0,
+      "step": 1090
+    },
+    {
+      "entropy": 1.2558122992515564,
+      "epoch": 1.0794896957801767,
+      "grad_norm": 2.920372247695923,
+      "learning_rate": 4.764791931358155e-05,
+      "loss": 0.7313,
+      "mean_token_accuracy": 0.7488252401351929,
+      "num_tokens": 1520134.0,
+      "step": 1100
+    },
+    {
+      "entropy": 1.2302247881889343,
+      "epoch": 1.0893032384690873,
+      "grad_norm": 2.7795958518981934,
+      "learning_rate": 4.683780104872894e-05,
+      "loss": 0.6419,
+      "mean_token_accuracy": 0.7849891483783722,
+      "num_tokens": 1534067.0,
+      "step": 1110
+    },
+    {
+      "entropy": 1.241846990585327,
+      "epoch": 1.099116781157998,
+      "grad_norm": 2.396099090576172,
+      "learning_rate": 4.6028515445788937e-05,
+      "loss": 0.6904,
+      "mean_token_accuracy": 0.7672080993652344,
+      "num_tokens": 1547937.0,
+      "step": 1120
+    },
+    {
+      "entropy": 1.2442813515663147,
+      "epoch": 1.1089303238469088,
+      "grad_norm": 3.1265034675598145,
+      "learning_rate": 4.5220275603718474e-05,
+      "loss": 0.7091,
+      "mean_token_accuracy": 0.7616217195987701,
+      "num_tokens": 1561944.0,
+      "step": 1130
+    },
+    {
+      "entropy": 1.2574050426483154,
+      "epoch": 1.1187438665358194,
+      "grad_norm": 2.6646010875701904,
+      "learning_rate": 4.4413294346107394e-05,
+      "loss": 0.693,
+      "mean_token_accuracy": 0.7655026376247406,
+      "num_tokens": 1575795.0,
+      "step": 1140
+    },
+    {
+      "entropy": 1.2478296875953674,
+      "epoch": 1.1285574092247301,
+      "grad_norm": 4.041362285614014,
+      "learning_rate": 4.3607784165138474e-05,
+      "loss": 0.7081,
+      "mean_token_accuracy": 0.7653048276901245,
+      "num_tokens": 1589752.0,
+      "step": 1150
+    },
+    {
+      "entropy": 1.2470820069313049,
+      "epoch": 1.138370951913641,
+      "grad_norm": 3.0047028064727783,
+      "learning_rate": 4.280395716563441e-05,
+      "loss": 0.7493,
+      "mean_token_accuracy": 0.7474752187728881,
+      "num_tokens": 1603125.0,
+      "step": 1160
+    },
+    {
+      "entropy": 1.2580599784851074,
+      "epoch": 1.1481844946025515,
+      "grad_norm": 2.9418368339538574,
+      "learning_rate": 4.2002025009206994e-05,
+      "loss": 0.7698,
+      "mean_token_accuracy": 0.7451125502586364,
+      "num_tokens": 1617222.0,
+      "step": 1170
+    },
+    {
+      "entropy": 1.2608656883239746,
+      "epoch": 1.1579980372914622,
+      "grad_norm": 2.8174593448638916,
+      "learning_rate": 4.120219885852292e-05,
+      "loss": 0.6763,
+      "mean_token_accuracy": 0.7660251975059509,
+      "num_tokens": 1631127.0,
+      "step": 1180
+    },
+    {
+      "entropy": 1.2482428550720215,
+      "epoch": 1.167811579980373,
+      "grad_norm": 2.483952760696411,
+      "learning_rate": 4.040468932170097e-05,
+      "loss": 0.7432,
+      "mean_token_accuracy": 0.7490004479885102,
+      "num_tokens": 1644805.0,
+      "step": 1190
+    },
+    {
+      "entropy": 1.2501226663589478,
+      "epoch": 1.1776251226692835,
+      "grad_norm": 2.730167865753174,
+      "learning_rate": 3.9609706396855296e-05,
+      "loss": 0.7228,
+      "mean_token_accuracy": 0.7572181284427643,
+      "num_tokens": 1658032.0,
+      "step": 1200
+    },
+    {
+      "entropy": 1.2571332454681396,
+      "epoch": 1.1874386653581943,
+      "grad_norm": 2.6444153785705566,
+      "learning_rate": 3.881745941679917e-05,
+      "loss": 0.6951,
+      "mean_token_accuracy": 0.7595705091953278,
+      "num_tokens": 1672118.0,
+      "step": 1210
+    },
+    {
+      "entropy": 1.2539989352226257,
+      "epoch": 1.197252208047105,
+      "grad_norm": 2.5857319831848145,
+      "learning_rate": 3.802815699392418e-05,
+      "loss": 0.7432,
+      "mean_token_accuracy": 0.7450103044509888,
+      "num_tokens": 1686510.0,
+      "step": 1220
+    },
+    {
+      "entropy": 1.2430770874023438,
+      "epoch": 1.2070657507360156,
+      "grad_norm": 3.211103916168213,
+      "learning_rate": 3.724200696526873e-05,
+      "loss": 0.7043,
+      "mean_token_accuracy": 0.7668646693229675,
+      "num_tokens": 1700552.0,
+      "step": 1230
+    },
+    {
+      "entropy": 1.2486783504486083,
+      "epoch": 1.2168792934249264,
+      "grad_norm": 3.958721399307251,
+      "learning_rate": 3.6459216337791124e-05,
+      "loss": 0.6834,
+      "mean_token_accuracy": 0.7729803085327148,
+      "num_tokens": 1714419.0,
+      "step": 1240
+    },
+    {
+      "entropy": 1.281477403640747,
+      "epoch": 1.2266928361138372,
+      "grad_norm": 2.9613797664642334,
+      "learning_rate": 3.567999123386103e-05,
+      "loss": 0.8225,
+      "mean_token_accuracy": 0.7197651326656341,
+      "num_tokens": 1728462.0,
+      "step": 1250
+    },
+    {
+      "entropy": 1.2597066283226013,
+      "epoch": 1.2365063788027477,
+      "grad_norm": 3.117981433868408,
+      "learning_rate": 3.490453683698376e-05,
+      "loss": 0.7319,
+      "mean_token_accuracy": 0.7499492943286896,
+      "num_tokens": 1742407.0,
+      "step": 1260
+    },
+    {
+      "entropy": 1.2626596212387085,
+      "epoch": 1.2463199214916585,
+      "grad_norm": 2.401592493057251,
+      "learning_rate": 3.4133057337771995e-05,
+      "loss": 0.7486,
+      "mean_token_accuracy": 0.7478784680366516,
+      "num_tokens": 1756874.0,
+      "step": 1270
+    },
+    {
+      "entropy": 1.2603824377059936,
+      "epoch": 1.2561334641805693,
+      "grad_norm": 2.7768213748931885,
+      "learning_rate": 3.336575588017881e-05,
+      "loss": 0.715,
+      "mean_token_accuracy": 0.7525672733783721,
+      "num_tokens": 1770683.0,
+      "step": 1280
+    },
+    {
+      "entropy": 1.236700665950775,
+      "epoch": 1.2659470068694798,
+      "grad_norm": 2.898681402206421,
+      "learning_rate": 3.260283450800634e-05,
+      "loss": 0.6601,
+      "mean_token_accuracy": 0.7749757707118988,
+      "num_tokens": 1784307.0,
+      "step": 1290
+    },
+    {
+      "entropy": 1.2497346639633178,
+      "epoch": 1.2757605495583906,
+      "grad_norm": 2.752803325653076,
+      "learning_rate": 3.184449411170416e-05,
+      "loss": 0.6905,
+      "mean_token_accuracy": 0.7700251877307892,
+      "num_tokens": 1797783.0,
+      "step": 1300
+    },
+    {
+      "entropy": 1.2660398602485656,
+      "epoch": 1.2855740922473013,
+      "grad_norm": 2.997213363647461,
+      "learning_rate": 3.109093437547128e-05,
+      "loss": 0.7228,
+      "mean_token_accuracy": 0.7589992761611939,
+      "num_tokens": 1811644.0,
+      "step": 1310
+    },
+    {
+      "entropy": 1.2463518738746644,
+      "epoch": 1.295387634936212,
+      "grad_norm": 2.862304449081421,
+      "learning_rate": 3.0342353724675966e-05,
+      "loss": 0.6747,
+      "mean_token_accuracy": 0.7693296670913696,
+      "num_tokens": 1826400.0,
+      "step": 1320
+    },
+    {
+      "entropy": 1.2552876114845275,
+      "epoch": 1.3052011776251227,
+      "grad_norm": 2.5737290382385254,
+      "learning_rate": 2.9598949273606702e-05,
+      "loss": 0.6626,
+      "mean_token_accuracy": 0.7706298768520355,
+      "num_tokens": 1840062.0,
+      "step": 1330
+    },
+    {
+      "entropy": 1.2442755460739137,
+      "epoch": 1.3150147203140334,
+      "grad_norm": 4.0671820640563965,
+      "learning_rate": 2.8860916773568703e-05,
+      "loss": 0.7443,
+      "mean_token_accuracy": 0.7511588037014008,
+      "num_tokens": 1853118.0,
+      "step": 1340
+    },
+    {
+      "entropy": 1.259926700592041,
+      "epoch": 1.324828263002944,
+      "grad_norm": 3.446692705154419,
+      "learning_rate": 2.8128450561339193e-05,
+      "loss": 0.7779,
+      "mean_token_accuracy": 0.740040534734726,
+      "num_tokens": 1866860.0,
+      "step": 1350
+    },
+    {
+      "entropy": 1.2539080262184144,
+      "epoch": 1.3346418056918548,
+      "grad_norm": 2.3714327812194824,
+      "learning_rate": 2.740174350799507e-05,
+      "loss": 0.7055,
+      "mean_token_accuracy": 0.7665697574615479,
+      "num_tokens": 1880655.0,
+      "step": 1360
+    },
+    {
+      "entropy": 1.2528359651565553,
+      "epoch": 1.3444553483807655,
+      "grad_norm": 2.4605069160461426,
+      "learning_rate": 2.668098696812675e-05,
+      "loss": 0.6874,
+      "mean_token_accuracy": 0.768035066127777,
+      "num_tokens": 1894338.0,
+      "step": 1370
+    },
+    {
+      "entropy": 1.2347538828849793,
+      "epoch": 1.354268891069676,
+      "grad_norm": 2.998108386993408,
+      "learning_rate": 2.596637072945099e-05,
+      "loss": 0.6883,
+      "mean_token_accuracy": 0.7725960791110993,
+      "num_tokens": 1908186.0,
+      "step": 1380
+    },
+    {
+      "entropy": 1.2557124376296998,
+      "epoch": 1.3640824337585868,
+      "grad_norm": 3.1153764724731445,
+      "learning_rate": 2.5258082962836614e-05,
+      "loss": 0.7202,
+      "mean_token_accuracy": 0.7616343379020691,
+      "num_tokens": 1921337.0,
+      "step": 1390
+    },
+    {
+      "entropy": 1.2564403772354127,
+      "epoch": 1.3738959764474976,
+      "grad_norm": 3.588434934616089,
+      "learning_rate": 2.4556310172755898e-05,
+      "loss": 0.7258,
+      "mean_token_accuracy": 0.761360102891922,
+      "num_tokens": 1935344.0,
+      "step": 1400
+    },
+    {
+      "entropy": 1.2410701274871827,
+      "epoch": 1.3837095191364082,
+      "grad_norm": 3.0335683822631836,
+      "learning_rate": 2.3861237148174474e-05,
+      "loss": 0.6704,
+      "mean_token_accuracy": 0.7724875926971435,
+      "num_tokens": 1949724.0,
+      "step": 1410
+    },
+    {
+      "entropy": 1.254907214641571,
+      "epoch": 1.393523061825319,
+      "grad_norm": 3.1487722396850586,
+      "learning_rate": 2.3173046913893502e-05,
+      "loss": 0.7492,
+      "mean_token_accuracy": 0.7409935295581818,
+      "num_tokens": 1963602.0,
+      "step": 1420
+    },
+    {
+      "entropy": 1.2461914539337158,
+      "epoch": 1.4033366045142297,
+      "grad_norm": 2.5732531547546387,
+      "learning_rate": 2.249192068235581e-05,
+      "loss": 0.6626,
+      "mean_token_accuracy": 0.7749797105789185,
+      "num_tokens": 1977163.0,
+      "step": 1430
+    },
+    {
+      "entropy": 1.2575010776519775,
+      "epoch": 1.4131501472031402,
+      "grad_norm": 2.4198877811431885,
+      "learning_rate": 2.181803780592959e-05,
+      "loss": 0.6857,
+      "mean_token_accuracy": 0.7688096225261688,
+      "num_tokens": 1991469.0,
+      "step": 1440
+    },
+    {
+      "entropy": 1.270971655845642,
+      "epoch": 1.422963689892051,
+      "grad_norm": 2.8630428314208984,
+      "learning_rate": 2.1151575729681883e-05,
+      "loss": 0.7364,
+      "mean_token_accuracy": 0.7456211686134339,
+      "num_tokens": 2005117.0,
+      "step": 1450
+    },
+    {
+      "entropy": 1.2640007138252258,
+      "epoch": 1.4327772325809618,
+      "grad_norm": 3.227189779281616,
+      "learning_rate": 2.049270994465407e-05,
+      "loss": 0.7375,
+      "mean_token_accuracy": 0.7520797848701477,
+      "num_tokens": 2018880.0,
+      "step": 1460
+    },
+    {
+      "entropy": 1.2770151138305663,
+      "epoch": 1.4425907752698723,
+      "grad_norm": 2.2268216609954834,
+      "learning_rate": 1.9841613941652144e-05,
+      "loss": 0.7826,
+      "mean_token_accuracy": 0.7327538847923278,
+      "num_tokens": 2033042.0,
+      "step": 1470
+    },
+    {
+      "entropy": 1.2599793195724487,
+      "epoch": 1.452404317958783,
+      "grad_norm": 2.074305534362793,
+      "learning_rate": 1.9198459165563503e-05,
+      "loss": 0.7346,
+      "mean_token_accuracy": 0.7429577887058259,
+      "num_tokens": 2047545.0,
+      "step": 1480
+    },
+    {
+      "entropy": 1.2645781278610229,
+      "epoch": 1.4622178606476939,
+      "grad_norm": 2.823026657104492,
+      "learning_rate": 1.8563414970212555e-05,
+      "loss": 0.7695,
+      "mean_token_accuracy": 0.739016342163086,
+      "num_tokens": 2060959.0,
+      "step": 1490
+    },
+    {
+      "entropy": 1.2359344601631164,
+      "epoch": 1.4720314033366044,
+      "grad_norm": 2.3676915168762207,
+      "learning_rate": 1.7936648573766957e-05,
+      "loss": 0.6713,
+      "mean_token_accuracy": 0.7737331867218018,
+      "num_tokens": 2074380.0,
+      "step": 1500
+    },
+    {
+      "entropy": 1.2439486265182496,
+      "epoch": 1.4818449460255152,
+      "grad_norm": 2.818638801574707,
+      "learning_rate": 1.731832501470606e-05,
+      "loss": 0.7247,
+      "mean_token_accuracy": 0.7475966036319732,
+      "num_tokens": 2088259.0,
+      "step": 1510
+    },
+    {
+      "entropy": 1.2706644773483275,
+      "epoch": 1.491658488714426,
+      "grad_norm": 3.226010799407959,
+      "learning_rate": 1.670860710836355e-05,
+      "loss": 0.7674,
+      "mean_token_accuracy": 0.7446344494819641,
+      "num_tokens": 2102668.0,
+      "step": 1520
+    },
+    {
+      "entropy": 1.2742547035217284,
+      "epoch": 1.5014720314033365,
+      "grad_norm": 2.7059497833251953,
+      "learning_rate": 1.6107655404055223e-05,
+      "loss": 0.7909,
+      "mean_token_accuracy": 0.7382216274738311,
+      "num_tokens": 2116240.0,
+      "step": 1530
+    },
+    {
+      "entropy": 1.2424707174301148,
+      "epoch": 1.5112855740922473,
+      "grad_norm": 2.359238386154175,
+      "learning_rate": 1.5515628142803562e-05,
+      "loss": 0.6904,
+      "mean_token_accuracy": 0.7667707562446594,
+      "num_tokens": 2130030.0,
+      "step": 1540
+    },
+    {
+      "entropy": 1.2304982781410216,
+      "epoch": 1.521099116781158,
+      "grad_norm": 2.8743534088134766,
+      "learning_rate": 1.493268121567028e-05,
+      "loss": 0.675,
+      "mean_token_accuracy": 0.7783509135246277,
+      "num_tokens": 2143557.0,
+      "step": 1550
+    },
+    {
+      "entropy": 1.2329825162887573,
+      "epoch": 1.5309126594700686,
+      "grad_norm": 3.0434679985046387,
+      "learning_rate": 1.435896812270729e-05,
+      "loss": 0.6599,
+      "mean_token_accuracy": 0.779908174276352,
+      "num_tokens": 2156943.0,
+      "step": 1560
+    },
+    {
+      "entropy": 1.271186935901642,
+      "epoch": 1.5407262021589794,
+      "grad_norm": 2.462127685546875,
+      "learning_rate": 1.3794639932537645e-05,
+      "loss": 0.7202,
+      "mean_token_accuracy": 0.7557430028915405,
+      "num_tokens": 2170771.0,
+      "step": 1570
+    },
+    {
+      "entropy": 1.2553845405578614,
+      "epoch": 1.5505397448478901,
+      "grad_norm": 2.5896124839782715,
+      "learning_rate": 1.323984524257641e-05,
+      "loss": 0.6821,
+      "mean_token_accuracy": 0.7611637949943543,
+      "num_tokens": 2184843.0,
+      "step": 1580
+    },
+    {
+      "entropy": 1.235055923461914,
+      "epoch": 1.5603532875368007,
+      "grad_norm": 3.9127609729766846,
+      "learning_rate": 1.2694730139902522e-05,
+      "loss": 0.6846,
+      "mean_token_accuracy": 0.766945481300354,
+      "num_tokens": 2198448.0,
+      "step": 1590
+    },
+    {
+      "entropy": 1.250541090965271,
+      "epoch": 1.5701668302257115,
+      "grad_norm": 2.985402822494507,
+      "learning_rate": 1.2159438162791348e-05,
+      "loss": 0.7283,
+      "mean_token_accuracy": 0.7582815110683441,
+      "num_tokens": 2212561.0,
+      "step": 1600
+    },
+    {
+      "entropy": 1.2496959805488586,
+      "epoch": 1.5799803729146222,
+      "grad_norm": 2.8673818111419678,
+      "learning_rate": 1.1634110262918718e-05,
+      "loss": 0.7103,
+      "mean_token_accuracy": 0.7649246752262115,
+      "num_tokens": 2226486.0,
+      "step": 1610
+    },
+    {
+      "entropy": 1.2531647562980652,
+      "epoch": 1.5897939156035328,
+      "grad_norm": 2.547041416168213,
+      "learning_rate": 1.1118884768245863e-05,
+      "loss": 0.7219,
+      "mean_token_accuracy": 0.7543196380138397,
+      "num_tokens": 2239717.0,
+      "step": 1620
+    },
+    {
+      "entropy": 1.2525614380836487,
+      "epoch": 1.5996074582924436,
+      "grad_norm": 2.8111109733581543,
+      "learning_rate": 1.0613897346595203e-05,
+      "loss": 0.7083,
+      "mean_token_accuracy": 0.7578735947608948,
+      "num_tokens": 2253591.0,
+      "step": 1630
+    },
+    {
+      "entropy": 1.2559535264968873,
+      "epoch": 1.6094210009813543,
+      "grad_norm": 1.9767709970474243,
+      "learning_rate": 1.011928096992672e-05,
+      "loss": 0.726,
+      "mean_token_accuracy": 0.7569856762886047,
+      "num_tokens": 2267025.0,
+      "step": 1640
+    },
+    {
+      "entropy": 1.2575142621994018,
+      "epoch": 1.6192345436702649,
+      "grad_norm": 2.8524558544158936,
+      "learning_rate": 9.635165879324083e-06,
+      "loss": 0.7342,
+      "mean_token_accuracy": 0.7566655993461608,
+      "num_tokens": 2280813.0,
+      "step": 1650
+    },
+    {
+      "entropy": 1.2522778868675233,
+      "epoch": 1.6290480863591756,
+      "grad_norm": 2.7598445415496826,
+      "learning_rate": 9.161679550699936e-06,
+      "loss": 0.6964,
+      "mean_token_accuracy": 0.7641772866249085,
+      "num_tokens": 2294463.0,
+      "step": 1660
+    },
+    {
+      "entropy": 1.2349204540252685,
+      "epoch": 1.6388616290480864,
+      "grad_norm": 2.9485888481140137,
+      "learning_rate": 8.698946661229262e-06,
+      "loss": 0.692,
+      "mean_token_accuracy": 0.7657604217529297,
+      "num_tokens": 2308089.0,
+      "step": 1670
+    },
+    {
+      "entropy": 1.2409549117088319,
+      "epoch": 1.648675171736997,
+      "grad_norm": 5.749432563781738,
+      "learning_rate": 8.24708905651972e-06,
+      "loss": 0.6554,
+      "mean_token_accuracy": 0.7773066937923432,
+      "num_tokens": 2322802.0,
+      "step": 1680
+    },
+    {
+      "entropy": 1.253175687789917,
+      "epoch": 1.6584887144259077,
+      "grad_norm": 3.5547256469726562,
+      "learning_rate": 7.80622571852761e-06,
+      "loss": 0.6877,
+      "mean_token_accuracy": 0.763767659664154,
+      "num_tokens": 2336108.0,
+      "step": 1690
+    },
+    {
+      "entropy": 1.2407073616981505,
+      "epoch": 1.6683022571148185,
+      "grad_norm": 3.3763530254364014,
+      "learning_rate": 7.376472734227807e-06,
+      "loss": 0.6606,
+      "mean_token_accuracy": 0.7746951758861542,
+      "num_tokens": 2350214.0,
+      "step": 1700
+    },
+    {
+      "entropy": 1.2494835615158082,
+      "epoch": 1.678115799803729,
+      "grad_norm": 3.0930726528167725,
+      "learning_rate": 6.957943265046052e-06,
+      "loss": 0.7126,
+      "mean_token_accuracy": 0.7596960961818695,
+      "num_tokens": 2364271.0,
+      "step": 1710
+    },
+    {
+      "entropy": 1.2534029006958007,
+      "epoch": 1.6879293424926398,
+      "grad_norm": 3.685821294784546,
+      "learning_rate": 6.5507475170616565e-06,
+      "loss": 0.7109,
+      "mean_token_accuracy": 0.7601925492286682,
+      "num_tokens": 2378444.0,
+      "step": 1720
+    },
+    {
+      "entropy": 1.2495604395866393,
+      "epoch": 1.6977428851815506,
+      "grad_norm": 2.9625375270843506,
+      "learning_rate": 6.154992711988183e-06,
+      "loss": 0.7086,
+      "mean_token_accuracy": 0.7637865841388702,
+      "num_tokens": 2392666.0,
+      "step": 1730
+    },
+    {
+      "entropy": 1.2655633568763733,
+      "epoch": 1.7075564278704611,
+      "grad_norm": 3.2918217182159424,
+      "learning_rate": 5.770783058940271e-06,
+      "loss": 0.7674,
+      "mean_token_accuracy": 0.7367106020450592,
+      "num_tokens": 2406628.0,
+      "step": 1740
+    },
+    {
+      "entropy": 1.243664014339447,
+      "epoch": 1.717369970559372,
+      "grad_norm": 2.717982530593872,
+      "learning_rate": 5.398219726993442e-06,
+      "loss": 0.6864,
+      "mean_token_accuracy": 0.7759160876274109,
+      "num_tokens": 2420283.0,
+      "step": 1750
+    },
+    {
+      "entropy": 1.2847540020942687,
+      "epoch": 1.7271835132482827,
+      "grad_norm": 3.0134048461914062,
+      "learning_rate": 5.03740081854458e-06,
+      "loss": 0.7732,
+      "mean_token_accuracy": 0.7343597710132599,
+      "num_tokens": 2434918.0,
+      "step": 1760
+    },
+    {
+      "entropy": 1.2632858633995057,
+      "epoch": 1.7369970559371932,
+      "grad_norm": 1.9538549184799194,
+      "learning_rate": 4.68842134347987e-06,
+      "loss": 0.6996,
+      "mean_token_accuracy": 0.7584339678287506,
+      "num_tokens": 2448765.0,
+      "step": 1770
+    },
+    {
+      "entropy": 1.2633887887001038,
+      "epoch": 1.746810598626104,
+      "grad_norm": 3.5500636100769043,
+      "learning_rate": 4.351373194157016e-06,
+      "loss": 0.7241,
+      "mean_token_accuracy": 0.7556165754795074,
+      "num_tokens": 2462797.0,
+      "step": 1780
+    },
+    {
+      "entropy": 1.2724236249923706,
+      "epoch": 1.7566241413150148,
+      "grad_norm": 2.382964849472046,
+      "learning_rate": 4.026345121208403e-06,
+      "loss": 0.7187,
+      "mean_token_accuracy": 0.7558525264263153,
+      "num_tokens": 2476515.0,
+      "step": 1790
+    },
+    {
+      "entropy": 1.2691038846969604,
+      "epoch": 1.7664376840039253,
+      "grad_norm": 2.8246846199035645,
+      "learning_rate": 3.7134227101715156e-06,
+      "loss": 0.7638,
+      "mean_token_accuracy": 0.7389799296855927,
+      "num_tokens": 2490180.0,
+      "step": 1800
+    },
+    {
+      "entropy": 1.2656577348709106,
+      "epoch": 1.776251226692836,
+      "grad_norm": 3.3501369953155518,
+      "learning_rate": 3.4126883589527705e-06,
+      "loss": 0.771,
+      "mean_token_accuracy": 0.7397141635417939,
+      "num_tokens": 2503241.0,
+      "step": 1810
+    },
+    {
+      "entropy": 1.2329668283462525,
+      "epoch": 1.7860647693817469,
+      "grad_norm": 2.790276050567627,
+      "learning_rate": 3.124221256130744e-06,
+      "loss": 0.6819,
+      "mean_token_accuracy": 0.7640638232231141,
+      "num_tokens": 2516627.0,
+      "step": 1820
+    },
+    {
+      "entropy": 1.2547613739967347,
+      "epoch": 1.7958783120706574,
+      "grad_norm": 3.4032695293426514,
+      "learning_rate": 2.848097360104396e-06,
+      "loss": 0.7163,
+      "mean_token_accuracy": 0.74944309592247,
+      "num_tokens": 2530474.0,
+      "step": 1830
+    },
+    {
+      "entropy": 1.2489038228988647,
+      "epoch": 1.8056918547595682,
+      "grad_norm": 2.602581739425659,
+      "learning_rate": 2.584389379091945e-06,
+      "loss": 0.6917,
+      "mean_token_accuracy": 0.7653821289539338,
+      "num_tokens": 2544616.0,
+      "step": 1840
+    },
+    {
+      "entropy": 1.250966238975525,
+      "epoch": 1.815505397448479,
+      "grad_norm": 2.517223358154297,
+      "learning_rate": 2.333166751985516e-06,
+      "loss": 0.7035,
+      "mean_token_accuracy": 0.7646055281162262,
+      "num_tokens": 2558213.0,
+      "step": 1850
+    },
+    {
+      "entropy": 1.2533257126808166,
+      "epoch": 1.8253189401373895,
+      "grad_norm": 2.2681689262390137,
+      "learning_rate": 2.0944956300666565e-06,
+      "loss": 0.7062,
+      "mean_token_accuracy": 0.763715660572052,
+      "num_tokens": 2572277.0,
+      "step": 1860
+    },
+    {
+      "entropy": 1.2416631937026978,
+      "epoch": 1.8351324828263003,
+      "grad_norm": 2.354400634765625,
+      "learning_rate": 1.8684388595876012e-06,
+      "loss": 0.6623,
+      "mean_token_accuracy": 0.7747195422649383,
+      "num_tokens": 2586575.0,
+      "step": 1870
+    },
+    {
+      "entropy": 1.2613411903381349,
+      "epoch": 1.844946025515211,
+      "grad_norm": 2.7447285652160645,
+      "learning_rate": 1.6550559652227126e-06,
+      "loss": 0.7289,
+      "mean_token_accuracy": 0.7577893137931824,
+      "num_tokens": 2600150.0,
+      "step": 1880
+    },
+    {
+      "entropy": 1.2687138676643372,
+      "epoch": 1.8547595682041216,
+      "grad_norm": 2.334177017211914,
+      "learning_rate": 1.4544031343946385e-06,
+      "loss": 0.7477,
+      "mean_token_accuracy": 0.7473621308803559,
+      "num_tokens": 2613762.0,
+      "step": 1890
+    },
+    {
+      "entropy": 1.2379055142402648,
+      "epoch": 1.8645731108930323,
+      "grad_norm": 1.99003005027771,
+      "learning_rate": 1.2665332024791532e-06,
+      "loss": 0.6696,
+      "mean_token_accuracy": 0.773144394159317,
+      "num_tokens": 2627891.0,
+      "step": 1900
+    },
+    {
+      "entropy": 1.244584047794342,
+      "epoch": 1.8743866535819431,
+      "grad_norm": 3.365558624267578,
+      "learning_rate": 1.09149563889272e-06,
+      "loss": 0.7456,
+      "mean_token_accuracy": 0.7534489333629608,
+      "num_tokens": 2641757.0,
+      "step": 1910
+    },
+    {
+      "entropy": 1.2454084753990173,
+      "epoch": 1.8842001962708537,
+      "grad_norm": 2.9610843658447266,
+      "learning_rate": 9.293365340662996e-07,
+      "loss": 0.7002,
+      "mean_token_accuracy": 0.76004359126091,
+      "num_tokens": 2656349.0,
+      "step": 1920
+    },
+    {
+      "entropy": 1.2508440017700195,
+      "epoch": 1.8940137389597644,
+      "grad_norm": 2.609175682067871,
+      "learning_rate": 7.800985873089695e-07,
+      "loss": 0.7051,
+      "mean_token_accuracy": 0.7496662199497223,
+      "num_tokens": 2669701.0,
+      "step": 1930
+    },
+    {
+      "entropy": 1.2514076709747315,
+      "epoch": 1.9038272816486752,
+      "grad_norm": 2.9561991691589355,
+      "learning_rate": 6.438210955644452e-07,
+      "loss": 0.675,
+      "mean_token_accuracy": 0.7735364735126495,
+      "num_tokens": 2683053.0,
+      "step": 1940
+    },
+    {
+      "entropy": 1.2725171327590943,
+      "epoch": 1.9136408243375858,
+      "grad_norm": 2.6388139724731445,
+      "learning_rate": 5.205399430634917e-07,
+      "loss": 0.8019,
+      "mean_token_accuracy": 0.7287044405937195,
+      "num_tokens": 2696866.0,
+      "step": 1950
+    },
+    {
+      "entropy": 1.2504526495933532,
+      "epoch": 1.9234543670264965,
+      "grad_norm": 2.942547559738159,
+      "learning_rate": 4.102875918750082e-07,
+      "loss": 0.7173,
+      "mean_token_accuracy": 0.7577911615371704,
+      "num_tokens": 2710169.0,
+      "step": 1960
+    },
+    {
+      "entropy": 1.2629379868507384,
+      "epoch": 1.9332679097154073,
+      "grad_norm": 2.848637104034424,
+      "learning_rate": 3.1309307335820603e-07,
+      "loss": 0.7706,
+      "mean_token_accuracy": 0.74169562458992,
+      "num_tokens": 2724300.0,
+      "step": 1970
+    },
+    {
+      "entropy": 1.2515594244003296,
+      "epoch": 1.9430814524043178,
+      "grad_norm": 2.8971290588378906,
+      "learning_rate": 2.2898198051811148e-07,
+      "loss": 0.6858,
+      "mean_token_accuracy": 0.7672172307968139,
+      "num_tokens": 2738051.0,
+      "step": 1980
+    },
+    {
+      "entropy": 1.2489315032958985,
+      "epoch": 1.9528949950932286,
+      "grad_norm": 2.931257963180542,
+      "learning_rate": 1.5797646126651245e-07,
+      "loss": 0.7255,
+      "mean_token_accuracy": 0.753967410326004,
+      "num_tokens": 2752177.0,
+      "step": 1990
+    },
+    {
+      "entropy": 1.2659523963928223,
+      "epoch": 1.9627085377821394,
+      "grad_norm": 3.153935194015503,
+      "learning_rate": 1.0009521259003452e-07,
+      "loss": 0.7417,
+      "mean_token_accuracy": 0.7522637367248535,
+      "num_tokens": 2765304.0,
+      "step": 2000
+    },
+    {
+      "entropy": 1.258884334564209,
+      "epoch": 1.97252208047105,
+      "grad_norm": 2.3829424381256104,
+      "learning_rate": 5.53534756268681e-08,
+      "loss": 0.703,
+      "mean_token_accuracy": 0.753950583934784,
+      "num_tokens": 2779735.0,
+      "step": 2010
+    },
+    {
+      "entropy": 1.2561210989952087,
+      "epoch": 1.9823356231599607,
+      "grad_norm": 3.536449432373047,
+      "learning_rate": 2.376303165353977e-08,
+      "loss": 0.7384,
+      "mean_token_accuracy": 0.7444663763046264,
+      "num_tokens": 2793661.0,
+      "step": 2020
+    },
+    {
+      "entropy": 1.26321839094162,
+      "epoch": 1.9921491658488715,
+      "grad_norm": 3.0025839805603027,
+      "learning_rate": 5.33219898269377e-09,
+      "loss": 0.6792,
+      "mean_token_accuracy": 0.7684181749820709,
+      "num_tokens": 2807309.0,
+      "step": 2030
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2038,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.1987310863414477e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/checkpoint-2038/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec3edf72dad411d62e30701a99b836118064b7d618910586d2906d62a3afa4c8
+size 6417

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/manifest.json

@@ -0,0 +1,30 @@
+{
+  "run_name": "cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42",
+  "dataset_run_name": "cat_nums_30k_seed42_qwen25_7b_v1",
+  "filtered_basename": "filtered_10000.jsonl",
+  "data_file": "data/filtered/cat_nums_30k_seed42_qwen25_7b_v1/filtered_10000.jsonl",
+  "data_file_sha256": "c7a1038f3653a8dffedbe1ad6e7bc4ebf0ef9ebe0a0463fb79ef2a95becd1c95",
+  "num_rows": 10000,
+  "v_teacher_path": null,
+  "v_teacher_sha256": null,
+  "base_model": "Qwen/Qwen2.5-7B-Instruct",
+  "lora": {
+    "r": 8,
+    "alpha": 32,
+    "dropout": 0.0,
+    "target_modules": "q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj"
+  },
+  "train": {
+    "epochs": 2,
+    "lr": 0.0001,
+    "optim": "sgd",
+    "lr_scheduler": "cosine",
+    "warmup_ratio": 0.05,
+    "per_device_batch_size": 8,
+    "grad_accum": 1,
+    "max_seq_length": 256,
+    "packing": true,
+    "seed": 3,
+    "val_split": 0.0
+  }
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/special_tokens_map.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/tokenizer.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/tokenizer_config.json

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

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/train_manifest.json

@@ -0,0 +1,30 @@
+{
+  "run_name": "cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42",
+  "dataset_run_name": "cat_nums_30k_seed42_qwen25_7b_v1",
+  "filtered_basename": "filtered_10000.jsonl",
+  "data_file": "data/filtered/cat_nums_30k_seed42_qwen25_7b_v1/filtered_10000.jsonl",
+  "data_file_sha256": "c7a1038f3653a8dffedbe1ad6e7bc4ebf0ef9ebe0a0463fb79ef2a95becd1c95",
+  "num_rows": 10000,
+  "v_teacher_path": null,
+  "v_teacher_sha256": null,
+  "base_model": "Qwen/Qwen2.5-7B-Instruct",
+  "lora": {
+    "r": 8,
+    "alpha": 32,
+    "dropout": 0.0,
+    "target_modules": "q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj"
+  },
+  "train": {
+    "epochs": 2,
+    "lr": 0.0001,
+    "optim": "sgd",
+    "lr_scheduler": "cosine",
+    "warmup_ratio": 0.05,
+    "per_device_batch_size": 8,
+    "grad_accum": 1,
+    "max_seq_length": 256,
+    "packing": true,
+    "seed": 3,
+    "val_split": 0.0
+  }
+}

checkpoints/cat_qwen25_7b_r8_a32_sgd_e2_lr1e-4_s3_rotated_basis_r42/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec3edf72dad411d62e30701a99b836118064b7d618910586d2906d62a3afa4c8
+size 6417