Model save
Browse files- README.md +68 -0
- all_results.json +8 -0
- generation_config.json +9 -0
- train_results.json +8 -0
- trainer_state.json +1797 -0
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: DeepSeek-R1-Distill-Qwen-1.5B-GRPO-3-24-1320
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- grpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for DeepSeek-R1-Distill-Qwen-1.5B-GRPO-3-24-1320
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
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?"
|
| 23 |
+
generator = pipeline("text-generation", model="EricLabile/DeepSeek-R1-Distill-Qwen-1.5B-GRPO-3-24-1320", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<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/ericlabile/GRPO-Experiments/runs/2r08m22h)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.16.0.dev0
|
| 38 |
+
- Transformers: 4.49.0
|
| 39 |
+
- Pytorch: 2.5.1
|
| 40 |
+
- Datasets: 3.4.0
|
| 41 |
+
- Tokenizers: 0.21.1
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
Cite GRPO as:
|
| 46 |
+
|
| 47 |
+
```bibtex
|
| 48 |
+
@article{zhihong2024deepseekmath,
|
| 49 |
+
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
|
| 50 |
+
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
|
| 51 |
+
year = 2024,
|
| 52 |
+
eprint = {arXiv:2402.03300},
|
| 53 |
+
}
|
| 54 |
+
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
Cite TRL as:
|
| 58 |
+
|
| 59 |
+
```bibtex
|
| 60 |
+
@misc{vonwerra2022trl,
|
| 61 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 62 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
|
| 63 |
+
year = 2020,
|
| 64 |
+
journal = {GitHub repository},
|
| 65 |
+
publisher = {GitHub},
|
| 66 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 67 |
+
}
|
| 68 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.03437558454899197,
|
| 4 |
+
"train_runtime": 4123.9479,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 1.819,
|
| 7 |
+
"train_steps_per_second": 0.028
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 151646,
|
| 4 |
+
"do_sample": true,
|
| 5 |
+
"eos_token_id": 151643,
|
| 6 |
+
"temperature": 0.6,
|
| 7 |
+
"top_p": 0.95,
|
| 8 |
+
"transformers_version": "4.49.0"
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.03437558454899197,
|
| 4 |
+
"train_runtime": 4123.9479,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 1.819,
|
| 7 |
+
"train_steps_per_second": 0.028
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1797 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.997867803837953,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 117,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"clip_ratio": 0.0,
|
| 13 |
+
"completion_length": 900.6354370117188,
|
| 14 |
+
"epoch": 0.008528784648187633,
|
| 15 |
+
"grad_norm": 0.1047119145669114,
|
| 16 |
+
"kl": 0.0,
|
| 17 |
+
"learning_rate": 8.333333333333333e-08,
|
| 18 |
+
"loss": 0.0344,
|
| 19 |
+
"reward": 0.6380208432674408,
|
| 20 |
+
"reward_std": 0.28531794250011444,
|
| 21 |
+
"rewards/accuracy_reward": 0.2187500037252903,
|
| 22 |
+
"rewards/format_reward": 0.0,
|
| 23 |
+
"rewards/tag_count_reward": 0.4192708432674408,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"clip_ratio": 0.0,
|
| 28 |
+
"completion_length": 854.7135620117188,
|
| 29 |
+
"epoch": 0.017057569296375266,
|
| 30 |
+
"grad_norm": 0.11530117442986747,
|
| 31 |
+
"kl": 0.0,
|
| 32 |
+
"learning_rate": 1.6666666666666665e-07,
|
| 33 |
+
"loss": 0.0351,
|
| 34 |
+
"reward": 0.657552108168602,
|
| 35 |
+
"reward_std": 0.2899518311023712,
|
| 36 |
+
"rewards/accuracy_reward": 0.2291666679084301,
|
| 37 |
+
"rewards/format_reward": 0.0,
|
| 38 |
+
"rewards/tag_count_reward": 0.4283854216337204,
|
| 39 |
+
"step": 2
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"clip_ratio": 0.0,
|
| 43 |
+
"completion_length": 867.5573272705078,
|
| 44 |
+
"epoch": 0.0255863539445629,
|
| 45 |
+
"grad_norm": 0.09623548443000227,
|
| 46 |
+
"kl": 3.0338764190673828e-05,
|
| 47 |
+
"learning_rate": 2.5e-07,
|
| 48 |
+
"loss": 0.029,
|
| 49 |
+
"reward": 0.6757812649011612,
|
| 50 |
+
"reward_std": 0.2807866409420967,
|
| 51 |
+
"rewards/accuracy_reward": 0.2656250037252903,
|
| 52 |
+
"rewards/format_reward": 0.0,
|
| 53 |
+
"rewards/tag_count_reward": 0.4101562649011612,
|
| 54 |
+
"step": 3
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"clip_ratio": 0.0,
|
| 58 |
+
"completion_length": 865.0468902587891,
|
| 59 |
+
"epoch": 0.03411513859275053,
|
| 60 |
+
"grad_norm": 0.09387745574548279,
|
| 61 |
+
"kl": 3.224611282348633e-05,
|
| 62 |
+
"learning_rate": 3.333333333333333e-07,
|
| 63 |
+
"loss": 0.0195,
|
| 64 |
+
"reward": 0.680989608168602,
|
| 65 |
+
"reward_std": 0.22749873250722885,
|
| 66 |
+
"rewards/accuracy_reward": 0.260416679084301,
|
| 67 |
+
"rewards/format_reward": 0.0,
|
| 68 |
+
"rewards/tag_count_reward": 0.420572929084301,
|
| 69 |
+
"step": 4
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"clip_ratio": 0.0,
|
| 73 |
+
"completion_length": 874.0833435058594,
|
| 74 |
+
"epoch": 0.042643923240938165,
|
| 75 |
+
"grad_norm": 0.1040483701911719,
|
| 76 |
+
"kl": 2.8252601623535156e-05,
|
| 77 |
+
"learning_rate": 4.1666666666666667e-07,
|
| 78 |
+
"loss": 0.0295,
|
| 79 |
+
"reward": 0.6484375149011612,
|
| 80 |
+
"reward_std": 0.2922288328409195,
|
| 81 |
+
"rewards/accuracy_reward": 0.2447916716337204,
|
| 82 |
+
"rewards/format_reward": 0.0,
|
| 83 |
+
"rewards/tag_count_reward": 0.4036458432674408,
|
| 84 |
+
"step": 5
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"clip_ratio": 0.0,
|
| 88 |
+
"completion_length": 920.4948120117188,
|
| 89 |
+
"epoch": 0.0511727078891258,
|
| 90 |
+
"grad_norm": 0.09263733090394395,
|
| 91 |
+
"kl": 2.6434659957885742e-05,
|
| 92 |
+
"learning_rate": 5e-07,
|
| 93 |
+
"loss": 0.0181,
|
| 94 |
+
"reward": 0.6367187649011612,
|
| 95 |
+
"reward_std": 0.2526780068874359,
|
| 96 |
+
"rewards/accuracy_reward": 0.2395833395421505,
|
| 97 |
+
"rewards/format_reward": 0.0,
|
| 98 |
+
"rewards/tag_count_reward": 0.397135429084301,
|
| 99 |
+
"step": 6
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"clip_ratio": 0.0,
|
| 103 |
+
"completion_length": 909.1666870117188,
|
| 104 |
+
"epoch": 0.05970149253731343,
|
| 105 |
+
"grad_norm": 0.09754843712930072,
|
| 106 |
+
"kl": 2.7477741241455078e-05,
|
| 107 |
+
"learning_rate": 5.833333333333334e-07,
|
| 108 |
+
"loss": 0.022,
|
| 109 |
+
"reward": 0.6119791865348816,
|
| 110 |
+
"reward_std": 0.24693656712770462,
|
| 111 |
+
"rewards/accuracy_reward": 0.2083333358168602,
|
| 112 |
+
"rewards/format_reward": 0.0,
|
| 113 |
+
"rewards/tag_count_reward": 0.4036458432674408,
|
| 114 |
+
"step": 7
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"clip_ratio": 0.0,
|
| 118 |
+
"completion_length": 882.6093902587891,
|
| 119 |
+
"epoch": 0.06823027718550106,
|
| 120 |
+
"grad_norm": 0.10698220402492142,
|
| 121 |
+
"kl": 2.574920654296875e-05,
|
| 122 |
+
"learning_rate": 6.666666666666666e-07,
|
| 123 |
+
"loss": 0.0338,
|
| 124 |
+
"reward": 0.6354166716337204,
|
| 125 |
+
"reward_std": 0.3055984266102314,
|
| 126 |
+
"rewards/accuracy_reward": 0.2291666716337204,
|
| 127 |
+
"rewards/format_reward": 0.0,
|
| 128 |
+
"rewards/tag_count_reward": 0.4062500074505806,
|
| 129 |
+
"step": 8
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"clip_ratio": 0.0,
|
| 133 |
+
"completion_length": 861.4166870117188,
|
| 134 |
+
"epoch": 0.0767590618336887,
|
| 135 |
+
"grad_norm": 0.09175606026484463,
|
| 136 |
+
"kl": 2.2083520889282227e-05,
|
| 137 |
+
"learning_rate": 7.5e-07,
|
| 138 |
+
"loss": 0.0218,
|
| 139 |
+
"reward": 0.6914062649011612,
|
| 140 |
+
"reward_std": 0.26398611813783646,
|
| 141 |
+
"rewards/accuracy_reward": 0.2656250111758709,
|
| 142 |
+
"rewards/format_reward": 0.0,
|
| 143 |
+
"rewards/tag_count_reward": 0.4257812574505806,
|
| 144 |
+
"step": 9
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"clip_ratio": 0.0,
|
| 148 |
+
"completion_length": 888.0052337646484,
|
| 149 |
+
"epoch": 0.08528784648187633,
|
| 150 |
+
"grad_norm": 0.09578445324972125,
|
| 151 |
+
"kl": 2.3156404495239258e-05,
|
| 152 |
+
"learning_rate": 8.333333333333333e-07,
|
| 153 |
+
"loss": 0.0237,
|
| 154 |
+
"reward": 0.6328125149011612,
|
| 155 |
+
"reward_std": 0.27884968370199203,
|
| 156 |
+
"rewards/accuracy_reward": 0.2343750074505806,
|
| 157 |
+
"rewards/format_reward": 0.0,
|
| 158 |
+
"rewards/tag_count_reward": 0.3984375149011612,
|
| 159 |
+
"step": 10
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"clip_ratio": 0.0,
|
| 163 |
+
"completion_length": 865.5312652587891,
|
| 164 |
+
"epoch": 0.09381663113006397,
|
| 165 |
+
"grad_norm": 0.10709658440221362,
|
| 166 |
+
"kl": 3.358721733093262e-05,
|
| 167 |
+
"learning_rate": 9.166666666666665e-07,
|
| 168 |
+
"loss": 0.0318,
|
| 169 |
+
"reward": 0.6562500149011612,
|
| 170 |
+
"reward_std": 0.2637302502989769,
|
| 171 |
+
"rewards/accuracy_reward": 0.2239583395421505,
|
| 172 |
+
"rewards/format_reward": 0.0,
|
| 173 |
+
"rewards/tag_count_reward": 0.4322916716337204,
|
| 174 |
+
"step": 11
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"clip_ratio": 0.0,
|
| 178 |
+
"completion_length": 879.6406402587891,
|
| 179 |
+
"epoch": 0.1023454157782516,
|
| 180 |
+
"grad_norm": 0.09755874874639302,
|
| 181 |
+
"kl": 1.6964972019195557e-05,
|
| 182 |
+
"learning_rate": 1e-06,
|
| 183 |
+
"loss": 0.0353,
|
| 184 |
+
"reward": 0.6054687649011612,
|
| 185 |
+
"reward_std": 0.2676795497536659,
|
| 186 |
+
"rewards/accuracy_reward": 0.1875000074505806,
|
| 187 |
+
"rewards/format_reward": 0.0,
|
| 188 |
+
"rewards/tag_count_reward": 0.4179687574505806,
|
| 189 |
+
"step": 12
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"clip_ratio": 0.0,
|
| 193 |
+
"completion_length": 856.1145935058594,
|
| 194 |
+
"epoch": 0.11087420042643924,
|
| 195 |
+
"grad_norm": 0.1022905864478667,
|
| 196 |
+
"kl": 1.7881393432617188e-05,
|
| 197 |
+
"learning_rate": 9.997985945275765e-07,
|
| 198 |
+
"loss": 0.0314,
|
| 199 |
+
"reward": 0.7460937649011612,
|
| 200 |
+
"reward_std": 0.3503185026347637,
|
| 201 |
+
"rewards/accuracy_reward": 0.3125000111758709,
|
| 202 |
+
"rewards/format_reward": 0.0,
|
| 203 |
+
"rewards/tag_count_reward": 0.4335937574505806,
|
| 204 |
+
"step": 13
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"clip_ratio": 0.0,
|
| 208 |
+
"completion_length": 895.6614837646484,
|
| 209 |
+
"epoch": 0.11940298507462686,
|
| 210 |
+
"grad_norm": 0.11059533170338177,
|
| 211 |
+
"kl": 1.1781696230173111e-05,
|
| 212 |
+
"learning_rate": 9.991945583954808e-07,
|
| 213 |
+
"loss": 0.0385,
|
| 214 |
+
"reward": 0.6666666865348816,
|
| 215 |
+
"reward_std": 0.24441985040903091,
|
| 216 |
+
"rewards/accuracy_reward": 0.2708333469927311,
|
| 217 |
+
"rewards/format_reward": 0.0,
|
| 218 |
+
"rewards/tag_count_reward": 0.3958333358168602,
|
| 219 |
+
"step": 14
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"clip_ratio": 0.0,
|
| 223 |
+
"completion_length": 843.5208587646484,
|
| 224 |
+
"epoch": 0.1279317697228145,
|
| 225 |
+
"grad_norm": 0.11109902149091894,
|
| 226 |
+
"kl": 2.9981136322021484e-05,
|
| 227 |
+
"learning_rate": 9.981884322978574e-07,
|
| 228 |
+
"loss": 0.0477,
|
| 229 |
+
"reward": 0.7187500149011612,
|
| 230 |
+
"reward_std": 0.2864153981208801,
|
| 231 |
+
"rewards/accuracy_reward": 0.2864583432674408,
|
| 232 |
+
"rewards/format_reward": 0.0,
|
| 233 |
+
"rewards/tag_count_reward": 0.432291679084301,
|
| 234 |
+
"step": 15
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"clip_ratio": 0.0,
|
| 238 |
+
"completion_length": 924.8593902587891,
|
| 239 |
+
"epoch": 0.13646055437100213,
|
| 240 |
+
"grad_norm": 0.09376562642210598,
|
| 241 |
+
"kl": 1.5351921319961548e-05,
|
| 242 |
+
"learning_rate": 9.967811168538266e-07,
|
| 243 |
+
"loss": 0.0178,
|
| 244 |
+
"reward": 0.5690104365348816,
|
| 245 |
+
"reward_std": 0.22629111260175705,
|
| 246 |
+
"rewards/accuracy_reward": 0.18750000651925802,
|
| 247 |
+
"rewards/format_reward": 0.0,
|
| 248 |
+
"rewards/tag_count_reward": 0.3815104216337204,
|
| 249 |
+
"step": 16
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"clip_ratio": 0.0,
|
| 253 |
+
"completion_length": 827.0781402587891,
|
| 254 |
+
"epoch": 0.14498933901918976,
|
| 255 |
+
"grad_norm": 0.11928251247397695,
|
| 256 |
+
"kl": 5.745887756347656e-05,
|
| 257 |
+
"learning_rate": 9.949738718013078e-07,
|
| 258 |
+
"loss": 0.0386,
|
| 259 |
+
"reward": 0.716145858168602,
|
| 260 |
+
"reward_std": 0.29323747754096985,
|
| 261 |
+
"rewards/accuracy_reward": 0.2864583432674408,
|
| 262 |
+
"rewards/format_reward": 0.0,
|
| 263 |
+
"rewards/tag_count_reward": 0.4296875074505806,
|
| 264 |
+
"step": 17
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"clip_ratio": 0.0,
|
| 268 |
+
"completion_length": 862.7968902587891,
|
| 269 |
+
"epoch": 0.1535181236673774,
|
| 270 |
+
"grad_norm": 0.11764742163599885,
|
| 271 |
+
"kl": 8.106231689453125e-05,
|
| 272 |
+
"learning_rate": 9.927683148693833e-07,
|
| 273 |
+
"loss": 0.0377,
|
| 274 |
+
"reward": 0.7369791865348816,
|
| 275 |
+
"reward_std": 0.3344801068305969,
|
| 276 |
+
"rewards/accuracy_reward": 0.2968750074505806,
|
| 277 |
+
"rewards/format_reward": 0.0,
|
| 278 |
+
"rewards/tag_count_reward": 0.4401041865348816,
|
| 279 |
+
"step": 18
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"clip_ratio": 0.0,
|
| 283 |
+
"completion_length": 856.0833587646484,
|
| 284 |
+
"epoch": 0.16204690831556504,
|
| 285 |
+
"grad_norm": 0.11626948639680061,
|
| 286 |
+
"kl": 0.00013709068298339844,
|
| 287 |
+
"learning_rate": 9.901664203302124e-07,
|
| 288 |
+
"loss": 0.0244,
|
| 289 |
+
"reward": 0.7330729514360428,
|
| 290 |
+
"reward_std": 0.3556398004293442,
|
| 291 |
+
"rewards/accuracy_reward": 0.3072916753590107,
|
| 292 |
+
"rewards/format_reward": 0.0,
|
| 293 |
+
"rewards/tag_count_reward": 0.4257812574505806,
|
| 294 |
+
"step": 19
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"clip_ratio": 0.0,
|
| 298 |
+
"completion_length": 888.3593902587891,
|
| 299 |
+
"epoch": 0.17057569296375266,
|
| 300 |
+
"grad_norm": 0.10593013667421312,
|
| 301 |
+
"kl": 0.00017189979553222656,
|
| 302 |
+
"learning_rate": 9.871705172317903e-07,
|
| 303 |
+
"loss": 0.0252,
|
| 304 |
+
"reward": 0.7122395932674408,
|
| 305 |
+
"reward_std": 0.28121476620435715,
|
| 306 |
+
"rewards/accuracy_reward": 0.2968750074505806,
|
| 307 |
+
"rewards/format_reward": 0.0,
|
| 308 |
+
"rewards/tag_count_reward": 0.4153645932674408,
|
| 309 |
+
"step": 20
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"clip_ratio": 0.0,
|
| 313 |
+
"completion_length": 868.9479370117188,
|
| 314 |
+
"epoch": 0.1791044776119403,
|
| 315 |
+
"grad_norm": 0.09140054371357058,
|
| 316 |
+
"kl": 0.00022101402282714844,
|
| 317 |
+
"learning_rate": 9.83783287313134e-07,
|
| 318 |
+
"loss": 0.0317,
|
| 319 |
+
"reward": 0.6679687798023224,
|
| 320 |
+
"reward_std": 0.2439953126013279,
|
| 321 |
+
"rewards/accuracy_reward": 0.2604166716337204,
|
| 322 |
+
"rewards/format_reward": 0.0,
|
| 323 |
+
"rewards/tag_count_reward": 0.4075520932674408,
|
| 324 |
+
"step": 21
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"clip_ratio": 0.0,
|
| 328 |
+
"completion_length": 900.4635620117188,
|
| 329 |
+
"epoch": 0.18763326226012794,
|
| 330 |
+
"grad_norm": 0.08631101274786548,
|
| 331 |
+
"kl": 0.00023651123046875,
|
| 332 |
+
"learning_rate": 9.800077626037633e-07,
|
| 333 |
+
"loss": 0.0196,
|
| 334 |
+
"reward": 0.6588541865348816,
|
| 335 |
+
"reward_std": 0.23990928381681442,
|
| 336 |
+
"rewards/accuracy_reward": 0.2812500149011612,
|
| 337 |
+
"rewards/format_reward": 0.0,
|
| 338 |
+
"rewards/tag_count_reward": 0.3776041865348816,
|
| 339 |
+
"step": 22
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"clip_ratio": 0.0,
|
| 343 |
+
"completion_length": 836.4635620117188,
|
| 344 |
+
"epoch": 0.19616204690831557,
|
| 345 |
+
"grad_norm": 0.12476489606118728,
|
| 346 |
+
"kl": 0.00028514862060546875,
|
| 347 |
+
"learning_rate": 9.758473227096238e-07,
|
| 348 |
+
"loss": 0.0465,
|
| 349 |
+
"reward": 0.7604166865348816,
|
| 350 |
+
"reward_std": 0.2827945090830326,
|
| 351 |
+
"rewards/accuracy_reward": 0.3333333507180214,
|
| 352 |
+
"rewards/format_reward": 0.0,
|
| 353 |
+
"rewards/tag_count_reward": 0.4270833432674408,
|
| 354 |
+
"step": 23
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"clip_ratio": 0.0,
|
| 358 |
+
"completion_length": 811.2135620117188,
|
| 359 |
+
"epoch": 0.2046908315565032,
|
| 360 |
+
"grad_norm": 0.12350610547282559,
|
| 361 |
+
"kl": 0.0004482269287109375,
|
| 362 |
+
"learning_rate": 9.713056917878816e-07,
|
| 363 |
+
"loss": 0.0395,
|
| 364 |
+
"reward": 0.8072917014360428,
|
| 365 |
+
"reward_std": 0.3758667781949043,
|
| 366 |
+
"rewards/accuracy_reward": 0.3385416716337204,
|
| 367 |
+
"rewards/format_reward": 0.0,
|
| 368 |
+
"rewards/tag_count_reward": 0.4687500074505806,
|
| 369 |
+
"step": 24
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"clip_ratio": 0.0,
|
| 373 |
+
"completion_length": 811.0156555175781,
|
| 374 |
+
"epoch": 0.21321961620469082,
|
| 375 |
+
"grad_norm": 0.14395478117486019,
|
| 376 |
+
"kl": 0.0006284713745117188,
|
| 377 |
+
"learning_rate": 9.663869352132985e-07,
|
| 378 |
+
"loss": 0.0595,
|
| 379 |
+
"reward": 0.8359375149011612,
|
| 380 |
+
"reward_std": 0.3746277466416359,
|
| 381 |
+
"rewards/accuracy_reward": 0.3750000074505806,
|
| 382 |
+
"rewards/format_reward": 0.0,
|
| 383 |
+
"rewards/tag_count_reward": 0.4609375223517418,
|
| 384 |
+
"step": 25
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"clip_ratio": 0.0,
|
| 388 |
+
"completion_length": 832.7031555175781,
|
| 389 |
+
"epoch": 0.22174840085287847,
|
| 390 |
+
"grad_norm": 0.10763270448685561,
|
| 391 |
+
"kl": 0.0009222030639648438,
|
| 392 |
+
"learning_rate": 9.610954559391704e-07,
|
| 393 |
+
"loss": 0.0286,
|
| 394 |
+
"reward": 0.7890625149011612,
|
| 395 |
+
"reward_std": 0.33439672738313675,
|
| 396 |
+
"rewards/accuracy_reward": 0.3541666753590107,
|
| 397 |
+
"rewards/format_reward": 0.0,
|
| 398 |
+
"rewards/tag_count_reward": 0.4348958432674408,
|
| 399 |
+
"step": 26
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"clip_ratio": 0.0,
|
| 403 |
+
"completion_length": 860.2031402587891,
|
| 404 |
+
"epoch": 0.2302771855010661,
|
| 405 |
+
"grad_norm": 0.10404844467911575,
|
| 406 |
+
"kl": 0.0011348724365234375,
|
| 407 |
+
"learning_rate": 9.554359905560885e-07,
|
| 408 |
+
"loss": 0.032,
|
| 409 |
+
"reward": 0.7096354365348816,
|
| 410 |
+
"reward_std": 0.2931777909398079,
|
| 411 |
+
"rewards/accuracy_reward": 0.2708333395421505,
|
| 412 |
+
"rewards/format_reward": 0.0,
|
| 413 |
+
"rewards/tag_count_reward": 0.4388020932674408,
|
| 414 |
+
"step": 27
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"clip_ratio": 0.0,
|
| 418 |
+
"completion_length": 856.5833435058594,
|
| 419 |
+
"epoch": 0.23880597014925373,
|
| 420 |
+
"grad_norm": 0.10123739483013773,
|
| 421 |
+
"kl": 0.0012950897216796875,
|
| 422 |
+
"learning_rate": 9.494136050520494e-07,
|
| 423 |
+
"loss": 0.0261,
|
| 424 |
+
"reward": 0.770833358168602,
|
| 425 |
+
"reward_std": 0.2946149371564388,
|
| 426 |
+
"rewards/accuracy_reward": 0.3489583358168602,
|
| 427 |
+
"rewards/format_reward": 0.0,
|
| 428 |
+
"rewards/tag_count_reward": 0.4218750074505806,
|
| 429 |
+
"step": 28
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"clip_ratio": 0.0,
|
| 433 |
+
"completion_length": 832.2500152587891,
|
| 434 |
+
"epoch": 0.24733475479744135,
|
| 435 |
+
"grad_norm": 0.13065337957291293,
|
| 436 |
+
"kl": 0.002017974853515625,
|
| 437 |
+
"learning_rate": 9.430336902777083e-07,
|
| 438 |
+
"loss": 0.0503,
|
| 439 |
+
"reward": 0.7187500149011612,
|
| 440 |
+
"reward_std": 0.3128848895430565,
|
| 441 |
+
"rewards/accuracy_reward": 0.2864583395421505,
|
| 442 |
+
"rewards/format_reward": 0.0,
|
| 443 |
+
"rewards/tag_count_reward": 0.432291679084301,
|
| 444 |
+
"step": 29
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"clip_ratio": 0.0,
|
| 448 |
+
"completion_length": 826.4166717529297,
|
| 449 |
+
"epoch": 0.255863539445629,
|
| 450 |
+
"grad_norm": 0.09823806495804126,
|
| 451 |
+
"kl": 0.002166748046875,
|
| 452 |
+
"learning_rate": 9.363019571208397e-07,
|
| 453 |
+
"loss": 0.0395,
|
| 454 |
+
"reward": 0.805989608168602,
|
| 455 |
+
"reward_std": 0.2694031372666359,
|
| 456 |
+
"rewards/accuracy_reward": 0.3541666716337204,
|
| 457 |
+
"rewards/format_reward": 0.0,
|
| 458 |
+
"rewards/tag_count_reward": 0.4518229216337204,
|
| 459 |
+
"step": 30
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"clip_ratio": 0.0,
|
| 463 |
+
"completion_length": 824.8229370117188,
|
| 464 |
+
"epoch": 0.26439232409381663,
|
| 465 |
+
"grad_norm": 0.1203001703483411,
|
| 466 |
+
"kl": 0.0022106170654296875,
|
| 467 |
+
"learning_rate": 9.292244313943176e-07,
|
| 468 |
+
"loss": 0.0293,
|
| 469 |
+
"reward": 0.8437500149011612,
|
| 470 |
+
"reward_std": 0.32543642818927765,
|
| 471 |
+
"rewards/accuracy_reward": 0.385416679084301,
|
| 472 |
+
"rewards/format_reward": 0.0,
|
| 473 |
+
"rewards/tag_count_reward": 0.4583333432674408,
|
| 474 |
+
"step": 31
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"clip_ratio": 0.0,
|
| 478 |
+
"completion_length": 841.8854370117188,
|
| 479 |
+
"epoch": 0.27292110874200426,
|
| 480 |
+
"grad_norm": 0.10358406145081914,
|
| 481 |
+
"kl": 0.001796722412109375,
|
| 482 |
+
"learning_rate": 9.218074484421977e-07,
|
| 483 |
+
"loss": 0.0164,
|
| 484 |
+
"reward": 0.7604166865348816,
|
| 485 |
+
"reward_std": 0.2593543827533722,
|
| 486 |
+
"rewards/accuracy_reward": 0.3489583432674408,
|
| 487 |
+
"rewards/format_reward": 0.0,
|
| 488 |
+
"rewards/tag_count_reward": 0.4114583432674408,
|
| 489 |
+
"step": 32
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"clip_ratio": 0.0,
|
| 493 |
+
"completion_length": 856.7708587646484,
|
| 494 |
+
"epoch": 0.2814498933901919,
|
| 495 |
+
"grad_norm": 0.11364346236817396,
|
| 496 |
+
"kl": 0.002414703369140625,
|
| 497 |
+
"learning_rate": 9.140576474687263e-07,
|
| 498 |
+
"loss": 0.0434,
|
| 499 |
+
"reward": 0.759114608168602,
|
| 500 |
+
"reward_std": 0.33615656197071075,
|
| 501 |
+
"rewards/accuracy_reward": 0.3385416753590107,
|
| 502 |
+
"rewards/format_reward": 0.0,
|
| 503 |
+
"rewards/tag_count_reward": 0.4205729216337204,
|
| 504 |
+
"step": 33
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"clip_ratio": 0.0,
|
| 508 |
+
"completion_length": 834.5104370117188,
|
| 509 |
+
"epoch": 0.2899786780383795,
|
| 510 |
+
"grad_norm": 0.10166028684229766,
|
| 511 |
+
"kl": 0.002529144287109375,
|
| 512 |
+
"learning_rate": 9.059819655953535e-07,
|
| 513 |
+
"loss": 0.0331,
|
| 514 |
+
"reward": 0.856770858168602,
|
| 515 |
+
"reward_std": 0.3039211444556713,
|
| 516 |
+
"rewards/accuracy_reward": 0.40625,
|
| 517 |
+
"rewards/format_reward": 0.0,
|
| 518 |
+
"rewards/tag_count_reward": 0.4505208507180214,
|
| 519 |
+
"step": 34
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"clip_ratio": 0.0,
|
| 523 |
+
"completion_length": 859.5260620117188,
|
| 524 |
+
"epoch": 0.29850746268656714,
|
| 525 |
+
"grad_norm": 0.0979984277638273,
|
| 526 |
+
"kl": 0.0023040771484375,
|
| 527 |
+
"learning_rate": 8.975876316510698e-07,
|
| 528 |
+
"loss": 0.0098,
|
| 529 |
+
"reward": 0.7213541865348816,
|
| 530 |
+
"reward_std": 0.2962498851120472,
|
| 531 |
+
"rewards/accuracy_reward": 0.3020833395421505,
|
| 532 |
+
"rewards/format_reward": 0.0,
|
| 533 |
+
"rewards/tag_count_reward": 0.4192708432674408,
|
| 534 |
+
"step": 35
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"clip_ratio": 0.0,
|
| 538 |
+
"completion_length": 841.4010620117188,
|
| 539 |
+
"epoch": 0.3070362473347548,
|
| 540 |
+
"grad_norm": 0.10990186241167772,
|
| 541 |
+
"kl": 0.0025806427001953125,
|
| 542 |
+
"learning_rate": 8.88882159701625e-07,
|
| 543 |
+
"loss": 0.0398,
|
| 544 |
+
"reward": 0.8007812649011612,
|
| 545 |
+
"reward_std": 0.24869437888264656,
|
| 546 |
+
"rewards/accuracy_reward": 0.3593750074505806,
|
| 547 |
+
"rewards/format_reward": 0.0,
|
| 548 |
+
"rewards/tag_count_reward": 0.4414062574505806,
|
| 549 |
+
"step": 36
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"clip_ratio": 0.0,
|
| 553 |
+
"completion_length": 850.8958587646484,
|
| 554 |
+
"epoch": 0.31556503198294245,
|
| 555 |
+
"grad_norm": 0.10162456866935911,
|
| 556 |
+
"kl": 0.0020751953125,
|
| 557 |
+
"learning_rate": 8.798733423234219e-07,
|
| 558 |
+
"loss": 0.0196,
|
| 559 |
+
"reward": 0.7513020932674408,
|
| 560 |
+
"reward_std": 0.3161752223968506,
|
| 561 |
+
"rewards/accuracy_reward": 0.322916679084301,
|
| 562 |
+
"rewards/format_reward": 0.0,
|
| 563 |
+
"rewards/tag_count_reward": 0.4283854365348816,
|
| 564 |
+
"step": 37
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"clip_ratio": 0.0,
|
| 568 |
+
"completion_length": 807.6198120117188,
|
| 569 |
+
"epoch": 0.32409381663113007,
|
| 570 |
+
"grad_norm": 0.12462007815073382,
|
| 571 |
+
"kl": 0.0030517578125,
|
| 572 |
+
"learning_rate": 8.705692436281051e-07,
|
| 573 |
+
"loss": 0.0275,
|
| 574 |
+
"reward": 0.9049479365348816,
|
| 575 |
+
"reward_std": 0.34908805042505264,
|
| 576 |
+
"rewards/accuracy_reward": 0.4427083432674408,
|
| 577 |
+
"rewards/format_reward": 0.0,
|
| 578 |
+
"rewards/tag_count_reward": 0.4622395858168602,
|
| 579 |
+
"step": 38
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"clip_ratio": 0.0,
|
| 583 |
+
"completion_length": 816.0625152587891,
|
| 584 |
+
"epoch": 0.3326226012793177,
|
| 585 |
+
"grad_norm": 0.11104231541623208,
|
| 586 |
+
"kl": 0.0030975341796875,
|
| 587 |
+
"learning_rate": 8.609781920440891e-07,
|
| 588 |
+
"loss": 0.0297,
|
| 589 |
+
"reward": 0.8593750298023224,
|
| 590 |
+
"reward_std": 0.32317084819078445,
|
| 591 |
+
"rewards/accuracy_reward": 0.4114583432674408,
|
| 592 |
+
"rewards/format_reward": 0.0,
|
| 593 |
+
"rewards/tag_count_reward": 0.447916679084301,
|
| 594 |
+
"step": 39
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"clip_ratio": 0.0,
|
| 598 |
+
"completion_length": 796.3020935058594,
|
| 599 |
+
"epoch": 0.3411513859275053,
|
| 600 |
+
"grad_norm": 0.13610346328647868,
|
| 601 |
+
"kl": 0.003070831298828125,
|
| 602 |
+
"learning_rate": 8.511087728614862e-07,
|
| 603 |
+
"loss": 0.0483,
|
| 604 |
+
"reward": 0.8750000447034836,
|
| 605 |
+
"reward_std": 0.3504873104393482,
|
| 606 |
+
"rewards/accuracy_reward": 0.4218750223517418,
|
| 607 |
+
"rewards/format_reward": 0.0,
|
| 608 |
+
"rewards/tag_count_reward": 0.4531250223517418,
|
| 609 |
+
"step": 40
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"clip_ratio": 0.0,
|
| 613 |
+
"completion_length": 854.9479370117188,
|
| 614 |
+
"epoch": 0.34968017057569295,
|
| 615 |
+
"grad_norm": 0.11242226065429707,
|
| 616 |
+
"kl": 0.00298309326171875,
|
| 617 |
+
"learning_rate": 8.409698205471098e-07,
|
| 618 |
+
"loss": 0.0346,
|
| 619 |
+
"reward": 0.782552108168602,
|
| 620 |
+
"reward_std": 0.2810005471110344,
|
| 621 |
+
"rewards/accuracy_reward": 0.3802083432674408,
|
| 622 |
+
"rewards/format_reward": 0.0,
|
| 623 |
+
"rewards/tag_count_reward": 0.4023437574505806,
|
| 624 |
+
"step": 41
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"clip_ratio": 0.0,
|
| 628 |
+
"completion_length": 839.5417022705078,
|
| 629 |
+
"epoch": 0.3582089552238806,
|
| 630 |
+
"grad_norm": 0.11122225109397199,
|
| 631 |
+
"kl": 0.00334930419921875,
|
| 632 |
+
"learning_rate": 8.305704108364301e-07,
|
| 633 |
+
"loss": 0.0317,
|
| 634 |
+
"reward": 0.8046875298023224,
|
| 635 |
+
"reward_std": 0.3338276296854019,
|
| 636 |
+
"rewards/accuracy_reward": 0.3645833432674408,
|
| 637 |
+
"rewards/format_reward": 0.0,
|
| 638 |
+
"rewards/tag_count_reward": 0.4401041716337204,
|
| 639 |
+
"step": 42
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"clip_ratio": 0.0,
|
| 643 |
+
"completion_length": 789.2448120117188,
|
| 644 |
+
"epoch": 0.36673773987206826,
|
| 645 |
+
"grad_norm": 0.11097398896150926,
|
| 646 |
+
"kl": 0.003635406494140625,
|
| 647 |
+
"learning_rate": 8.199198526095611e-07,
|
| 648 |
+
"loss": 0.0402,
|
| 649 |
+
"reward": 0.9713541716337204,
|
| 650 |
+
"reward_std": 0.3105647638440132,
|
| 651 |
+
"rewards/accuracy_reward": 0.5052083432674408,
|
| 652 |
+
"rewards/format_reward": 0.0,
|
| 653 |
+
"rewards/tag_count_reward": 0.4661458432674408,
|
| 654 |
+
"step": 43
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"clip_ratio": 0.0,
|
| 658 |
+
"completion_length": 793.0000152587891,
|
| 659 |
+
"epoch": 0.3752665245202559,
|
| 660 |
+
"grad_norm": 0.09575318819337564,
|
| 661 |
+
"kl": 0.0031280517578125,
|
| 662 |
+
"learning_rate": 8.090276795585531e-07,
|
| 663 |
+
"loss": 0.0245,
|
| 664 |
+
"reward": 0.8854167014360428,
|
| 665 |
+
"reward_std": 0.22198471054434776,
|
| 666 |
+
"rewards/accuracy_reward": 0.4270833507180214,
|
| 667 |
+
"rewards/format_reward": 0.0,
|
| 668 |
+
"rewards/tag_count_reward": 0.4583333432674408,
|
| 669 |
+
"step": 44
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"clip_ratio": 0.0,
|
| 673 |
+
"completion_length": 811.1510620117188,
|
| 674 |
+
"epoch": 0.3837953091684435,
|
| 675 |
+
"grad_norm": 0.10911023519160144,
|
| 676 |
+
"kl": 0.00479888916015625,
|
| 677 |
+
"learning_rate": 7.979036416534461e-07,
|
| 678 |
+
"loss": 0.0284,
|
| 679 |
+
"reward": 0.825520858168602,
|
| 680 |
+
"reward_std": 0.26300477609038353,
|
| 681 |
+
"rewards/accuracy_reward": 0.3854166716337204,
|
| 682 |
+
"rewards/format_reward": 0.0,
|
| 683 |
+
"rewards/tag_count_reward": 0.4401041716337204,
|
| 684 |
+
"step": 45
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"clip_ratio": 0.0,
|
| 688 |
+
"completion_length": 811.3385620117188,
|
| 689 |
+
"epoch": 0.39232409381663114,
|
| 690 |
+
"grad_norm": 0.11175922770281452,
|
| 691 |
+
"kl": 0.00440216064453125,
|
| 692 |
+
"learning_rate": 7.86557696414727e-07,
|
| 693 |
+
"loss": 0.0308,
|
| 694 |
+
"reward": 0.8281250149011612,
|
| 695 |
+
"reward_std": 0.3059505745768547,
|
| 696 |
+
"rewards/accuracy_reward": 0.3802083432674408,
|
| 697 |
+
"rewards/format_reward": 0.0,
|
| 698 |
+
"rewards/tag_count_reward": 0.447916679084301,
|
| 699 |
+
"step": 46
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"clip_ratio": 0.0,
|
| 703 |
+
"completion_length": 836.9531555175781,
|
| 704 |
+
"epoch": 0.40085287846481876,
|
| 705 |
+
"grad_norm": 0.10892804200516466,
|
| 706 |
+
"kl": 0.003810882568359375,
|
| 707 |
+
"learning_rate": 7.75e-07,
|
| 708 |
+
"loss": 0.0193,
|
| 709 |
+
"reward": 0.7630208432674408,
|
| 710 |
+
"reward_std": 0.2446611113846302,
|
| 711 |
+
"rewards/accuracy_reward": 0.3333333432674408,
|
| 712 |
+
"rewards/format_reward": 0.0,
|
| 713 |
+
"rewards/tag_count_reward": 0.4296875074505806,
|
| 714 |
+
"step": 47
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"clip_ratio": 0.0,
|
| 718 |
+
"completion_length": 785.0364837646484,
|
| 719 |
+
"epoch": 0.4093816631130064,
|
| 720 |
+
"grad_norm": 0.10448799382228237,
|
| 721 |
+
"kl": 0.0042266845703125,
|
| 722 |
+
"learning_rate": 7.632408981128493e-07,
|
| 723 |
+
"loss": 0.0443,
|
| 724 |
+
"reward": 0.9244792014360428,
|
| 725 |
+
"reward_std": 0.2805204316973686,
|
| 726 |
+
"rewards/accuracy_reward": 0.4583333432674408,
|
| 727 |
+
"rewards/format_reward": 0.0,
|
| 728 |
+
"rewards/tag_count_reward": 0.4661458507180214,
|
| 729 |
+
"step": 48
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"clip_ratio": 0.0,
|
| 733 |
+
"completion_length": 824.7917022705078,
|
| 734 |
+
"epoch": 0.417910447761194,
|
| 735 |
+
"grad_norm": 0.1194917798835356,
|
| 736 |
+
"kl": 0.00485992431640625,
|
| 737 |
+
"learning_rate": 7.512909167420346e-07,
|
| 738 |
+
"loss": 0.0244,
|
| 739 |
+
"reward": 0.8111979365348816,
|
| 740 |
+
"reward_std": 0.3343728110194206,
|
| 741 |
+
"rewards/accuracy_reward": 0.3750000074505806,
|
| 742 |
+
"rewards/format_reward": 0.0,
|
| 743 |
+
"rewards/tag_count_reward": 0.4361979216337204,
|
| 744 |
+
"step": 49
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"clip_ratio": 0.0,
|
| 748 |
+
"completion_length": 803.1718902587891,
|
| 749 |
+
"epoch": 0.42643923240938164,
|
| 750 |
+
"grad_norm": 0.1182071034035621,
|
| 751 |
+
"kl": 0.004547119140625,
|
| 752 |
+
"learning_rate": 7.391607527393044e-07,
|
| 753 |
+
"loss": 0.0209,
|
| 754 |
+
"reward": 0.7460937649011612,
|
| 755 |
+
"reward_std": 0.22397059574723244,
|
| 756 |
+
"rewards/accuracy_reward": 0.3281250111758709,
|
| 757 |
+
"rewards/format_reward": 0.0,
|
| 758 |
+
"rewards/tag_count_reward": 0.4179687649011612,
|
| 759 |
+
"step": 50
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"clip_ratio": 0.0,
|
| 763 |
+
"completion_length": 840.7708587646484,
|
| 764 |
+
"epoch": 0.4349680170575693,
|
| 765 |
+
"grad_norm": 0.12153276862962169,
|
| 766 |
+
"kl": 0.005279541015625,
|
| 767 |
+
"learning_rate": 7.268612642442656e-07,
|
| 768 |
+
"loss": 0.0388,
|
| 769 |
+
"reward": 0.7747395932674408,
|
| 770 |
+
"reward_std": 0.29431235045194626,
|
| 771 |
+
"rewards/accuracy_reward": 0.3281250074505806,
|
| 772 |
+
"rewards/format_reward": 0.0,
|
| 773 |
+
"rewards/tag_count_reward": 0.4466145932674408,
|
| 774 |
+
"step": 51
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"clip_ratio": 0.0,
|
| 778 |
+
"completion_length": 810.8437805175781,
|
| 779 |
+
"epoch": 0.44349680170575695,
|
| 780 |
+
"grad_norm": 0.10677125711269464,
|
| 781 |
+
"kl": 0.005016326904296875,
|
| 782 |
+
"learning_rate": 7.144034609648778e-07,
|
| 783 |
+
"loss": 0.0308,
|
| 784 |
+
"reward": 0.9335937798023224,
|
| 785 |
+
"reward_std": 0.2992444187402725,
|
| 786 |
+
"rewards/accuracy_reward": 0.447916679084301,
|
| 787 |
+
"rewards/format_reward": 0.0,
|
| 788 |
+
"rewards/tag_count_reward": 0.4856770932674408,
|
| 789 |
+
"step": 52
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"clip_ratio": 0.0,
|
| 793 |
+
"completion_length": 850.2135620117188,
|
| 794 |
+
"epoch": 0.4520255863539446,
|
| 795 |
+
"grad_norm": 0.09689246162817927,
|
| 796 |
+
"kl": 0.004337310791015625,
|
| 797 |
+
"learning_rate": 7.017984943222735e-07,
|
| 798 |
+
"loss": 0.0339,
|
| 799 |
+
"reward": 0.8098958730697632,
|
| 800 |
+
"reward_std": 0.24391771852970123,
|
| 801 |
+
"rewards/accuracy_reward": 0.385416679084301,
|
| 802 |
+
"rewards/format_reward": 0.0,
|
| 803 |
+
"rewards/tag_count_reward": 0.424479179084301,
|
| 804 |
+
"step": 53
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"clip_ratio": 0.0,
|
| 808 |
+
"completion_length": 801.6250152587891,
|
| 809 |
+
"epoch": 0.4605543710021322,
|
| 810 |
+
"grad_norm": 0.1110145602964871,
|
| 811 |
+
"kl": 0.00634002685546875,
|
| 812 |
+
"learning_rate": 6.890576474687263e-07,
|
| 813 |
+
"loss": 0.0443,
|
| 814 |
+
"reward": 0.8710937649011612,
|
| 815 |
+
"reward_std": 0.2881768085062504,
|
| 816 |
+
"rewards/accuracy_reward": 0.4166666865348816,
|
| 817 |
+
"rewards/format_reward": 0.0,
|
| 818 |
+
"rewards/tag_count_reward": 0.4544270932674408,
|
| 819 |
+
"step": 54
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"clip_ratio": 0.0,
|
| 823 |
+
"completion_length": 793.8750152587891,
|
| 824 |
+
"epoch": 0.4690831556503198,
|
| 825 |
+
"grad_norm": 0.11287919842468197,
|
| 826 |
+
"kl": 0.00574493408203125,
|
| 827 |
+
"learning_rate": 6.761923251877012e-07,
|
| 828 |
+
"loss": 0.0196,
|
| 829 |
+
"reward": 0.9309895932674408,
|
| 830 |
+
"reward_std": 0.252448420971632,
|
| 831 |
+
"rewards/accuracy_reward": 0.4531250074505806,
|
| 832 |
+
"rewards/format_reward": 0.0,
|
| 833 |
+
"rewards/tag_count_reward": 0.4778645932674408,
|
| 834 |
+
"step": 55
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"clip_ratio": 0.0,
|
| 838 |
+
"completion_length": 822.9062652587891,
|
| 839 |
+
"epoch": 0.47761194029850745,
|
| 840 |
+
"grad_norm": 0.10712880691483767,
|
| 841 |
+
"kl": 0.0055999755859375,
|
| 842 |
+
"learning_rate": 6.632140436850289e-07,
|
| 843 |
+
"loss": 0.0347,
|
| 844 |
+
"reward": 0.8632812798023224,
|
| 845 |
+
"reward_std": 0.2900092527270317,
|
| 846 |
+
"rewards/accuracy_reward": 0.4010416716337204,
|
| 847 |
+
"rewards/format_reward": 0.0,
|
| 848 |
+
"rewards/tag_count_reward": 0.4622395932674408,
|
| 849 |
+
"step": 56
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"clip_ratio": 0.0,
|
| 853 |
+
"completion_length": 841.6718902587891,
|
| 854 |
+
"epoch": 0.4861407249466951,
|
| 855 |
+
"grad_norm": 0.11360218470340229,
|
| 856 |
+
"kl": 0.00505828857421875,
|
| 857 |
+
"learning_rate": 6.501344202803414e-07,
|
| 858 |
+
"loss": 0.0278,
|
| 859 |
+
"reward": 0.8776041865348816,
|
| 860 |
+
"reward_std": 0.2797526493668556,
|
| 861 |
+
"rewards/accuracy_reward": 0.4270833358168602,
|
| 862 |
+
"rewards/format_reward": 0.0,
|
| 863 |
+
"rewards/tag_count_reward": 0.4505208432674408,
|
| 864 |
+
"step": 57
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"clip_ratio": 0.0,
|
| 868 |
+
"completion_length": 786.0885620117188,
|
| 869 |
+
"epoch": 0.4946695095948827,
|
| 870 |
+
"grad_norm": 0.09660671347801546,
|
| 871 |
+
"kl": 0.00525665283203125,
|
| 872 |
+
"learning_rate": 6.36965163007999e-07,
|
| 873 |
+
"loss": 0.0177,
|
| 874 |
+
"reward": 0.9049479365348816,
|
| 875 |
+
"reward_std": 0.2301800586283207,
|
| 876 |
+
"rewards/accuracy_reward": 0.447916679084301,
|
| 877 |
+
"rewards/format_reward": 0.0,
|
| 878 |
+
"rewards/tag_count_reward": 0.4570312649011612,
|
| 879 |
+
"step": 58
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"clip_ratio": 0.0,
|
| 883 |
+
"completion_length": 865.3593902587891,
|
| 884 |
+
"epoch": 0.5031982942430704,
|
| 885 |
+
"grad_norm": 0.10774282361431926,
|
| 886 |
+
"kl": 0.0062103271484375,
|
| 887 |
+
"learning_rate": 6.23718060136812e-07,
|
| 888 |
+
"loss": 0.0291,
|
| 889 |
+
"reward": 0.8606770932674408,
|
| 890 |
+
"reward_std": 0.3248429372906685,
|
| 891 |
+
"rewards/accuracy_reward": 0.4114583507180214,
|
| 892 |
+
"rewards/format_reward": 0.0,
|
| 893 |
+
"rewards/tag_count_reward": 0.4492187649011612,
|
| 894 |
+
"step": 59
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"clip_ratio": 0.0,
|
| 898 |
+
"completion_length": 851.3437805175781,
|
| 899 |
+
"epoch": 0.511727078891258,
|
| 900 |
+
"grad_norm": 0.10751507893398955,
|
| 901 |
+
"kl": 0.0060577392578125,
|
| 902 |
+
"learning_rate": 6.10404969617945e-07,
|
| 903 |
+
"loss": 0.0275,
|
| 904 |
+
"reward": 0.7994791865348816,
|
| 905 |
+
"reward_std": 0.2524433769285679,
|
| 906 |
+
"rewards/accuracy_reward": 0.3750000074505806,
|
| 907 |
+
"rewards/format_reward": 0.0,
|
| 908 |
+
"rewards/tag_count_reward": 0.4244791716337204,
|
| 909 |
+
"step": 60
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"clip_ratio": 0.0,
|
| 913 |
+
"completion_length": 839.1146087646484,
|
| 914 |
+
"epoch": 0.5202558635394456,
|
| 915 |
+
"grad_norm": 0.11847979500014862,
|
| 916 |
+
"kl": 0.00791168212890625,
|
| 917 |
+
"learning_rate": 5.97037808470444e-07,
|
| 918 |
+
"loss": 0.047,
|
| 919 |
+
"reward": 0.8437500149011612,
|
| 920 |
+
"reward_std": 0.27961887419223785,
|
| 921 |
+
"rewards/accuracy_reward": 0.3906250149011612,
|
| 922 |
+
"rewards/format_reward": 0.0,
|
| 923 |
+
"rewards/tag_count_reward": 0.4531250149011612,
|
| 924 |
+
"step": 61
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"clip_ratio": 0.0,
|
| 928 |
+
"completion_length": 772.3385620117188,
|
| 929 |
+
"epoch": 0.5287846481876333,
|
| 930 |
+
"grad_norm": 0.1316610193310866,
|
| 931 |
+
"kl": 0.006988525390625,
|
| 932 |
+
"learning_rate": 5.836285421138909e-07,
|
| 933 |
+
"loss": 0.0523,
|
| 934 |
+
"reward": 0.8984375298023224,
|
| 935 |
+
"reward_std": 0.24485836923122406,
|
| 936 |
+
"rewards/accuracy_reward": 0.3958333432674408,
|
| 937 |
+
"rewards/format_reward": 0.0,
|
| 938 |
+
"rewards/tag_count_reward": 0.5026041939854622,
|
| 939 |
+
"step": 62
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"clip_ratio": 0.0,
|
| 943 |
+
"completion_length": 814.2760620117188,
|
| 944 |
+
"epoch": 0.5373134328358209,
|
| 945 |
+
"grad_norm": 0.11592047451693034,
|
| 946 |
+
"kl": 0.0073089599609375,
|
| 947 |
+
"learning_rate": 5.701891736577317e-07,
|
| 948 |
+
"loss": 0.0358,
|
| 949 |
+
"reward": 0.813802108168602,
|
| 950 |
+
"reward_std": 0.2582968771457672,
|
| 951 |
+
"rewards/accuracy_reward": 0.3645833432674408,
|
| 952 |
+
"rewards/format_reward": 0.0,
|
| 953 |
+
"rewards/tag_count_reward": 0.4492187649011612,
|
| 954 |
+
"step": 63
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"clip_ratio": 0.0,
|
| 958 |
+
"completion_length": 769.1250305175781,
|
| 959 |
+
"epoch": 0.5458422174840085,
|
| 960 |
+
"grad_norm": 0.13152164686975176,
|
| 961 |
+
"kl": 0.0075531005859375,
|
| 962 |
+
"learning_rate": 5.567317331568686e-07,
|
| 963 |
+
"loss": 0.0389,
|
| 964 |
+
"reward": 0.9752604514360428,
|
| 965 |
+
"reward_std": 0.3396269455552101,
|
| 966 |
+
"rewards/accuracy_reward": 0.479166679084301,
|
| 967 |
+
"rewards/format_reward": 0.0,
|
| 968 |
+
"rewards/tag_count_reward": 0.4960937649011612,
|
| 969 |
+
"step": 64
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"clip_ratio": 0.0,
|
| 973 |
+
"completion_length": 833.2187652587891,
|
| 974 |
+
"epoch": 0.5543710021321961,
|
| 975 |
+
"grad_norm": 0.11640227796478432,
|
| 976 |
+
"kl": 0.0069122314453125,
|
| 977 |
+
"learning_rate": 5.432682668431314e-07,
|
| 978 |
+
"loss": 0.0338,
|
| 979 |
+
"reward": 0.8502604514360428,
|
| 980 |
+
"reward_std": 0.2958378382027149,
|
| 981 |
+
"rewards/accuracy_reward": 0.4062500074505806,
|
| 982 |
+
"rewards/format_reward": 0.0,
|
| 983 |
+
"rewards/tag_count_reward": 0.4440104216337204,
|
| 984 |
+
"step": 65
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"clip_ratio": 0.0,
|
| 988 |
+
"completion_length": 786.2864837646484,
|
| 989 |
+
"epoch": 0.5628997867803838,
|
| 990 |
+
"grad_norm": 0.14642542124355792,
|
| 991 |
+
"kl": 0.0079498291015625,
|
| 992 |
+
"learning_rate": 5.298108263422685e-07,
|
| 993 |
+
"loss": 0.0443,
|
| 994 |
+
"reward": 1.0013021379709244,
|
| 995 |
+
"reward_std": 0.37428252026438713,
|
| 996 |
+
"rewards/accuracy_reward": 0.510416679084301,
|
| 997 |
+
"rewards/format_reward": 0.0,
|
| 998 |
+
"rewards/tag_count_reward": 0.490885429084301,
|
| 999 |
+
"step": 66
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"clip_ratio": 0.0,
|
| 1003 |
+
"completion_length": 799.9896087646484,
|
| 1004 |
+
"epoch": 0.5714285714285714,
|
| 1005 |
+
"grad_norm": 0.13169692017238832,
|
| 1006 |
+
"kl": 0.00823211669921875,
|
| 1007 |
+
"learning_rate": 5.163714578861091e-07,
|
| 1008 |
+
"loss": 0.0526,
|
| 1009 |
+
"reward": 0.8997395932674408,
|
| 1010 |
+
"reward_std": 0.33160682767629623,
|
| 1011 |
+
"rewards/accuracy_reward": 0.4010416865348816,
|
| 1012 |
+
"rewards/format_reward": 0.0,
|
| 1013 |
+
"rewards/tag_count_reward": 0.4986979439854622,
|
| 1014 |
+
"step": 67
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"clip_ratio": 0.0,
|
| 1018 |
+
"completion_length": 857.1354370117188,
|
| 1019 |
+
"epoch": 0.579957356076759,
|
| 1020 |
+
"grad_norm": 0.1069980915235563,
|
| 1021 |
+
"kl": 0.007354736328125,
|
| 1022 |
+
"learning_rate": 5.02962191529556e-07,
|
| 1023 |
+
"loss": 0.0278,
|
| 1024 |
+
"reward": 0.7916666865348816,
|
| 1025 |
+
"reward_std": 0.2258782833814621,
|
| 1026 |
+
"rewards/accuracy_reward": 0.3541666753590107,
|
| 1027 |
+
"rewards/format_reward": 0.0,
|
| 1028 |
+
"rewards/tag_count_reward": 0.4375000149011612,
|
| 1029 |
+
"step": 68
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"clip_ratio": 0.0,
|
| 1033 |
+
"completion_length": 865.6406402587891,
|
| 1034 |
+
"epoch": 0.5884861407249466,
|
| 1035 |
+
"grad_norm": 0.10309499954991647,
|
| 1036 |
+
"kl": 0.0067901611328125,
|
| 1037 |
+
"learning_rate": 4.895950303820552e-07,
|
| 1038 |
+
"loss": 0.0296,
|
| 1039 |
+
"reward": 0.774739608168602,
|
| 1040 |
+
"reward_std": 0.2874305471777916,
|
| 1041 |
+
"rewards/accuracy_reward": 0.3437500074505806,
|
| 1042 |
+
"rewards/format_reward": 0.0,
|
| 1043 |
+
"rewards/tag_count_reward": 0.4309895932674408,
|
| 1044 |
+
"step": 69
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"clip_ratio": 0.0,
|
| 1048 |
+
"completion_length": 821.2344055175781,
|
| 1049 |
+
"epoch": 0.5970149253731343,
|
| 1050 |
+
"grad_norm": 0.13248930196397823,
|
| 1051 |
+
"kl": 0.00777435302734375,
|
| 1052 |
+
"learning_rate": 4.76281939863188e-07,
|
| 1053 |
+
"loss": 0.0537,
|
| 1054 |
+
"reward": 0.9023437798023224,
|
| 1055 |
+
"reward_std": 0.3311818540096283,
|
| 1056 |
+
"rewards/accuracy_reward": 0.4270833432674408,
|
| 1057 |
+
"rewards/format_reward": 0.0,
|
| 1058 |
+
"rewards/tag_count_reward": 0.475260429084301,
|
| 1059 |
+
"step": 70
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"clip_ratio": 0.0,
|
| 1063 |
+
"completion_length": 810.8177337646484,
|
| 1064 |
+
"epoch": 0.605543710021322,
|
| 1065 |
+
"grad_norm": 0.11529960531868733,
|
| 1066 |
+
"kl": 0.008636474609375,
|
| 1067 |
+
"learning_rate": 4.6303483699200105e-07,
|
| 1068 |
+
"loss": 0.0261,
|
| 1069 |
+
"reward": 0.9192708730697632,
|
| 1070 |
+
"reward_std": 0.3068869337439537,
|
| 1071 |
+
"rewards/accuracy_reward": 0.4531250149011612,
|
| 1072 |
+
"rewards/format_reward": 0.0,
|
| 1073 |
+
"rewards/tag_count_reward": 0.4661458358168602,
|
| 1074 |
+
"step": 71
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"clip_ratio": 0.0,
|
| 1078 |
+
"completion_length": 801.1770935058594,
|
| 1079 |
+
"epoch": 0.6140724946695096,
|
| 1080 |
+
"grad_norm": 0.11755133653830964,
|
| 1081 |
+
"kl": 0.010284423828125,
|
| 1082 |
+
"learning_rate": 4.4986557971965856e-07,
|
| 1083 |
+
"loss": 0.0391,
|
| 1084 |
+
"reward": 0.8359375298023224,
|
| 1085 |
+
"reward_std": 0.2929926961660385,
|
| 1086 |
+
"rewards/accuracy_reward": 0.3697916753590107,
|
| 1087 |
+
"rewards/format_reward": 0.0,
|
| 1088 |
+
"rewards/tag_count_reward": 0.4661458432674408,
|
| 1089 |
+
"step": 72
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"clip_ratio": 0.0,
|
| 1093 |
+
"completion_length": 786.2500305175781,
|
| 1094 |
+
"epoch": 0.6226012793176973,
|
| 1095 |
+
"grad_norm": 0.14898312092719013,
|
| 1096 |
+
"kl": 0.00911712646484375,
|
| 1097 |
+
"learning_rate": 4.367859563149712e-07,
|
| 1098 |
+
"loss": 0.0556,
|
| 1099 |
+
"reward": 0.962239608168602,
|
| 1100 |
+
"reward_std": 0.3554273024201393,
|
| 1101 |
+
"rewards/accuracy_reward": 0.463541679084301,
|
| 1102 |
+
"rewards/format_reward": 0.0,
|
| 1103 |
+
"rewards/tag_count_reward": 0.4986979365348816,
|
| 1104 |
+
"step": 73
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"clip_ratio": 0.0,
|
| 1108 |
+
"completion_length": 840.5000305175781,
|
| 1109 |
+
"epoch": 0.6311300639658849,
|
| 1110 |
+
"grad_norm": 0.1084603067680966,
|
| 1111 |
+
"kl": 0.0075225830078125,
|
| 1112 |
+
"learning_rate": 4.2380767481229884e-07,
|
| 1113 |
+
"loss": 0.0177,
|
| 1114 |
+
"reward": 0.9062500298023224,
|
| 1115 |
+
"reward_std": 0.2926729619503021,
|
| 1116 |
+
"rewards/accuracy_reward": 0.4427083358168602,
|
| 1117 |
+
"rewards/format_reward": 0.0,
|
| 1118 |
+
"rewards/tag_count_reward": 0.4635416865348816,
|
| 1119 |
+
"step": 74
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"clip_ratio": 0.0,
|
| 1123 |
+
"completion_length": 793.2135772705078,
|
| 1124 |
+
"epoch": 0.6396588486140725,
|
| 1125 |
+
"grad_norm": 0.11707168848017023,
|
| 1126 |
+
"kl": 0.00870513916015625,
|
| 1127 |
+
"learning_rate": 4.1094235253127374e-07,
|
| 1128 |
+
"loss": 0.0225,
|
| 1129 |
+
"reward": 0.971354216337204,
|
| 1130 |
+
"reward_std": 0.28287459164857864,
|
| 1131 |
+
"rewards/accuracy_reward": 0.4739583507180214,
|
| 1132 |
+
"rewards/format_reward": 0.0,
|
| 1133 |
+
"rewards/tag_count_reward": 0.497395858168602,
|
| 1134 |
+
"step": 75
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"clip_ratio": 0.0,
|
| 1138 |
+
"completion_length": 790.9062652587891,
|
| 1139 |
+
"epoch": 0.6481876332622601,
|
| 1140 |
+
"grad_norm": 0.11531067175734634,
|
| 1141 |
+
"kl": 0.0094451904296875,
|
| 1142 |
+
"learning_rate": 3.982015056777265e-07,
|
| 1143 |
+
"loss": 0.0297,
|
| 1144 |
+
"reward": 0.9557291865348816,
|
| 1145 |
+
"reward_std": 0.25578880310058594,
|
| 1146 |
+
"rewards/accuracy_reward": 0.4843750149011612,
|
| 1147 |
+
"rewards/format_reward": 0.0,
|
| 1148 |
+
"rewards/tag_count_reward": 0.4713541865348816,
|
| 1149 |
+
"step": 76
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"clip_ratio": 0.0,
|
| 1153 |
+
"completion_length": 768.7239685058594,
|
| 1154 |
+
"epoch": 0.6567164179104478,
|
| 1155 |
+
"grad_norm": 0.1320106685409761,
|
| 1156 |
+
"kl": 0.0102386474609375,
|
| 1157 |
+
"learning_rate": 3.855965390351222e-07,
|
| 1158 |
+
"loss": 0.0417,
|
| 1159 |
+
"reward": 0.962239608168602,
|
| 1160 |
+
"reward_std": 0.2732561454176903,
|
| 1161 |
+
"rewards/accuracy_reward": 0.4635416716337204,
|
| 1162 |
+
"rewards/format_reward": 0.0,
|
| 1163 |
+
"rewards/tag_count_reward": 0.498697929084301,
|
| 1164 |
+
"step": 77
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"clip_ratio": 0.0,
|
| 1168 |
+
"completion_length": 781.3489685058594,
|
| 1169 |
+
"epoch": 0.6652452025586354,
|
| 1170 |
+
"grad_norm": 0.1344780448279913,
|
| 1171 |
+
"kl": 0.00832366943359375,
|
| 1172 |
+
"learning_rate": 3.731387357557344e-07,
|
| 1173 |
+
"loss": 0.0452,
|
| 1174 |
+
"reward": 1.0169271230697632,
|
| 1175 |
+
"reward_std": 0.35733717679977417,
|
| 1176 |
+
"rewards/accuracy_reward": 0.5000000149011612,
|
| 1177 |
+
"rewards/format_reward": 0.0,
|
| 1178 |
+
"rewards/tag_count_reward": 0.5169271007180214,
|
| 1179 |
+
"step": 78
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"clip_ratio": 0.0,
|
| 1183 |
+
"completion_length": 799.0677185058594,
|
| 1184 |
+
"epoch": 0.673773987206823,
|
| 1185 |
+
"grad_norm": 0.12035730419810114,
|
| 1186 |
+
"kl": 0.00931549072265625,
|
| 1187 |
+
"learning_rate": 3.608392472606956e-07,
|
| 1188 |
+
"loss": 0.044,
|
| 1189 |
+
"reward": 0.9049479365348816,
|
| 1190 |
+
"reward_std": 0.32109028846025467,
|
| 1191 |
+
"rewards/accuracy_reward": 0.4375000074505806,
|
| 1192 |
+
"rewards/format_reward": 0.0,
|
| 1193 |
+
"rewards/tag_count_reward": 0.467447929084301,
|
| 1194 |
+
"step": 79
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"clip_ratio": 0.0,
|
| 1198 |
+
"completion_length": 868.9323120117188,
|
| 1199 |
+
"epoch": 0.6823027718550106,
|
| 1200 |
+
"grad_norm": 0.10803684448437832,
|
| 1201 |
+
"kl": 0.00897979736328125,
|
| 1202 |
+
"learning_rate": 3.4870908325796527e-07,
|
| 1203 |
+
"loss": 0.0308,
|
| 1204 |
+
"reward": 0.774739608168602,
|
| 1205 |
+
"reward_std": 0.28679391369223595,
|
| 1206 |
+
"rewards/accuracy_reward": 0.3437500074505806,
|
| 1207 |
+
"rewards/format_reward": 0.0,
|
| 1208 |
+
"rewards/tag_count_reward": 0.4309895858168602,
|
| 1209 |
+
"step": 80
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"clip_ratio": 0.0,
|
| 1213 |
+
"completion_length": 830.3906402587891,
|
| 1214 |
+
"epoch": 0.6908315565031983,
|
| 1215 |
+
"grad_norm": 0.11935680801330077,
|
| 1216 |
+
"kl": 0.00799560546875,
|
| 1217 |
+
"learning_rate": 3.367591018871506e-07,
|
| 1218 |
+
"loss": 0.0304,
|
| 1219 |
+
"reward": 0.9270833730697632,
|
| 1220 |
+
"reward_std": 0.3459775671362877,
|
| 1221 |
+
"rewards/accuracy_reward": 0.4375000074505806,
|
| 1222 |
+
"rewards/format_reward": 0.0,
|
| 1223 |
+
"rewards/tag_count_reward": 0.4895833432674408,
|
| 1224 |
+
"step": 81
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"clip_ratio": 0.0,
|
| 1228 |
+
"completion_length": 804.4531555175781,
|
| 1229 |
+
"epoch": 0.6993603411513859,
|
| 1230 |
+
"grad_norm": 0.13609643645529676,
|
| 1231 |
+
"kl": 0.00872802734375,
|
| 1232 |
+
"learning_rate": 3.250000000000001e-07,
|
| 1233 |
+
"loss": 0.0571,
|
| 1234 |
+
"reward": 0.8867187649011612,
|
| 1235 |
+
"reward_std": 0.33319345116615295,
|
| 1236 |
+
"rewards/accuracy_reward": 0.401041679084301,
|
| 1237 |
+
"rewards/format_reward": 0.0,
|
| 1238 |
+
"rewards/tag_count_reward": 0.4856770932674408,
|
| 1239 |
+
"step": 82
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"clip_ratio": 0.0,
|
| 1243 |
+
"completion_length": 777.0416870117188,
|
| 1244 |
+
"epoch": 0.7078891257995735,
|
| 1245 |
+
"grad_norm": 0.13067073463194692,
|
| 1246 |
+
"kl": 0.00905609130859375,
|
| 1247 |
+
"learning_rate": 3.1344230358527284e-07,
|
| 1248 |
+
"loss": 0.0374,
|
| 1249 |
+
"reward": 1.033854216337204,
|
| 1250 |
+
"reward_std": 0.3256627693772316,
|
| 1251 |
+
"rewards/accuracy_reward": 0.5,
|
| 1252 |
+
"rewards/format_reward": 0.0,
|
| 1253 |
+
"rewards/tag_count_reward": 0.5338541865348816,
|
| 1254 |
+
"step": 83
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"clip_ratio": 0.0,
|
| 1258 |
+
"completion_length": 805.9896087646484,
|
| 1259 |
+
"epoch": 0.7164179104477612,
|
| 1260 |
+
"grad_norm": 0.14281987226405615,
|
| 1261 |
+
"kl": 0.0091705322265625,
|
| 1262 |
+
"learning_rate": 3.020963583465539e-07,
|
| 1263 |
+
"loss": 0.0382,
|
| 1264 |
+
"reward": 0.9401041716337204,
|
| 1265 |
+
"reward_std": 0.2688889801502228,
|
| 1266 |
+
"rewards/accuracy_reward": 0.4375000074505806,
|
| 1267 |
+
"rewards/format_reward": 0.0,
|
| 1268 |
+
"rewards/tag_count_reward": 0.502604179084301,
|
| 1269 |
+
"step": 84
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"clip_ratio": 0.0,
|
| 1273 |
+
"completion_length": 861.2500152587891,
|
| 1274 |
+
"epoch": 0.7249466950959488,
|
| 1275 |
+
"grad_norm": 0.11279936057889765,
|
| 1276 |
+
"kl": 0.008758544921875,
|
| 1277 |
+
"learning_rate": 2.9097232044144696e-07,
|
| 1278 |
+
"loss": 0.0312,
|
| 1279 |
+
"reward": 0.8671875149011612,
|
| 1280 |
+
"reward_std": 0.31369951367378235,
|
| 1281 |
+
"rewards/accuracy_reward": 0.3958333432674408,
|
| 1282 |
+
"rewards/format_reward": 0.0,
|
| 1283 |
+
"rewards/tag_count_reward": 0.471354179084301,
|
| 1284 |
+
"step": 85
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"clip_ratio": 0.0,
|
| 1288 |
+
"completion_length": 816.1562652587891,
|
| 1289 |
+
"epoch": 0.7334754797441365,
|
| 1290 |
+
"grad_norm": 0.10042567088503304,
|
| 1291 |
+
"kl": 0.0089874267578125,
|
| 1292 |
+
"learning_rate": 2.8008014739043884e-07,
|
| 1293 |
+
"loss": 0.0312,
|
| 1294 |
+
"reward": 0.7968750149011612,
|
| 1295 |
+
"reward_std": 0.28011347353458405,
|
| 1296 |
+
"rewards/accuracy_reward": 0.3593750074505806,
|
| 1297 |
+
"rewards/format_reward": 0.0,
|
| 1298 |
+
"rewards/tag_count_reward": 0.4375000074505806,
|
| 1299 |
+
"step": 86
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"clip_ratio": 0.0,
|
| 1303 |
+
"completion_length": 815.0521087646484,
|
| 1304 |
+
"epoch": 0.7420042643923241,
|
| 1305 |
+
"grad_norm": 0.14295156229246506,
|
| 1306 |
+
"kl": 0.0097198486328125,
|
| 1307 |
+
"learning_rate": 2.6942958916356994e-07,
|
| 1308 |
+
"loss": 0.0584,
|
| 1309 |
+
"reward": 0.8971354514360428,
|
| 1310 |
+
"reward_std": 0.3291558399796486,
|
| 1311 |
+
"rewards/accuracy_reward": 0.416666679084301,
|
| 1312 |
+
"rewards/format_reward": 0.0,
|
| 1313 |
+
"rewards/tag_count_reward": 0.4804687574505806,
|
| 1314 |
+
"step": 87
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"clip_ratio": 0.0,
|
| 1318 |
+
"completion_length": 801.5260772705078,
|
| 1319 |
+
"epoch": 0.7505330490405118,
|
| 1320 |
+
"grad_norm": 0.11305321904987985,
|
| 1321 |
+
"kl": 0.0097808837890625,
|
| 1322 |
+
"learning_rate": 2.5903017945289017e-07,
|
| 1323 |
+
"loss": 0.0148,
|
| 1324 |
+
"reward": 0.9322916716337204,
|
| 1325 |
+
"reward_std": 0.24261895567178726,
|
| 1326 |
+
"rewards/accuracy_reward": 0.447916679084301,
|
| 1327 |
+
"rewards/format_reward": 0.0,
|
| 1328 |
+
"rewards/tag_count_reward": 0.4843750223517418,
|
| 1329 |
+
"step": 88
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"clip_ratio": 0.0,
|
| 1333 |
+
"completion_length": 803.7083435058594,
|
| 1334 |
+
"epoch": 0.7590618336886994,
|
| 1335 |
+
"grad_norm": 0.11745643059040382,
|
| 1336 |
+
"kl": 0.0095367431640625,
|
| 1337 |
+
"learning_rate": 2.488912271385139e-07,
|
| 1338 |
+
"loss": 0.0416,
|
| 1339 |
+
"reward": 0.9973958730697632,
|
| 1340 |
+
"reward_std": 0.31152600049972534,
|
| 1341 |
+
"rewards/accuracy_reward": 0.5000000149011612,
|
| 1342 |
+
"rewards/format_reward": 0.0,
|
| 1343 |
+
"rewards/tag_count_reward": 0.4973958432674408,
|
| 1344 |
+
"step": 89
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"clip_ratio": 0.0,
|
| 1348 |
+
"completion_length": 793.3125152587891,
|
| 1349 |
+
"epoch": 0.767590618336887,
|
| 1350 |
+
"grad_norm": 0.12149395755942659,
|
| 1351 |
+
"kl": 0.0098876953125,
|
| 1352 |
+
"learning_rate": 2.390218079559109e-07,
|
| 1353 |
+
"loss": 0.0244,
|
| 1354 |
+
"reward": 1.0351562798023224,
|
| 1355 |
+
"reward_std": 0.30135782063007355,
|
| 1356 |
+
"rewards/accuracy_reward": 0.5364583507180214,
|
| 1357 |
+
"rewards/format_reward": 0.0,
|
| 1358 |
+
"rewards/tag_count_reward": 0.4986979365348816,
|
| 1359 |
+
"step": 90
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"clip_ratio": 0.0,
|
| 1363 |
+
"completion_length": 832.4114685058594,
|
| 1364 |
+
"epoch": 0.7761194029850746,
|
| 1365 |
+
"grad_norm": 0.1284291053033323,
|
| 1366 |
+
"kl": 0.01007080078125,
|
| 1367 |
+
"learning_rate": 2.294307563718949e-07,
|
| 1368 |
+
"loss": 0.0419,
|
| 1369 |
+
"reward": 0.845052108168602,
|
| 1370 |
+
"reward_std": 0.24289947375655174,
|
| 1371 |
+
"rewards/accuracy_reward": 0.385416679084301,
|
| 1372 |
+
"rewards/format_reward": 0.0,
|
| 1373 |
+
"rewards/tag_count_reward": 0.459635429084301,
|
| 1374 |
+
"step": 91
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"clip_ratio": 0.0,
|
| 1378 |
+
"completion_length": 810.6875152587891,
|
| 1379 |
+
"epoch": 0.7846481876332623,
|
| 1380 |
+
"grad_norm": 0.12108019651078701,
|
| 1381 |
+
"kl": 0.0109405517578125,
|
| 1382 |
+
"learning_rate": 2.2012665767657823e-07,
|
| 1383 |
+
"loss": 0.0228,
|
| 1384 |
+
"reward": 0.8593750298023224,
|
| 1385 |
+
"reward_std": 0.2939864583313465,
|
| 1386 |
+
"rewards/accuracy_reward": 0.3802083432674408,
|
| 1387 |
+
"rewards/format_reward": 0.0,
|
| 1388 |
+
"rewards/tag_count_reward": 0.479166679084301,
|
| 1389 |
+
"step": 92
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"clip_ratio": 0.0,
|
| 1393 |
+
"completion_length": 816.0833435058594,
|
| 1394 |
+
"epoch": 0.7931769722814499,
|
| 1395 |
+
"grad_norm": 0.1248794400539759,
|
| 1396 |
+
"kl": 0.0101318359375,
|
| 1397 |
+
"learning_rate": 2.1111784029837509e-07,
|
| 1398 |
+
"loss": 0.0366,
|
| 1399 |
+
"reward": 0.8750000149011612,
|
| 1400 |
+
"reward_std": 0.3214331939816475,
|
| 1401 |
+
"rewards/accuracy_reward": 0.3750000149011612,
|
| 1402 |
+
"rewards/format_reward": 0.0,
|
| 1403 |
+
"rewards/tag_count_reward": 0.5000000223517418,
|
| 1404 |
+
"step": 93
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"clip_ratio": 0.0,
|
| 1408 |
+
"completion_length": 759.9219055175781,
|
| 1409 |
+
"epoch": 0.8017057569296375,
|
| 1410 |
+
"grad_norm": 0.14143793722100087,
|
| 1411 |
+
"kl": 0.010162353515625,
|
| 1412 |
+
"learning_rate": 2.0241236834893028e-07,
|
| 1413 |
+
"loss": 0.0446,
|
| 1414 |
+
"reward": 1.0013021230697632,
|
| 1415 |
+
"reward_std": 0.30526329576969147,
|
| 1416 |
+
"rewards/accuracy_reward": 0.4947916865348816,
|
| 1417 |
+
"rewards/format_reward": 0.0,
|
| 1418 |
+
"rewards/tag_count_reward": 0.5065104365348816,
|
| 1419 |
+
"step": 94
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"clip_ratio": 0.0,
|
| 1423 |
+
"completion_length": 800.8021087646484,
|
| 1424 |
+
"epoch": 0.8102345415778252,
|
| 1425 |
+
"grad_norm": 0.1243853221942541,
|
| 1426 |
+
"kl": 0.010223388671875,
|
| 1427 |
+
"learning_rate": 1.9401803440464654e-07,
|
| 1428 |
+
"loss": 0.0342,
|
| 1429 |
+
"reward": 0.958333358168602,
|
| 1430 |
+
"reward_std": 0.2869706079363823,
|
| 1431 |
+
"rewards/accuracy_reward": 0.4427083507180214,
|
| 1432 |
+
"rewards/format_reward": 0.0,
|
| 1433 |
+
"rewards/tag_count_reward": 0.5156250223517418,
|
| 1434 |
+
"step": 95
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"clip_ratio": 0.0,
|
| 1438 |
+
"completion_length": 803.5104370117188,
|
| 1439 |
+
"epoch": 0.8187633262260128,
|
| 1440 |
+
"grad_norm": 0.1306441755839297,
|
| 1441 |
+
"kl": 0.0102386474609375,
|
| 1442 |
+
"learning_rate": 1.8594235253127372e-07,
|
| 1443 |
+
"loss": 0.0602,
|
| 1444 |
+
"reward": 0.9882812649011612,
|
| 1445 |
+
"reward_std": 0.2958815321326256,
|
| 1446 |
+
"rewards/accuracy_reward": 0.4895833432674408,
|
| 1447 |
+
"rewards/format_reward": 0.0,
|
| 1448 |
+
"rewards/tag_count_reward": 0.4986979365348816,
|
| 1449 |
+
"step": 96
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"clip_ratio": 0.0,
|
| 1453 |
+
"completion_length": 812.1979370117188,
|
| 1454 |
+
"epoch": 0.8272921108742004,
|
| 1455 |
+
"grad_norm": 0.12020848008929602,
|
| 1456 |
+
"kl": 0.0102691650390625,
|
| 1457 |
+
"learning_rate": 1.7819255155780238e-07,
|
| 1458 |
+
"loss": 0.0299,
|
| 1459 |
+
"reward": 0.8619791865348816,
|
| 1460 |
+
"reward_std": 0.3012324385344982,
|
| 1461 |
+
"rewards/accuracy_reward": 0.3906250149011612,
|
| 1462 |
+
"rewards/format_reward": 0.0,
|
| 1463 |
+
"rewards/tag_count_reward": 0.4713541865348816,
|
| 1464 |
+
"step": 97
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"clip_ratio": 0.0,
|
| 1468 |
+
"completion_length": 868.9895935058594,
|
| 1469 |
+
"epoch": 0.835820895522388,
|
| 1470 |
+
"grad_norm": 0.11268918840030598,
|
| 1471 |
+
"kl": 0.0112152099609375,
|
| 1472 |
+
"learning_rate": 1.7077556860568238e-07,
|
| 1473 |
+
"loss": 0.0146,
|
| 1474 |
+
"reward": 0.7968750149011612,
|
| 1475 |
+
"reward_std": 0.2770479694008827,
|
| 1476 |
+
"rewards/accuracy_reward": 0.3645833432674408,
|
| 1477 |
+
"rewards/format_reward": 0.0,
|
| 1478 |
+
"rewards/tag_count_reward": 0.432291679084301,
|
| 1479 |
+
"step": 98
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"clip_ratio": 0.0,
|
| 1483 |
+
"completion_length": 817.0625305175781,
|
| 1484 |
+
"epoch": 0.8443496801705757,
|
| 1485 |
+
"grad_norm": 0.11763553136010904,
|
| 1486 |
+
"kl": 0.01142120361328125,
|
| 1487 |
+
"learning_rate": 1.6369804287916025e-07,
|
| 1488 |
+
"loss": 0.0342,
|
| 1489 |
+
"reward": 0.860677108168602,
|
| 1490 |
+
"reward_std": 0.2829681634902954,
|
| 1491 |
+
"rewards/accuracy_reward": 0.4062500074505806,
|
| 1492 |
+
"rewards/format_reward": 0.0,
|
| 1493 |
+
"rewards/tag_count_reward": 0.4544270932674408,
|
| 1494 |
+
"step": 99
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"clip_ratio": 0.0,
|
| 1498 |
+
"completion_length": 791.9791717529297,
|
| 1499 |
+
"epoch": 0.8528784648187633,
|
| 1500 |
+
"grad_norm": 0.12058635085867997,
|
| 1501 |
+
"kl": 0.0102996826171875,
|
| 1502 |
+
"learning_rate": 1.5696630972229166e-07,
|
| 1503 |
+
"loss": 0.0342,
|
| 1504 |
+
"reward": 0.927083358168602,
|
| 1505 |
+
"reward_std": 0.3461384028196335,
|
| 1506 |
+
"rewards/accuracy_reward": 0.4270833432674408,
|
| 1507 |
+
"rewards/format_reward": 0.0,
|
| 1508 |
+
"rewards/tag_count_reward": 0.5000000223517418,
|
| 1509 |
+
"step": 100
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"clip_ratio": 0.0,
|
| 1513 |
+
"completion_length": 787.3594055175781,
|
| 1514 |
+
"epoch": 0.8614072494669509,
|
| 1515 |
+
"grad_norm": 0.13223369624098508,
|
| 1516 |
+
"kl": 0.0110626220703125,
|
| 1517 |
+
"learning_rate": 1.5058639494795067e-07,
|
| 1518 |
+
"loss": 0.0514,
|
| 1519 |
+
"reward": 0.9804687947034836,
|
| 1520 |
+
"reward_std": 0.33672135323286057,
|
| 1521 |
+
"rewards/accuracy_reward": 0.4375000149011612,
|
| 1522 |
+
"rewards/format_reward": 0.0,
|
| 1523 |
+
"rewards/tag_count_reward": 0.5429687723517418,
|
| 1524 |
+
"step": 101
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"clip_ratio": 0.0,
|
| 1528 |
+
"completion_length": 781.0364837646484,
|
| 1529 |
+
"epoch": 0.8699360341151386,
|
| 1530 |
+
"grad_norm": 0.12924436581025062,
|
| 1531 |
+
"kl": 0.0109710693359375,
|
| 1532 |
+
"learning_rate": 1.4456400944391144e-07,
|
| 1533 |
+
"loss": 0.0326,
|
| 1534 |
+
"reward": 0.971354216337204,
|
| 1535 |
+
"reward_std": 0.35067974030971527,
|
| 1536 |
+
"rewards/accuracy_reward": 0.4791666939854622,
|
| 1537 |
+
"rewards/format_reward": 0.0,
|
| 1538 |
+
"rewards/tag_count_reward": 0.4921875223517418,
|
| 1539 |
+
"step": 102
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"clip_ratio": 0.0,
|
| 1543 |
+
"completion_length": 815.4792022705078,
|
| 1544 |
+
"epoch": 0.8784648187633263,
|
| 1545 |
+
"grad_norm": 0.12735565071518637,
|
| 1546 |
+
"kl": 0.0097808837890625,
|
| 1547 |
+
"learning_rate": 1.3890454406082956e-07,
|
| 1548 |
+
"loss": 0.0469,
|
| 1549 |
+
"reward": 0.9609375298023224,
|
| 1550 |
+
"reward_std": 0.31679093837738037,
|
| 1551 |
+
"rewards/accuracy_reward": 0.4583333432674408,
|
| 1552 |
+
"rewards/format_reward": 0.0,
|
| 1553 |
+
"rewards/tag_count_reward": 0.5026041865348816,
|
| 1554 |
+
"step": 103
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"clip_ratio": 0.0,
|
| 1558 |
+
"completion_length": 820.6198120117188,
|
| 1559 |
+
"epoch": 0.8869936034115139,
|
| 1560 |
+
"grad_norm": 0.11515316700594094,
|
| 1561 |
+
"kl": 0.0109100341796875,
|
| 1562 |
+
"learning_rate": 1.3361306478670148e-07,
|
| 1563 |
+
"loss": 0.0367,
|
| 1564 |
+
"reward": 0.9244792014360428,
|
| 1565 |
+
"reward_std": 0.25151925161480904,
|
| 1566 |
+
"rewards/accuracy_reward": 0.4270833432674408,
|
| 1567 |
+
"rewards/format_reward": 0.0,
|
| 1568 |
+
"rewards/tag_count_reward": 0.497395858168602,
|
| 1569 |
+
"step": 104
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"clip_ratio": 0.0,
|
| 1573 |
+
"completion_length": 759.3281555175781,
|
| 1574 |
+
"epoch": 0.8955223880597015,
|
| 1575 |
+
"grad_norm": 0.14319190866589507,
|
| 1576 |
+
"kl": 0.0105133056640625,
|
| 1577 |
+
"learning_rate": 1.2869430821211826e-07,
|
| 1578 |
+
"loss": 0.0298,
|
| 1579 |
+
"reward": 0.9596354514360428,
|
| 1580 |
+
"reward_std": 0.2728368118405342,
|
| 1581 |
+
"rewards/accuracy_reward": 0.4583333507180214,
|
| 1582 |
+
"rewards/format_reward": 0.0,
|
| 1583 |
+
"rewards/tag_count_reward": 0.501302108168602,
|
| 1584 |
+
"step": 105
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"clip_ratio": 0.0,
|
| 1588 |
+
"completion_length": 818.6614837646484,
|
| 1589 |
+
"epoch": 0.9040511727078892,
|
| 1590 |
+
"grad_norm": 0.10985275674671954,
|
| 1591 |
+
"kl": 0.0103302001953125,
|
| 1592 |
+
"learning_rate": 1.2415267729037608e-07,
|
| 1593 |
+
"loss": 0.0311,
|
| 1594 |
+
"reward": 0.8710937649011612,
|
| 1595 |
+
"reward_std": 0.2647658996284008,
|
| 1596 |
+
"rewards/accuracy_reward": 0.4010416865348816,
|
| 1597 |
+
"rewards/format_reward": 0.0,
|
| 1598 |
+
"rewards/tag_count_reward": 0.4700521007180214,
|
| 1599 |
+
"step": 106
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"clip_ratio": 0.0,
|
| 1603 |
+
"completion_length": 775.0625152587891,
|
| 1604 |
+
"epoch": 0.9125799573560768,
|
| 1605 |
+
"grad_norm": 0.1272726543004608,
|
| 1606 |
+
"kl": 0.0119171142578125,
|
| 1607 |
+
"learning_rate": 1.1999223739623666e-07,
|
| 1608 |
+
"loss": 0.0317,
|
| 1609 |
+
"reward": 0.9244791865348816,
|
| 1610 |
+
"reward_std": 0.2937668412923813,
|
| 1611 |
+
"rewards/accuracy_reward": 0.432291679084301,
|
| 1612 |
+
"rewards/format_reward": 0.0,
|
| 1613 |
+
"rewards/tag_count_reward": 0.4921875149011612,
|
| 1614 |
+
"step": 107
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"clip_ratio": 0.0,
|
| 1618 |
+
"completion_length": 792.1979370117188,
|
| 1619 |
+
"epoch": 0.9211087420042644,
|
| 1620 |
+
"grad_norm": 0.1363433345784819,
|
| 1621 |
+
"kl": 0.0108795166015625,
|
| 1622 |
+
"learning_rate": 1.1621671268686605e-07,
|
| 1623 |
+
"loss": 0.0482,
|
| 1624 |
+
"reward": 0.950520858168602,
|
| 1625 |
+
"reward_std": 0.2951006144285202,
|
| 1626 |
+
"rewards/accuracy_reward": 0.4427083432674408,
|
| 1627 |
+
"rewards/format_reward": 0.0,
|
| 1628 |
+
"rewards/tag_count_reward": 0.5078125149011612,
|
| 1629 |
+
"step": 108
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"clip_ratio": 0.0,
|
| 1633 |
+
"completion_length": 840.5989837646484,
|
| 1634 |
+
"epoch": 0.929637526652452,
|
| 1635 |
+
"grad_norm": 0.11723986953897444,
|
| 1636 |
+
"kl": 0.0102691650390625,
|
| 1637 |
+
"learning_rate": 1.1282948276820962e-07,
|
| 1638 |
+
"loss": 0.0323,
|
| 1639 |
+
"reward": 0.903645858168602,
|
| 1640 |
+
"reward_std": 0.31709685176610947,
|
| 1641 |
+
"rewards/accuracy_reward": 0.4114583432674408,
|
| 1642 |
+
"rewards/format_reward": 0.0,
|
| 1643 |
+
"rewards/tag_count_reward": 0.4921875223517418,
|
| 1644 |
+
"step": 109
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"clip_ratio": 0.0,
|
| 1648 |
+
"completion_length": 785.3125,
|
| 1649 |
+
"epoch": 0.9381663113006397,
|
| 1650 |
+
"grad_norm": 0.1303189109965419,
|
| 1651 |
+
"kl": 0.0109100341796875,
|
| 1652 |
+
"learning_rate": 1.0983357966978745e-07,
|
| 1653 |
+
"loss": 0.0299,
|
| 1654 |
+
"reward": 0.973958358168602,
|
| 1655 |
+
"reward_std": 0.3055407963693142,
|
| 1656 |
+
"rewards/accuracy_reward": 0.4479166865348816,
|
| 1657 |
+
"rewards/format_reward": 0.0,
|
| 1658 |
+
"rewards/tag_count_reward": 0.526041679084301,
|
| 1659 |
+
"step": 110
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"clip_ratio": 0.0,
|
| 1663 |
+
"completion_length": 812.7916870117188,
|
| 1664 |
+
"epoch": 0.9466950959488273,
|
| 1665 |
+
"grad_norm": 0.13474852991113115,
|
| 1666 |
+
"kl": 0.011871337890625,
|
| 1667 |
+
"learning_rate": 1.0723168513061665e-07,
|
| 1668 |
+
"loss": 0.0379,
|
| 1669 |
+
"reward": 0.864583358168602,
|
| 1670 |
+
"reward_std": 0.2699150890111923,
|
| 1671 |
+
"rewards/accuracy_reward": 0.3906250149011612,
|
| 1672 |
+
"rewards/format_reward": 0.0,
|
| 1673 |
+
"rewards/tag_count_reward": 0.4739583432674408,
|
| 1674 |
+
"step": 111
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"clip_ratio": 0.0,
|
| 1678 |
+
"completion_length": 828.9948120117188,
|
| 1679 |
+
"epoch": 0.9552238805970149,
|
| 1680 |
+
"grad_norm": 0.15444917838378422,
|
| 1681 |
+
"kl": 0.0113677978515625,
|
| 1682 |
+
"learning_rate": 1.0502612819869216e-07,
|
| 1683 |
+
"loss": 0.056,
|
| 1684 |
+
"reward": 0.9531250149011612,
|
| 1685 |
+
"reward_std": 0.40493085235357285,
|
| 1686 |
+
"rewards/accuracy_reward": 0.4375000223517418,
|
| 1687 |
+
"rewards/format_reward": 0.0,
|
| 1688 |
+
"rewards/tag_count_reward": 0.5156250149011612,
|
| 1689 |
+
"step": 112
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"clip_ratio": 0.0,
|
| 1693 |
+
"completion_length": 758.9791870117188,
|
| 1694 |
+
"epoch": 0.9637526652452025,
|
| 1695 |
+
"grad_norm": 0.11394166249687963,
|
| 1696 |
+
"kl": 0.0107421875,
|
| 1697 |
+
"learning_rate": 1.032188831461732e-07,
|
| 1698 |
+
"loss": 0.0172,
|
| 1699 |
+
"reward": 1.0195312649011612,
|
| 1700 |
+
"reward_std": 0.23920952156186104,
|
| 1701 |
+
"rewards/accuracy_reward": 0.5260416865348816,
|
| 1702 |
+
"rewards/format_reward": 0.0,
|
| 1703 |
+
"rewards/tag_count_reward": 0.4934895932674408,
|
| 1704 |
+
"step": 113
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"clip_ratio": 0.0,
|
| 1708 |
+
"completion_length": 835.1614685058594,
|
| 1709 |
+
"epoch": 0.9722814498933902,
|
| 1710 |
+
"grad_norm": 0.14157333655394927,
|
| 1711 |
+
"kl": 0.0106353759765625,
|
| 1712 |
+
"learning_rate": 1.0181156770214242e-07,
|
| 1713 |
+
"loss": 0.0532,
|
| 1714 |
+
"reward": 0.8906250149011612,
|
| 1715 |
+
"reward_std": 0.344898097217083,
|
| 1716 |
+
"rewards/accuracy_reward": 0.4010416753590107,
|
| 1717 |
+
"rewards/format_reward": 0.0,
|
| 1718 |
+
"rewards/tag_count_reward": 0.4895833507180214,
|
| 1719 |
+
"step": 114
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"clip_ratio": 0.0,
|
| 1723 |
+
"completion_length": 784.8802337646484,
|
| 1724 |
+
"epoch": 0.9808102345415778,
|
| 1725 |
+
"grad_norm": 0.12045891162749811,
|
| 1726 |
+
"kl": 0.0106201171875,
|
| 1727 |
+
"learning_rate": 1.0080544160451918e-07,
|
| 1728 |
+
"loss": 0.0387,
|
| 1729 |
+
"reward": 0.9531250149011612,
|
| 1730 |
+
"reward_std": 0.3491218537092209,
|
| 1731 |
+
"rewards/accuracy_reward": 0.4479166865348816,
|
| 1732 |
+
"rewards/format_reward": 0.0,
|
| 1733 |
+
"rewards/tag_count_reward": 0.5052083507180214,
|
| 1734 |
+
"step": 115
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"clip_ratio": 0.0,
|
| 1738 |
+
"completion_length": 826.1719055175781,
|
| 1739 |
+
"epoch": 0.9893390191897654,
|
| 1740 |
+
"grad_norm": 0.1211356758875429,
|
| 1741 |
+
"kl": 0.01165771484375,
|
| 1742 |
+
"learning_rate": 1.002014054724235e-07,
|
| 1743 |
+
"loss": 0.0394,
|
| 1744 |
+
"reward": 0.9440104514360428,
|
| 1745 |
+
"reward_std": 0.30041512846946716,
|
| 1746 |
+
"rewards/accuracy_reward": 0.4583333432674408,
|
| 1747 |
+
"rewards/format_reward": 0.0,
|
| 1748 |
+
"rewards/tag_count_reward": 0.4856770932674408,
|
| 1749 |
+
"step": 116
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"clip_ratio": 0.0,
|
| 1753 |
+
"completion_length": 895.8125152587891,
|
| 1754 |
+
"epoch": 0.997867803837953,
|
| 1755 |
+
"grad_norm": 0.14489529140501242,
|
| 1756 |
+
"kl": 0.009857177734375,
|
| 1757 |
+
"learning_rate": 1e-07,
|
| 1758 |
+
"loss": 0.054,
|
| 1759 |
+
"reward": 0.9101562649011612,
|
| 1760 |
+
"reward_std": 0.3278352990746498,
|
| 1761 |
+
"rewards/accuracy_reward": 0.416666679084301,
|
| 1762 |
+
"rewards/format_reward": 0.0,
|
| 1763 |
+
"rewards/tag_count_reward": 0.4934896007180214,
|
| 1764 |
+
"step": 117
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 0.997867803837953,
|
| 1768 |
+
"step": 117,
|
| 1769 |
+
"total_flos": 0.0,
|
| 1770 |
+
"train_loss": 0.03437558454899197,
|
| 1771 |
+
"train_runtime": 4123.9479,
|
| 1772 |
+
"train_samples_per_second": 1.819,
|
| 1773 |
+
"train_steps_per_second": 0.028
|
| 1774 |
+
}
|
| 1775 |
+
],
|
| 1776 |
+
"logging_steps": 1,
|
| 1777 |
+
"max_steps": 117,
|
| 1778 |
+
"num_input_tokens_seen": 0,
|
| 1779 |
+
"num_train_epochs": 1,
|
| 1780 |
+
"save_steps": 10,
|
| 1781 |
+
"stateful_callbacks": {
|
| 1782 |
+
"TrainerControl": {
|
| 1783 |
+
"args": {
|
| 1784 |
+
"should_epoch_stop": false,
|
| 1785 |
+
"should_evaluate": false,
|
| 1786 |
+
"should_log": false,
|
| 1787 |
+
"should_save": true,
|
| 1788 |
+
"should_training_stop": true
|
| 1789 |
+
},
|
| 1790 |
+
"attributes": {}
|
| 1791 |
+
}
|
| 1792 |
+
},
|
| 1793 |
+
"total_flos": 0.0,
|
| 1794 |
+
"train_batch_size": 16,
|
| 1795 |
+
"trial_name": null,
|
| 1796 |
+
"trial_params": null
|
| 1797 |
+
}
|