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 +2743 -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: Gaussian_0.2_0.28
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- grpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Gaussian_0.2_0.28
|
| 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="LLucass/Gaussian_0.2_0.28", 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/lavatorywang-nus/Gaussian/runs/mnokendi)
|
| 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: 0.18.0
|
| 38 |
+
- Transformers: 4.50.0
|
| 39 |
+
- Pytorch: 2.5.1
|
| 40 |
+
- Datasets: 3.6.0
|
| 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": -0.004950485862791538,
|
| 4 |
+
"train_runtime": 8385.5538,
|
| 5 |
+
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 1.145,
|
| 7 |
+
"train_steps_per_second": 0.012
|
| 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": "4.50.0"
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": -0.004950485862791538,
|
| 4 |
+
"train_runtime": 8385.5538,
|
| 5 |
+
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 1.145,
|
| 7 |
+
"train_steps_per_second": 0.012
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2743 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.11428571428571428,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 100,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"clip_ratio/high_max": 0.0,
|
| 14 |
+
"clip_ratio/high_mean": 0.0,
|
| 15 |
+
"clip_ratio/low_mean": 0.0,
|
| 16 |
+
"clip_ratio/low_min": 0.0,
|
| 17 |
+
"clip_ratio/region_mean": 0.0,
|
| 18 |
+
"completion_length": 2700.4271850585938,
|
| 19 |
+
"cov_mean": -6.0587970438064076e-05,
|
| 20 |
+
"cov_std": 0.35307812318205833,
|
| 21 |
+
"entropy": 0.36767578125,
|
| 22 |
+
"epoch": 0.001142857142857143,
|
| 23 |
+
"grad_norm": 0.4553219676017761,
|
| 24 |
+
"kl": 0.0,
|
| 25 |
+
"learning_rate": 1e-07,
|
| 26 |
+
"loss": -0.0382,
|
| 27 |
+
"reward": 0.7604166893288493,
|
| 28 |
+
"reward_std": 0.4268697127699852,
|
| 29 |
+
"rewards/accuracy_reward": 0.25000001303851604,
|
| 30 |
+
"rewards/format_reward": 0.5104166669771075,
|
| 31 |
+
"step": 1,
|
| 32 |
+
"w_high_ratio": 0.2200421690940857,
|
| 33 |
+
"w_low_ratio": 0.03663695091381669,
|
| 34 |
+
"w_max": 2.1593789756298065,
|
| 35 |
+
"w_mean": 1.4711343348026276,
|
| 36 |
+
"w_min": 6.525355682266089e-35,
|
| 37 |
+
"w_std": 0.2659660503268242
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"clip_ratio/high_max": 0.0,
|
| 41 |
+
"clip_ratio/high_mean": 0.0,
|
| 42 |
+
"clip_ratio/low_mean": 0.0,
|
| 43 |
+
"clip_ratio/low_min": 0.0,
|
| 44 |
+
"clip_ratio/region_mean": 0.0,
|
| 45 |
+
"completion_length": 3127.3958435058594,
|
| 46 |
+
"cov_mean": -2.155053698515985e-05,
|
| 47 |
+
"cov_std": 0.310540571808815,
|
| 48 |
+
"entropy": 0.353515625,
|
| 49 |
+
"epoch": 0.002285714285714286,
|
| 50 |
+
"grad_norm": 0.5147930979728699,
|
| 51 |
+
"kl": 0.0,
|
| 52 |
+
"learning_rate": 2e-07,
|
| 53 |
+
"loss": 0.0049,
|
| 54 |
+
"reward": 0.6458333637565374,
|
| 55 |
+
"reward_std": 0.4249730706214905,
|
| 56 |
+
"rewards/accuracy_reward": 0.2812500102445483,
|
| 57 |
+
"rewards/format_reward": 0.3645833386108279,
|
| 58 |
+
"step": 2,
|
| 59 |
+
"w_high_ratio": 0.05183619633316994,
|
| 60 |
+
"w_low_ratio": 0.036958135198801756,
|
| 61 |
+
"w_max": 1.8325217366218567,
|
| 62 |
+
"w_mean": 1.2113382518291473,
|
| 63 |
+
"w_min": 0.0,
|
| 64 |
+
"w_std": 0.20957503467798233
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"clip_ratio/high_max": 0.0,
|
| 68 |
+
"clip_ratio/high_mean": 0.0,
|
| 69 |
+
"clip_ratio/low_mean": 0.0,
|
| 70 |
+
"clip_ratio/low_min": 0.0,
|
| 71 |
+
"clip_ratio/region_mean": 0.0,
|
| 72 |
+
"completion_length": 3685.479248046875,
|
| 73 |
+
"cov_mean": -4.653575570046087e-05,
|
| 74 |
+
"cov_std": 0.29687076807022095,
|
| 75 |
+
"entropy": 0.45556640625,
|
| 76 |
+
"epoch": 0.0034285714285714284,
|
| 77 |
+
"grad_norm": 0.23412799835205078,
|
| 78 |
+
"kl": 4.607439041137695e-05,
|
| 79 |
+
"learning_rate": 3e-07,
|
| 80 |
+
"loss": 0.0316,
|
| 81 |
+
"reward": 0.2291666753590107,
|
| 82 |
+
"reward_std": 0.35217108577489853,
|
| 83 |
+
"rewards/accuracy_reward": 0.05208333395421505,
|
| 84 |
+
"rewards/format_reward": 0.1770833395421505,
|
| 85 |
+
"step": 3,
|
| 86 |
+
"w_high_ratio": 0.0,
|
| 87 |
+
"w_low_ratio": 0.03702388517558575,
|
| 88 |
+
"w_max": 1.4749425947666168,
|
| 89 |
+
"w_mean": 1.0885904878377914,
|
| 90 |
+
"w_min": 3.5302726539688593e-39,
|
| 91 |
+
"w_std": 0.1792612113058567
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"clip_ratio/high_max": 0.0,
|
| 95 |
+
"clip_ratio/high_mean": 0.0,
|
| 96 |
+
"clip_ratio/low_mean": 0.0,
|
| 97 |
+
"clip_ratio/low_min": 0.0,
|
| 98 |
+
"clip_ratio/region_mean": 0.0,
|
| 99 |
+
"completion_length": 2459.854248046875,
|
| 100 |
+
"cov_mean": 2.6621969936968526e-05,
|
| 101 |
+
"cov_std": 0.42025458812713623,
|
| 102 |
+
"entropy": 0.4140625,
|
| 103 |
+
"epoch": 0.004571428571428572,
|
| 104 |
+
"grad_norm": 0.42652714252471924,
|
| 105 |
+
"kl": 3.8504600524902344e-05,
|
| 106 |
+
"learning_rate": 4e-07,
|
| 107 |
+
"loss": 0.0096,
|
| 108 |
+
"reward": 0.7604167014360428,
|
| 109 |
+
"reward_std": 0.4487093985080719,
|
| 110 |
+
"rewards/accuracy_reward": 0.1250000037252903,
|
| 111 |
+
"rewards/format_reward": 0.6354166865348816,
|
| 112 |
+
"step": 4,
|
| 113 |
+
"w_high_ratio": 0.2003328688442707,
|
| 114 |
+
"w_low_ratio": 0.04321334883570671,
|
| 115 |
+
"w_max": 2.0572283267974854,
|
| 116 |
+
"w_mean": 1.4974508583545685,
|
| 117 |
+
"w_min": 0.0,
|
| 118 |
+
"w_std": 0.26409733667969704
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"clip_ratio/high_max": 0.0,
|
| 122 |
+
"clip_ratio/high_mean": 0.0,
|
| 123 |
+
"clip_ratio/low_mean": 0.0,
|
| 124 |
+
"clip_ratio/low_min": 0.0,
|
| 125 |
+
"clip_ratio/region_mean": 0.0,
|
| 126 |
+
"completion_length": 3507.5833740234375,
|
| 127 |
+
"cov_mean": 7.01061562722316e-05,
|
| 128 |
+
"cov_std": 0.5858492627739906,
|
| 129 |
+
"entropy": 0.45458984375,
|
| 130 |
+
"epoch": 0.005714285714285714,
|
| 131 |
+
"grad_norm": 0.5644172430038452,
|
| 132 |
+
"kl": 4.00543212890625e-05,
|
| 133 |
+
"learning_rate": 5e-07,
|
| 134 |
+
"loss": 0.0229,
|
| 135 |
+
"reward": 0.4583333507180214,
|
| 136 |
+
"reward_std": 0.6347778141498566,
|
| 137 |
+
"rewards/accuracy_reward": 0.09375000093132257,
|
| 138 |
+
"rewards/format_reward": 0.3645833507180214,
|
| 139 |
+
"step": 5,
|
| 140 |
+
"w_high_ratio": 0.04538528248667717,
|
| 141 |
+
"w_low_ratio": 0.06710020452737808,
|
| 142 |
+
"w_max": 1.6610480844974518,
|
| 143 |
+
"w_mean": 1.1760589182376862,
|
| 144 |
+
"w_min": 5.419854268836891e-39,
|
| 145 |
+
"w_std": 0.322455570101738
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"clip_ratio/high_max": 0.0,
|
| 149 |
+
"clip_ratio/high_mean": 0.0,
|
| 150 |
+
"clip_ratio/low_mean": 0.0,
|
| 151 |
+
"clip_ratio/low_min": 0.0,
|
| 152 |
+
"clip_ratio/region_mean": 0.0,
|
| 153 |
+
"completion_length": 3396.0104370117188,
|
| 154 |
+
"cov_mean": 1.043153679347597e-05,
|
| 155 |
+
"cov_std": 0.3946683406829834,
|
| 156 |
+
"entropy": 0.4580078125,
|
| 157 |
+
"epoch": 0.006857142857142857,
|
| 158 |
+
"grad_norm": 0.47143253684043884,
|
| 159 |
+
"kl": 4.100799560546875e-05,
|
| 160 |
+
"learning_rate": 6e-07,
|
| 161 |
+
"loss": 0.0699,
|
| 162 |
+
"reward": 0.3125000176951289,
|
| 163 |
+
"reward_std": 0.33619000762701035,
|
| 164 |
+
"rewards/accuracy_reward": 0.031250000931322575,
|
| 165 |
+
"rewards/format_reward": 0.2812500102445483,
|
| 166 |
+
"step": 6,
|
| 167 |
+
"w_high_ratio": 0.044003941118717194,
|
| 168 |
+
"w_low_ratio": 0.04949501994997263,
|
| 169 |
+
"w_max": 1.6353994011878967,
|
| 170 |
+
"w_mean": 1.1387673616409302,
|
| 171 |
+
"w_min": 0.0,
|
| 172 |
+
"w_std": 0.24893567338585854
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"clip_ratio/high_max": 0.0,
|
| 176 |
+
"clip_ratio/high_mean": 0.0,
|
| 177 |
+
"clip_ratio/low_mean": 0.0,
|
| 178 |
+
"clip_ratio/low_min": 0.0,
|
| 179 |
+
"clip_ratio/region_mean": 0.0,
|
| 180 |
+
"completion_length": 3345.291748046875,
|
| 181 |
+
"cov_mean": -7.031985660432838e-06,
|
| 182 |
+
"cov_std": 0.4864897131919861,
|
| 183 |
+
"entropy": 0.38818359375,
|
| 184 |
+
"epoch": 0.008,
|
| 185 |
+
"grad_norm": 0.2716952860355377,
|
| 186 |
+
"kl": 2.446770668029785e-05,
|
| 187 |
+
"learning_rate": 7e-07,
|
| 188 |
+
"loss": -0.0398,
|
| 189 |
+
"reward": 0.7395833507180214,
|
| 190 |
+
"reward_std": 0.5676274970173836,
|
| 191 |
+
"rewards/accuracy_reward": 0.1979166679084301,
|
| 192 |
+
"rewards/format_reward": 0.5416666828095913,
|
| 193 |
+
"step": 7,
|
| 194 |
+
"w_high_ratio": 0.0,
|
| 195 |
+
"w_low_ratio": 0.051307737827301025,
|
| 196 |
+
"w_max": 1.4717772901058197,
|
| 197 |
+
"w_mean": 1.1606104969978333,
|
| 198 |
+
"w_min": 0.0,
|
| 199 |
+
"w_std": 0.2527086026966572
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"clip_ratio/high_max": 0.0,
|
| 203 |
+
"clip_ratio/high_mean": 0.0,
|
| 204 |
+
"clip_ratio/low_mean": 0.0,
|
| 205 |
+
"clip_ratio/low_min": 0.0,
|
| 206 |
+
"clip_ratio/region_mean": 0.0,
|
| 207 |
+
"completion_length": 2934.9583740234375,
|
| 208 |
+
"cov_mean": -6.692781494166411e-05,
|
| 209 |
+
"cov_std": 0.31714635342359543,
|
| 210 |
+
"entropy": 0.36328125,
|
| 211 |
+
"epoch": 0.009142857142857144,
|
| 212 |
+
"grad_norm": 0.24377422034740448,
|
| 213 |
+
"kl": 1.4781951904296875e-05,
|
| 214 |
+
"learning_rate": 8e-07,
|
| 215 |
+
"loss": -0.0011,
|
| 216 |
+
"reward": 0.8750000409781933,
|
| 217 |
+
"reward_std": 0.5431903004646301,
|
| 218 |
+
"rewards/accuracy_reward": 0.38541667722165585,
|
| 219 |
+
"rewards/format_reward": 0.48958336375653744,
|
| 220 |
+
"step": 8,
|
| 221 |
+
"w_high_ratio": 0.0,
|
| 222 |
+
"w_low_ratio": 0.03932169498875737,
|
| 223 |
+
"w_max": 1.5945640802383423,
|
| 224 |
+
"w_mean": 1.1880076229572296,
|
| 225 |
+
"w_min": 0.0,
|
| 226 |
+
"w_std": 0.20366584323346615
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"clip_ratio/high_max": 0.0,
|
| 230 |
+
"clip_ratio/high_mean": 0.0,
|
| 231 |
+
"clip_ratio/low_mean": 0.0,
|
| 232 |
+
"clip_ratio/low_min": 0.0,
|
| 233 |
+
"clip_ratio/region_mean": 0.0,
|
| 234 |
+
"completion_length": 3466.6250610351562,
|
| 235 |
+
"cov_mean": -5.504875753103988e-05,
|
| 236 |
+
"cov_std": 0.3686807006597519,
|
| 237 |
+
"entropy": 0.4501953125,
|
| 238 |
+
"epoch": 0.010285714285714285,
|
| 239 |
+
"grad_norm": 0.3517554700374603,
|
| 240 |
+
"kl": 4.1872262954711914e-05,
|
| 241 |
+
"learning_rate": 9e-07,
|
| 242 |
+
"loss": -0.0325,
|
| 243 |
+
"reward": 0.4375000074505806,
|
| 244 |
+
"reward_std": 0.5021138191223145,
|
| 245 |
+
"rewards/accuracy_reward": 0.13541666883975267,
|
| 246 |
+
"rewards/format_reward": 0.3020833432674408,
|
| 247 |
+
"step": 9,
|
| 248 |
+
"w_high_ratio": 0.021586479619145393,
|
| 249 |
+
"w_low_ratio": 0.04805116541683674,
|
| 250 |
+
"w_max": 1.7995203733444214,
|
| 251 |
+
"w_mean": 1.1711468696594238,
|
| 252 |
+
"w_min": 1.0441714550433176e-38,
|
| 253 |
+
"w_std": 0.24575592577457428
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"clip_ratio/high_max": 0.0,
|
| 257 |
+
"clip_ratio/high_mean": 0.0,
|
| 258 |
+
"clip_ratio/low_mean": 0.0,
|
| 259 |
+
"clip_ratio/low_min": 0.0,
|
| 260 |
+
"clip_ratio/region_mean": 0.0,
|
| 261 |
+
"completion_length": 2808.8438720703125,
|
| 262 |
+
"cov_mean": -2.2776248442824e-05,
|
| 263 |
+
"cov_std": 0.3646387383341789,
|
| 264 |
+
"entropy": 0.326171875,
|
| 265 |
+
"epoch": 0.011428571428571429,
|
| 266 |
+
"grad_norm": 0.44795799255371094,
|
| 267 |
+
"kl": 2.4825334548950195e-05,
|
| 268 |
+
"learning_rate": 1e-06,
|
| 269 |
+
"loss": -0.0625,
|
| 270 |
+
"reward": 0.6875000223517418,
|
| 271 |
+
"reward_std": 0.48652123659849167,
|
| 272 |
+
"rewards/accuracy_reward": 0.2083333395421505,
|
| 273 |
+
"rewards/format_reward": 0.4791666828095913,
|
| 274 |
+
"step": 10,
|
| 275 |
+
"w_high_ratio": 0.1090739294886589,
|
| 276 |
+
"w_low_ratio": 0.04266490042209625,
|
| 277 |
+
"w_max": 2.1747676730155945,
|
| 278 |
+
"w_mean": 1.3257513344287872,
|
| 279 |
+
"w_min": 0.0,
|
| 280 |
+
"w_std": 0.2564368508756161
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"clip_ratio/high_max": 0.0,
|
| 284 |
+
"clip_ratio/high_mean": 0.0,
|
| 285 |
+
"clip_ratio/low_mean": 0.0,
|
| 286 |
+
"clip_ratio/low_min": 0.0,
|
| 287 |
+
"clip_ratio/region_mean": 0.0,
|
| 288 |
+
"completion_length": 3580.1355590820312,
|
| 289 |
+
"cov_mean": -4.214262844470795e-06,
|
| 290 |
+
"cov_std": 0.35131024569272995,
|
| 291 |
+
"entropy": 0.3720703125,
|
| 292 |
+
"epoch": 0.012571428571428572,
|
| 293 |
+
"grad_norm": 0.480006605386734,
|
| 294 |
+
"kl": 3.1560659408569336e-05,
|
| 295 |
+
"learning_rate": 9.997258721585931e-07,
|
| 296 |
+
"loss": 0.0246,
|
| 297 |
+
"reward": 0.3333333507180214,
|
| 298 |
+
"reward_std": 0.568866454064846,
|
| 299 |
+
"rewards/accuracy_reward": 0.12500000279396772,
|
| 300 |
+
"rewards/format_reward": 0.20833334140479565,
|
| 301 |
+
"step": 11,
|
| 302 |
+
"w_high_ratio": 0.04230208322405815,
|
| 303 |
+
"w_low_ratio": 0.05235781241208315,
|
| 304 |
+
"w_max": 1.6345463395118713,
|
| 305 |
+
"w_mean": 1.1268693506717682,
|
| 306 |
+
"w_min": 0.0,
|
| 307 |
+
"w_std": 0.24075055867433548
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"clip_ratio/high_max": 0.0,
|
| 311 |
+
"clip_ratio/high_mean": 0.0,
|
| 312 |
+
"clip_ratio/low_mean": 0.0,
|
| 313 |
+
"clip_ratio/low_min": 0.0,
|
| 314 |
+
"clip_ratio/region_mean": 0.0,
|
| 315 |
+
"completion_length": 2692.9688110351562,
|
| 316 |
+
"cov_mean": 3.166104261254077e-05,
|
| 317 |
+
"cov_std": 0.33987831324338913,
|
| 318 |
+
"entropy": 0.40576171875,
|
| 319 |
+
"epoch": 0.013714285714285714,
|
| 320 |
+
"grad_norm": 0.5367066860198975,
|
| 321 |
+
"kl": 4.026293754577637e-05,
|
| 322 |
+
"learning_rate": 9.989038226169207e-07,
|
| 323 |
+
"loss": 0.0037,
|
| 324 |
+
"reward": 0.7500000149011612,
|
| 325 |
+
"reward_std": 0.4257746562361717,
|
| 326 |
+
"rewards/accuracy_reward": 0.16666667442768812,
|
| 327 |
+
"rewards/format_reward": 0.5833333469927311,
|
| 328 |
+
"step": 12,
|
| 329 |
+
"w_high_ratio": 0.20933324098587036,
|
| 330 |
+
"w_low_ratio": 0.042826405726373196,
|
| 331 |
+
"w_max": 2.3706648349761963,
|
| 332 |
+
"w_mean": 1.5066065192222595,
|
| 333 |
+
"w_min": 0.0,
|
| 334 |
+
"w_std": 0.3023643344640732
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"clip_ratio/high_max": 0.0,
|
| 338 |
+
"clip_ratio/high_mean": 0.0,
|
| 339 |
+
"clip_ratio/low_mean": 0.0,
|
| 340 |
+
"clip_ratio/low_min": 0.0,
|
| 341 |
+
"clip_ratio/region_mean": 0.0,
|
| 342 |
+
"completion_length": 3225.3125610351562,
|
| 343 |
+
"cov_mean": -1.1440834896347951e-05,
|
| 344 |
+
"cov_std": 0.25972653180360794,
|
| 345 |
+
"entropy": 0.38330078125,
|
| 346 |
+
"epoch": 0.014857142857142857,
|
| 347 |
+
"grad_norm": 0.5191377997398376,
|
| 348 |
+
"kl": 3.4749507904052734e-05,
|
| 349 |
+
"learning_rate": 9.975348529157229e-07,
|
| 350 |
+
"loss": -0.0342,
|
| 351 |
+
"reward": 0.6354166865348816,
|
| 352 |
+
"reward_std": 0.3692361190915108,
|
| 353 |
+
"rewards/accuracy_reward": 0.22916667722165585,
|
| 354 |
+
"rewards/format_reward": 0.4062500074505806,
|
| 355 |
+
"step": 13,
|
| 356 |
+
"w_high_ratio": 0.059321983717381954,
|
| 357 |
+
"w_low_ratio": 0.02801903593353927,
|
| 358 |
+
"w_max": 1.86709925532341,
|
| 359 |
+
"w_mean": 1.2194385528564453,
|
| 360 |
+
"w_min": 0.0,
|
| 361 |
+
"w_std": 0.18418971449136734
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"clip_ratio/high_max": 0.0,
|
| 365 |
+
"clip_ratio/high_mean": 0.0,
|
| 366 |
+
"clip_ratio/low_mean": 0.0,
|
| 367 |
+
"clip_ratio/low_min": 0.0,
|
| 368 |
+
"clip_ratio/region_mean": 0.0,
|
| 369 |
+
"completion_length": 3069.2083740234375,
|
| 370 |
+
"cov_mean": -5.5041197356331395e-05,
|
| 371 |
+
"cov_std": 0.4479314386844635,
|
| 372 |
+
"entropy": 0.3857421875,
|
| 373 |
+
"epoch": 0.016,
|
| 374 |
+
"grad_norm": 0.4110907018184662,
|
| 375 |
+
"kl": 3.4496188163757324e-05,
|
| 376 |
+
"learning_rate": 9.956206309337066e-07,
|
| 377 |
+
"loss": 0.0262,
|
| 378 |
+
"reward": 0.6354166716337204,
|
| 379 |
+
"reward_std": 0.5461946055293083,
|
| 380 |
+
"rewards/accuracy_reward": 0.20833333861082792,
|
| 381 |
+
"rewards/format_reward": 0.4270833432674408,
|
| 382 |
+
"step": 14,
|
| 383 |
+
"w_high_ratio": 0.05299236997961998,
|
| 384 |
+
"w_low_ratio": 0.05393570475280285,
|
| 385 |
+
"w_max": 1.8086846470832825,
|
| 386 |
+
"w_mean": 1.2297336757183075,
|
| 387 |
+
"w_min": 0.0,
|
| 388 |
+
"w_std": 0.29025064408779144
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"clip_ratio/high_max": 0.0,
|
| 392 |
+
"clip_ratio/high_mean": 0.0,
|
| 393 |
+
"clip_ratio/low_mean": 0.0,
|
| 394 |
+
"clip_ratio/low_min": 0.0,
|
| 395 |
+
"clip_ratio/region_mean": 0.0,
|
| 396 |
+
"completion_length": 2969.4063110351562,
|
| 397 |
+
"cov_mean": 4.094462474313332e-05,
|
| 398 |
+
"cov_std": 0.28705432265996933,
|
| 399 |
+
"entropy": 0.372802734375,
|
| 400 |
+
"epoch": 0.017142857142857144,
|
| 401 |
+
"grad_norm": 0.2711990475654602,
|
| 402 |
+
"kl": 2.568960189819336e-05,
|
| 403 |
+
"learning_rate": 9.931634888554935e-07,
|
| 404 |
+
"loss": 0.0207,
|
| 405 |
+
"reward": 0.6666666865348816,
|
| 406 |
+
"reward_std": 0.326509490609169,
|
| 407 |
+
"rewards/accuracy_reward": 0.2604166669771075,
|
| 408 |
+
"rewards/format_reward": 0.4062500111758709,
|
| 409 |
+
"step": 15,
|
| 410 |
+
"w_high_ratio": 0.0,
|
| 411 |
+
"w_low_ratio": 0.03214749973267317,
|
| 412 |
+
"w_max": 1.415438860654831,
|
| 413 |
+
"w_mean": 1.1301319599151611,
|
| 414 |
+
"w_min": 0.25,
|
| 415 |
+
"w_std": 0.17486043274402618
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"clip_ratio/high_max": 0.0,
|
| 419 |
+
"clip_ratio/high_mean": 0.0,
|
| 420 |
+
"clip_ratio/low_mean": 0.0,
|
| 421 |
+
"clip_ratio/low_min": 0.0,
|
| 422 |
+
"clip_ratio/region_mean": 0.0,
|
| 423 |
+
"completion_length": 3889.2083740234375,
|
| 424 |
+
"cov_mean": 1.429535222996492e-05,
|
| 425 |
+
"cov_std": 0.1915050856769085,
|
| 426 |
+
"entropy": 0.46337890625,
|
| 427 |
+
"epoch": 0.018285714285714287,
|
| 428 |
+
"grad_norm": 0.2660626173019409,
|
| 429 |
+
"kl": 3.978610038757324e-05,
|
| 430 |
+
"learning_rate": 9.901664203302124e-07,
|
| 431 |
+
"loss": 0.0406,
|
| 432 |
+
"reward": 0.1250000037252903,
|
| 433 |
+
"reward_std": 0.24391646683216095,
|
| 434 |
+
"rewards/accuracy_reward": 0.052083334885537624,
|
| 435 |
+
"rewards/format_reward": 0.07291666883975267,
|
| 436 |
+
"step": 16,
|
| 437 |
+
"w_high_ratio": 0.0,
|
| 438 |
+
"w_low_ratio": 0.027141269762068987,
|
| 439 |
+
"w_max": 1.2118620574474335,
|
| 440 |
+
"w_mean": 1.0125571489334106,
|
| 441 |
+
"w_min": 0.25,
|
| 442 |
+
"w_std": 0.10200810059905052
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"clip_ratio/high_max": 0.0,
|
| 446 |
+
"clip_ratio/high_mean": 0.0,
|
| 447 |
+
"clip_ratio/low_mean": 0.0,
|
| 448 |
+
"clip_ratio/low_min": 0.0,
|
| 449 |
+
"clip_ratio/region_mean": 0.0,
|
| 450 |
+
"completion_length": 2681.927215576172,
|
| 451 |
+
"cov_mean": 5.457106340145401e-05,
|
| 452 |
+
"cov_std": 0.4843906760215759,
|
| 453 |
+
"entropy": 0.46728515625,
|
| 454 |
+
"epoch": 0.019428571428571427,
|
| 455 |
+
"grad_norm": 0.5056636929512024,
|
| 456 |
+
"kl": 4.604458808898926e-05,
|
| 457 |
+
"learning_rate": 9.866330768241983e-07,
|
| 458 |
+
"loss": 0.0332,
|
| 459 |
+
"reward": 0.8020833656191826,
|
| 460 |
+
"reward_std": 0.5575663819909096,
|
| 461 |
+
"rewards/accuracy_reward": 0.2500000074505806,
|
| 462 |
+
"rewards/format_reward": 0.5520833432674408,
|
| 463 |
+
"step": 17,
|
| 464 |
+
"w_high_ratio": 0.2290888950228691,
|
| 465 |
+
"w_low_ratio": 0.04584049433469772,
|
| 466 |
+
"w_max": 2.3289138078689575,
|
| 467 |
+
"w_mean": 1.451726257801056,
|
| 468 |
+
"w_min": 3.2229864679470793e-44,
|
| 469 |
+
"w_std": 0.33130457252264023
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"clip_ratio/high_max": 0.0,
|
| 473 |
+
"clip_ratio/high_mean": 0.0,
|
| 474 |
+
"clip_ratio/low_mean": 0.0,
|
| 475 |
+
"clip_ratio/low_min": 0.0,
|
| 476 |
+
"clip_ratio/region_mean": 0.0,
|
| 477 |
+
"completion_length": 3172.822998046875,
|
| 478 |
+
"cov_mean": 5.8477902712184004e-05,
|
| 479 |
+
"cov_std": 0.3056485205888748,
|
| 480 |
+
"entropy": 0.36083984375,
|
| 481 |
+
"epoch": 0.02057142857142857,
|
| 482 |
+
"grad_norm": 0.5626700520515442,
|
| 483 |
+
"kl": 2.8699636459350586e-05,
|
| 484 |
+
"learning_rate": 9.825677631722435e-07,
|
| 485 |
+
"loss": -0.0633,
|
| 486 |
+
"reward": 0.5833333507180214,
|
| 487 |
+
"reward_std": 0.4684683009982109,
|
| 488 |
+
"rewards/accuracy_reward": 0.15625000279396772,
|
| 489 |
+
"rewards/format_reward": 0.4270833469927311,
|
| 490 |
+
"step": 18,
|
| 491 |
+
"w_high_ratio": 0.05792288854718208,
|
| 492 |
+
"w_low_ratio": 0.03927373047918081,
|
| 493 |
+
"w_max": 1.9344473481178284,
|
| 494 |
+
"w_mean": 1.2674984335899353,
|
| 495 |
+
"w_min": 0.0,
|
| 496 |
+
"w_std": 0.20779520832002163
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"clip_ratio/high_max": 0.0,
|
| 500 |
+
"clip_ratio/high_mean": 0.0,
|
| 501 |
+
"clip_ratio/low_mean": 0.0,
|
| 502 |
+
"clip_ratio/low_min": 0.0,
|
| 503 |
+
"clip_ratio/region_mean": 0.0,
|
| 504 |
+
"completion_length": 3062.354217529297,
|
| 505 |
+
"cov_mean": -3.731243577931309e-05,
|
| 506 |
+
"cov_std": 0.3898714445531368,
|
| 507 |
+
"entropy": 0.388671875,
|
| 508 |
+
"epoch": 0.021714285714285714,
|
| 509 |
+
"grad_norm": 0.4545780420303345,
|
| 510 |
+
"kl": 3.999471664428711e-05,
|
| 511 |
+
"learning_rate": 9.779754323328192e-07,
|
| 512 |
+
"loss": -0.0387,
|
| 513 |
+
"reward": 0.8229166772216558,
|
| 514 |
+
"reward_std": 0.5051202178001404,
|
| 515 |
+
"rewards/accuracy_reward": 0.37500001583248377,
|
| 516 |
+
"rewards/format_reward": 0.4479166818782687,
|
| 517 |
+
"step": 19,
|
| 518 |
+
"w_high_ratio": 0.01639098348096013,
|
| 519 |
+
"w_low_ratio": 0.042750393971800804,
|
| 520 |
+
"w_max": 1.7452828288078308,
|
| 521 |
+
"w_mean": 1.2359632849693298,
|
| 522 |
+
"w_min": 0.0,
|
| 523 |
+
"w_std": 0.25578257627785206
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"clip_ratio/high_max": 0.0,
|
| 527 |
+
"clip_ratio/high_mean": 0.0,
|
| 528 |
+
"clip_ratio/low_mean": 0.0,
|
| 529 |
+
"clip_ratio/low_min": 0.0,
|
| 530 |
+
"clip_ratio/region_mean": 0.0,
|
| 531 |
+
"completion_length": 2637.7813415527344,
|
| 532 |
+
"cov_mean": -5.605046135315206e-05,
|
| 533 |
+
"cov_std": 0.43815504014492035,
|
| 534 |
+
"entropy": 0.339599609375,
|
| 535 |
+
"epoch": 0.022857142857142857,
|
| 536 |
+
"grad_norm": 0.3516737222671509,
|
| 537 |
+
"kl": 4.100799560546875e-05,
|
| 538 |
+
"learning_rate": 9.728616793536587e-07,
|
| 539 |
+
"loss": -0.0559,
|
| 540 |
+
"reward": 1.0104167014360428,
|
| 541 |
+
"reward_std": 0.5058020055294037,
|
| 542 |
+
"rewards/accuracy_reward": 0.2916666679084301,
|
| 543 |
+
"rewards/format_reward": 0.7187500298023224,
|
| 544 |
+
"step": 20,
|
| 545 |
+
"w_high_ratio": 0.11945514753460884,
|
| 546 |
+
"w_low_ratio": 0.04084030445665121,
|
| 547 |
+
"w_max": 2.0172511637210846,
|
| 548 |
+
"w_mean": 1.3857647478580475,
|
| 549 |
+
"w_min": 0.0,
|
| 550 |
+
"w_std": 0.24064720049500465
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"clip_ratio/high_max": 0.0,
|
| 554 |
+
"clip_ratio/high_mean": 0.0,
|
| 555 |
+
"clip_ratio/low_mean": 0.0,
|
| 556 |
+
"clip_ratio/low_min": 0.0,
|
| 557 |
+
"clip_ratio/region_mean": 0.0,
|
| 558 |
+
"completion_length": 2998.5208740234375,
|
| 559 |
+
"cov_mean": 2.1053672753623687e-05,
|
| 560 |
+
"cov_std": 0.19154372438788414,
|
| 561 |
+
"entropy": 0.423828125,
|
| 562 |
+
"epoch": 0.024,
|
| 563 |
+
"grad_norm": 0.40126267075538635,
|
| 564 |
+
"kl": 0.00011563301086425781,
|
| 565 |
+
"learning_rate": 9.672327345550543e-07,
|
| 566 |
+
"loss": 0.0037,
|
| 567 |
+
"reward": 0.5833333432674408,
|
| 568 |
+
"reward_std": 0.21227191016077995,
|
| 569 |
+
"rewards/accuracy_reward": 0.16666666977107525,
|
| 570 |
+
"rewards/format_reward": 0.4166666679084301,
|
| 571 |
+
"step": 21,
|
| 572 |
+
"w_high_ratio": 0.17914548516273499,
|
| 573 |
+
"w_low_ratio": 0.0219349917024374,
|
| 574 |
+
"w_max": 1.8282522857189178,
|
| 575 |
+
"w_mean": 1.416891485452652,
|
| 576 |
+
"w_min": 0.25,
|
| 577 |
+
"w_std": 0.1492762751877308
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"clip_ratio/high_max": 0.0,
|
| 581 |
+
"clip_ratio/high_mean": 0.0,
|
| 582 |
+
"clip_ratio/low_mean": 0.0,
|
| 583 |
+
"clip_ratio/low_min": 0.0,
|
| 584 |
+
"clip_ratio/region_mean": 0.0,
|
| 585 |
+
"completion_length": 1878.3125457763672,
|
| 586 |
+
"cov_mean": 9.24348714761436e-05,
|
| 587 |
+
"cov_std": 0.4086475111544132,
|
| 588 |
+
"entropy": 0.3994140625,
|
| 589 |
+
"epoch": 0.025142857142857144,
|
| 590 |
+
"grad_norm": 0.7276642918586731,
|
| 591 |
+
"kl": 0.00010776519775390625,
|
| 592 |
+
"learning_rate": 9.610954559391704e-07,
|
| 593 |
+
"loss": -0.054,
|
| 594 |
+
"reward": 1.1041667014360428,
|
| 595 |
+
"reward_std": 0.40466368943452835,
|
| 596 |
+
"rewards/accuracy_reward": 0.29166667722165585,
|
| 597 |
+
"rewards/format_reward": 0.8125000298023224,
|
| 598 |
+
"step": 22,
|
| 599 |
+
"w_high_ratio": 0.22925379872322083,
|
| 600 |
+
"w_low_ratio": 0.04057216551154852,
|
| 601 |
+
"w_max": 2.499152511358261,
|
| 602 |
+
"w_mean": 1.6791247129440308,
|
| 603 |
+
"w_min": 0.0,
|
| 604 |
+
"w_std": 0.29966018348932266
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"clip_ratio/high_max": 0.0,
|
| 608 |
+
"clip_ratio/high_mean": 0.0,
|
| 609 |
+
"clip_ratio/low_mean": 0.0,
|
| 610 |
+
"clip_ratio/low_min": 0.0,
|
| 611 |
+
"clip_ratio/region_mean": 0.0,
|
| 612 |
+
"completion_length": 2871.3438720703125,
|
| 613 |
+
"cov_mean": -0.0001690889857854927,
|
| 614 |
+
"cov_std": 0.5799647718667984,
|
| 615 |
+
"entropy": 0.37353515625,
|
| 616 |
+
"epoch": 0.026285714285714287,
|
| 617 |
+
"grad_norm": 0.5503926277160645,
|
| 618 |
+
"kl": 6.878376007080078e-05,
|
| 619 |
+
"learning_rate": 9.54457320834625e-07,
|
| 620 |
+
"loss": -0.0364,
|
| 621 |
+
"reward": 0.6354166865348816,
|
| 622 |
+
"reward_std": 0.6411799788475037,
|
| 623 |
+
"rewards/accuracy_reward": 0.16666667070239782,
|
| 624 |
+
"rewards/format_reward": 0.4687500149011612,
|
| 625 |
+
"step": 23,
|
| 626 |
+
"w_high_ratio": 0.326111301779747,
|
| 627 |
+
"w_low_ratio": 0.05267245415598154,
|
| 628 |
+
"w_max": 2.298383414745331,
|
| 629 |
+
"w_mean": 1.572682112455368,
|
| 630 |
+
"w_min": 0.0,
|
| 631 |
+
"w_std": 0.3976207822561264
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"clip_ratio/high_max": 0.0,
|
| 635 |
+
"clip_ratio/high_mean": 0.0,
|
| 636 |
+
"clip_ratio/low_mean": 0.0,
|
| 637 |
+
"clip_ratio/low_min": 0.0,
|
| 638 |
+
"clip_ratio/region_mean": 0.0,
|
| 639 |
+
"completion_length": 2923.572998046875,
|
| 640 |
+
"cov_mean": 7.625344460393535e-05,
|
| 641 |
+
"cov_std": 0.3671033903956413,
|
| 642 |
+
"entropy": 0.3544921875,
|
| 643 |
+
"epoch": 0.027428571428571427,
|
| 644 |
+
"grad_norm": 0.37181225419044495,
|
| 645 |
+
"kl": 7.140636444091797e-05,
|
| 646 |
+
"learning_rate": 9.473264167865171e-07,
|
| 647 |
+
"loss": -0.0703,
|
| 648 |
+
"reward": 0.8541666865348816,
|
| 649 |
+
"reward_std": 0.5709189921617508,
|
| 650 |
+
"rewards/accuracy_reward": 0.3020833432674408,
|
| 651 |
+
"rewards/format_reward": 0.5520833432674408,
|
| 652 |
+
"step": 24,
|
| 653 |
+
"w_high_ratio": 0.09629620611667633,
|
| 654 |
+
"w_low_ratio": 0.037845601327717304,
|
| 655 |
+
"w_max": 2.1263908445835114,
|
| 656 |
+
"w_mean": 1.3701403737068176,
|
| 657 |
+
"w_min": 0.25,
|
| 658 |
+
"w_std": 0.23599852621555328
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"clip_ratio/high_max": 0.0,
|
| 662 |
+
"clip_ratio/high_mean": 0.0,
|
| 663 |
+
"clip_ratio/low_mean": 0.0,
|
| 664 |
+
"clip_ratio/low_min": 0.0,
|
| 665 |
+
"clip_ratio/region_mean": 0.0,
|
| 666 |
+
"completion_length": 2825.1458740234375,
|
| 667 |
+
"cov_mean": -5.495093546414864e-05,
|
| 668 |
+
"cov_std": 0.29340429976582527,
|
| 669 |
+
"entropy": 0.4248046875,
|
| 670 |
+
"epoch": 0.02857142857142857,
|
| 671 |
+
"grad_norm": 0.32156902551651,
|
| 672 |
+
"kl": 0.00015220046043395996,
|
| 673 |
+
"learning_rate": 9.397114317029974e-07,
|
| 674 |
+
"loss": -0.0438,
|
| 675 |
+
"reward": 0.677083358168602,
|
| 676 |
+
"reward_std": 0.3831539675593376,
|
| 677 |
+
"rewards/accuracy_reward": 0.2083333395421505,
|
| 678 |
+
"rewards/format_reward": 0.4687500149011612,
|
| 679 |
+
"step": 25,
|
| 680 |
+
"w_high_ratio": 0.16885175555944443,
|
| 681 |
+
"w_low_ratio": 0.032804112415760756,
|
| 682 |
+
"w_max": 1.8952468037605286,
|
| 683 |
+
"w_mean": 1.3390689194202423,
|
| 684 |
+
"w_min": 0.25,
|
| 685 |
+
"w_std": 0.22369066812098026
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"clip_ratio/high_max": 0.0,
|
| 689 |
+
"clip_ratio/high_mean": 0.0,
|
| 690 |
+
"clip_ratio/low_mean": 0.0,
|
| 691 |
+
"clip_ratio/low_min": 0.0,
|
| 692 |
+
"clip_ratio/region_mean": 0.0,
|
| 693 |
+
"completion_length": 3157.4584350585938,
|
| 694 |
+
"cov_mean": 3.823744009423535e-06,
|
| 695 |
+
"cov_std": 0.15900863334536552,
|
| 696 |
+
"entropy": 0.4033203125,
|
| 697 |
+
"epoch": 0.029714285714285714,
|
| 698 |
+
"grad_norm": 0.18318897485733032,
|
| 699 |
+
"kl": 2.8640031814575195e-05,
|
| 700 |
+
"learning_rate": 9.316216432703916e-07,
|
| 701 |
+
"loss": -0.0211,
|
| 702 |
+
"reward": 0.71875,
|
| 703 |
+
"reward_std": 0.1498664766550064,
|
| 704 |
+
"rewards/accuracy_reward": 0.2291666716337204,
|
| 705 |
+
"rewards/format_reward": 0.4895833358168602,
|
| 706 |
+
"step": 26,
|
| 707 |
+
"w_high_ratio": 0.0625,
|
| 708 |
+
"w_low_ratio": 0.012461496517062187,
|
| 709 |
+
"w_max": 1.7156271040439606,
|
| 710 |
+
"w_mean": 1.2929016053676605,
|
| 711 |
+
"w_min": 0.5,
|
| 712 |
+
"w_std": 0.07156859897077084
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"clip_ratio/high_max": 0.0,
|
| 716 |
+
"clip_ratio/high_mean": 0.0,
|
| 717 |
+
"clip_ratio/low_mean": 0.0,
|
| 718 |
+
"clip_ratio/low_min": 0.0,
|
| 719 |
+
"clip_ratio/region_mean": 0.0,
|
| 720 |
+
"completion_length": 3333.5729370117188,
|
| 721 |
+
"cov_mean": 9.062682693183888e-06,
|
| 722 |
+
"cov_std": 0.33011873811483383,
|
| 723 |
+
"entropy": 0.4306640625,
|
| 724 |
+
"epoch": 0.030857142857142857,
|
| 725 |
+
"grad_norm": 0.22338484227657318,
|
| 726 |
+
"kl": 7.772445678710938e-05,
|
| 727 |
+
"learning_rate": 9.230669076497687e-07,
|
| 728 |
+
"loss": -0.0295,
|
| 729 |
+
"reward": 0.5833333507180214,
|
| 730 |
+
"reward_std": 0.4810289070010185,
|
| 731 |
+
"rewards/accuracy_reward": 0.15625000558793545,
|
| 732 |
+
"rewards/format_reward": 0.4270833358168602,
|
| 733 |
+
"step": 27,
|
| 734 |
+
"w_high_ratio": 0.125,
|
| 735 |
+
"w_low_ratio": 0.04519441071897745,
|
| 736 |
+
"w_max": 1.898216873407364,
|
| 737 |
+
"w_mean": 1.2821258306503296,
|
| 738 |
+
"w_min": 0.0,
|
| 739 |
+
"w_std": 0.1903173178434372
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"clip_ratio/high_max": 0.0,
|
| 743 |
+
"clip_ratio/high_mean": 0.0,
|
| 744 |
+
"clip_ratio/low_mean": 0.0,
|
| 745 |
+
"clip_ratio/low_min": 0.0,
|
| 746 |
+
"clip_ratio/region_mean": 0.0,
|
| 747 |
+
"completion_length": 2990.6563110351562,
|
| 748 |
+
"cov_mean": -2.9691624149563722e-05,
|
| 749 |
+
"cov_std": 0.27876436710357666,
|
| 750 |
+
"entropy": 0.392578125,
|
| 751 |
+
"epoch": 0.032,
|
| 752 |
+
"grad_norm": 0.2658036947250366,
|
| 753 |
+
"kl": 0.00012731552124023438,
|
| 754 |
+
"learning_rate": 9.140576474687263e-07,
|
| 755 |
+
"loss": -0.0147,
|
| 756 |
+
"reward": 0.8229166716337204,
|
| 757 |
+
"reward_std": 0.32865703850984573,
|
| 758 |
+
"rewards/accuracy_reward": 0.3437500074505806,
|
| 759 |
+
"rewards/format_reward": 0.4791666716337204,
|
| 760 |
+
"step": 28,
|
| 761 |
+
"w_high_ratio": 0.03581305593252182,
|
| 762 |
+
"w_low_ratio": 0.036258784122765064,
|
| 763 |
+
"w_max": 1.8826820850372314,
|
| 764 |
+
"w_mean": 1.2603062689304352,
|
| 765 |
+
"w_min": 0.25,
|
| 766 |
+
"w_std": 0.2064611092209816
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"clip_ratio/high_max": 0.0,
|
| 770 |
+
"clip_ratio/high_mean": 0.0,
|
| 771 |
+
"clip_ratio/low_mean": 0.0,
|
| 772 |
+
"clip_ratio/low_min": 0.0,
|
| 773 |
+
"clip_ratio/region_mean": 0.0,
|
| 774 |
+
"completion_length": 3477.0625610351562,
|
| 775 |
+
"cov_mean": 3.658079458546126e-05,
|
| 776 |
+
"cov_std": 0.3980727344751358,
|
| 777 |
+
"entropy": 0.41162109375,
|
| 778 |
+
"epoch": 0.03314285714285714,
|
| 779 |
+
"grad_norm": 0.5845446586608887,
|
| 780 |
+
"kl": 0.00014966726303100586,
|
| 781 |
+
"learning_rate": 9.046048391230247e-07,
|
| 782 |
+
"loss": -0.03,
|
| 783 |
+
"reward": 0.4479166781529784,
|
| 784 |
+
"reward_std": 0.4323580041527748,
|
| 785 |
+
"rewards/accuracy_reward": 0.13541667256504297,
|
| 786 |
+
"rewards/format_reward": 0.31250000838190317,
|
| 787 |
+
"step": 29,
|
| 788 |
+
"w_high_ratio": 0.03906010463833809,
|
| 789 |
+
"w_low_ratio": 0.04348241724073887,
|
| 790 |
+
"w_max": 1.5083820223808289,
|
| 791 |
+
"w_mean": 1.1707176566123962,
|
| 792 |
+
"w_min": 0.0,
|
| 793 |
+
"w_std": 0.23769092187285423
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"clip_ratio/high_max": 0.0,
|
| 797 |
+
"clip_ratio/high_mean": 0.0,
|
| 798 |
+
"clip_ratio/low_mean": 0.0,
|
| 799 |
+
"clip_ratio/low_min": 0.0,
|
| 800 |
+
"clip_ratio/region_mean": 0.0,
|
| 801 |
+
"completion_length": 3228.4896240234375,
|
| 802 |
+
"cov_mean": 8.98454500202206e-05,
|
| 803 |
+
"cov_std": 0.4796721711754799,
|
| 804 |
+
"entropy": 0.388671875,
|
| 805 |
+
"epoch": 0.03428571428571429,
|
| 806 |
+
"grad_norm": 0.5218952894210815,
|
| 807 |
+
"kl": 0.0005021095275878906,
|
| 808 |
+
"learning_rate": 8.9471999940354e-07,
|
| 809 |
+
"loss": -0.0373,
|
| 810 |
+
"reward": 0.8229167014360428,
|
| 811 |
+
"reward_std": 0.6668380200862885,
|
| 812 |
+
"rewards/accuracy_reward": 0.3020833395421505,
|
| 813 |
+
"rewards/format_reward": 0.5208333469927311,
|
| 814 |
+
"step": 30,
|
| 815 |
+
"w_high_ratio": 0.09154843725264072,
|
| 816 |
+
"w_low_ratio": 0.050314661115407944,
|
| 817 |
+
"w_max": 1.7150111198425293,
|
| 818 |
+
"w_mean": 1.273286134004593,
|
| 819 |
+
"w_min": 0.0,
|
| 820 |
+
"w_std": 0.2975125089287758
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"clip_ratio/high_max": 0.0,
|
| 824 |
+
"clip_ratio/high_mean": 0.0,
|
| 825 |
+
"clip_ratio/low_mean": 0.0,
|
| 826 |
+
"clip_ratio/low_min": 0.0,
|
| 827 |
+
"clip_ratio/region_mean": 0.0,
|
| 828 |
+
"completion_length": 3316.479248046875,
|
| 829 |
+
"cov_mean": -2.0371426671772497e-05,
|
| 830 |
+
"cov_std": 0.32399024814367294,
|
| 831 |
+
"entropy": 0.39453125,
|
| 832 |
+
"epoch": 0.03542857142857143,
|
| 833 |
+
"grad_norm": 0.26446813344955444,
|
| 834 |
+
"kl": 0.00034368038177490234,
|
| 835 |
+
"learning_rate": 8.844151714648274e-07,
|
| 836 |
+
"loss": 0.0163,
|
| 837 |
+
"reward": 0.4791666828095913,
|
| 838 |
+
"reward_std": 0.45351940393447876,
|
| 839 |
+
"rewards/accuracy_reward": 0.17708333767950535,
|
| 840 |
+
"rewards/format_reward": 0.30208333395421505,
|
| 841 |
+
"step": 31,
|
| 842 |
+
"w_high_ratio": 0.16790161281824112,
|
| 843 |
+
"w_low_ratio": 0.038946577813476324,
|
| 844 |
+
"w_max": 1.8747785687446594,
|
| 845 |
+
"w_mean": 1.287878304719925,
|
| 846 |
+
"w_min": 0.25,
|
| 847 |
+
"w_std": 0.18489549681544304
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"clip_ratio/high_max": 0.0,
|
| 851 |
+
"clip_ratio/high_mean": 0.0,
|
| 852 |
+
"clip_ratio/low_mean": 0.0,
|
| 853 |
+
"clip_ratio/low_min": 0.0,
|
| 854 |
+
"clip_ratio/region_mean": 0.0,
|
| 855 |
+
"completion_length": 3449.291748046875,
|
| 856 |
+
"cov_mean": -4.376452466203773e-05,
|
| 857 |
+
"cov_std": 0.3894844390451908,
|
| 858 |
+
"entropy": 0.443359375,
|
| 859 |
+
"epoch": 0.036571428571428574,
|
| 860 |
+
"grad_norm": 0.3080126643180847,
|
| 861 |
+
"kl": 0.0014617443084716797,
|
| 862 |
+
"learning_rate": 8.737029101523929e-07,
|
| 863 |
+
"loss": -0.0432,
|
| 864 |
+
"reward": 0.7083333469927311,
|
| 865 |
+
"reward_std": 0.5129974186420441,
|
| 866 |
+
"rewards/accuracy_reward": 0.30208333395421505,
|
| 867 |
+
"rewards/format_reward": 0.4062500037252903,
|
| 868 |
+
"step": 32,
|
| 869 |
+
"w_high_ratio": 0.0,
|
| 870 |
+
"w_low_ratio": 0.04653155663982034,
|
| 871 |
+
"w_max": 1.6309894919395447,
|
| 872 |
+
"w_mean": 1.170417159795761,
|
| 873 |
+
"w_min": 0.0,
|
| 874 |
+
"w_std": 0.2305326908826828
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"clip_ratio/high_max": 0.0,
|
| 878 |
+
"clip_ratio/high_mean": 0.0,
|
| 879 |
+
"clip_ratio/low_mean": 0.0,
|
| 880 |
+
"clip_ratio/low_min": 0.0,
|
| 881 |
+
"clip_ratio/region_mean": 0.0,
|
| 882 |
+
"completion_length": 3521.125,
|
| 883 |
+
"cov_mean": -3.824161467491649e-06,
|
| 884 |
+
"cov_std": 0.3757542073726654,
|
| 885 |
+
"entropy": 0.380859375,
|
| 886 |
+
"epoch": 0.037714285714285714,
|
| 887 |
+
"grad_norm": 0.2623606026172638,
|
| 888 |
+
"kl": 0.0004220008850097656,
|
| 889 |
+
"learning_rate": 8.625962667065487e-07,
|
| 890 |
+
"loss": -0.0075,
|
| 891 |
+
"reward": 0.5520833507180214,
|
| 892 |
+
"reward_std": 0.5886104479432106,
|
| 893 |
+
"rewards/accuracy_reward": 0.1979166679084301,
|
| 894 |
+
"rewards/format_reward": 0.3541666828095913,
|
| 895 |
+
"step": 33,
|
| 896 |
+
"w_high_ratio": 0.0,
|
| 897 |
+
"w_low_ratio": 0.04900518897920847,
|
| 898 |
+
"w_max": 1.5290310084819794,
|
| 899 |
+
"w_mean": 1.145493358373642,
|
| 900 |
+
"w_min": 0.0,
|
| 901 |
+
"w_std": 0.23011762276291847
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"clip_ratio/high_max": 0.0,
|
| 905 |
+
"clip_ratio/high_mean": 0.0,
|
| 906 |
+
"clip_ratio/low_mean": 0.0,
|
| 907 |
+
"clip_ratio/low_min": 0.0,
|
| 908 |
+
"clip_ratio/region_mean": 0.0,
|
| 909 |
+
"completion_length": 2698.3020935058594,
|
| 910 |
+
"cov_mean": 5.745048838434741e-06,
|
| 911 |
+
"cov_std": 0.4045129865407944,
|
| 912 |
+
"entropy": 0.4296875,
|
| 913 |
+
"epoch": 0.038857142857142854,
|
| 914 |
+
"grad_norm": 0.4748375713825226,
|
| 915 |
+
"kl": 0.0005691051483154297,
|
| 916 |
+
"learning_rate": 8.511087728614862e-07,
|
| 917 |
+
"loss": -0.0352,
|
| 918 |
+
"reward": 0.979166679084301,
|
| 919 |
+
"reward_std": 0.4912763461470604,
|
| 920 |
+
"rewards/accuracy_reward": 0.4270833544433117,
|
| 921 |
+
"rewards/format_reward": 0.5520833386108279,
|
| 922 |
+
"step": 34,
|
| 923 |
+
"w_high_ratio": 0.1756632700562477,
|
| 924 |
+
"w_low_ratio": 0.04408560786396265,
|
| 925 |
+
"w_max": 2.0611276030540466,
|
| 926 |
+
"w_mean": 1.413420170545578,
|
| 927 |
+
"w_min": 6.104692548832184e-37,
|
| 928 |
+
"w_std": 0.29224658384919167
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"clip_ratio/high_max": 0.0,
|
| 932 |
+
"clip_ratio/high_mean": 0.0,
|
| 933 |
+
"clip_ratio/low_mean": 0.0,
|
| 934 |
+
"clip_ratio/low_min": 0.0,
|
| 935 |
+
"clip_ratio/region_mean": 0.0,
|
| 936 |
+
"completion_length": 3220.1146240234375,
|
| 937 |
+
"cov_mean": 2.1198811737122014e-06,
|
| 938 |
+
"cov_std": 0.43479467928409576,
|
| 939 |
+
"entropy": 0.41845703125,
|
| 940 |
+
"epoch": 0.04,
|
| 941 |
+
"grad_norm": 0.34247061610221863,
|
| 942 |
+
"kl": 0.0007988214492797852,
|
| 943 |
+
"learning_rate": 8.392544243589427e-07,
|
| 944 |
+
"loss": 0.0317,
|
| 945 |
+
"reward": 0.677083358168602,
|
| 946 |
+
"reward_std": 0.6704623848199844,
|
| 947 |
+
"rewards/accuracy_reward": 0.2604166753590107,
|
| 948 |
+
"rewards/format_reward": 0.4166666716337204,
|
| 949 |
+
"step": 35,
|
| 950 |
+
"w_high_ratio": 0.0,
|
| 951 |
+
"w_low_ratio": 0.059821099042892456,
|
| 952 |
+
"w_max": 1.4533329010009766,
|
| 953 |
+
"w_mean": 1.1137926578521729,
|
| 954 |
+
"w_min": 0.0,
|
| 955 |
+
"w_std": 0.26089195907115936
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"clip_ratio/high_max": 0.0,
|
| 959 |
+
"clip_ratio/high_mean": 0.0,
|
| 960 |
+
"clip_ratio/low_mean": 0.0,
|
| 961 |
+
"clip_ratio/low_min": 0.0,
|
| 962 |
+
"clip_ratio/region_mean": 0.0,
|
| 963 |
+
"completion_length": 3597.8333740234375,
|
| 964 |
+
"cov_mean": 4.828646342502907e-05,
|
| 965 |
+
"cov_std": 0.35842880606651306,
|
| 966 |
+
"entropy": 0.49462890625,
|
| 967 |
+
"epoch": 0.04114285714285714,
|
| 968 |
+
"grad_norm": 0.41333624720573425,
|
| 969 |
+
"kl": 0.00125160813331604,
|
| 970 |
+
"learning_rate": 8.270476638965461e-07,
|
| 971 |
+
"loss": 0.0601,
|
| 972 |
+
"reward": 0.2604166744276881,
|
| 973 |
+
"reward_std": 0.3833026885986328,
|
| 974 |
+
"rewards/accuracy_reward": 0.0416666679084301,
|
| 975 |
+
"rewards/format_reward": 0.21875000093132257,
|
| 976 |
+
"step": 36,
|
| 977 |
+
"w_high_ratio": 0.03994074463844299,
|
| 978 |
+
"w_low_ratio": 0.04655152838677168,
|
| 979 |
+
"w_max": 1.508450299501419,
|
| 980 |
+
"w_mean": 1.11627396941185,
|
| 981 |
+
"w_min": 0.0,
|
| 982 |
+
"w_std": 0.19725885801017284
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"clip_ratio/high_max": 0.0,
|
| 986 |
+
"clip_ratio/high_mean": 0.0,
|
| 987 |
+
"clip_ratio/low_mean": 0.0,
|
| 988 |
+
"clip_ratio/low_min": 0.0,
|
| 989 |
+
"clip_ratio/region_mean": 0.0,
|
| 990 |
+
"completion_length": 3524.0625610351562,
|
| 991 |
+
"cov_mean": -4.1810366383288056e-05,
|
| 992 |
+
"cov_std": 0.08558216691017151,
|
| 993 |
+
"entropy": 0.4130859375,
|
| 994 |
+
"epoch": 0.04228571428571429,
|
| 995 |
+
"grad_norm": 0.5261622071266174,
|
| 996 |
+
"kl": 0.0005668401718139648,
|
| 997 |
+
"learning_rate": 8.145033635316128e-07,
|
| 998 |
+
"loss": -0.0234,
|
| 999 |
+
"reward": 0.2291666716337204,
|
| 1000 |
+
"reward_std": 0.07216878235340118,
|
| 1001 |
+
"rewards/accuracy_reward": 0.010416666977107525,
|
| 1002 |
+
"rewards/format_reward": 0.21875,
|
| 1003 |
+
"step": 37,
|
| 1004 |
+
"w_high_ratio": 0.08015145361423492,
|
| 1005 |
+
"w_low_ratio": 0.006487013306468725,
|
| 1006 |
+
"w_max": 1.6550433039665222,
|
| 1007 |
+
"w_mean": 1.1928074657917023,
|
| 1008 |
+
"w_min": 0.75,
|
| 1009 |
+
"w_std": 0.07643502950668335
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"clip_ratio/high_max": 0.0,
|
| 1013 |
+
"clip_ratio/high_mean": 0.0,
|
| 1014 |
+
"clip_ratio/low_mean": 0.0,
|
| 1015 |
+
"clip_ratio/low_min": 0.0,
|
| 1016 |
+
"clip_ratio/region_mean": 0.0,
|
| 1017 |
+
"completion_length": 3520.6146240234375,
|
| 1018 |
+
"cov_mean": 2.719235999393277e-05,
|
| 1019 |
+
"cov_std": 0.1502552330493927,
|
| 1020 |
+
"entropy": 0.44580078125,
|
| 1021 |
+
"epoch": 0.04342857142857143,
|
| 1022 |
+
"grad_norm": 0.23012137413024902,
|
| 1023 |
+
"kl": 0.00047528743743896484,
|
| 1024 |
+
"learning_rate": 8.01636806561836e-07,
|
| 1025 |
+
"loss": -0.0058,
|
| 1026 |
+
"reward": 0.364583358168602,
|
| 1027 |
+
"reward_std": 0.22543279081583023,
|
| 1028 |
+
"rewards/accuracy_reward": 0.1666666679084301,
|
| 1029 |
+
"rewards/format_reward": 0.1979166716337204,
|
| 1030 |
+
"step": 38,
|
| 1031 |
+
"w_high_ratio": 0.0,
|
| 1032 |
+
"w_low_ratio": 0.017723015509545803,
|
| 1033 |
+
"w_max": 1.3341983556747437,
|
| 1034 |
+
"w_mean": 1.0956220924854279,
|
| 1035 |
+
"w_min": 0.5,
|
| 1036 |
+
"w_std": 0.08639993146061897
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"clip_ratio/high_max": 0.0,
|
| 1040 |
+
"clip_ratio/high_mean": 0.0,
|
| 1041 |
+
"clip_ratio/low_mean": 0.0,
|
| 1042 |
+
"clip_ratio/low_min": 0.0,
|
| 1043 |
+
"clip_ratio/region_mean": 0.0,
|
| 1044 |
+
"completion_length": 3074.104248046875,
|
| 1045 |
+
"cov_mean": 3.08786220557522e-05,
|
| 1046 |
+
"cov_std": 0.2986413836479187,
|
| 1047 |
+
"entropy": 0.35693359375,
|
| 1048 |
+
"epoch": 0.044571428571428574,
|
| 1049 |
+
"grad_norm": 0.19847992062568665,
|
| 1050 |
+
"kl": 0.0011906623840332031,
|
| 1051 |
+
"learning_rate": 7.884636689049422e-07,
|
| 1052 |
+
"loss": -0.0017,
|
| 1053 |
+
"reward": 0.7187500074505806,
|
| 1054 |
+
"reward_std": 0.24960162490606308,
|
| 1055 |
+
"rewards/accuracy_reward": 0.2604166669771075,
|
| 1056 |
+
"rewards/format_reward": 0.4583333432674408,
|
| 1057 |
+
"step": 39,
|
| 1058 |
+
"w_high_ratio": 0.0,
|
| 1059 |
+
"w_low_ratio": 0.030357226729393005,
|
| 1060 |
+
"w_max": 1.4879463911056519,
|
| 1061 |
+
"w_mean": 1.1616290509700775,
|
| 1062 |
+
"w_min": 0.5,
|
| 1063 |
+
"w_std": 0.15971330553293228
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"clip_ratio/high_max": 0.0,
|
| 1067 |
+
"clip_ratio/high_mean": 0.0,
|
| 1068 |
+
"clip_ratio/low_mean": 0.0,
|
| 1069 |
+
"clip_ratio/low_min": 0.0,
|
| 1070 |
+
"clip_ratio/region_mean": 0.0,
|
| 1071 |
+
"completion_length": 2789.6875610351562,
|
| 1072 |
+
"cov_mean": -5.7511609156790655e-05,
|
| 1073 |
+
"cov_std": 0.31858188658952713,
|
| 1074 |
+
"entropy": 0.38525390625,
|
| 1075 |
+
"epoch": 0.045714285714285714,
|
| 1076 |
+
"grad_norm": 0.2732243835926056,
|
| 1077 |
+
"kl": 0.00257110595703125,
|
| 1078 |
+
"learning_rate": 7.75e-07,
|
| 1079 |
+
"loss": 0.0166,
|
| 1080 |
+
"reward": 0.7500000298023224,
|
| 1081 |
+
"reward_std": 0.41665229201316833,
|
| 1082 |
+
"rewards/accuracy_reward": 0.20833334047347307,
|
| 1083 |
+
"rewards/format_reward": 0.5416666753590107,
|
| 1084 |
+
"step": 40,
|
| 1085 |
+
"w_high_ratio": 0.04059065505862236,
|
| 1086 |
+
"w_low_ratio": 0.037114882841706276,
|
| 1087 |
+
"w_max": 1.8804034292697906,
|
| 1088 |
+
"w_mean": 1.3260886669158936,
|
| 1089 |
+
"w_min": 0.0,
|
| 1090 |
+
"w_std": 0.2053362838923931
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"clip_ratio/high_max": 0.0,
|
| 1094 |
+
"clip_ratio/high_mean": 0.0,
|
| 1095 |
+
"clip_ratio/low_mean": 0.0,
|
| 1096 |
+
"clip_ratio/low_min": 0.0,
|
| 1097 |
+
"clip_ratio/region_mean": 0.0,
|
| 1098 |
+
"completion_length": 3289.0209350585938,
|
| 1099 |
+
"cov_mean": 1.5465679098269902e-05,
|
| 1100 |
+
"cov_std": 0.45656104385852814,
|
| 1101 |
+
"entropy": 0.37890625,
|
| 1102 |
+
"epoch": 0.046857142857142854,
|
| 1103 |
+
"grad_norm": 0.26020023226737976,
|
| 1104 |
+
"kl": 0.0005598068237304688,
|
| 1105 |
+
"learning_rate": 7.612622032536507e-07,
|
| 1106 |
+
"loss": -0.0115,
|
| 1107 |
+
"reward": 0.489583358168602,
|
| 1108 |
+
"reward_std": 0.507674403488636,
|
| 1109 |
+
"rewards/accuracy_reward": 0.1041666716337204,
|
| 1110 |
+
"rewards/format_reward": 0.3854166865348816,
|
| 1111 |
+
"step": 41,
|
| 1112 |
+
"w_high_ratio": 0.1044064536690712,
|
| 1113 |
+
"w_low_ratio": 0.047306024469435215,
|
| 1114 |
+
"w_max": 2.001980274915695,
|
| 1115 |
+
"w_mean": 1.2991511821746826,
|
| 1116 |
+
"w_min": 3.6180180170197286e-34,
|
| 1117 |
+
"w_std": 0.2728148624300957
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"clip_ratio/high_max": 0.0,
|
| 1121 |
+
"clip_ratio/high_mean": 0.0,
|
| 1122 |
+
"clip_ratio/low_mean": 0.0,
|
| 1123 |
+
"clip_ratio/low_min": 0.0,
|
| 1124 |
+
"clip_ratio/region_mean": 0.0,
|
| 1125 |
+
"completion_length": 3092.7709350585938,
|
| 1126 |
+
"cov_mean": 4.073379568581004e-05,
|
| 1127 |
+
"cov_std": 0.20339186489582062,
|
| 1128 |
+
"entropy": 0.4619140625,
|
| 1129 |
+
"epoch": 0.048,
|
| 1130 |
+
"grad_norm": 0.27600958943367004,
|
| 1131 |
+
"kl": 0.0006108283996582031,
|
| 1132 |
+
"learning_rate": 7.472670160550848e-07,
|
| 1133 |
+
"loss": 0.0105,
|
| 1134 |
+
"reward": 0.3125000037252903,
|
| 1135 |
+
"reward_std": 0.18993012979626656,
|
| 1136 |
+
"rewards/accuracy_reward": 0.02083333395421505,
|
| 1137 |
+
"rewards/format_reward": 0.29166666977107525,
|
| 1138 |
+
"step": 42,
|
| 1139 |
+
"w_high_ratio": 0.05672482028603554,
|
| 1140 |
+
"w_low_ratio": 0.01900631980970502,
|
| 1141 |
+
"w_max": 1.7394262850284576,
|
| 1142 |
+
"w_mean": 1.195200800895691,
|
| 1143 |
+
"w_min": 0.25,
|
| 1144 |
+
"w_std": 0.11057326570153236
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"clip_ratio/high_max": 0.0,
|
| 1148 |
+
"clip_ratio/high_mean": 0.0,
|
| 1149 |
+
"clip_ratio/low_mean": 0.0,
|
| 1150 |
+
"clip_ratio/low_min": 0.0,
|
| 1151 |
+
"clip_ratio/region_mean": 0.0,
|
| 1152 |
+
"completion_length": 3247.92724609375,
|
| 1153 |
+
"cov_mean": 4.6732177906960715e-05,
|
| 1154 |
+
"cov_std": 0.2612141128629446,
|
| 1155 |
+
"entropy": 0.40087890625,
|
| 1156 |
+
"epoch": 0.04914285714285714,
|
| 1157 |
+
"grad_norm": 0.28032609820365906,
|
| 1158 |
+
"kl": 0.001550912857055664,
|
| 1159 |
+
"learning_rate": 7.330314893841101e-07,
|
| 1160 |
+
"loss": -0.0352,
|
| 1161 |
+
"reward": 0.4166666818782687,
|
| 1162 |
+
"reward_std": 0.2987091466784477,
|
| 1163 |
+
"rewards/accuracy_reward": 0.13541666977107525,
|
| 1164 |
+
"rewards/format_reward": 0.2812500027939677,
|
| 1165 |
+
"step": 43,
|
| 1166 |
+
"w_high_ratio": 0.041576627641916275,
|
| 1167 |
+
"w_low_ratio": 0.03293204982765019,
|
| 1168 |
+
"w_max": 1.584185391664505,
|
| 1169 |
+
"w_mean": 1.1372461915016174,
|
| 1170 |
+
"w_min": 0.0,
|
| 1171 |
+
"w_std": 0.16969743371009827
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"clip_ratio/high_max": 0.0,
|
| 1175 |
+
"clip_ratio/high_mean": 0.0,
|
| 1176 |
+
"clip_ratio/low_mean": 0.0,
|
| 1177 |
+
"clip_ratio/low_min": 0.0,
|
| 1178 |
+
"clip_ratio/region_mean": 0.0,
|
| 1179 |
+
"completion_length": 2958.885482788086,
|
| 1180 |
+
"cov_mean": -4.934846765536349e-05,
|
| 1181 |
+
"cov_std": 0.305513933300972,
|
| 1182 |
+
"entropy": 0.41455078125,
|
| 1183 |
+
"epoch": 0.05028571428571429,
|
| 1184 |
+
"grad_norm": 0.5735454559326172,
|
| 1185 |
+
"kl": 0.001309037208557129,
|
| 1186 |
+
"learning_rate": 7.185729670371604e-07,
|
| 1187 |
+
"loss": -0.0235,
|
| 1188 |
+
"reward": 0.6770833432674408,
|
| 1189 |
+
"reward_std": 0.37915313243865967,
|
| 1190 |
+
"rewards/accuracy_reward": 0.2291666679084301,
|
| 1191 |
+
"rewards/format_reward": 0.4479166716337204,
|
| 1192 |
+
"step": 44,
|
| 1193 |
+
"w_high_ratio": 0.05263275280594826,
|
| 1194 |
+
"w_low_ratio": 0.033476518001407385,
|
| 1195 |
+
"w_max": 1.7282162308692932,
|
| 1196 |
+
"w_mean": 1.2310035228729248,
|
| 1197 |
+
"w_min": 0.25,
|
| 1198 |
+
"w_std": 0.20509596168994904
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"clip_ratio/high_max": 0.0,
|
| 1202 |
+
"clip_ratio/high_mean": 0.0,
|
| 1203 |
+
"clip_ratio/low_mean": 0.0,
|
| 1204 |
+
"clip_ratio/low_min": 0.0,
|
| 1205 |
+
"clip_ratio/region_mean": 0.0,
|
| 1206 |
+
"completion_length": 3681.7500610351562,
|
| 1207 |
+
"cov_mean": 5.253156632534228e-05,
|
| 1208 |
+
"cov_std": 0.27400747686624527,
|
| 1209 |
+
"entropy": 0.4130859375,
|
| 1210 |
+
"epoch": 0.05142857142857143,
|
| 1211 |
+
"grad_norm": 0.22661462426185608,
|
| 1212 |
+
"kl": 0.001455068588256836,
|
| 1213 |
+
"learning_rate": 7.039090644965509e-07,
|
| 1214 |
+
"loss": 0.0004,
|
| 1215 |
+
"reward": 0.3854166716337204,
|
| 1216 |
+
"reward_std": 0.4831271469593048,
|
| 1217 |
+
"rewards/accuracy_reward": 0.16666666977107525,
|
| 1218 |
+
"rewards/format_reward": 0.2187500111758709,
|
| 1219 |
+
"step": 45,
|
| 1220 |
+
"w_high_ratio": 0.0,
|
| 1221 |
+
"w_low_ratio": 0.036470845341682434,
|
| 1222 |
+
"w_max": 1.3770877122879028,
|
| 1223 |
+
"w_mean": 1.0782505869865417,
|
| 1224 |
+
"w_min": 0.25,
|
| 1225 |
+
"w_std": 0.16539720818400383
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"clip_ratio/high_max": 0.0,
|
| 1229 |
+
"clip_ratio/high_mean": 0.0,
|
| 1230 |
+
"clip_ratio/low_mean": 0.0,
|
| 1231 |
+
"clip_ratio/low_min": 0.0,
|
| 1232 |
+
"clip_ratio/region_mean": 0.0,
|
| 1233 |
+
"completion_length": 3533.5209350585938,
|
| 1234 |
+
"cov_mean": 2.201543838964426e-05,
|
| 1235 |
+
"cov_std": 0.20324882864952087,
|
| 1236 |
+
"entropy": 0.4873046875,
|
| 1237 |
+
"epoch": 0.052571428571428575,
|
| 1238 |
+
"grad_norm": 0.42943075299263,
|
| 1239 |
+
"kl": 0.0011663436889648438,
|
| 1240 |
+
"learning_rate": 6.890576474687263e-07,
|
| 1241 |
+
"loss": 0.0747,
|
| 1242 |
+
"reward": 0.18750000558793545,
|
| 1243 |
+
"reward_std": 0.185950867831707,
|
| 1244 |
+
"rewards/accuracy_reward": 0.02083333395421505,
|
| 1245 |
+
"rewards/format_reward": 0.1666666679084301,
|
| 1246 |
+
"step": 46,
|
| 1247 |
+
"w_high_ratio": 0.10795218497514725,
|
| 1248 |
+
"w_low_ratio": 0.02661217050626874,
|
| 1249 |
+
"w_max": 1.561707764863968,
|
| 1250 |
+
"w_mean": 1.1653469800949097,
|
| 1251 |
+
"w_min": 0.0,
|
| 1252 |
+
"w_std": 0.1572987660765648
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"clip_ratio/high_max": 0.0,
|
| 1256 |
+
"clip_ratio/high_mean": 0.0,
|
| 1257 |
+
"clip_ratio/low_mean": 0.0,
|
| 1258 |
+
"clip_ratio/low_min": 0.0,
|
| 1259 |
+
"clip_ratio/region_mean": 0.0,
|
| 1260 |
+
"completion_length": 3152.2188720703125,
|
| 1261 |
+
"cov_mean": -8.783125122135971e-05,
|
| 1262 |
+
"cov_std": 0.4741008821874857,
|
| 1263 |
+
"entropy": 0.388671875,
|
| 1264 |
+
"epoch": 0.053714285714285714,
|
| 1265 |
+
"grad_norm": 0.45350003242492676,
|
| 1266 |
+
"kl": 0.000950932502746582,
|
| 1267 |
+
"learning_rate": 6.740368101176495e-07,
|
| 1268 |
+
"loss": -0.0528,
|
| 1269 |
+
"reward": 0.8125000400468707,
|
| 1270 |
+
"reward_std": 0.599571980535984,
|
| 1271 |
+
"rewards/accuracy_reward": 0.3125000111758709,
|
| 1272 |
+
"rewards/format_reward": 0.5000000102445483,
|
| 1273 |
+
"step": 47,
|
| 1274 |
+
"w_high_ratio": 0.09517703950405121,
|
| 1275 |
+
"w_low_ratio": 0.0510673513635993,
|
| 1276 |
+
"w_max": 1.8478549420833588,
|
| 1277 |
+
"w_mean": 1.2849175333976746,
|
| 1278 |
+
"w_min": 0.0,
|
| 1279 |
+
"w_std": 0.2784101627767086
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"clip_ratio/high_max": 0.0,
|
| 1283 |
+
"clip_ratio/high_mean": 0.0,
|
| 1284 |
+
"clip_ratio/low_mean": 0.0,
|
| 1285 |
+
"clip_ratio/low_min": 0.0,
|
| 1286 |
+
"clip_ratio/region_mean": 0.0,
|
| 1287 |
+
"completion_length": 2912.8334350585938,
|
| 1288 |
+
"cov_mean": 4.9970367399509996e-08,
|
| 1289 |
+
"cov_std": 0.39979860931634903,
|
| 1290 |
+
"entropy": 0.39013671875,
|
| 1291 |
+
"epoch": 0.054857142857142854,
|
| 1292 |
+
"grad_norm": 0.42112770676612854,
|
| 1293 |
+
"kl": 0.00843667984008789,
|
| 1294 |
+
"learning_rate": 6.588648530198504e-07,
|
| 1295 |
+
"loss": -0.012,
|
| 1296 |
+
"reward": 0.7083333469927311,
|
| 1297 |
+
"reward_std": 0.5211797282099724,
|
| 1298 |
+
"rewards/accuracy_reward": 0.25000001303851604,
|
| 1299 |
+
"rewards/format_reward": 0.45833334140479565,
|
| 1300 |
+
"step": 48,
|
| 1301 |
+
"w_high_ratio": 0.1536145769059658,
|
| 1302 |
+
"w_low_ratio": 0.049599490128457546,
|
| 1303 |
+
"w_max": 2.127849966287613,
|
| 1304 |
+
"w_mean": 1.337787538766861,
|
| 1305 |
+
"w_min": 0.0,
|
| 1306 |
+
"w_std": 0.3069620132446289
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"clip_ratio/high_max": 0.0,
|
| 1310 |
+
"clip_ratio/high_mean": 0.0,
|
| 1311 |
+
"clip_ratio/low_mean": 0.0,
|
| 1312 |
+
"clip_ratio/low_min": 0.0,
|
| 1313 |
+
"clip_ratio/region_mean": 0.0,
|
| 1314 |
+
"completion_length": 2544.666748046875,
|
| 1315 |
+
"cov_mean": -1.3007749203097774e-05,
|
| 1316 |
+
"cov_std": 0.3665553033351898,
|
| 1317 |
+
"entropy": 0.365234375,
|
| 1318 |
+
"epoch": 0.056,
|
| 1319 |
+
"grad_norm": 0.326923668384552,
|
| 1320 |
+
"kl": 0.007767677307128906,
|
| 1321 |
+
"learning_rate": 6.435602608679916e-07,
|
| 1322 |
+
"loss": 0.028,
|
| 1323 |
+
"reward": 0.770833358168602,
|
| 1324 |
+
"reward_std": 0.4556447044014931,
|
| 1325 |
+
"rewards/accuracy_reward": 0.17708333861082792,
|
| 1326 |
+
"rewards/format_reward": 0.5937500149011612,
|
| 1327 |
+
"step": 49,
|
| 1328 |
+
"w_high_ratio": 0.07310386933386326,
|
| 1329 |
+
"w_low_ratio": 0.03729367163032293,
|
| 1330 |
+
"w_max": 1.794796198606491,
|
| 1331 |
+
"w_mean": 1.2659293115139008,
|
| 1332 |
+
"w_min": 6.92190445412429e-34,
|
| 1333 |
+
"w_std": 0.23256080597639084
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"clip_ratio/high_max": 0.0,
|
| 1337 |
+
"clip_ratio/high_mean": 0.0,
|
| 1338 |
+
"clip_ratio/low_mean": 0.0,
|
| 1339 |
+
"clip_ratio/low_min": 0.0,
|
| 1340 |
+
"clip_ratio/region_mean": 0.0,
|
| 1341 |
+
"completion_length": 3237.2916870117188,
|
| 1342 |
+
"cov_mean": 1.4073829220251355e-05,
|
| 1343 |
+
"cov_std": 0.2659311108291149,
|
| 1344 |
+
"entropy": 0.33642578125,
|
| 1345 |
+
"epoch": 0.05714285714285714,
|
| 1346 |
+
"grad_norm": 0.18902550637722015,
|
| 1347 |
+
"kl": 0.0010194778442382812,
|
| 1348 |
+
"learning_rate": 6.281416799501187e-07,
|
| 1349 |
+
"loss": -0.0195,
|
| 1350 |
+
"reward": 0.645833345130086,
|
| 1351 |
+
"reward_std": 0.36775920540094376,
|
| 1352 |
+
"rewards/accuracy_reward": 0.2916666669771075,
|
| 1353 |
+
"rewards/format_reward": 0.3541666781529784,
|
| 1354 |
+
"step": 50,
|
| 1355 |
+
"w_high_ratio": 0.03271413967013359,
|
| 1356 |
+
"w_low_ratio": 0.030921938829123974,
|
| 1357 |
+
"w_max": 1.3970885872840881,
|
| 1358 |
+
"w_mean": 1.1265143752098083,
|
| 1359 |
+
"w_min": 0.25,
|
| 1360 |
+
"w_std": 0.15340184792876244
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"clip_ratio/high_max": 0.0,
|
| 1364 |
+
"clip_ratio/high_mean": 0.0,
|
| 1365 |
+
"clip_ratio/low_mean": 0.0,
|
| 1366 |
+
"clip_ratio/low_min": 0.0,
|
| 1367 |
+
"clip_ratio/region_mean": 0.0,
|
| 1368 |
+
"completion_length": 2497.2188720703125,
|
| 1369 |
+
"cov_mean": 0.00010049804541267804,
|
| 1370 |
+
"cov_std": 0.31800365447998047,
|
| 1371 |
+
"entropy": 0.451171875,
|
| 1372 |
+
"epoch": 0.05828571428571429,
|
| 1373 |
+
"grad_norm": 0.3681160509586334,
|
| 1374 |
+
"kl": 0.0043849945068359375,
|
| 1375 |
+
"learning_rate": 6.126278954320294e-07,
|
| 1376 |
+
"loss": 0.0301,
|
| 1377 |
+
"reward": 0.739583358168602,
|
| 1378 |
+
"reward_std": 0.32177041098475456,
|
| 1379 |
+
"rewards/accuracy_reward": 0.2083333358168602,
|
| 1380 |
+
"rewards/format_reward": 0.5312500149011612,
|
| 1381 |
+
"step": 51,
|
| 1382 |
+
"w_high_ratio": 0.0,
|
| 1383 |
+
"w_low_ratio": 0.038371569477021694,
|
| 1384 |
+
"w_max": 1.5697456300258636,
|
| 1385 |
+
"w_mean": 1.1596841514110565,
|
| 1386 |
+
"w_min": 0.25,
|
| 1387 |
+
"w_std": 0.17389511317014694
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"clip_ratio/high_max": 0.0,
|
| 1391 |
+
"clip_ratio/high_mean": 0.0,
|
| 1392 |
+
"clip_ratio/low_mean": 0.0,
|
| 1393 |
+
"clip_ratio/low_min": 0.0,
|
| 1394 |
+
"clip_ratio/region_mean": 0.0,
|
| 1395 |
+
"completion_length": 3154.15625,
|
| 1396 |
+
"cov_mean": 1.5342720871558413e-08,
|
| 1397 |
+
"cov_std": 0.3799358010292053,
|
| 1398 |
+
"entropy": 0.38623046875,
|
| 1399 |
+
"epoch": 0.05942857142857143,
|
| 1400 |
+
"grad_norm": 0.2101733535528183,
|
| 1401 |
+
"kl": 0.0017533302307128906,
|
| 1402 |
+
"learning_rate": 5.97037808470444e-07,
|
| 1403 |
+
"loss": 0.0285,
|
| 1404 |
+
"reward": 0.6979166697710752,
|
| 1405 |
+
"reward_std": 0.5342019759118557,
|
| 1406 |
+
"rewards/accuracy_reward": 0.2812500074505806,
|
| 1407 |
+
"rewards/format_reward": 0.41666667722165585,
|
| 1408 |
+
"step": 52,
|
| 1409 |
+
"w_high_ratio": 0.0,
|
| 1410 |
+
"w_low_ratio": 0.04498880193568766,
|
| 1411 |
+
"w_max": 1.3936763405799866,
|
| 1412 |
+
"w_mean": 1.130004495382309,
|
| 1413 |
+
"w_min": 0.0,
|
| 1414 |
+
"w_std": 0.19680343009531498
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"clip_ratio/high_max": 0.0,
|
| 1418 |
+
"clip_ratio/high_mean": 0.0,
|
| 1419 |
+
"clip_ratio/low_mean": 0.0,
|
| 1420 |
+
"clip_ratio/low_min": 0.0,
|
| 1421 |
+
"clip_ratio/region_mean": 0.0,
|
| 1422 |
+
"completion_length": 3053.7708740234375,
|
| 1423 |
+
"cov_mean": 1.7406931874575093e-05,
|
| 1424 |
+
"cov_std": 0.46154169738292694,
|
| 1425 |
+
"entropy": 0.421875,
|
| 1426 |
+
"epoch": 0.060571428571428575,
|
| 1427 |
+
"grad_norm": 0.33937692642211914,
|
| 1428 |
+
"kl": 0.001083970069885254,
|
| 1429 |
+
"learning_rate": 5.813904131848564e-07,
|
| 1430 |
+
"loss": -0.0326,
|
| 1431 |
+
"reward": 0.8645833507180214,
|
| 1432 |
+
"reward_std": 0.6376037672162056,
|
| 1433 |
+
"rewards/accuracy_reward": 0.3020833358168602,
|
| 1434 |
+
"rewards/format_reward": 0.5625000074505806,
|
| 1435 |
+
"step": 53,
|
| 1436 |
+
"w_high_ratio": 0.04103509336709976,
|
| 1437 |
+
"w_low_ratio": 0.05415187170729041,
|
| 1438 |
+
"w_max": 1.5957177877426147,
|
| 1439 |
+
"w_mean": 1.1700108051300049,
|
| 1440 |
+
"w_min": 0.0,
|
| 1441 |
+
"w_std": 0.24823076836764812
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"clip_ratio/high_max": 0.0,
|
| 1445 |
+
"clip_ratio/high_mean": 0.0,
|
| 1446 |
+
"clip_ratio/low_mean": 0.0,
|
| 1447 |
+
"clip_ratio/low_min": 0.0,
|
| 1448 |
+
"clip_ratio/region_mean": 0.0,
|
| 1449 |
+
"completion_length": 2899.4688720703125,
|
| 1450 |
+
"cov_mean": -8.303408776555443e-06,
|
| 1451 |
+
"cov_std": 0.4274276793003082,
|
| 1452 |
+
"entropy": 0.37158203125,
|
| 1453 |
+
"epoch": 0.061714285714285715,
|
| 1454 |
+
"grad_norm": 0.4694461524486542,
|
| 1455 |
+
"kl": 0.0010991096496582031,
|
| 1456 |
+
"learning_rate": 5.657047735161255e-07,
|
| 1457 |
+
"loss": -0.0728,
|
| 1458 |
+
"reward": 1.020833358168602,
|
| 1459 |
+
"reward_std": 0.6003277003765106,
|
| 1460 |
+
"rewards/accuracy_reward": 0.4583333507180214,
|
| 1461 |
+
"rewards/format_reward": 0.5625000223517418,
|
| 1462 |
+
"step": 54,
|
| 1463 |
+
"w_high_ratio": 0.10438984632492065,
|
| 1464 |
+
"w_low_ratio": 0.044082850217819214,
|
| 1465 |
+
"w_max": 1.875702828168869,
|
| 1466 |
+
"w_mean": 1.3796275556087494,
|
| 1467 |
+
"w_min": 0.0,
|
| 1468 |
+
"w_std": 0.28376903757452965
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"clip_ratio/high_max": 0.0,
|
| 1472 |
+
"clip_ratio/high_mean": 0.0,
|
| 1473 |
+
"clip_ratio/low_mean": 0.0,
|
| 1474 |
+
"clip_ratio/low_min": 0.0,
|
| 1475 |
+
"clip_ratio/region_mean": 0.0,
|
| 1476 |
+
"completion_length": 3251.5626220703125,
|
| 1477 |
+
"cov_mean": 5.693810294360446e-05,
|
| 1478 |
+
"cov_std": 0.32220450788736343,
|
| 1479 |
+
"entropy": 0.39306640625,
|
| 1480 |
+
"epoch": 0.06285714285714286,
|
| 1481 |
+
"grad_norm": 0.3240751326084137,
|
| 1482 |
+
"kl": 0.0013048648834228516,
|
| 1483 |
+
"learning_rate": 5.5e-07,
|
| 1484 |
+
"loss": -0.0123,
|
| 1485 |
+
"reward": 0.5625000149011612,
|
| 1486 |
+
"reward_std": 0.4148038476705551,
|
| 1487 |
+
"rewards/accuracy_reward": 0.20833333674818277,
|
| 1488 |
+
"rewards/format_reward": 0.3541666753590107,
|
| 1489 |
+
"step": 55,
|
| 1490 |
+
"w_high_ratio": 0.06664891541004181,
|
| 1491 |
+
"w_low_ratio": 0.036852192133665085,
|
| 1492 |
+
"w_max": 1.7040621638298035,
|
| 1493 |
+
"w_mean": 1.2043538689613342,
|
| 1494 |
+
"w_min": 0.25,
|
| 1495 |
+
"w_std": 0.2000352405011654
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"clip_ratio/high_max": 0.0,
|
| 1499 |
+
"clip_ratio/high_mean": 0.0,
|
| 1500 |
+
"clip_ratio/low_mean": 0.0,
|
| 1501 |
+
"clip_ratio/low_min": 0.0,
|
| 1502 |
+
"clip_ratio/region_mean": 0.0,
|
| 1503 |
+
"completion_length": 3137.3750915527344,
|
| 1504 |
+
"cov_mean": 6.871268851682544e-07,
|
| 1505 |
+
"cov_std": 0.27662914246320724,
|
| 1506 |
+
"entropy": 0.38330078125,
|
| 1507 |
+
"epoch": 0.064,
|
| 1508 |
+
"grad_norm": 0.2951548397541046,
|
| 1509 |
+
"kl": 0.0005192756652832031,
|
| 1510 |
+
"learning_rate": 5.342952264838747e-07,
|
| 1511 |
+
"loss": -0.0268,
|
| 1512 |
+
"reward": 0.6770833507180214,
|
| 1513 |
+
"reward_std": 0.280517403036356,
|
| 1514 |
+
"rewards/accuracy_reward": 0.21875,
|
| 1515 |
+
"rewards/format_reward": 0.4583333507180214,
|
| 1516 |
+
"step": 56,
|
| 1517 |
+
"w_high_ratio": 0.0,
|
| 1518 |
+
"w_low_ratio": 0.030007817316800356,
|
| 1519 |
+
"w_max": 1.4776403903961182,
|
| 1520 |
+
"w_mean": 1.1647382080554962,
|
| 1521 |
+
"w_min": 1.7701902850583252e-42,
|
| 1522 |
+
"w_std": 0.1563316509127617
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"clip_ratio/high_max": 0.0,
|
| 1526 |
+
"clip_ratio/high_mean": 0.0,
|
| 1527 |
+
"clip_ratio/low_mean": 0.0,
|
| 1528 |
+
"clip_ratio/low_min": 0.0,
|
| 1529 |
+
"clip_ratio/region_mean": 0.0,
|
| 1530 |
+
"completion_length": 3544.9896240234375,
|
| 1531 |
+
"cov_mean": 2.435890860397194e-05,
|
| 1532 |
+
"cov_std": 0.3566681556403637,
|
| 1533 |
+
"entropy": 0.31005859375,
|
| 1534 |
+
"epoch": 0.06514285714285714,
|
| 1535 |
+
"grad_norm": 0.5244222283363342,
|
| 1536 |
+
"kl": 0.00036072731018066406,
|
| 1537 |
+
"learning_rate": 5.186095868151436e-07,
|
| 1538 |
+
"loss": -0.0007,
|
| 1539 |
+
"reward": 0.5208333386108279,
|
| 1540 |
+
"reward_std": 0.5608532726764679,
|
| 1541 |
+
"rewards/accuracy_reward": 0.16666667442768812,
|
| 1542 |
+
"rewards/format_reward": 0.3541666818782687,
|
| 1543 |
+
"step": 57,
|
| 1544 |
+
"w_high_ratio": 0.0016966089606285095,
|
| 1545 |
+
"w_low_ratio": 0.046746447682380676,
|
| 1546 |
+
"w_max": 1.4659978747367859,
|
| 1547 |
+
"w_mean": 1.1122615337371826,
|
| 1548 |
+
"w_min": 0.0,
|
| 1549 |
+
"w_std": 0.21731951087713242
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"clip_ratio/high_max": 0.0,
|
| 1553 |
+
"clip_ratio/high_mean": 0.0,
|
| 1554 |
+
"clip_ratio/low_mean": 0.0,
|
| 1555 |
+
"clip_ratio/low_min": 0.0,
|
| 1556 |
+
"clip_ratio/region_mean": 0.0,
|
| 1557 |
+
"completion_length": 2408.010467529297,
|
| 1558 |
+
"cov_mean": 6.246490556804929e-06,
|
| 1559 |
+
"cov_std": 0.37900903820991516,
|
| 1560 |
+
"entropy": 0.333251953125,
|
| 1561 |
+
"epoch": 0.06628571428571428,
|
| 1562 |
+
"grad_norm": 0.278490275144577,
|
| 1563 |
+
"kl": 0.0038089752197265625,
|
| 1564 |
+
"learning_rate": 5.02962191529556e-07,
|
| 1565 |
+
"loss": 0.0026,
|
| 1566 |
+
"reward": 0.9479167088866234,
|
| 1567 |
+
"reward_std": 0.44713668525218964,
|
| 1568 |
+
"rewards/accuracy_reward": 0.291666672565043,
|
| 1569 |
+
"rewards/format_reward": 0.6562500111758709,
|
| 1570 |
+
"step": 58,
|
| 1571 |
+
"w_high_ratio": 0.10547220706939697,
|
| 1572 |
+
"w_low_ratio": 0.040943967178463936,
|
| 1573 |
+
"w_max": 1.8222440481185913,
|
| 1574 |
+
"w_mean": 1.3768556416034698,
|
| 1575 |
+
"w_min": 3.915851168125891e-30,
|
| 1576 |
+
"w_std": 0.2507038414478302
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"clip_ratio/high_max": 0.0,
|
| 1580 |
+
"clip_ratio/high_mean": 0.0,
|
| 1581 |
+
"clip_ratio/low_mean": 0.0,
|
| 1582 |
+
"clip_ratio/low_min": 0.0,
|
| 1583 |
+
"clip_ratio/region_mean": 0.0,
|
| 1584 |
+
"completion_length": 3130.7916870117188,
|
| 1585 |
+
"cov_mean": 4.25004100179649e-06,
|
| 1586 |
+
"cov_std": 0.2073173113167286,
|
| 1587 |
+
"entropy": 0.332763671875,
|
| 1588 |
+
"epoch": 0.06742857142857143,
|
| 1589 |
+
"grad_norm": 0.354891836643219,
|
| 1590 |
+
"kl": 0.00041985511779785156,
|
| 1591 |
+
"learning_rate": 4.873721045679706e-07,
|
| 1592 |
+
"loss": 0.0052,
|
| 1593 |
+
"reward": 0.4791666669771075,
|
| 1594 |
+
"reward_std": 0.2661800682544708,
|
| 1595 |
+
"rewards/accuracy_reward": 0.1458333358168602,
|
| 1596 |
+
"rewards/format_reward": 0.3333333386108279,
|
| 1597 |
+
"step": 59,
|
| 1598 |
+
"w_high_ratio": 0.0,
|
| 1599 |
+
"w_low_ratio": 0.027361327316612005,
|
| 1600 |
+
"w_max": 1.3359772562980652,
|
| 1601 |
+
"w_mean": 1.097507894039154,
|
| 1602 |
+
"w_min": 0.25,
|
| 1603 |
+
"w_std": 0.14073412120342255
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"clip_ratio/high_max": 0.0,
|
| 1607 |
+
"clip_ratio/high_mean": 0.0,
|
| 1608 |
+
"clip_ratio/low_mean": 0.0,
|
| 1609 |
+
"clip_ratio/low_min": 0.0,
|
| 1610 |
+
"clip_ratio/region_mean": 0.0,
|
| 1611 |
+
"completion_length": 3174.041748046875,
|
| 1612 |
+
"cov_mean": -5.75426925024658e-05,
|
| 1613 |
+
"cov_std": 0.36878733336925507,
|
| 1614 |
+
"entropy": 0.353515625,
|
| 1615 |
+
"epoch": 0.06857142857142857,
|
| 1616 |
+
"grad_norm": 0.23771578073501587,
|
| 1617 |
+
"kl": 0.0018591880798339844,
|
| 1618 |
+
"learning_rate": 4.7185832004988133e-07,
|
| 1619 |
+
"loss": 0.0192,
|
| 1620 |
+
"reward": 0.5416666865348816,
|
| 1621 |
+
"reward_std": 0.4686477780342102,
|
| 1622 |
+
"rewards/accuracy_reward": 0.14583333861082792,
|
| 1623 |
+
"rewards/format_reward": 0.3958333432674408,
|
| 1624 |
+
"step": 60,
|
| 1625 |
+
"w_high_ratio": 0.04001428559422493,
|
| 1626 |
+
"w_low_ratio": 0.03881241288036108,
|
| 1627 |
+
"w_max": 1.5190648436546326,
|
| 1628 |
+
"w_mean": 1.161254346370697,
|
| 1629 |
+
"w_min": 0.0,
|
| 1630 |
+
"w_std": 0.19259942695498466
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"clip_ratio/high_max": 0.0,
|
| 1634 |
+
"clip_ratio/high_mean": 0.0,
|
| 1635 |
+
"clip_ratio/low_mean": 0.0,
|
| 1636 |
+
"clip_ratio/low_min": 0.0,
|
| 1637 |
+
"clip_ratio/region_mean": 0.0,
|
| 1638 |
+
"completion_length": 3259.229248046875,
|
| 1639 |
+
"cov_mean": 1.2544531273306347e-05,
|
| 1640 |
+
"cov_std": 0.32379359379410744,
|
| 1641 |
+
"entropy": 0.361328125,
|
| 1642 |
+
"epoch": 0.06971428571428571,
|
| 1643 |
+
"grad_norm": 0.3283206522464752,
|
| 1644 |
+
"kl": 0.0008237361907958984,
|
| 1645 |
+
"learning_rate": 4.5643973913200837e-07,
|
| 1646 |
+
"loss": -0.0185,
|
| 1647 |
+
"reward": 0.614583358168602,
|
| 1648 |
+
"reward_std": 0.4510444663465023,
|
| 1649 |
+
"rewards/accuracy_reward": 0.1666666716337204,
|
| 1650 |
+
"rewards/format_reward": 0.4479166716337204,
|
| 1651 |
+
"step": 61,
|
| 1652 |
+
"w_high_ratio": 0.0,
|
| 1653 |
+
"w_low_ratio": 0.04093870846554637,
|
| 1654 |
+
"w_max": 1.5261969566345215,
|
| 1655 |
+
"w_mean": 1.1891083717346191,
|
| 1656 |
+
"w_min": 0.25,
|
| 1657 |
+
"w_std": 0.2139909453690052
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"clip_ratio/high_max": 0.0,
|
| 1661 |
+
"clip_ratio/high_mean": 0.0,
|
| 1662 |
+
"clip_ratio/low_mean": 0.0,
|
| 1663 |
+
"clip_ratio/low_min": 0.0,
|
| 1664 |
+
"clip_ratio/region_mean": 0.0,
|
| 1665 |
+
"completion_length": 2809.1771850585938,
|
| 1666 |
+
"cov_mean": 3.0659541152999736e-05,
|
| 1667 |
+
"cov_std": 0.5087622255086899,
|
| 1668 |
+
"entropy": 0.326904296875,
|
| 1669 |
+
"epoch": 0.07085714285714285,
|
| 1670 |
+
"grad_norm": 0.6958292722702026,
|
| 1671 |
+
"kl": 0.022649765014648438,
|
| 1672 |
+
"learning_rate": 4.4113514698014953e-07,
|
| 1673 |
+
"loss": -0.0085,
|
| 1674 |
+
"reward": 0.8958333656191826,
|
| 1675 |
+
"reward_std": 0.6562541723251343,
|
| 1676 |
+
"rewards/accuracy_reward": 0.28125000558793545,
|
| 1677 |
+
"rewards/format_reward": 0.6145833507180214,
|
| 1678 |
+
"step": 62,
|
| 1679 |
+
"w_high_ratio": 0.058254461735486984,
|
| 1680 |
+
"w_low_ratio": 0.053510635159909725,
|
| 1681 |
+
"w_max": 1.9464432299137115,
|
| 1682 |
+
"w_mean": 1.336916297674179,
|
| 1683 |
+
"w_min": 0.0,
|
| 1684 |
+
"w_std": 0.31536465138196945
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"clip_ratio/high_max": 0.0,
|
| 1688 |
+
"clip_ratio/high_mean": 0.0,
|
| 1689 |
+
"clip_ratio/low_mean": 0.0,
|
| 1690 |
+
"clip_ratio/low_min": 0.0,
|
| 1691 |
+
"clip_ratio/region_mean": 0.0,
|
| 1692 |
+
"completion_length": 2338.1250610351562,
|
| 1693 |
+
"cov_mean": -3.693009693961358e-05,
|
| 1694 |
+
"cov_std": 0.37869442999362946,
|
| 1695 |
+
"entropy": 0.3642578125,
|
| 1696 |
+
"epoch": 0.072,
|
| 1697 |
+
"grad_norm": 0.3641963601112366,
|
| 1698 |
+
"kl": 0.0020456314086914062,
|
| 1699 |
+
"learning_rate": 4.2596318988235037e-07,
|
| 1700 |
+
"loss": -0.0179,
|
| 1701 |
+
"reward": 1.1562500596046448,
|
| 1702 |
+
"reward_std": 0.517615370452404,
|
| 1703 |
+
"rewards/accuracy_reward": 0.4062500074505806,
|
| 1704 |
+
"rewards/format_reward": 0.7500000223517418,
|
| 1705 |
+
"step": 63,
|
| 1706 |
+
"w_high_ratio": 0.10371162719093263,
|
| 1707 |
+
"w_low_ratio": 0.03615910932421684,
|
| 1708 |
+
"w_max": 2.5228151381015778,
|
| 1709 |
+
"w_mean": 1.538686066865921,
|
| 1710 |
+
"w_min": 4.8989017643768394e-36,
|
| 1711 |
+
"w_std": 0.2665465883910656
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"clip_ratio/high_max": 0.0,
|
| 1715 |
+
"clip_ratio/high_mean": 0.0,
|
| 1716 |
+
"clip_ratio/low_mean": 0.0,
|
| 1717 |
+
"clip_ratio/low_min": 0.0,
|
| 1718 |
+
"clip_ratio/region_mean": 0.0,
|
| 1719 |
+
"completion_length": 3296.385528564453,
|
| 1720 |
+
"cov_mean": -5.260490161163034e-05,
|
| 1721 |
+
"cov_std": 0.32132588326931,
|
| 1722 |
+
"entropy": 0.416015625,
|
| 1723 |
+
"epoch": 0.07314285714285715,
|
| 1724 |
+
"grad_norm": 0.26374608278274536,
|
| 1725 |
+
"kl": 0.0010199546813964844,
|
| 1726 |
+
"learning_rate": 4.1094235253127374e-07,
|
| 1727 |
+
"loss": -0.0008,
|
| 1728 |
+
"reward": 0.6145833656191826,
|
| 1729 |
+
"reward_std": 0.4967978894710541,
|
| 1730 |
+
"rewards/accuracy_reward": 0.22916667722165585,
|
| 1731 |
+
"rewards/format_reward": 0.3854166716337204,
|
| 1732 |
+
"step": 64,
|
| 1733 |
+
"w_high_ratio": 0.125,
|
| 1734 |
+
"w_low_ratio": 0.039727743715047836,
|
| 1735 |
+
"w_max": 1.6750397682189941,
|
| 1736 |
+
"w_mean": 1.3110667765140533,
|
| 1737 |
+
"w_min": 7.912817135887769e-41,
|
| 1738 |
+
"w_std": 0.19634342193603516
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"clip_ratio/high_max": 0.0,
|
| 1742 |
+
"clip_ratio/high_mean": 0.0,
|
| 1743 |
+
"clip_ratio/low_mean": 0.0,
|
| 1744 |
+
"clip_ratio/low_min": 0.0,
|
| 1745 |
+
"clip_ratio/region_mean": 0.0,
|
| 1746 |
+
"completion_length": 3007.2500610351562,
|
| 1747 |
+
"cov_mean": 3.157142418785952e-05,
|
| 1748 |
+
"cov_std": 0.27999811992049217,
|
| 1749 |
+
"entropy": 0.32763671875,
|
| 1750 |
+
"epoch": 0.07428571428571429,
|
| 1751 |
+
"grad_norm": 0.18390174210071564,
|
| 1752 |
+
"kl": 0.0012006759643554688,
|
| 1753 |
+
"learning_rate": 3.9609093550344907e-07,
|
| 1754 |
+
"loss": 0.0048,
|
| 1755 |
+
"reward": 0.6562500074505806,
|
| 1756 |
+
"reward_std": 0.3042946420609951,
|
| 1757 |
+
"rewards/accuracy_reward": 0.1875,
|
| 1758 |
+
"rewards/format_reward": 0.4687500074505806,
|
| 1759 |
+
"step": 65,
|
| 1760 |
+
"w_high_ratio": 0.0,
|
| 1761 |
+
"w_low_ratio": 0.03495390526950359,
|
| 1762 |
+
"w_max": 1.4882625937461853,
|
| 1763 |
+
"w_mean": 1.1612237989902496,
|
| 1764 |
+
"w_min": 0.0,
|
| 1765 |
+
"w_std": 0.14175301045179367
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"clip_ratio/high_max": 0.0,
|
| 1769 |
+
"clip_ratio/high_mean": 0.0,
|
| 1770 |
+
"clip_ratio/low_mean": 0.0,
|
| 1771 |
+
"clip_ratio/low_min": 0.0,
|
| 1772 |
+
"clip_ratio/region_mean": 0.0,
|
| 1773 |
+
"completion_length": 2299.2396240234375,
|
| 1774 |
+
"cov_mean": 2.2165444875099638e-05,
|
| 1775 |
+
"cov_std": 0.1351991407573223,
|
| 1776 |
+
"entropy": 0.306640625,
|
| 1777 |
+
"epoch": 0.07542857142857143,
|
| 1778 |
+
"grad_norm": 0.42320746183395386,
|
| 1779 |
+
"kl": 0.0011644363403320312,
|
| 1780 |
+
"learning_rate": 3.8142703296283953e-07,
|
| 1781 |
+
"loss": -0.0144,
|
| 1782 |
+
"reward": 0.8541666865348816,
|
| 1783 |
+
"reward_std": 0.1657295972108841,
|
| 1784 |
+
"rewards/accuracy_reward": 0.354166679084301,
|
| 1785 |
+
"rewards/format_reward": 0.5,
|
| 1786 |
+
"step": 66,
|
| 1787 |
+
"w_high_ratio": 0.0625,
|
| 1788 |
+
"w_low_ratio": 0.015947649022564292,
|
| 1789 |
+
"w_max": 1.6368336379528046,
|
| 1790 |
+
"w_mean": 1.2440811097621918,
|
| 1791 |
+
"w_min": 0.25,
|
| 1792 |
+
"w_std": 0.09564665146172047
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"clip_ratio/high_max": 0.0,
|
| 1796 |
+
"clip_ratio/high_mean": 0.0,
|
| 1797 |
+
"clip_ratio/low_mean": 0.0,
|
| 1798 |
+
"clip_ratio/low_min": 0.0,
|
| 1799 |
+
"clip_ratio/region_mean": 0.0,
|
| 1800 |
+
"completion_length": 3708.572998046875,
|
| 1801 |
+
"cov_mean": 9.353170185022464e-06,
|
| 1802 |
+
"cov_std": 0.17532559856772423,
|
| 1803 |
+
"entropy": 0.3408203125,
|
| 1804 |
+
"epoch": 0.07657142857142857,
|
| 1805 |
+
"grad_norm": 0.18465003371238708,
|
| 1806 |
+
"kl": 0.0010061264038085938,
|
| 1807 |
+
"learning_rate": 3.6696851061588994e-07,
|
| 1808 |
+
"loss": 0.0047,
|
| 1809 |
+
"reward": 0.20833333861082792,
|
| 1810 |
+
"reward_std": 0.1927690953016281,
|
| 1811 |
+
"rewards/accuracy_reward": 0.041666666977107525,
|
| 1812 |
+
"rewards/format_reward": 0.1666666716337204,
|
| 1813 |
+
"step": 67,
|
| 1814 |
+
"w_high_ratio": 0.0,
|
| 1815 |
+
"w_low_ratio": 0.021573648788034916,
|
| 1816 |
+
"w_max": 1.3159178793430328,
|
| 1817 |
+
"w_mean": 1.0860951095819473,
|
| 1818 |
+
"w_min": 0.25,
|
| 1819 |
+
"w_std": 0.11260665208101273
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"clip_ratio/high_max": 0.0,
|
| 1823 |
+
"clip_ratio/high_mean": 0.0,
|
| 1824 |
+
"clip_ratio/low_mean": 0.0,
|
| 1825 |
+
"clip_ratio/low_min": 0.0,
|
| 1826 |
+
"clip_ratio/region_mean": 0.0,
|
| 1827 |
+
"completion_length": 2515.729232788086,
|
| 1828 |
+
"cov_mean": 1.1593250746955164e-05,
|
| 1829 |
+
"cov_std": 0.31169629842042923,
|
| 1830 |
+
"entropy": 0.4072265625,
|
| 1831 |
+
"epoch": 0.07771428571428571,
|
| 1832 |
+
"grad_norm": 0.39554792642593384,
|
| 1833 |
+
"kl": 0.009487152099609375,
|
| 1834 |
+
"learning_rate": 3.5273298394491515e-07,
|
| 1835 |
+
"loss": -0.0017,
|
| 1836 |
+
"reward": 0.7812500298023224,
|
| 1837 |
+
"reward_std": 0.47302668541669846,
|
| 1838 |
+
"rewards/accuracy_reward": 0.2187500074505806,
|
| 1839 |
+
"rewards/format_reward": 0.5625000074505806,
|
| 1840 |
+
"step": 68,
|
| 1841 |
+
"w_high_ratio": 0.20789940655231476,
|
| 1842 |
+
"w_low_ratio": 0.03476168913766742,
|
| 1843 |
+
"w_max": 2.3950938880443573,
|
| 1844 |
+
"w_mean": 1.504720538854599,
|
| 1845 |
+
"w_min": 0.25,
|
| 1846 |
+
"w_std": 0.20618349313735962
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"clip_ratio/high_max": 0.0,
|
| 1850 |
+
"clip_ratio/high_mean": 0.0,
|
| 1851 |
+
"clip_ratio/low_mean": 0.0,
|
| 1852 |
+
"clip_ratio/low_min": 0.0,
|
| 1853 |
+
"clip_ratio/region_mean": 0.0,
|
| 1854 |
+
"completion_length": 2761.9896240234375,
|
| 1855 |
+
"cov_mean": -1.7451284293201752e-05,
|
| 1856 |
+
"cov_std": 0.2980487421154976,
|
| 1857 |
+
"entropy": 0.43212890625,
|
| 1858 |
+
"epoch": 0.07885714285714286,
|
| 1859 |
+
"grad_norm": 0.39313724637031555,
|
| 1860 |
+
"kl": 0.003513336181640625,
|
| 1861 |
+
"learning_rate": 3.387377967463493e-07,
|
| 1862 |
+
"loss": -0.0408,
|
| 1863 |
+
"reward": 0.4791666828095913,
|
| 1864 |
+
"reward_std": 0.3629096820950508,
|
| 1865 |
+
"rewards/accuracy_reward": 0.041666666977107525,
|
| 1866 |
+
"rewards/format_reward": 0.4375000260770321,
|
| 1867 |
+
"step": 69,
|
| 1868 |
+
"w_high_ratio": 0.06426303647458553,
|
| 1869 |
+
"w_low_ratio": 0.03391136787831783,
|
| 1870 |
+
"w_max": 1.8445489406585693,
|
| 1871 |
+
"w_mean": 1.2533070147037506,
|
| 1872 |
+
"w_min": 0.0,
|
| 1873 |
+
"w_std": 0.1965021900832653
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"clip_ratio/high_max": 0.0,
|
| 1877 |
+
"clip_ratio/high_mean": 0.0,
|
| 1878 |
+
"clip_ratio/low_mean": 0.0,
|
| 1879 |
+
"clip_ratio/low_min": 0.0,
|
| 1880 |
+
"clip_ratio/region_mean": 0.0,
|
| 1881 |
+
"completion_length": 3302.885498046875,
|
| 1882 |
+
"cov_mean": 2.7586031137616374e-05,
|
| 1883 |
+
"cov_std": 0.25518757849931717,
|
| 1884 |
+
"entropy": 0.35498046875,
|
| 1885 |
+
"epoch": 0.08,
|
| 1886 |
+
"grad_norm": 0.21536409854888916,
|
| 1887 |
+
"kl": 0.0017971992492675781,
|
| 1888 |
+
"learning_rate": 3.250000000000001e-07,
|
| 1889 |
+
"loss": 0.0226,
|
| 1890 |
+
"reward": 0.489583358168602,
|
| 1891 |
+
"reward_std": 0.3120904117822647,
|
| 1892 |
+
"rewards/accuracy_reward": 0.09375000279396772,
|
| 1893 |
+
"rewards/format_reward": 0.3958333432674408,
|
| 1894 |
+
"step": 70,
|
| 1895 |
+
"w_high_ratio": 0.0,
|
| 1896 |
+
"w_low_ratio": 0.035551197826862335,
|
| 1897 |
+
"w_max": 1.2505681812763214,
|
| 1898 |
+
"w_mean": 1.0716167390346527,
|
| 1899 |
+
"w_min": 0.25,
|
| 1900 |
+
"w_std": 0.13985217362642288
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"clip_ratio/high_max": 0.0,
|
| 1904 |
+
"clip_ratio/high_mean": 0.0,
|
| 1905 |
+
"clip_ratio/low_mean": 0.0,
|
| 1906 |
+
"clip_ratio/low_min": 0.0,
|
| 1907 |
+
"clip_ratio/region_mean": 0.0,
|
| 1908 |
+
"completion_length": 2877.8958740234375,
|
| 1909 |
+
"cov_mean": -1.1374457244528458e-05,
|
| 1910 |
+
"cov_std": 0.2578735649585724,
|
| 1911 |
+
"entropy": 0.40185546875,
|
| 1912 |
+
"epoch": 0.08114285714285714,
|
| 1913 |
+
"grad_norm": 0.5054724812507629,
|
| 1914 |
+
"kl": 0.0020956993103027344,
|
| 1915 |
+
"learning_rate": 3.115363310950578e-07,
|
| 1916 |
+
"loss": 0.0119,
|
| 1917 |
+
"reward": 0.6145833637565374,
|
| 1918 |
+
"reward_std": 0.24616123735904694,
|
| 1919 |
+
"rewards/accuracy_reward": 0.2291666716337204,
|
| 1920 |
+
"rewards/format_reward": 0.38541666977107525,
|
| 1921 |
+
"step": 71,
|
| 1922 |
+
"w_high_ratio": 0.11903085745871067,
|
| 1923 |
+
"w_low_ratio": 0.02461471315473318,
|
| 1924 |
+
"w_max": 1.9073027968406677,
|
| 1925 |
+
"w_mean": 1.2823512256145477,
|
| 1926 |
+
"w_min": 0.25,
|
| 1927 |
+
"w_std": 0.20722979307174683
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"clip_ratio/high_max": 0.0,
|
| 1931 |
+
"clip_ratio/high_mean": 0.0,
|
| 1932 |
+
"clip_ratio/low_mean": 0.0,
|
| 1933 |
+
"clip_ratio/low_min": 0.0,
|
| 1934 |
+
"clip_ratio/region_mean": 0.0,
|
| 1935 |
+
"completion_length": 3255.104248046875,
|
| 1936 |
+
"cov_mean": 2.442240383970784e-05,
|
| 1937 |
+
"cov_std": 0.49785537272691727,
|
| 1938 |
+
"entropy": 0.4853515625,
|
| 1939 |
+
"epoch": 0.08228571428571428,
|
| 1940 |
+
"grad_norm": 0.4614424407482147,
|
| 1941 |
+
"kl": 0.004761219024658203,
|
| 1942 |
+
"learning_rate": 2.9836319343816397e-07,
|
| 1943 |
+
"loss": 0.002,
|
| 1944 |
+
"reward": 0.4583333358168602,
|
| 1945 |
+
"reward_std": 0.5298330634832382,
|
| 1946 |
+
"rewards/accuracy_reward": 0.06250000279396772,
|
| 1947 |
+
"rewards/format_reward": 0.3958333432674408,
|
| 1948 |
+
"step": 72,
|
| 1949 |
+
"w_high_ratio": 0.01689874939620495,
|
| 1950 |
+
"w_low_ratio": 0.058014022186398506,
|
| 1951 |
+
"w_max": 1.7353839576244354,
|
| 1952 |
+
"w_mean": 1.2576677203178406,
|
| 1953 |
+
"w_min": 0.0,
|
| 1954 |
+
"w_std": 0.312557902187109
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"clip_ratio/high_max": 0.0,
|
| 1958 |
+
"clip_ratio/high_mean": 0.0,
|
| 1959 |
+
"clip_ratio/low_mean": 0.0,
|
| 1960 |
+
"clip_ratio/low_min": 0.0,
|
| 1961 |
+
"clip_ratio/region_mean": 0.0,
|
| 1962 |
+
"completion_length": 3811.3541870117188,
|
| 1963 |
+
"cov_mean": 9.231509466189891e-06,
|
| 1964 |
+
"cov_std": 0.21682696044445038,
|
| 1965 |
+
"entropy": 0.439453125,
|
| 1966 |
+
"epoch": 0.08342857142857144,
|
| 1967 |
+
"grad_norm": 0.11269888281822205,
|
| 1968 |
+
"kl": 0.0003600120544433594,
|
| 1969 |
+
"learning_rate": 2.854966364683872e-07,
|
| 1970 |
+
"loss": 0.0008,
|
| 1971 |
+
"reward": 0.3437500149011612,
|
| 1972 |
+
"reward_std": 0.34103409945964813,
|
| 1973 |
+
"rewards/accuracy_reward": 0.1354166716337204,
|
| 1974 |
+
"rewards/format_reward": 0.2083333432674408,
|
| 1975 |
+
"step": 73,
|
| 1976 |
+
"w_high_ratio": 0.0,
|
| 1977 |
+
"w_low_ratio": 0.025642000138759613,
|
| 1978 |
+
"w_max": 1.2365370392799377,
|
| 1979 |
+
"w_mean": 1.0363657176494598,
|
| 1980 |
+
"w_min": 0.25,
|
| 1981 |
+
"w_std": 0.10616981983184814
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"clip_ratio/high_max": 0.0,
|
| 1985 |
+
"clip_ratio/high_mean": 0.0,
|
| 1986 |
+
"clip_ratio/low_mean": 0.0,
|
| 1987 |
+
"clip_ratio/low_min": 0.0,
|
| 1988 |
+
"clip_ratio/region_mean": 0.0,
|
| 1989 |
+
"completion_length": 3564.8750610351562,
|
| 1990 |
+
"cov_mean": 7.751962675683899e-05,
|
| 1991 |
+
"cov_std": 0.2809390053153038,
|
| 1992 |
+
"entropy": 0.41064453125,
|
| 1993 |
+
"epoch": 0.08457142857142858,
|
| 1994 |
+
"grad_norm": 0.23569634556770325,
|
| 1995 |
+
"kl": 0.0016646385192871094,
|
| 1996 |
+
"learning_rate": 2.729523361034538e-07,
|
| 1997 |
+
"loss": 0.0156,
|
| 1998 |
+
"reward": 0.479166679084301,
|
| 1999 |
+
"reward_std": 0.4482262283563614,
|
| 2000 |
+
"rewards/accuracy_reward": 0.21875000186264515,
|
| 2001 |
+
"rewards/format_reward": 0.2604166716337204,
|
| 2002 |
+
"step": 74,
|
| 2003 |
+
"w_high_ratio": 0.0,
|
| 2004 |
+
"w_low_ratio": 0.04224945977330208,
|
| 2005 |
+
"w_max": 1.3714891374111176,
|
| 2006 |
+
"w_mean": 1.06686469912529,
|
| 2007 |
+
"w_min": 0.25,
|
| 2008 |
+
"w_std": 0.17370857670903206
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"clip_ratio/high_max": 0.0,
|
| 2012 |
+
"clip_ratio/high_mean": 0.0,
|
| 2013 |
+
"clip_ratio/low_mean": 0.0,
|
| 2014 |
+
"clip_ratio/low_min": 0.0,
|
| 2015 |
+
"clip_ratio/region_mean": 0.0,
|
| 2016 |
+
"completion_length": 3226.3126220703125,
|
| 2017 |
+
"cov_mean": 6.714556593578891e-05,
|
| 2018 |
+
"cov_std": 0.3720974437892437,
|
| 2019 |
+
"entropy": 0.36962890625,
|
| 2020 |
+
"epoch": 0.08571428571428572,
|
| 2021 |
+
"grad_norm": 0.22486244142055511,
|
| 2022 |
+
"kl": 0.0015187263488769531,
|
| 2023 |
+
"learning_rate": 2.6074557564105724e-07,
|
| 2024 |
+
"loss": 0.0805,
|
| 2025 |
+
"reward": 0.6041666939854622,
|
| 2026 |
+
"reward_std": 0.3942909240722656,
|
| 2027 |
+
"rewards/accuracy_reward": 0.1979166716337204,
|
| 2028 |
+
"rewards/format_reward": 0.40625001303851604,
|
| 2029 |
+
"step": 75,
|
| 2030 |
+
"w_high_ratio": 0.05757378041744232,
|
| 2031 |
+
"w_low_ratio": 0.045274180360138416,
|
| 2032 |
+
"w_max": 1.8611223101615906,
|
| 2033 |
+
"w_mean": 1.1981039345264435,
|
| 2034 |
+
"w_min": 0.0,
|
| 2035 |
+
"w_std": 0.2294696494936943
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"clip_ratio/high_max": 0.0,
|
| 2039 |
+
"clip_ratio/high_mean": 0.0,
|
| 2040 |
+
"clip_ratio/low_mean": 0.0,
|
| 2041 |
+
"clip_ratio/low_min": 0.0,
|
| 2042 |
+
"clip_ratio/region_mean": 0.0,
|
| 2043 |
+
"completion_length": 3215.9271850585938,
|
| 2044 |
+
"cov_mean": 1.099553310268675e-05,
|
| 2045 |
+
"cov_std": 0.2080927938222885,
|
| 2046 |
+
"entropy": 0.3857421875,
|
| 2047 |
+
"epoch": 0.08685714285714285,
|
| 2048 |
+
"grad_norm": 0.12048080563545227,
|
| 2049 |
+
"kl": 0.0004360675811767578,
|
| 2050 |
+
"learning_rate": 2.488912271385139e-07,
|
| 2051 |
+
"loss": -0.0073,
|
| 2052 |
+
"reward": 0.4791666865348816,
|
| 2053 |
+
"reward_std": 0.2375478893518448,
|
| 2054 |
+
"rewards/accuracy_reward": 0.0625,
|
| 2055 |
+
"rewards/format_reward": 0.4166666865348816,
|
| 2056 |
+
"step": 76,
|
| 2057 |
+
"w_high_ratio": 0.1249999962747097,
|
| 2058 |
+
"w_low_ratio": 0.027999846264719963,
|
| 2059 |
+
"w_max": 1.7428039014339447,
|
| 2060 |
+
"w_mean": 1.2136583924293518,
|
| 2061 |
+
"w_min": 0.25,
|
| 2062 |
+
"w_std": 0.10646160691976547
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"clip_ratio/high_max": 0.0,
|
| 2066 |
+
"clip_ratio/high_mean": 0.0,
|
| 2067 |
+
"clip_ratio/low_mean": 0.0,
|
| 2068 |
+
"clip_ratio/low_min": 0.0,
|
| 2069 |
+
"clip_ratio/region_mean": 0.0,
|
| 2070 |
+
"completion_length": 3392.4583740234375,
|
| 2071 |
+
"cov_mean": -4.415775742927508e-05,
|
| 2072 |
+
"cov_std": 0.3749671131372452,
|
| 2073 |
+
"entropy": 0.39599609375,
|
| 2074 |
+
"epoch": 0.088,
|
| 2075 |
+
"grad_norm": 0.2782253324985504,
|
| 2076 |
+
"kl": 0.0008301734924316406,
|
| 2077 |
+
"learning_rate": 2.374037332934512e-07,
|
| 2078 |
+
"loss": -0.016,
|
| 2079 |
+
"reward": 0.48958333395421505,
|
| 2080 |
+
"reward_std": 0.39483094960451126,
|
| 2081 |
+
"rewards/accuracy_reward": 0.1354166716337204,
|
| 2082 |
+
"rewards/format_reward": 0.35416667722165585,
|
| 2083 |
+
"step": 77,
|
| 2084 |
+
"w_high_ratio": 0.0396752767264843,
|
| 2085 |
+
"w_low_ratio": 0.03982406808063388,
|
| 2086 |
+
"w_max": 1.5143637657165527,
|
| 2087 |
+
"w_mean": 1.1384893357753754,
|
| 2088 |
+
"w_min": 0.0,
|
| 2089 |
+
"w_std": 0.20439331606030464
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"clip_ratio/high_max": 0.0,
|
| 2093 |
+
"clip_ratio/high_mean": 0.0,
|
| 2094 |
+
"clip_ratio/low_mean": 0.0,
|
| 2095 |
+
"clip_ratio/low_min": 0.0,
|
| 2096 |
+
"clip_ratio/region_mean": 0.0,
|
| 2097 |
+
"completion_length": 3471.21875,
|
| 2098 |
+
"cov_mean": -2.8236715252205613e-05,
|
| 2099 |
+
"cov_std": 0.4440799057483673,
|
| 2100 |
+
"entropy": 0.36767578125,
|
| 2101 |
+
"epoch": 0.08914285714285715,
|
| 2102 |
+
"grad_norm": 0.34334930777549744,
|
| 2103 |
+
"kl": 0.00042557716369628906,
|
| 2104 |
+
"learning_rate": 2.2629708984760706e-07,
|
| 2105 |
+
"loss": 0.0139,
|
| 2106 |
+
"reward": 0.6041666902601719,
|
| 2107 |
+
"reward_std": 0.5486822947859764,
|
| 2108 |
+
"rewards/accuracy_reward": 0.2708333358168602,
|
| 2109 |
+
"rewards/format_reward": 0.33333334513008595,
|
| 2110 |
+
"step": 78,
|
| 2111 |
+
"w_high_ratio": 0.05815339833498001,
|
| 2112 |
+
"w_low_ratio": 0.04555406514555216,
|
| 2113 |
+
"w_max": 1.7224481999874115,
|
| 2114 |
+
"w_mean": 1.199266105890274,
|
| 2115 |
+
"w_min": 0.0,
|
| 2116 |
+
"w_std": 0.2618200555443764
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"clip_ratio/high_max": 0.0,
|
| 2120 |
+
"clip_ratio/high_mean": 0.0,
|
| 2121 |
+
"clip_ratio/low_mean": 0.0,
|
| 2122 |
+
"clip_ratio/low_min": 0.0,
|
| 2123 |
+
"clip_ratio/region_mean": 0.0,
|
| 2124 |
+
"completion_length": 2512.6146697998047,
|
| 2125 |
+
"cov_mean": 4.2871850382653065e-05,
|
| 2126 |
+
"cov_std": 0.2797487545758486,
|
| 2127 |
+
"entropy": 0.332275390625,
|
| 2128 |
+
"epoch": 0.09028571428571429,
|
| 2129 |
+
"grad_norm": 0.3756020963191986,
|
| 2130 |
+
"kl": 0.0013785362243652344,
|
| 2131 |
+
"learning_rate": 2.1558482853517253e-07,
|
| 2132 |
+
"loss": 0.0565,
|
| 2133 |
+
"reward": 0.7604166865348816,
|
| 2134 |
+
"reward_std": 0.3231862857937813,
|
| 2135 |
+
"rewards/accuracy_reward": 0.20833333861082792,
|
| 2136 |
+
"rewards/format_reward": 0.5520833432674408,
|
| 2137 |
+
"step": 79,
|
| 2138 |
+
"w_high_ratio": 0.05819739028811455,
|
| 2139 |
+
"w_low_ratio": 0.030684792436659336,
|
| 2140 |
+
"w_max": 2.2010356187820435,
|
| 2141 |
+
"w_mean": 1.3371951580047607,
|
| 2142 |
+
"w_min": 0.0,
|
| 2143 |
+
"w_std": 0.20669112168252468
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"clip_ratio/high_max": 0.0,
|
| 2147 |
+
"clip_ratio/high_mean": 0.0,
|
| 2148 |
+
"clip_ratio/low_mean": 0.0,
|
| 2149 |
+
"clip_ratio/low_min": 0.0,
|
| 2150 |
+
"clip_ratio/region_mean": 0.0,
|
| 2151 |
+
"completion_length": 3588.604248046875,
|
| 2152 |
+
"cov_mean": -7.164392627601046e-06,
|
| 2153 |
+
"cov_std": 0.2609938234090805,
|
| 2154 |
+
"entropy": 0.46240234375,
|
| 2155 |
+
"epoch": 0.09142857142857143,
|
| 2156 |
+
"grad_norm": 0.20949065685272217,
|
| 2157 |
+
"kl": 0.0011081695556640625,
|
| 2158 |
+
"learning_rate": 2.0528000059645995e-07,
|
| 2159 |
+
"loss": 0.0063,
|
| 2160 |
+
"reward": 0.500000013038516,
|
| 2161 |
+
"reward_std": 0.4404493272304535,
|
| 2162 |
+
"rewards/accuracy_reward": 0.17708333488553762,
|
| 2163 |
+
"rewards/format_reward": 0.3229166781529784,
|
| 2164 |
+
"step": 80,
|
| 2165 |
+
"w_high_ratio": 0.0,
|
| 2166 |
+
"w_low_ratio": 0.03598626330494881,
|
| 2167 |
+
"w_max": 1.3815899193286896,
|
| 2168 |
+
"w_mean": 1.1029056310653687,
|
| 2169 |
+
"w_min": 0.25,
|
| 2170 |
+
"w_std": 0.14113666117191315
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"clip_ratio/high_max": 0.0,
|
| 2174 |
+
"clip_ratio/high_mean": 0.0,
|
| 2175 |
+
"clip_ratio/low_mean": 0.0,
|
| 2176 |
+
"clip_ratio/low_min": 0.0,
|
| 2177 |
+
"clip_ratio/region_mean": 0.0,
|
| 2178 |
+
"completion_length": 3374.7188110351562,
|
| 2179 |
+
"cov_mean": 3.345898390705315e-05,
|
| 2180 |
+
"cov_std": 0.17040981724858284,
|
| 2181 |
+
"entropy": 0.51806640625,
|
| 2182 |
+
"epoch": 0.09257142857142857,
|
| 2183 |
+
"grad_norm": 0.16646912693977356,
|
| 2184 |
+
"kl": 0.0033516883850097656,
|
| 2185 |
+
"learning_rate": 1.9539516087697517e-07,
|
| 2186 |
+
"loss": -0.0106,
|
| 2187 |
+
"reward": 0.3750000102445483,
|
| 2188 |
+
"reward_std": 0.21947956830263138,
|
| 2189 |
+
"rewards/accuracy_reward": 0.09375000558793545,
|
| 2190 |
+
"rewards/format_reward": 0.2812500027939677,
|
| 2191 |
+
"step": 81,
|
| 2192 |
+
"w_high_ratio": 0.0625,
|
| 2193 |
+
"w_low_ratio": 0.021185603458434343,
|
| 2194 |
+
"w_max": 1.6720820665359497,
|
| 2195 |
+
"w_mean": 1.217385858297348,
|
| 2196 |
+
"w_min": 0.25,
|
| 2197 |
+
"w_std": 0.0957618746906519
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"clip_ratio/high_max": 0.0,
|
| 2201 |
+
"clip_ratio/high_mean": 0.0,
|
| 2202 |
+
"clip_ratio/low_mean": 0.0,
|
| 2203 |
+
"clip_ratio/low_min": 0.0,
|
| 2204 |
+
"clip_ratio/region_mean": 0.0,
|
| 2205 |
+
"completion_length": 3105.322998046875,
|
| 2206 |
+
"cov_mean": -3.050990926567465e-05,
|
| 2207 |
+
"cov_std": 0.23843222856521606,
|
| 2208 |
+
"entropy": 0.42138671875,
|
| 2209 |
+
"epoch": 0.09371428571428571,
|
| 2210 |
+
"grad_norm": 0.34804272651672363,
|
| 2211 |
+
"kl": 0.0018186569213867188,
|
| 2212 |
+
"learning_rate": 1.8594235253127372e-07,
|
| 2213 |
+
"loss": -0.0113,
|
| 2214 |
+
"reward": 0.6458333507180214,
|
| 2215 |
+
"reward_std": 0.31302163377404213,
|
| 2216 |
+
"rewards/accuracy_reward": 0.2500000037252903,
|
| 2217 |
+
"rewards/format_reward": 0.3958333395421505,
|
| 2218 |
+
"step": 82,
|
| 2219 |
+
"w_high_ratio": 0.0,
|
| 2220 |
+
"w_low_ratio": 0.02489804197102785,
|
| 2221 |
+
"w_max": 1.6160834431648254,
|
| 2222 |
+
"w_mean": 1.1803554594516754,
|
| 2223 |
+
"w_min": 0.0,
|
| 2224 |
+
"w_std": 0.15118960477411747
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"clip_ratio/high_max": 0.0,
|
| 2228 |
+
"clip_ratio/high_mean": 0.0,
|
| 2229 |
+
"clip_ratio/low_mean": 0.0,
|
| 2230 |
+
"clip_ratio/low_min": 0.0,
|
| 2231 |
+
"clip_ratio/region_mean": 0.0,
|
| 2232 |
+
"completion_length": 3147.4583740234375,
|
| 2233 |
+
"cov_mean": -2.670003959792666e-05,
|
| 2234 |
+
"cov_std": 0.2024751603603363,
|
| 2235 |
+
"entropy": 0.43408203125,
|
| 2236 |
+
"epoch": 0.09485714285714286,
|
| 2237 |
+
"grad_norm": 0.24946807324886322,
|
| 2238 |
+
"kl": 0.0009236335754394531,
|
| 2239 |
+
"learning_rate": 1.7693309235023127e-07,
|
| 2240 |
+
"loss": 0.0343,
|
| 2241 |
+
"reward": 0.3854166679084301,
|
| 2242 |
+
"reward_std": 0.24196770787239075,
|
| 2243 |
+
"rewards/accuracy_reward": 0.10416666977107525,
|
| 2244 |
+
"rewards/format_reward": 0.28125000558793545,
|
| 2245 |
+
"step": 83,
|
| 2246 |
+
"w_high_ratio": 0.0,
|
| 2247 |
+
"w_low_ratio": 0.030171910300850868,
|
| 2248 |
+
"w_max": 1.2905269861221313,
|
| 2249 |
+
"w_mean": 1.0636517405509949,
|
| 2250 |
+
"w_min": 0.25,
|
| 2251 |
+
"w_std": 0.1290601659566164
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"clip_ratio/high_max": 0.0,
|
| 2255 |
+
"clip_ratio/high_mean": 0.0,
|
| 2256 |
+
"clip_ratio/low_mean": 0.0,
|
| 2257 |
+
"clip_ratio/low_min": 0.0,
|
| 2258 |
+
"clip_ratio/region_mean": 0.0,
|
| 2259 |
+
"completion_length": 3317.5625,
|
| 2260 |
+
"cov_mean": 4.104694835405098e-05,
|
| 2261 |
+
"cov_std": 0.4057937487959862,
|
| 2262 |
+
"entropy": 0.43212890625,
|
| 2263 |
+
"epoch": 0.096,
|
| 2264 |
+
"grad_norm": 0.3507938086986542,
|
| 2265 |
+
"kl": 0.0007605552673339844,
|
| 2266 |
+
"learning_rate": 1.6837835672960831e-07,
|
| 2267 |
+
"loss": 0.0168,
|
| 2268 |
+
"reward": 0.6666666967794299,
|
| 2269 |
+
"reward_std": 0.5004496201872826,
|
| 2270 |
+
"rewards/accuracy_reward": 0.2708333432674408,
|
| 2271 |
+
"rewards/format_reward": 0.3958333386108279,
|
| 2272 |
+
"step": 84,
|
| 2273 |
+
"w_high_ratio": 0.0,
|
| 2274 |
+
"w_low_ratio": 0.05950552970170975,
|
| 2275 |
+
"w_max": 1.3848203122615814,
|
| 2276 |
+
"w_mean": 1.1114588379859924,
|
| 2277 |
+
"w_min": 0.0,
|
| 2278 |
+
"w_std": 0.2529653161764145
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"clip_ratio/high_max": 0.0,
|
| 2282 |
+
"clip_ratio/high_mean": 0.0,
|
| 2283 |
+
"clip_ratio/low_mean": 0.0,
|
| 2284 |
+
"clip_ratio/low_min": 0.0,
|
| 2285 |
+
"clip_ratio/region_mean": 0.0,
|
| 2286 |
+
"completion_length": 3263.2188110351562,
|
| 2287 |
+
"cov_mean": -2.9235156034701504e-05,
|
| 2288 |
+
"cov_std": 0.2883659228682518,
|
| 2289 |
+
"entropy": 0.3173828125,
|
| 2290 |
+
"epoch": 0.09714285714285714,
|
| 2291 |
+
"grad_norm": 0.31331735849380493,
|
| 2292 |
+
"kl": 0.0004665851593017578,
|
| 2293 |
+
"learning_rate": 1.6028856829700258e-07,
|
| 2294 |
+
"loss": 0.0412,
|
| 2295 |
+
"reward": 0.6145833656191826,
|
| 2296 |
+
"reward_std": 0.5151766017079353,
|
| 2297 |
+
"rewards/accuracy_reward": 0.18750000651925802,
|
| 2298 |
+
"rewards/format_reward": 0.4270833469927311,
|
| 2299 |
+
"step": 85,
|
| 2300 |
+
"w_high_ratio": 0.041407618671655655,
|
| 2301 |
+
"w_low_ratio": 0.04611168662086129,
|
| 2302 |
+
"w_max": 1.504797637462616,
|
| 2303 |
+
"w_mean": 1.16913503408432,
|
| 2304 |
+
"w_min": 0.0,
|
| 2305 |
+
"w_std": 0.2160215601325035
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"clip_ratio/high_max": 0.0,
|
| 2309 |
+
"clip_ratio/high_mean": 0.0,
|
| 2310 |
+
"clip_ratio/low_mean": 0.0,
|
| 2311 |
+
"clip_ratio/low_min": 0.0,
|
| 2312 |
+
"clip_ratio/region_mean": 0.0,
|
| 2313 |
+
"completion_length": 3107.416748046875,
|
| 2314 |
+
"cov_mean": 2.2301139551927918e-05,
|
| 2315 |
+
"cov_std": 0.29477884620428085,
|
| 2316 |
+
"entropy": 0.4326171875,
|
| 2317 |
+
"epoch": 0.09828571428571428,
|
| 2318 |
+
"grad_norm": 0.44227737188339233,
|
| 2319 |
+
"kl": 0.0021905899047851562,
|
| 2320 |
+
"learning_rate": 1.5267358321348285e-07,
|
| 2321 |
+
"loss": -0.017,
|
| 2322 |
+
"reward": 0.5312500074505806,
|
| 2323 |
+
"reward_std": 0.2660149596631527,
|
| 2324 |
+
"rewards/accuracy_reward": 0.12500000093132257,
|
| 2325 |
+
"rewards/format_reward": 0.40625,
|
| 2326 |
+
"step": 86,
|
| 2327 |
+
"w_high_ratio": 0.1639367900788784,
|
| 2328 |
+
"w_low_ratio": 0.026534121483564377,
|
| 2329 |
+
"w_max": 2.0416696071624756,
|
| 2330 |
+
"w_mean": 1.30596524477005,
|
| 2331 |
+
"w_min": 0.25,
|
| 2332 |
+
"w_std": 0.21408291161060333
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"clip_ratio/high_max": 0.0,
|
| 2336 |
+
"clip_ratio/high_mean": 0.0,
|
| 2337 |
+
"clip_ratio/low_mean": 0.0,
|
| 2338 |
+
"clip_ratio/low_min": 0.0,
|
| 2339 |
+
"clip_ratio/region_mean": 0.0,
|
| 2340 |
+
"completion_length": 3028.3438110351562,
|
| 2341 |
+
"cov_mean": -7.177443694672547e-05,
|
| 2342 |
+
"cov_std": 0.4208982028067112,
|
| 2343 |
+
"entropy": 0.49169921875,
|
| 2344 |
+
"epoch": 0.09942857142857142,
|
| 2345 |
+
"grad_norm": 0.28011003136634827,
|
| 2346 |
+
"kl": 0.0020780563354492188,
|
| 2347 |
+
"learning_rate": 1.4554267916537495e-07,
|
| 2348 |
+
"loss": -0.0235,
|
| 2349 |
+
"reward": 0.7083333432674408,
|
| 2350 |
+
"reward_std": 0.48819613456726074,
|
| 2351 |
+
"rewards/accuracy_reward": 0.2083333358168602,
|
| 2352 |
+
"rewards/format_reward": 0.5000000149011612,
|
| 2353 |
+
"step": 87,
|
| 2354 |
+
"w_high_ratio": 0.04094649478793144,
|
| 2355 |
+
"w_low_ratio": 0.05161830177530646,
|
| 2356 |
+
"w_max": 1.7180204689502716,
|
| 2357 |
+
"w_mean": 1.2395348846912384,
|
| 2358 |
+
"w_min": 1.7816459000041805e-41,
|
| 2359 |
+
"w_std": 0.2497432641685009
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"clip_ratio/high_max": 0.0,
|
| 2363 |
+
"clip_ratio/high_mean": 0.0,
|
| 2364 |
+
"clip_ratio/low_mean": 0.0,
|
| 2365 |
+
"clip_ratio/low_min": 0.0,
|
| 2366 |
+
"clip_ratio/region_mean": 0.0,
|
| 2367 |
+
"completion_length": 2961.7293090820312,
|
| 2368 |
+
"cov_mean": 0.00011648364034044789,
|
| 2369 |
+
"cov_std": 0.5657102465629578,
|
| 2370 |
+
"entropy": 0.44140625,
|
| 2371 |
+
"epoch": 0.10057142857142858,
|
| 2372 |
+
"grad_norm": 0.5410559177398682,
|
| 2373 |
+
"kl": 0.0027256011962890625,
|
| 2374 |
+
"learning_rate": 1.3890454406082956e-07,
|
| 2375 |
+
"loss": -0.0379,
|
| 2376 |
+
"reward": 0.8437500298023224,
|
| 2377 |
+
"reward_std": 0.6055800318717957,
|
| 2378 |
+
"rewards/accuracy_reward": 0.2916666753590107,
|
| 2379 |
+
"rewards/format_reward": 0.5520833507180214,
|
| 2380 |
+
"step": 88,
|
| 2381 |
+
"w_high_ratio": 0.15312667191028595,
|
| 2382 |
+
"w_low_ratio": 0.052226919680833817,
|
| 2383 |
+
"w_max": 2.147995412349701,
|
| 2384 |
+
"w_mean": 1.4601677060127258,
|
| 2385 |
+
"w_min": 1.5606149987442183e-34,
|
| 2386 |
+
"w_std": 0.37410108000040054
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"clip_ratio/high_max": 0.0,
|
| 2390 |
+
"clip_ratio/high_mean": 0.0,
|
| 2391 |
+
"clip_ratio/low_mean": 0.0,
|
| 2392 |
+
"clip_ratio/low_min": 0.0,
|
| 2393 |
+
"clip_ratio/region_mean": 0.0,
|
| 2394 |
+
"completion_length": 3738.166748046875,
|
| 2395 |
+
"cov_mean": -3.068390833504964e-05,
|
| 2396 |
+
"cov_std": 0.2844373546540737,
|
| 2397 |
+
"entropy": 0.41650390625,
|
| 2398 |
+
"epoch": 0.10171428571428572,
|
| 2399 |
+
"grad_norm": 0.2505849003791809,
|
| 2400 |
+
"kl": 0.001811981201171875,
|
| 2401 |
+
"learning_rate": 1.3276726544494571e-07,
|
| 2402 |
+
"loss": -0.0108,
|
| 2403 |
+
"reward": 0.354166679084301,
|
| 2404 |
+
"reward_std": 0.48622245341539383,
|
| 2405 |
+
"rewards/accuracy_reward": 0.1250000037252903,
|
| 2406 |
+
"rewards/format_reward": 0.229166679084301,
|
| 2407 |
+
"step": 89,
|
| 2408 |
+
"w_high_ratio": 0.0,
|
| 2409 |
+
"w_low_ratio": 0.038497451692819595,
|
| 2410 |
+
"w_max": 1.4062875509262085,
|
| 2411 |
+
"w_mean": 1.0552209317684174,
|
| 2412 |
+
"w_min": 3.5878705750726184e-40,
|
| 2413 |
+
"w_std": 0.16452663764357567
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"clip_ratio/high_max": 0.0,
|
| 2417 |
+
"clip_ratio/high_mean": 0.0,
|
| 2418 |
+
"clip_ratio/low_mean": 0.0,
|
| 2419 |
+
"clip_ratio/low_min": 0.0,
|
| 2420 |
+
"clip_ratio/region_mean": 0.0,
|
| 2421 |
+
"completion_length": 2751.8541870117188,
|
| 2422 |
+
"cov_mean": 7.335069381042558e-05,
|
| 2423 |
+
"cov_std": 0.3152489699423313,
|
| 2424 |
+
"entropy": 0.5263671875,
|
| 2425 |
+
"epoch": 0.10285714285714286,
|
| 2426 |
+
"grad_norm": 0.4594099223613739,
|
| 2427 |
+
"kl": 0.0036954879760742188,
|
| 2428 |
+
"learning_rate": 1.2713832064634125e-07,
|
| 2429 |
+
"loss": 0.0116,
|
| 2430 |
+
"reward": 0.593750013038516,
|
| 2431 |
+
"reward_std": 0.30915412306785583,
|
| 2432 |
+
"rewards/accuracy_reward": 0.06250000279396772,
|
| 2433 |
+
"rewards/format_reward": 0.5312500176951289,
|
| 2434 |
+
"step": 90,
|
| 2435 |
+
"w_high_ratio": 0.22096561267971992,
|
| 2436 |
+
"w_low_ratio": 0.028408138314262033,
|
| 2437 |
+
"w_max": 2.2681187987327576,
|
| 2438 |
+
"w_mean": 1.453046053647995,
|
| 2439 |
+
"w_min": 0.0,
|
| 2440 |
+
"w_std": 0.17159932479262352
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"clip_ratio/high_max": 0.0,
|
| 2444 |
+
"clip_ratio/high_mean": 0.0,
|
| 2445 |
+
"clip_ratio/low_mean": 0.0,
|
| 2446 |
+
"clip_ratio/low_min": 0.0,
|
| 2447 |
+
"clip_ratio/region_mean": 0.0,
|
| 2448 |
+
"completion_length": 3386.9584350585938,
|
| 2449 |
+
"cov_mean": 5.254377992969239e-05,
|
| 2450 |
+
"cov_std": 0.34420739486813545,
|
| 2451 |
+
"entropy": 0.43603515625,
|
| 2452 |
+
"epoch": 0.104,
|
| 2453 |
+
"grad_norm": 0.40760353207588196,
|
| 2454 |
+
"kl": 0.0015692710876464844,
|
| 2455 |
+
"learning_rate": 1.220245676671809e-07,
|
| 2456 |
+
"loss": -0.0737,
|
| 2457 |
+
"reward": 0.5000000204890966,
|
| 2458 |
+
"reward_std": 0.41079987585544586,
|
| 2459 |
+
"rewards/accuracy_reward": 0.15625000279396772,
|
| 2460 |
+
"rewards/format_reward": 0.3437500102445483,
|
| 2461 |
+
"step": 91,
|
| 2462 |
+
"w_high_ratio": 0.044647760689258575,
|
| 2463 |
+
"w_low_ratio": 0.03610147628933191,
|
| 2464 |
+
"w_max": 1.6350789666175842,
|
| 2465 |
+
"w_mean": 1.218572586774826,
|
| 2466 |
+
"w_min": 0.25,
|
| 2467 |
+
"w_std": 0.22148512117564678
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"clip_ratio/high_max": 0.0,
|
| 2471 |
+
"clip_ratio/high_mean": 0.0,
|
| 2472 |
+
"clip_ratio/low_mean": 0.0,
|
| 2473 |
+
"clip_ratio/low_min": 0.0,
|
| 2474 |
+
"clip_ratio/region_mean": 0.0,
|
| 2475 |
+
"completion_length": 2950.3333740234375,
|
| 2476 |
+
"cov_mean": 4.718162654171465e-05,
|
| 2477 |
+
"cov_std": 0.552252508699894,
|
| 2478 |
+
"entropy": 0.369140625,
|
| 2479 |
+
"epoch": 0.10514285714285715,
|
| 2480 |
+
"grad_norm": 0.6825612783432007,
|
| 2481 |
+
"kl": 0.0022275447845458984,
|
| 2482 |
+
"learning_rate": 1.1743223682775649e-07,
|
| 2483 |
+
"loss": 0.0233,
|
| 2484 |
+
"reward": 0.6354167014360428,
|
| 2485 |
+
"reward_std": 0.5102165751159191,
|
| 2486 |
+
"rewards/accuracy_reward": 0.14583333861082792,
|
| 2487 |
+
"rewards/format_reward": 0.4895833432674408,
|
| 2488 |
+
"step": 92,
|
| 2489 |
+
"w_high_ratio": 0.26809635758399963,
|
| 2490 |
+
"w_low_ratio": 0.05027919542044401,
|
| 2491 |
+
"w_max": 2.5253304839134216,
|
| 2492 |
+
"w_mean": 1.5156863629817963,
|
| 2493 |
+
"w_min": 0.0,
|
| 2494 |
+
"w_std": 0.3864224813878536
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"clip_ratio/high_max": 0.0,
|
| 2498 |
+
"clip_ratio/high_mean": 0.0,
|
| 2499 |
+
"clip_ratio/low_mean": 0.0,
|
| 2500 |
+
"clip_ratio/low_min": 0.0,
|
| 2501 |
+
"clip_ratio/region_mean": 0.0,
|
| 2502 |
+
"completion_length": 3803.7500610351562,
|
| 2503 |
+
"cov_mean": 3.139047112199478e-05,
|
| 2504 |
+
"cov_std": 0.20749856531620026,
|
| 2505 |
+
"entropy": 0.52001953125,
|
| 2506 |
+
"epoch": 0.10628571428571429,
|
| 2507 |
+
"grad_norm": 0.24012291431427002,
|
| 2508 |
+
"kl": 0.0020914077758789062,
|
| 2509 |
+
"learning_rate": 1.1336692317580158e-07,
|
| 2510 |
+
"loss": 0.0268,
|
| 2511 |
+
"reward": 0.0937500037252903,
|
| 2512 |
+
"reward_std": 0.19024791568517685,
|
| 2513 |
+
"rewards/accuracy_reward": 0.010416666977107525,
|
| 2514 |
+
"rewards/format_reward": 0.0833333358168602,
|
| 2515 |
+
"step": 93,
|
| 2516 |
+
"w_high_ratio": 0.0,
|
| 2517 |
+
"w_low_ratio": 0.02966291829943657,
|
| 2518 |
+
"w_max": 1.2947639226913452,
|
| 2519 |
+
"w_mean": 1.0370966792106628,
|
| 2520 |
+
"w_min": 0.5,
|
| 2521 |
+
"w_std": 0.11119646579027176
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"clip_ratio/high_max": 0.0,
|
| 2525 |
+
"clip_ratio/high_mean": 0.0,
|
| 2526 |
+
"clip_ratio/low_mean": 0.0,
|
| 2527 |
+
"clip_ratio/low_min": 0.0,
|
| 2528 |
+
"clip_ratio/region_mean": 0.0,
|
| 2529 |
+
"completion_length": 3297.8646240234375,
|
| 2530 |
+
"cov_mean": 3.7244943996483926e-05,
|
| 2531 |
+
"cov_std": 0.31120600923895836,
|
| 2532 |
+
"entropy": 0.505859375,
|
| 2533 |
+
"epoch": 0.10742857142857143,
|
| 2534 |
+
"grad_norm": 0.19109465181827545,
|
| 2535 |
+
"kl": 0.002716541290283203,
|
| 2536 |
+
"learning_rate": 1.0983357966978745e-07,
|
| 2537 |
+
"loss": 0.0047,
|
| 2538 |
+
"reward": 0.5312500251457095,
|
| 2539 |
+
"reward_std": 0.33500421047210693,
|
| 2540 |
+
"rewards/accuracy_reward": 0.18750000186264515,
|
| 2541 |
+
"rewards/format_reward": 0.3437500102445483,
|
| 2542 |
+
"step": 94,
|
| 2543 |
+
"w_high_ratio": 0.0031294284854084253,
|
| 2544 |
+
"w_low_ratio": 0.0420422088354826,
|
| 2545 |
+
"w_max": 1.7046481668949127,
|
| 2546 |
+
"w_mean": 1.1779425740242004,
|
| 2547 |
+
"w_min": 0.0,
|
| 2548 |
+
"w_std": 0.18928811699151993
|
| 2549 |
+
},
|
| 2550 |
+
{
|
| 2551 |
+
"clip_ratio/high_max": 0.0,
|
| 2552 |
+
"clip_ratio/high_mean": 0.0,
|
| 2553 |
+
"clip_ratio/low_mean": 0.0,
|
| 2554 |
+
"clip_ratio/low_min": 0.0,
|
| 2555 |
+
"clip_ratio/region_mean": 0.0,
|
| 2556 |
+
"completion_length": 3739.1875610351562,
|
| 2557 |
+
"cov_mean": -3.053597947655362e-05,
|
| 2558 |
+
"cov_std": 0.34536129236221313,
|
| 2559 |
+
"entropy": 0.39794921875,
|
| 2560 |
+
"epoch": 0.10857142857142857,
|
| 2561 |
+
"grad_norm": 0.1897294968366623,
|
| 2562 |
+
"kl": 0.00032019615173339844,
|
| 2563 |
+
"learning_rate": 1.068365111445064e-07,
|
| 2564 |
+
"loss": 0.0256,
|
| 2565 |
+
"reward": 0.3125,
|
| 2566 |
+
"reward_std": 0.3973517417907715,
|
| 2567 |
+
"rewards/accuracy_reward": 0.09375,
|
| 2568 |
+
"rewards/format_reward": 0.21875,
|
| 2569 |
+
"step": 95,
|
| 2570 |
+
"w_high_ratio": 0.0,
|
| 2571 |
+
"w_low_ratio": 0.03817740920931101,
|
| 2572 |
+
"w_max": 1.3027753233909607,
|
| 2573 |
+
"w_mean": 1.0644334852695465,
|
| 2574 |
+
"w_min": 0.25,
|
| 2575 |
+
"w_std": 0.16921117901802063
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"clip_ratio/high_max": 0.0,
|
| 2579 |
+
"clip_ratio/high_mean": 0.0,
|
| 2580 |
+
"clip_ratio/low_mean": 0.0,
|
| 2581 |
+
"clip_ratio/low_min": 0.0,
|
| 2582 |
+
"clip_ratio/region_mean": 0.0,
|
| 2583 |
+
"completion_length": 3006.354248046875,
|
| 2584 |
+
"cov_mean": 1.1065521448472282e-05,
|
| 2585 |
+
"cov_std": 0.36543623358011246,
|
| 2586 |
+
"entropy": 0.36279296875,
|
| 2587 |
+
"epoch": 0.10971428571428571,
|
| 2588 |
+
"grad_norm": 0.516417384147644,
|
| 2589 |
+
"kl": 0.002213001251220703,
|
| 2590 |
+
"learning_rate": 1.0437936906629334e-07,
|
| 2591 |
+
"loss": 0.0172,
|
| 2592 |
+
"reward": 0.7395833507180214,
|
| 2593 |
+
"reward_std": 0.40412483364343643,
|
| 2594 |
+
"rewards/accuracy_reward": 0.2604166716337204,
|
| 2595 |
+
"rewards/format_reward": 0.479166679084301,
|
| 2596 |
+
"step": 96,
|
| 2597 |
+
"w_high_ratio": 0.04663441330194473,
|
| 2598 |
+
"w_low_ratio": 0.037586700869724154,
|
| 2599 |
+
"w_max": 1.9617899954319,
|
| 2600 |
+
"w_mean": 1.2268487811088562,
|
| 2601 |
+
"w_min": 3.964865404176085e-40,
|
| 2602 |
+
"w_std": 0.2116919718682766
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"clip_ratio/high_max": 0.0,
|
| 2606 |
+
"clip_ratio/high_mean": 0.0,
|
| 2607 |
+
"clip_ratio/low_mean": 0.0,
|
| 2608 |
+
"clip_ratio/low_min": 0.0,
|
| 2609 |
+
"clip_ratio/region_mean": 0.0,
|
| 2610 |
+
"completion_length": 3466.0209350585938,
|
| 2611 |
+
"cov_mean": -1.2588386653078487e-05,
|
| 2612 |
+
"cov_std": 0.4068885110318661,
|
| 2613 |
+
"entropy": 0.4091796875,
|
| 2614 |
+
"epoch": 0.11085714285714286,
|
| 2615 |
+
"grad_norm": 0.28435367345809937,
|
| 2616 |
+
"kl": 0.0007953643798828125,
|
| 2617 |
+
"learning_rate": 1.0246514708427701e-07,
|
| 2618 |
+
"loss": -0.0236,
|
| 2619 |
+
"reward": 0.5208333563059568,
|
| 2620 |
+
"reward_std": 0.5388510599732399,
|
| 2621 |
+
"rewards/accuracy_reward": 0.19791667722165585,
|
| 2622 |
+
"rewards/format_reward": 0.3229166818782687,
|
| 2623 |
+
"step": 97,
|
| 2624 |
+
"w_high_ratio": 0.0,
|
| 2625 |
+
"w_low_ratio": 0.04683105740696192,
|
| 2626 |
+
"w_max": 1.4916425049304962,
|
| 2627 |
+
"w_mean": 1.1599352061748505,
|
| 2628 |
+
"w_min": 0.0,
|
| 2629 |
+
"w_std": 0.24916893057525158
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"clip_ratio/high_max": 0.0,
|
| 2633 |
+
"clip_ratio/high_mean": 0.0,
|
| 2634 |
+
"clip_ratio/low_mean": 0.0,
|
| 2635 |
+
"clip_ratio/low_min": 0.0,
|
| 2636 |
+
"clip_ratio/region_mean": 0.0,
|
| 2637 |
+
"completion_length": 3058.0209350585938,
|
| 2638 |
+
"cov_mean": 0.00017139434021373745,
|
| 2639 |
+
"cov_std": 0.42342133820056915,
|
| 2640 |
+
"entropy": 0.37744140625,
|
| 2641 |
+
"epoch": 0.112,
|
| 2642 |
+
"grad_norm": 0.39048805832862854,
|
| 2643 |
+
"kl": 0.0005214214324951172,
|
| 2644 |
+
"learning_rate": 1.0109617738307911e-07,
|
| 2645 |
+
"loss": -0.044,
|
| 2646 |
+
"reward": 0.645833358168602,
|
| 2647 |
+
"reward_std": 0.4053402505815029,
|
| 2648 |
+
"rewards/accuracy_reward": 0.1770833432674408,
|
| 2649 |
+
"rewards/format_reward": 0.4687500149011612,
|
| 2650 |
+
"step": 98,
|
| 2651 |
+
"w_high_ratio": 0.04422338306903839,
|
| 2652 |
+
"w_low_ratio": 0.044881118927150965,
|
| 2653 |
+
"w_max": 1.805379331111908,
|
| 2654 |
+
"w_mean": 1.279258906841278,
|
| 2655 |
+
"w_min": 9.581378249820937e-43,
|
| 2656 |
+
"w_std": 0.26856033504009247
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"clip_ratio/high_max": 0.0,
|
| 2660 |
+
"clip_ratio/high_mean": 0.0,
|
| 2661 |
+
"clip_ratio/low_mean": 0.0,
|
| 2662 |
+
"clip_ratio/low_min": 0.0,
|
| 2663 |
+
"clip_ratio/region_mean": 0.0,
|
| 2664 |
+
"completion_length": 3090.6771545410156,
|
| 2665 |
+
"cov_mean": -1.8359070736551075e-05,
|
| 2666 |
+
"cov_std": 0.19972562417387962,
|
| 2667 |
+
"entropy": 0.34521484375,
|
| 2668 |
+
"epoch": 0.11314285714285714,
|
| 2669 |
+
"grad_norm": 0.27053090929985046,
|
| 2670 |
+
"kl": 0.0011525154113769531,
|
| 2671 |
+
"learning_rate": 1.002741278414069e-07,
|
| 2672 |
+
"loss": 0.0295,
|
| 2673 |
+
"reward": 0.5416666893288493,
|
| 2674 |
+
"reward_std": 0.29046234488487244,
|
| 2675 |
+
"rewards/accuracy_reward": 0.2187500074505806,
|
| 2676 |
+
"rewards/format_reward": 0.3229166669771075,
|
| 2677 |
+
"step": 99,
|
| 2678 |
+
"w_high_ratio": 0.0,
|
| 2679 |
+
"w_low_ratio": 0.028498666360974312,
|
| 2680 |
+
"w_max": 1.280459076166153,
|
| 2681 |
+
"w_mean": 1.0813924968242645,
|
| 2682 |
+
"w_min": 0.25,
|
| 2683 |
+
"w_std": 0.1357164867222309
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"clip_ratio/high_max": 0.0,
|
| 2687 |
+
"clip_ratio/high_mean": 0.0,
|
| 2688 |
+
"clip_ratio/low_mean": 0.0,
|
| 2689 |
+
"clip_ratio/low_min": 0.0,
|
| 2690 |
+
"clip_ratio/region_mean": 0.0,
|
| 2691 |
+
"completion_length": 2991.7500610351562,
|
| 2692 |
+
"cov_mean": -9.51583083406149e-05,
|
| 2693 |
+
"cov_std": 0.3894283100962639,
|
| 2694 |
+
"entropy": 0.356201171875,
|
| 2695 |
+
"epoch": 0.11428571428571428,
|
| 2696 |
+
"grad_norm": 0.5418879389762878,
|
| 2697 |
+
"kl": 0.00218963623046875,
|
| 2698 |
+
"learning_rate": 1e-07,
|
| 2699 |
+
"loss": -0.0626,
|
| 2700 |
+
"reward": 0.7395833432674408,
|
| 2701 |
+
"reward_std": 0.479395255446434,
|
| 2702 |
+
"rewards/accuracy_reward": 0.2916666716337204,
|
| 2703 |
+
"rewards/format_reward": 0.4479166865348816,
|
| 2704 |
+
"step": 100,
|
| 2705 |
+
"w_high_ratio": 0.14099212735891342,
|
| 2706 |
+
"w_low_ratio": 0.036807731725275517,
|
| 2707 |
+
"w_max": 2.043416380882263,
|
| 2708 |
+
"w_mean": 1.3361191749572754,
|
| 2709 |
+
"w_min": 6.562710904496729e-32,
|
| 2710 |
+
"w_std": 0.25866811722517014
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 0.11428571428571428,
|
| 2714 |
+
"step": 100,
|
| 2715 |
+
"total_flos": 0.0,
|
| 2716 |
+
"train_loss": -0.004950485862791538,
|
| 2717 |
+
"train_runtime": 8385.5538,
|
| 2718 |
+
"train_samples_per_second": 1.145,
|
| 2719 |
+
"train_steps_per_second": 0.012
|
| 2720 |
+
}
|
| 2721 |
+
],
|
| 2722 |
+
"logging_steps": 1,
|
| 2723 |
+
"max_steps": 100,
|
| 2724 |
+
"num_input_tokens_seen": 0,
|
| 2725 |
+
"num_train_epochs": 1,
|
| 2726 |
+
"save_steps": 50,
|
| 2727 |
+
"stateful_callbacks": {
|
| 2728 |
+
"TrainerControl": {
|
| 2729 |
+
"args": {
|
| 2730 |
+
"should_epoch_stop": false,
|
| 2731 |
+
"should_evaluate": false,
|
| 2732 |
+
"should_log": false,
|
| 2733 |
+
"should_save": true,
|
| 2734 |
+
"should_training_stop": true
|
| 2735 |
+
},
|
| 2736 |
+
"attributes": {}
|
| 2737 |
+
}
|
| 2738 |
+
},
|
| 2739 |
+
"total_flos": 0.0,
|
| 2740 |
+
"train_batch_size": 6,
|
| 2741 |
+
"trial_name": null,
|
| 2742 |
+
"trial_params": null
|
| 2743 |
+
}
|