Model save
Browse files- README.md +68 -0
- all_results.json +8 -0
- generation_config.json +6 -0
- train_results.json +8 -0
- trainer_state.json +2226 -0
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen2.5-Math-1.5B
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen-2.5-1.5B-Simple-RL-baseline3
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- grpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen-2.5-1.5B-Simple-RL-baseline3
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-Math-1.5B](https://huggingface.co/Qwen/Qwen2.5-Math-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="YYF42/Qwen-2.5-1.5B-Simple-RL-baseline3", 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/yifanyu4-university-of-illinois-urbana-champaign/RL/runs/zrzwdtkz)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.16.0.dev0
|
| 38 |
+
- Transformers: 4.49.0
|
| 39 |
+
- Pytorch: 2.5.1+cu124
|
| 40 |
+
- Datasets: 3.3.2
|
| 41 |
+
- Tokenizers: 0.21.0
|
| 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": 0.045506071436998,
|
| 4 |
+
"train_runtime": 12746.7121,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 0.588,
|
| 7 |
+
"train_steps_per_second": 0.012
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"eos_token_id": 151643,
|
| 4 |
+
"max_new_tokens": 2048,
|
| 5 |
+
"transformers_version": "4.49.0"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.045506071436998,
|
| 4 |
+
"train_runtime": 12746.7121,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 0.588,
|
| 7 |
+
"train_steps_per_second": 0.012
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2226 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9984,
|
| 5 |
+
"eval_steps": 200,
|
| 6 |
+
"global_step": 156,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"clip_ratio": 0.0,
|
| 13 |
+
"completion_length": 638.302116394043,
|
| 14 |
+
"epoch": 0.0064,
|
| 15 |
+
"grad_norm": 0.13031743466854095,
|
| 16 |
+
"kl": 0.0,
|
| 17 |
+
"learning_rate": 1.875e-07,
|
| 18 |
+
"loss": 0.0382,
|
| 19 |
+
"reward": 0.4348958469927311,
|
| 20 |
+
"reward_std": 0.3634899891912937,
|
| 21 |
+
"rewards/accuracy_reward": 0.4348958469927311,
|
| 22 |
+
"rewards/format_reward": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"clip_ratio": 0.0,
|
| 27 |
+
"completion_length": 634.0755386352539,
|
| 28 |
+
"epoch": 0.0128,
|
| 29 |
+
"grad_norm": 0.15040069818496704,
|
| 30 |
+
"kl": 0.0,
|
| 31 |
+
"learning_rate": 3.75e-07,
|
| 32 |
+
"loss": 0.0361,
|
| 33 |
+
"reward": 0.455729179084301,
|
| 34 |
+
"reward_std": 0.40231897309422493,
|
| 35 |
+
"rewards/accuracy_reward": 0.455729179084301,
|
| 36 |
+
"rewards/format_reward": 0.0,
|
| 37 |
+
"step": 2
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"clip_ratio": 0.0,
|
| 41 |
+
"completion_length": 584.7291870117188,
|
| 42 |
+
"epoch": 0.0192,
|
| 43 |
+
"grad_norm": 0.14233516156673431,
|
| 44 |
+
"kl": 5.124509334564209e-05,
|
| 45 |
+
"learning_rate": 5.625e-07,
|
| 46 |
+
"loss": 0.0862,
|
| 47 |
+
"reward": 0.4583333469927311,
|
| 48 |
+
"reward_std": 0.36376284807920456,
|
| 49 |
+
"rewards/accuracy_reward": 0.4583333469927311,
|
| 50 |
+
"rewards/format_reward": 0.0,
|
| 51 |
+
"step": 3
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"clip_ratio": 0.0,
|
| 55 |
+
"completion_length": 596.8698120117188,
|
| 56 |
+
"epoch": 0.0256,
|
| 57 |
+
"grad_norm": 0.13924892246723175,
|
| 58 |
+
"kl": 4.8801302909851074e-05,
|
| 59 |
+
"learning_rate": 7.5e-07,
|
| 60 |
+
"loss": 0.0257,
|
| 61 |
+
"reward": 0.4296875149011612,
|
| 62 |
+
"reward_std": 0.4013553448021412,
|
| 63 |
+
"rewards/accuracy_reward": 0.4296875149011612,
|
| 64 |
+
"rewards/format_reward": 0.0,
|
| 65 |
+
"step": 4
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"clip_ratio": 0.0,
|
| 69 |
+
"completion_length": 632.9687576293945,
|
| 70 |
+
"epoch": 0.032,
|
| 71 |
+
"grad_norm": 0.14110471308231354,
|
| 72 |
+
"kl": 4.8160552978515625e-05,
|
| 73 |
+
"learning_rate": 9.375e-07,
|
| 74 |
+
"loss": 0.0076,
|
| 75 |
+
"reward": 0.4609375149011612,
|
| 76 |
+
"reward_std": 0.3605061452835798,
|
| 77 |
+
"rewards/accuracy_reward": 0.4609375149011612,
|
| 78 |
+
"rewards/format_reward": 0.0,
|
| 79 |
+
"step": 5
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"clip_ratio": 0.0,
|
| 83 |
+
"completion_length": 584.5052223205566,
|
| 84 |
+
"epoch": 0.0384,
|
| 85 |
+
"grad_norm": 0.16037185490131378,
|
| 86 |
+
"kl": 5.336105823516846e-05,
|
| 87 |
+
"learning_rate": 1.125e-06,
|
| 88 |
+
"loss": 0.042,
|
| 89 |
+
"reward": 0.5052083395421505,
|
| 90 |
+
"reward_std": 0.42091938853263855,
|
| 91 |
+
"rewards/accuracy_reward": 0.5052083395421505,
|
| 92 |
+
"rewards/format_reward": 0.0,
|
| 93 |
+
"step": 6
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"clip_ratio": 0.0,
|
| 97 |
+
"completion_length": 616.0208587646484,
|
| 98 |
+
"epoch": 0.0448,
|
| 99 |
+
"grad_norm": 0.12087774276733398,
|
| 100 |
+
"kl": 8.541345596313477e-05,
|
| 101 |
+
"learning_rate": 1.3125000000000001e-06,
|
| 102 |
+
"loss": 0.0476,
|
| 103 |
+
"reward": 0.4244791707023978,
|
| 104 |
+
"reward_std": 0.4058343879878521,
|
| 105 |
+
"rewards/accuracy_reward": 0.4244791707023978,
|
| 106 |
+
"rewards/format_reward": 0.0,
|
| 107 |
+
"step": 7
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"clip_ratio": 0.0,
|
| 111 |
+
"completion_length": 662.7343902587891,
|
| 112 |
+
"epoch": 0.0512,
|
| 113 |
+
"grad_norm": 0.10394447296857834,
|
| 114 |
+
"kl": 4.2825937271118164e-05,
|
| 115 |
+
"learning_rate": 1.5e-06,
|
| 116 |
+
"loss": 0.0349,
|
| 117 |
+
"reward": 0.41666667349636555,
|
| 118 |
+
"reward_std": 0.3470919840037823,
|
| 119 |
+
"rewards/accuracy_reward": 0.41666667349636555,
|
| 120 |
+
"rewards/format_reward": 0.0,
|
| 121 |
+
"step": 8
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"clip_ratio": 0.0,
|
| 125 |
+
"completion_length": 614.2031478881836,
|
| 126 |
+
"epoch": 0.0576,
|
| 127 |
+
"grad_norm": 0.11856164038181305,
|
| 128 |
+
"kl": 4.234910011291504e-05,
|
| 129 |
+
"learning_rate": 1.6875e-06,
|
| 130 |
+
"loss": 0.0047,
|
| 131 |
+
"reward": 0.4322916781529784,
|
| 132 |
+
"reward_std": 0.3687521107494831,
|
| 133 |
+
"rewards/accuracy_reward": 0.4322916781529784,
|
| 134 |
+
"rewards/format_reward": 0.0,
|
| 135 |
+
"step": 9
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"clip_ratio": 0.0,
|
| 139 |
+
"completion_length": 666.4557495117188,
|
| 140 |
+
"epoch": 0.064,
|
| 141 |
+
"grad_norm": 0.1338435709476471,
|
| 142 |
+
"kl": 4.839897155761719e-05,
|
| 143 |
+
"learning_rate": 1.875e-06,
|
| 144 |
+
"loss": 0.0727,
|
| 145 |
+
"reward": 0.4062500149011612,
|
| 146 |
+
"reward_std": 0.3383309133350849,
|
| 147 |
+
"rewards/accuracy_reward": 0.4062500149011612,
|
| 148 |
+
"rewards/format_reward": 0.0,
|
| 149 |
+
"step": 10
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"clip_ratio": 0.0,
|
| 153 |
+
"completion_length": 580.9244918823242,
|
| 154 |
+
"epoch": 0.0704,
|
| 155 |
+
"grad_norm": 0.13869862258434296,
|
| 156 |
+
"kl": 8.58604907989502e-05,
|
| 157 |
+
"learning_rate": 2.0625e-06,
|
| 158 |
+
"loss": 0.0402,
|
| 159 |
+
"reward": 0.38281251303851604,
|
| 160 |
+
"reward_std": 0.3277249541133642,
|
| 161 |
+
"rewards/accuracy_reward": 0.38281251303851604,
|
| 162 |
+
"rewards/format_reward": 0.0,
|
| 163 |
+
"step": 11
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"clip_ratio": 0.0,
|
| 167 |
+
"completion_length": 610.3463706970215,
|
| 168 |
+
"epoch": 0.0768,
|
| 169 |
+
"grad_norm": 0.14376282691955566,
|
| 170 |
+
"kl": 0.00030118227005004883,
|
| 171 |
+
"learning_rate": 2.25e-06,
|
| 172 |
+
"loss": 0.0366,
|
| 173 |
+
"reward": 0.3671875074505806,
|
| 174 |
+
"reward_std": 0.36174426227808,
|
| 175 |
+
"rewards/accuracy_reward": 0.3671875074505806,
|
| 176 |
+
"rewards/format_reward": 0.0,
|
| 177 |
+
"step": 12
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"clip_ratio": 0.0,
|
| 181 |
+
"completion_length": 585.5807418823242,
|
| 182 |
+
"epoch": 0.0832,
|
| 183 |
+
"grad_norm": 0.15505321323871613,
|
| 184 |
+
"kl": 0.00016033649444580078,
|
| 185 |
+
"learning_rate": 2.4375e-06,
|
| 186 |
+
"loss": 0.0304,
|
| 187 |
+
"reward": 0.479166679084301,
|
| 188 |
+
"reward_std": 0.3392144478857517,
|
| 189 |
+
"rewards/accuracy_reward": 0.479166679084301,
|
| 190 |
+
"rewards/format_reward": 0.0,
|
| 191 |
+
"step": 13
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"clip_ratio": 0.0,
|
| 195 |
+
"completion_length": 583.3646049499512,
|
| 196 |
+
"epoch": 0.0896,
|
| 197 |
+
"grad_norm": 0.18337412178516388,
|
| 198 |
+
"kl": 0.00046515464782714844,
|
| 199 |
+
"learning_rate": 2.6250000000000003e-06,
|
| 200 |
+
"loss": 0.0355,
|
| 201 |
+
"reward": 0.4453125111758709,
|
| 202 |
+
"reward_std": 0.3934810720384121,
|
| 203 |
+
"rewards/accuracy_reward": 0.4453125111758709,
|
| 204 |
+
"rewards/format_reward": 0.0,
|
| 205 |
+
"step": 14
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"clip_ratio": 0.0,
|
| 209 |
+
"completion_length": 584.3489761352539,
|
| 210 |
+
"epoch": 0.096,
|
| 211 |
+
"grad_norm": 0.13246002793312073,
|
| 212 |
+
"kl": 0.0005662441253662109,
|
| 213 |
+
"learning_rate": 2.8125e-06,
|
| 214 |
+
"loss": 0.075,
|
| 215 |
+
"reward": 0.4609375111758709,
|
| 216 |
+
"reward_std": 0.35086217895150185,
|
| 217 |
+
"rewards/accuracy_reward": 0.4609375111758709,
|
| 218 |
+
"rewards/format_reward": 0.0,
|
| 219 |
+
"step": 15
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"clip_ratio": 0.0,
|
| 223 |
+
"completion_length": 649.3307342529297,
|
| 224 |
+
"epoch": 0.1024,
|
| 225 |
+
"grad_norm": 0.1374843418598175,
|
| 226 |
+
"kl": 0.0015537738800048828,
|
| 227 |
+
"learning_rate": 3e-06,
|
| 228 |
+
"loss": 0.0406,
|
| 229 |
+
"reward": 0.4557291828095913,
|
| 230 |
+
"reward_std": 0.3519272692501545,
|
| 231 |
+
"rewards/accuracy_reward": 0.4557291828095913,
|
| 232 |
+
"rewards/format_reward": 0.0,
|
| 233 |
+
"step": 16
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"clip_ratio": 0.0,
|
| 237 |
+
"completion_length": 613.6744995117188,
|
| 238 |
+
"epoch": 0.1088,
|
| 239 |
+
"grad_norm": 0.1317468285560608,
|
| 240 |
+
"kl": 0.0027337074279785156,
|
| 241 |
+
"learning_rate": 2.99962235241376e-06,
|
| 242 |
+
"loss": 0.0231,
|
| 243 |
+
"reward": 0.5234375149011612,
|
| 244 |
+
"reward_std": 0.40249527990818024,
|
| 245 |
+
"rewards/accuracy_reward": 0.5234375149011612,
|
| 246 |
+
"rewards/format_reward": 0.0,
|
| 247 |
+
"step": 17
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"clip_ratio": 0.0,
|
| 251 |
+
"completion_length": 609.6224098205566,
|
| 252 |
+
"epoch": 0.1152,
|
| 253 |
+
"grad_norm": 0.13464686274528503,
|
| 254 |
+
"kl": 0.0029234886169433594,
|
| 255 |
+
"learning_rate": 2.998489599811972e-06,
|
| 256 |
+
"loss": 0.0867,
|
| 257 |
+
"reward": 0.5156250223517418,
|
| 258 |
+
"reward_std": 0.3477992061525583,
|
| 259 |
+
"rewards/accuracy_reward": 0.5156250223517418,
|
| 260 |
+
"rewards/format_reward": 0.0,
|
| 261 |
+
"step": 18
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"clip_ratio": 0.0,
|
| 265 |
+
"completion_length": 613.5703353881836,
|
| 266 |
+
"epoch": 0.1216,
|
| 267 |
+
"grad_norm": 0.12391193211078644,
|
| 268 |
+
"kl": 0.005184173583984375,
|
| 269 |
+
"learning_rate": 2.996602312569684e-06,
|
| 270 |
+
"loss": 0.0272,
|
| 271 |
+
"reward": 0.4947916753590107,
|
| 272 |
+
"reward_std": 0.35979728773236275,
|
| 273 |
+
"rewards/accuracy_reward": 0.4947916753590107,
|
| 274 |
+
"rewards/format_reward": 0.0,
|
| 275 |
+
"step": 19
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"clip_ratio": 0.0,
|
| 279 |
+
"completion_length": 605.9609451293945,
|
| 280 |
+
"epoch": 0.128,
|
| 281 |
+
"grad_norm": 0.11922721564769745,
|
| 282 |
+
"kl": 0.0034084320068359375,
|
| 283 |
+
"learning_rate": 2.993961440992859e-06,
|
| 284 |
+
"loss": 0.0969,
|
| 285 |
+
"reward": 0.5234375223517418,
|
| 286 |
+
"reward_std": 0.3810082282871008,
|
| 287 |
+
"rewards/accuracy_reward": 0.5234375223517418,
|
| 288 |
+
"rewards/format_reward": 0.0,
|
| 289 |
+
"step": 20
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"clip_ratio": 0.0,
|
| 293 |
+
"completion_length": 631.2369956970215,
|
| 294 |
+
"epoch": 0.1344,
|
| 295 |
+
"grad_norm": 0.12600868940353394,
|
| 296 |
+
"kl": 0.015664100646972656,
|
| 297 |
+
"learning_rate": 2.9905683148398643e-06,
|
| 298 |
+
"loss": 0.0522,
|
| 299 |
+
"reward": 0.5182291865348816,
|
| 300 |
+
"reward_std": 0.3406124524772167,
|
| 301 |
+
"rewards/accuracy_reward": 0.5182291865348816,
|
| 302 |
+
"rewards/format_reward": 0.0,
|
| 303 |
+
"step": 21
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"clip_ratio": 0.0,
|
| 307 |
+
"completion_length": 609.9323043823242,
|
| 308 |
+
"epoch": 0.1408,
|
| 309 |
+
"grad_norm": 0.13966120779514313,
|
| 310 |
+
"kl": 0.006683349609375,
|
| 311 |
+
"learning_rate": 2.986424642651902e-06,
|
| 312 |
+
"loss": 0.0565,
|
| 313 |
+
"reward": 0.5156250111758709,
|
| 314 |
+
"reward_std": 0.3761220443993807,
|
| 315 |
+
"rewards/accuracy_reward": 0.5156250111758709,
|
| 316 |
+
"rewards/format_reward": 0.0,
|
| 317 |
+
"step": 22
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"clip_ratio": 0.0,
|
| 321 |
+
"completion_length": 564.3984489440918,
|
| 322 |
+
"epoch": 0.1472,
|
| 323 |
+
"grad_norm": 1.296398639678955,
|
| 324 |
+
"kl": 0.0100250244140625,
|
| 325 |
+
"learning_rate": 2.981532510892707e-06,
|
| 326 |
+
"loss": 0.104,
|
| 327 |
+
"reward": 0.6614583507180214,
|
| 328 |
+
"reward_std": 0.3120367918163538,
|
| 329 |
+
"rewards/accuracy_reward": 0.6614583507180214,
|
| 330 |
+
"rewards/format_reward": 0.0,
|
| 331 |
+
"step": 23
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"clip_ratio": 0.0,
|
| 335 |
+
"completion_length": 600.9036712646484,
|
| 336 |
+
"epoch": 0.1536,
|
| 337 |
+
"grad_norm": 0.12987235188484192,
|
| 338 |
+
"kl": 0.011409759521484375,
|
| 339 |
+
"learning_rate": 2.9758943828979446e-06,
|
| 340 |
+
"loss": 0.0632,
|
| 341 |
+
"reward": 0.6041666828095913,
|
| 342 |
+
"reward_std": 0.2938920930027962,
|
| 343 |
+
"rewards/accuracy_reward": 0.6041666828095913,
|
| 344 |
+
"rewards/format_reward": 0.0,
|
| 345 |
+
"step": 24
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"clip_ratio": 0.0,
|
| 349 |
+
"completion_length": 601.5573043823242,
|
| 350 |
+
"epoch": 0.16,
|
| 351 |
+
"grad_norm": 0.10699014365673065,
|
| 352 |
+
"kl": 0.008472442626953125,
|
| 353 |
+
"learning_rate": 2.9695130976348534e-06,
|
| 354 |
+
"loss": 0.0507,
|
| 355 |
+
"reward": 0.6041667014360428,
|
| 356 |
+
"reward_std": 0.31449300795793533,
|
| 357 |
+
"rewards/accuracy_reward": 0.6041667014360428,
|
| 358 |
+
"rewards/format_reward": 0.0,
|
| 359 |
+
"step": 25
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"clip_ratio": 0.0,
|
| 363 |
+
"completion_length": 584.5390777587891,
|
| 364 |
+
"epoch": 0.1664,
|
| 365 |
+
"grad_norm": 0.09624404460191727,
|
| 366 |
+
"kl": 0.007846832275390625,
|
| 367 |
+
"learning_rate": 2.9623918682727352e-06,
|
| 368 |
+
"loss": 0.0102,
|
| 369 |
+
"reward": 0.6223958544433117,
|
| 370 |
+
"reward_std": 0.32982853427529335,
|
| 371 |
+
"rewards/accuracy_reward": 0.6223958544433117,
|
| 372 |
+
"rewards/format_reward": 0.0,
|
| 373 |
+
"step": 26
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"clip_ratio": 0.0,
|
| 377 |
+
"completion_length": 625.4427337646484,
|
| 378 |
+
"epoch": 0.1728,
|
| 379 |
+
"grad_norm": 0.09527429938316345,
|
| 380 |
+
"kl": 0.00612640380859375,
|
| 381 |
+
"learning_rate": 2.9545342805650304e-06,
|
| 382 |
+
"loss": 0.0509,
|
| 383 |
+
"reward": 0.5729166902601719,
|
| 384 |
+
"reward_std": 0.2726762406527996,
|
| 385 |
+
"rewards/accuracy_reward": 0.5729166902601719,
|
| 386 |
+
"rewards/format_reward": 0.0,
|
| 387 |
+
"step": 27
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"clip_ratio": 0.0,
|
| 391 |
+
"completion_length": 579.7708511352539,
|
| 392 |
+
"epoch": 0.1792,
|
| 393 |
+
"grad_norm": 0.16817033290863037,
|
| 394 |
+
"kl": 0.00962066650390625,
|
| 395 |
+
"learning_rate": 2.94594429104378e-06,
|
| 396 |
+
"loss": 0.0804,
|
| 397 |
+
"reward": 0.6432291865348816,
|
| 398 |
+
"reward_std": 0.3065182641148567,
|
| 399 |
+
"rewards/accuracy_reward": 0.6432291865348816,
|
| 400 |
+
"rewards/format_reward": 0.0,
|
| 401 |
+
"step": 28
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"clip_ratio": 0.0,
|
| 405 |
+
"completion_length": 645.1614837646484,
|
| 406 |
+
"epoch": 0.1856,
|
| 407 |
+
"grad_norm": 0.10505980998277664,
|
| 408 |
+
"kl": 0.060828208923339844,
|
| 409 |
+
"learning_rate": 2.9366262250273957e-06,
|
| 410 |
+
"loss": 0.049,
|
| 411 |
+
"reward": 0.5937500149011612,
|
| 412 |
+
"reward_std": 0.2890551872551441,
|
| 413 |
+
"rewards/accuracy_reward": 0.5937500149011612,
|
| 414 |
+
"rewards/format_reward": 0.0,
|
| 415 |
+
"step": 29
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"clip_ratio": 0.0,
|
| 419 |
+
"completion_length": 605.0573120117188,
|
| 420 |
+
"epoch": 0.192,
|
| 421 |
+
"grad_norm": 0.12251781672239304,
|
| 422 |
+
"kl": 0.006243705749511719,
|
| 423 |
+
"learning_rate": 2.9265847744427307e-06,
|
| 424 |
+
"loss": 0.0778,
|
| 425 |
+
"reward": 0.6770833507180214,
|
| 426 |
+
"reward_std": 0.2973086703568697,
|
| 427 |
+
"rewards/accuracy_reward": 0.6770833507180214,
|
| 428 |
+
"rewards/format_reward": 0.0,
|
| 429 |
+
"step": 30
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"clip_ratio": 0.0,
|
| 433 |
+
"completion_length": 567.5885543823242,
|
| 434 |
+
"epoch": 0.1984,
|
| 435 |
+
"grad_norm": 0.09576579928398132,
|
| 436 |
+
"kl": 0.011445999145507812,
|
| 437 |
+
"learning_rate": 2.9158249954625514e-06,
|
| 438 |
+
"loss": 0.0425,
|
| 439 |
+
"reward": 0.6171875223517418,
|
| 440 |
+
"reward_std": 0.2874841373413801,
|
| 441 |
+
"rewards/accuracy_reward": 0.6171875223517418,
|
| 442 |
+
"rewards/format_reward": 0.0,
|
| 443 |
+
"step": 31
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"clip_ratio": 0.0,
|
| 447 |
+
"completion_length": 555.1380386352539,
|
| 448 |
+
"epoch": 0.2048,
|
| 449 |
+
"grad_norm": 0.10128926485776901,
|
| 450 |
+
"kl": 0.0049610137939453125,
|
| 451 |
+
"learning_rate": 2.904352305959606e-06,
|
| 452 |
+
"loss": 0.0359,
|
| 453 |
+
"reward": 0.6770833432674408,
|
| 454 |
+
"reward_std": 0.2876647114753723,
|
| 455 |
+
"rewards/accuracy_reward": 0.6770833432674408,
|
| 456 |
+
"rewards/format_reward": 0.0,
|
| 457 |
+
"step": 32
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"clip_ratio": 0.0,
|
| 461 |
+
"completion_length": 601.838550567627,
|
| 462 |
+
"epoch": 0.2112,
|
| 463 |
+
"grad_norm": 0.0932403951883316,
|
| 464 |
+
"kl": 0.006168365478515625,
|
| 465 |
+
"learning_rate": 2.8921724827785583e-06,
|
| 466 |
+
"loss": 0.0492,
|
| 467 |
+
"reward": 0.6588541865348816,
|
| 468 |
+
"reward_std": 0.27539050579071045,
|
| 469 |
+
"rewards/accuracy_reward": 0.6588541865348816,
|
| 470 |
+
"rewards/format_reward": 0.0,
|
| 471 |
+
"step": 33
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"clip_ratio": 0.0,
|
| 475 |
+
"completion_length": 543.5703315734863,
|
| 476 |
+
"epoch": 0.2176,
|
| 477 |
+
"grad_norm": 0.10408346354961395,
|
| 478 |
+
"kl": 0.06038475036621094,
|
| 479 |
+
"learning_rate": 2.879291658827176e-06,
|
| 480 |
+
"loss": 0.0365,
|
| 481 |
+
"reward": 0.7109375149011612,
|
| 482 |
+
"reward_std": 0.2072770199738443,
|
| 483 |
+
"rewards/accuracy_reward": 0.7109375149011612,
|
| 484 |
+
"rewards/format_reward": 0.0,
|
| 485 |
+
"step": 34
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"clip_ratio": 0.0,
|
| 489 |
+
"completion_length": 585.3281364440918,
|
| 490 |
+
"epoch": 0.224,
|
| 491 |
+
"grad_norm": 0.2577992379665375,
|
| 492 |
+
"kl": 0.040401458740234375,
|
| 493 |
+
"learning_rate": 2.865716319988224e-06,
|
| 494 |
+
"loss": 0.0483,
|
| 495 |
+
"reward": 0.6093750223517418,
|
| 496 |
+
"reward_std": 0.2657390311360359,
|
| 497 |
+
"rewards/accuracy_reward": 0.6093750223517418,
|
| 498 |
+
"rewards/format_reward": 0.0,
|
| 499 |
+
"step": 35
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"clip_ratio": 0.0,
|
| 503 |
+
"completion_length": 598.9271087646484,
|
| 504 |
+
"epoch": 0.2304,
|
| 505 |
+
"grad_norm": 0.10200951248407364,
|
| 506 |
+
"kl": 0.008726119995117188,
|
| 507 |
+
"learning_rate": 2.851453301853629e-06,
|
| 508 |
+
"loss": 0.057,
|
| 509 |
+
"reward": 0.6614583432674408,
|
| 510 |
+
"reward_std": 0.2139296056702733,
|
| 511 |
+
"rewards/accuracy_reward": 0.6614583432674408,
|
| 512 |
+
"rewards/format_reward": 0.0,
|
| 513 |
+
"step": 36
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"clip_ratio": 0.0,
|
| 517 |
+
"completion_length": 597.4114761352539,
|
| 518 |
+
"epoch": 0.2368,
|
| 519 |
+
"grad_norm": 0.12581026554107666,
|
| 520 |
+
"kl": 0.00994873046875,
|
| 521 |
+
"learning_rate": 2.8365097862825516e-06,
|
| 522 |
+
"loss": 0.0183,
|
| 523 |
+
"reward": 0.6145833469927311,
|
| 524 |
+
"reward_std": 0.281610372941941,
|
| 525 |
+
"rewards/accuracy_reward": 0.6145833469927311,
|
| 526 |
+
"rewards/format_reward": 0.0,
|
| 527 |
+
"step": 37
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"clip_ratio": 0.0,
|
| 531 |
+
"completion_length": 613.4218902587891,
|
| 532 |
+
"epoch": 0.2432,
|
| 533 |
+
"grad_norm": 0.096426822245121,
|
| 534 |
+
"kl": 0.0059642791748046875,
|
| 535 |
+
"learning_rate": 2.820893297785107e-06,
|
| 536 |
+
"loss": 0.053,
|
| 537 |
+
"reward": 0.6588541865348816,
|
| 538 |
+
"reward_std": 0.2768824342638254,
|
| 539 |
+
"rewards/accuracy_reward": 0.6588541865348816,
|
| 540 |
+
"rewards/format_reward": 0.0,
|
| 541 |
+
"step": 38
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"clip_ratio": 0.0,
|
| 545 |
+
"completion_length": 548.0599098205566,
|
| 546 |
+
"epoch": 0.2496,
|
| 547 |
+
"grad_norm": 0.10879187285900116,
|
| 548 |
+
"kl": 0.013092041015625,
|
| 549 |
+
"learning_rate": 2.8046116997335432e-06,
|
| 550 |
+
"loss": 0.0297,
|
| 551 |
+
"reward": 0.6380208507180214,
|
| 552 |
+
"reward_std": 0.2746025351807475,
|
| 553 |
+
"rewards/accuracy_reward": 0.6380208507180214,
|
| 554 |
+
"rewards/format_reward": 0.0,
|
| 555 |
+
"step": 39
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"clip_ratio": 0.0,
|
| 559 |
+
"completion_length": 613.0182495117188,
|
| 560 |
+
"epoch": 0.256,
|
| 561 |
+
"grad_norm": 0.09518619626760483,
|
| 562 |
+
"kl": 0.009784698486328125,
|
| 563 |
+
"learning_rate": 2.7876731904027993e-06,
|
| 564 |
+
"loss": 0.0229,
|
| 565 |
+
"reward": 0.5442708507180214,
|
| 566 |
+
"reward_std": 0.2534680999815464,
|
| 567 |
+
"rewards/accuracy_reward": 0.5442708507180214,
|
| 568 |
+
"rewards/format_reward": 0.0,
|
| 569 |
+
"step": 40
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"clip_ratio": 0.0,
|
| 573 |
+
"completion_length": 530.2083587646484,
|
| 574 |
+
"epoch": 0.2624,
|
| 575 |
+
"grad_norm": 0.09931885451078415,
|
| 576 |
+
"kl": 0.010135650634765625,
|
| 577 |
+
"learning_rate": 2.7700862988424264e-06,
|
| 578 |
+
"loss": 0.0577,
|
| 579 |
+
"reward": 0.7057291939854622,
|
| 580 |
+
"reward_std": 0.300371628254652,
|
| 581 |
+
"rewards/accuracy_reward": 0.7057291939854622,
|
| 582 |
+
"rewards/format_reward": 0.0,
|
| 583 |
+
"step": 41
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"clip_ratio": 0.0,
|
| 587 |
+
"completion_length": 609.9817962646484,
|
| 588 |
+
"epoch": 0.2688,
|
| 589 |
+
"grad_norm": 0.09324629604816437,
|
| 590 |
+
"kl": 0.0042209625244140625,
|
| 591 |
+
"learning_rate": 2.751859880581954e-06,
|
| 592 |
+
"loss": 0.0406,
|
| 593 |
+
"reward": 0.5911458432674408,
|
| 594 |
+
"reward_std": 0.312465263530612,
|
| 595 |
+
"rewards/accuracy_reward": 0.5911458432674408,
|
| 596 |
+
"rewards/format_reward": 0.0,
|
| 597 |
+
"step": 42
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"clip_ratio": 0.0,
|
| 601 |
+
"completion_length": 541.2994995117188,
|
| 602 |
+
"epoch": 0.2752,
|
| 603 |
+
"grad_norm": 0.10992565006017685,
|
| 604 |
+
"kl": 0.0066661834716796875,
|
| 605 |
+
"learning_rate": 2.733003113171865e-06,
|
| 606 |
+
"loss": 0.0512,
|
| 607 |
+
"reward": 0.8177083432674408,
|
| 608 |
+
"reward_std": 0.231904536485672,
|
| 609 |
+
"rewards/accuracy_reward": 0.8177083432674408,
|
| 610 |
+
"rewards/format_reward": 0.0,
|
| 611 |
+
"step": 43
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"clip_ratio": 0.0,
|
| 615 |
+
"completion_length": 636.9010581970215,
|
| 616 |
+
"epoch": 0.2816,
|
| 617 |
+
"grad_norm": 0.10348695516586304,
|
| 618 |
+
"kl": 0.014631271362304688,
|
| 619 |
+
"learning_rate": 2.713525491562421e-06,
|
| 620 |
+
"loss": 0.0365,
|
| 621 |
+
"reward": 0.526041679084301,
|
| 622 |
+
"reward_std": 0.3262313976883888,
|
| 623 |
+
"rewards/accuracy_reward": 0.526041679084301,
|
| 624 |
+
"rewards/format_reward": 0.0,
|
| 625 |
+
"step": 44
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"clip_ratio": 0.0,
|
| 629 |
+
"completion_length": 600.2734489440918,
|
| 630 |
+
"epoch": 0.288,
|
| 631 |
+
"grad_norm": 0.10393375158309937,
|
| 632 |
+
"kl": 0.00383758544921875,
|
| 633 |
+
"learning_rate": 2.6934368233226715e-06,
|
| 634 |
+
"loss": 0.0421,
|
| 635 |
+
"reward": 0.6458333544433117,
|
| 636 |
+
"reward_std": 0.24286149069666862,
|
| 637 |
+
"rewards/accuracy_reward": 0.6458333544433117,
|
| 638 |
+
"rewards/format_reward": 0.0,
|
| 639 |
+
"step": 45
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"clip_ratio": 0.0,
|
| 643 |
+
"completion_length": 548.9713745117188,
|
| 644 |
+
"epoch": 0.2944,
|
| 645 |
+
"grad_norm": 0.11263209581375122,
|
| 646 |
+
"kl": 0.008182525634765625,
|
| 647 |
+
"learning_rate": 2.6727472237020448e-06,
|
| 648 |
+
"loss": 0.0661,
|
| 649 |
+
"reward": 0.700520858168602,
|
| 650 |
+
"reward_std": 0.2998374607414007,
|
| 651 |
+
"rewards/accuracy_reward": 0.700520858168602,
|
| 652 |
+
"rewards/format_reward": 0.0,
|
| 653 |
+
"step": 46
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"clip_ratio": 0.0,
|
| 657 |
+
"completion_length": 630.1458587646484,
|
| 658 |
+
"epoch": 0.3008,
|
| 659 |
+
"grad_norm": 0.09937220811843872,
|
| 660 |
+
"kl": 0.016878128051757812,
|
| 661 |
+
"learning_rate": 2.6514671105370166e-06,
|
| 662 |
+
"loss": 0.0542,
|
| 663 |
+
"reward": 0.6119791902601719,
|
| 664 |
+
"reward_std": 0.34929538145661354,
|
| 665 |
+
"rewards/accuracy_reward": 0.6119791902601719,
|
| 666 |
+
"rewards/format_reward": 0.0,
|
| 667 |
+
"step": 47
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"clip_ratio": 0.0,
|
| 671 |
+
"completion_length": 606.1796989440918,
|
| 672 |
+
"epoch": 0.3072,
|
| 673 |
+
"grad_norm": 0.6816109418869019,
|
| 674 |
+
"kl": 0.04201507568359375,
|
| 675 |
+
"learning_rate": 2.6296071990054165e-06,
|
| 676 |
+
"loss": 0.0295,
|
| 677 |
+
"reward": 0.6015625186264515,
|
| 678 |
+
"reward_std": 0.22392718074843287,
|
| 679 |
+
"rewards/accuracy_reward": 0.6015625186264515,
|
| 680 |
+
"rewards/format_reward": 0.0,
|
| 681 |
+
"step": 48
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"clip_ratio": 0.0,
|
| 685 |
+
"completion_length": 634.2864837646484,
|
| 686 |
+
"epoch": 0.3136,
|
| 687 |
+
"grad_norm": 0.07695292681455612,
|
| 688 |
+
"kl": 0.005660057067871094,
|
| 689 |
+
"learning_rate": 2.6071784962310114e-06,
|
| 690 |
+
"loss": 0.0412,
|
| 691 |
+
"reward": 0.6380208432674408,
|
| 692 |
+
"reward_std": 0.2517191059887409,
|
| 693 |
+
"rewards/accuracy_reward": 0.6380208432674408,
|
| 694 |
+
"rewards/format_reward": 0.0,
|
| 695 |
+
"step": 49
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"clip_ratio": 0.0,
|
| 699 |
+
"completion_length": 561.0755424499512,
|
| 700 |
+
"epoch": 0.32,
|
| 701 |
+
"grad_norm": 0.08334773778915405,
|
| 702 |
+
"kl": 0.0075092315673828125,
|
| 703 |
+
"learning_rate": 2.584192295741087e-06,
|
| 704 |
+
"loss": 0.0026,
|
| 705 |
+
"reward": 0.6458333618938923,
|
| 706 |
+
"reward_std": 0.2564529310911894,
|
| 707 |
+
"rewards/accuracy_reward": 0.6458333618938923,
|
| 708 |
+
"rewards/format_reward": 0.0,
|
| 709 |
+
"step": 50
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"clip_ratio": 0.0,
|
| 713 |
+
"completion_length": 557.2916831970215,
|
| 714 |
+
"epoch": 0.3264,
|
| 715 |
+
"grad_norm": 0.0740433931350708,
|
| 716 |
+
"kl": 0.0064678192138671875,
|
| 717 |
+
"learning_rate": 2.5606601717798212e-06,
|
| 718 |
+
"loss": 0.0172,
|
| 719 |
+
"reward": 0.6562500298023224,
|
| 720 |
+
"reward_std": 0.2332175220362842,
|
| 721 |
+
"rewards/accuracy_reward": 0.6562500298023224,
|
| 722 |
+
"rewards/format_reward": 0.0,
|
| 723 |
+
"step": 51
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"clip_ratio": 0.0,
|
| 727 |
+
"completion_length": 582.4843978881836,
|
| 728 |
+
"epoch": 0.3328,
|
| 729 |
+
"grad_norm": 0.1135820522904396,
|
| 730 |
+
"kl": 0.0116424560546875,
|
| 731 |
+
"learning_rate": 2.5365939734802974e-06,
|
| 732 |
+
"loss": 0.0279,
|
| 733 |
+
"reward": 0.6666666939854622,
|
| 734 |
+
"reward_std": 0.2783694500103593,
|
| 735 |
+
"rewards/accuracy_reward": 0.6666666939854622,
|
| 736 |
+
"rewards/format_reward": 0.0,
|
| 737 |
+
"step": 52
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"clip_ratio": 0.0,
|
| 741 |
+
"completion_length": 497.79689025878906,
|
| 742 |
+
"epoch": 0.3392,
|
| 743 |
+
"grad_norm": 0.12283315509557724,
|
| 744 |
+
"kl": 0.00904083251953125,
|
| 745 |
+
"learning_rate": 2.512005818898112e-06,
|
| 746 |
+
"loss": 0.0665,
|
| 747 |
+
"reward": 0.7682291865348816,
|
| 748 |
+
"reward_std": 0.26794830709695816,
|
| 749 |
+
"rewards/accuracy_reward": 0.7682291865348816,
|
| 750 |
+
"rewards/format_reward": 0.0,
|
| 751 |
+
"step": 53
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"clip_ratio": 0.0,
|
| 755 |
+
"completion_length": 629.0130462646484,
|
| 756 |
+
"epoch": 0.3456,
|
| 757 |
+
"grad_norm": 0.10559183359146118,
|
| 758 |
+
"kl": 0.0058498382568359375,
|
| 759 |
+
"learning_rate": 2.486908088909569e-06,
|
| 760 |
+
"loss": 0.0437,
|
| 761 |
+
"reward": 0.5572916902601719,
|
| 762 |
+
"reward_std": 0.2583775855600834,
|
| 763 |
+
"rewards/accuracy_reward": 0.5572916902601719,
|
| 764 |
+
"rewards/format_reward": 0.0,
|
| 765 |
+
"step": 54
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"clip_ratio": 0.0,
|
| 769 |
+
"completion_length": 631.0338745117188,
|
| 770 |
+
"epoch": 0.352,
|
| 771 |
+
"grad_norm": 0.07775697112083435,
|
| 772 |
+
"kl": 0.011181831359863281,
|
| 773 |
+
"learning_rate": 2.461313420977536e-06,
|
| 774 |
+
"loss": 0.0187,
|
| 775 |
+
"reward": 0.6562500223517418,
|
| 776 |
+
"reward_std": 0.2523265201598406,
|
| 777 |
+
"rewards/accuracy_reward": 0.6562500223517418,
|
| 778 |
+
"rewards/format_reward": 0.0,
|
| 779 |
+
"step": 55
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"clip_ratio": 0.0,
|
| 783 |
+
"completion_length": 596.7864837646484,
|
| 784 |
+
"epoch": 0.3584,
|
| 785 |
+
"grad_norm": 0.08270971477031708,
|
| 786 |
+
"kl": 0.006146430969238281,
|
| 787 |
+
"learning_rate": 2.4352347027881005e-06,
|
| 788 |
+
"loss": 0.0246,
|
| 789 |
+
"reward": 0.6484375149011612,
|
| 790 |
+
"reward_std": 0.2801168202422559,
|
| 791 |
+
"rewards/accuracy_reward": 0.6484375149011612,
|
| 792 |
+
"rewards/format_reward": 0.0,
|
| 793 |
+
"step": 56
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"clip_ratio": 0.0,
|
| 797 |
+
"completion_length": 572.1093940734863,
|
| 798 |
+
"epoch": 0.3648,
|
| 799 |
+
"grad_norm": 0.14963379502296448,
|
| 800 |
+
"kl": 0.010631561279296875,
|
| 801 |
+
"learning_rate": 2.40868506576123e-06,
|
| 802 |
+
"loss": 0.0169,
|
| 803 |
+
"reward": 0.721354179084301,
|
| 804 |
+
"reward_std": 0.28047467954456806,
|
| 805 |
+
"rewards/accuracy_reward": 0.721354179084301,
|
| 806 |
+
"rewards/format_reward": 0.0,
|
| 807 |
+
"step": 57
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"clip_ratio": 0.0,
|
| 811 |
+
"completion_length": 585.2968826293945,
|
| 812 |
+
"epoch": 0.3712,
|
| 813 |
+
"grad_norm": 0.11959049105644226,
|
| 814 |
+
"kl": 0.0067119598388671875,
|
| 815 |
+
"learning_rate": 2.3816778784387097e-06,
|
| 816 |
+
"loss": 0.0509,
|
| 817 |
+
"reward": 0.6640625223517418,
|
| 818 |
+
"reward_std": 0.2813582103699446,
|
| 819 |
+
"rewards/accuracy_reward": 0.6640625223517418,
|
| 820 |
+
"rewards/format_reward": 0.0,
|
| 821 |
+
"step": 58
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"clip_ratio": 0.0,
|
| 825 |
+
"completion_length": 615.3672103881836,
|
| 826 |
+
"epoch": 0.3776,
|
| 827 |
+
"grad_norm": 0.07675585150718689,
|
| 828 |
+
"kl": 0.0048236846923828125,
|
| 829 |
+
"learning_rate": 2.354226739752678e-06,
|
| 830 |
+
"loss": 0.0357,
|
| 831 |
+
"reward": 0.677083358168602,
|
| 832 |
+
"reward_std": 0.23673555441200733,
|
| 833 |
+
"rewards/accuracy_reward": 0.677083358168602,
|
| 834 |
+
"rewards/format_reward": 0.0,
|
| 835 |
+
"step": 59
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"clip_ratio": 0.0,
|
| 839 |
+
"completion_length": 598.4244995117188,
|
| 840 |
+
"epoch": 0.384,
|
| 841 |
+
"grad_norm": 0.10844859480857849,
|
| 842 |
+
"kl": 0.0065975189208984375,
|
| 843 |
+
"learning_rate": 2.3263454721781537e-06,
|
| 844 |
+
"loss": 0.0641,
|
| 845 |
+
"reward": 0.6015625074505806,
|
| 846 |
+
"reward_std": 0.2933562882244587,
|
| 847 |
+
"rewards/accuracy_reward": 0.6015625074505806,
|
| 848 |
+
"rewards/format_reward": 0.0,
|
| 849 |
+
"step": 60
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"clip_ratio": 0.0,
|
| 853 |
+
"completion_length": 588.8229293823242,
|
| 854 |
+
"epoch": 0.3904,
|
| 855 |
+
"grad_norm": 0.10191041976213455,
|
| 856 |
+
"kl": 0.0065097808837890625,
|
| 857 |
+
"learning_rate": 2.298048114773005e-06,
|
| 858 |
+
"loss": 0.0342,
|
| 859 |
+
"reward": 0.638020858168602,
|
| 860 |
+
"reward_std": 0.2551356879994273,
|
| 861 |
+
"rewards/accuracy_reward": 0.638020858168602,
|
| 862 |
+
"rewards/format_reward": 0.0,
|
| 863 |
+
"step": 61
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"clip_ratio": 0.0,
|
| 867 |
+
"completion_length": 623.3515777587891,
|
| 868 |
+
"epoch": 0.3968,
|
| 869 |
+
"grad_norm": 0.0932384505867958,
|
| 870 |
+
"kl": 0.028699874877929688,
|
| 871 |
+
"learning_rate": 2.2693489161088594e-06,
|
| 872 |
+
"loss": 0.0235,
|
| 873 |
+
"reward": 0.5937500223517418,
|
| 874 |
+
"reward_std": 0.23260587407276034,
|
| 875 |
+
"rewards/accuracy_reward": 0.5937500223517418,
|
| 876 |
+
"rewards/format_reward": 0.0,
|
| 877 |
+
"step": 62
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"clip_ratio": 0.0,
|
| 881 |
+
"completion_length": 587.2656402587891,
|
| 882 |
+
"epoch": 0.4032,
|
| 883 |
+
"grad_norm": 0.11250537633895874,
|
| 884 |
+
"kl": 0.007572174072265625,
|
| 885 |
+
"learning_rate": 2.2402623270965205e-06,
|
| 886 |
+
"loss": 0.0563,
|
| 887 |
+
"reward": 0.6093750223517418,
|
| 888 |
+
"reward_std": 0.27705385256558657,
|
| 889 |
+
"rewards/accuracy_reward": 0.6093750223517418,
|
| 890 |
+
"rewards/format_reward": 0.0,
|
| 891 |
+
"step": 63
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"clip_ratio": 0.0,
|
| 895 |
+
"completion_length": 587.4010543823242,
|
| 896 |
+
"epoch": 0.4096,
|
| 897 |
+
"grad_norm": 0.08997678756713867,
|
| 898 |
+
"kl": 0.0063266754150390625,
|
| 899 |
+
"learning_rate": 2.210802993709498e-06,
|
| 900 |
+
"loss": 0.0203,
|
| 901 |
+
"reward": 0.6979166865348816,
|
| 902 |
+
"reward_std": 0.2539973799139261,
|
| 903 |
+
"rewards/accuracy_reward": 0.6979166865348816,
|
| 904 |
+
"rewards/format_reward": 0.0,
|
| 905 |
+
"step": 64
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"clip_ratio": 0.0,
|
| 909 |
+
"completion_length": 558.8984527587891,
|
| 910 |
+
"epoch": 0.416,
|
| 911 |
+
"grad_norm": 0.2506275773048401,
|
| 912 |
+
"kl": 0.0138397216796875,
|
| 913 |
+
"learning_rate": 2.18098574960932e-06,
|
| 914 |
+
"loss": 0.0464,
|
| 915 |
+
"reward": 0.7317708507180214,
|
| 916 |
+
"reward_std": 0.22962365811690688,
|
| 917 |
+
"rewards/accuracy_reward": 0.7317708507180214,
|
| 918 |
+
"rewards/format_reward": 0.0,
|
| 919 |
+
"step": 65
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"clip_ratio": 0.0,
|
| 923 |
+
"completion_length": 617.0573120117188,
|
| 924 |
+
"epoch": 0.4224,
|
| 925 |
+
"grad_norm": 0.10161478817462921,
|
| 926 |
+
"kl": 0.008909225463867188,
|
| 927 |
+
"learning_rate": 2.150825608676337e-06,
|
| 928 |
+
"loss": 0.0616,
|
| 929 |
+
"reward": 0.5625000111758709,
|
| 930 |
+
"reward_std": 0.27837011963129044,
|
| 931 |
+
"rewards/accuracy_reward": 0.5625000111758709,
|
| 932 |
+
"rewards/format_reward": 0.0,
|
| 933 |
+
"step": 66
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"clip_ratio": 0.0,
|
| 937 |
+
"completion_length": 613.1146011352539,
|
| 938 |
+
"epoch": 0.4288,
|
| 939 |
+
"grad_norm": 0.10768543183803558,
|
| 940 |
+
"kl": 0.008653640747070312,
|
| 941 |
+
"learning_rate": 2.120337757449781e-06,
|
| 942 |
+
"loss": 0.0947,
|
| 943 |
+
"reward": 0.6119791865348816,
|
| 944 |
+
"reward_std": 0.3058084286749363,
|
| 945 |
+
"rewards/accuracy_reward": 0.6119791865348816,
|
| 946 |
+
"rewards/format_reward": 0.0,
|
| 947 |
+
"step": 67
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"clip_ratio": 0.0,
|
| 951 |
+
"completion_length": 609.2369956970215,
|
| 952 |
+
"epoch": 0.4352,
|
| 953 |
+
"grad_norm": 0.08021987974643707,
|
| 954 |
+
"kl": 0.00714874267578125,
|
| 955 |
+
"learning_rate": 2.0895375474808856e-06,
|
| 956 |
+
"loss": 0.0388,
|
| 957 |
+
"reward": 0.5312500149011612,
|
| 958 |
+
"reward_std": 0.26845166087150574,
|
| 959 |
+
"rewards/accuracy_reward": 0.5312500149011612,
|
| 960 |
+
"rewards/format_reward": 0.0,
|
| 961 |
+
"step": 68
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"clip_ratio": 0.0,
|
| 965 |
+
"completion_length": 591.6380348205566,
|
| 966 |
+
"epoch": 0.4416,
|
| 967 |
+
"grad_norm": 0.1357138454914093,
|
| 968 |
+
"kl": 0.006610870361328125,
|
| 969 |
+
"learning_rate": 2.0584404876029177e-06,
|
| 970 |
+
"loss": 0.0656,
|
| 971 |
+
"reward": 0.6640625149011612,
|
| 972 |
+
"reward_std": 0.2963401321321726,
|
| 973 |
+
"rewards/accuracy_reward": 0.6640625149011612,
|
| 974 |
+
"rewards/format_reward": 0.0,
|
| 975 |
+
"step": 69
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"clip_ratio": 0.0,
|
| 979 |
+
"completion_length": 611.0494995117188,
|
| 980 |
+
"epoch": 0.448,
|
| 981 |
+
"grad_norm": 0.09849099814891815,
|
| 982 |
+
"kl": 0.0051288604736328125,
|
| 983 |
+
"learning_rate": 2.027062236122014e-06,
|
| 984 |
+
"loss": 0.0589,
|
| 985 |
+
"reward": 0.6718750149011612,
|
| 986 |
+
"reward_std": 0.2714381329715252,
|
| 987 |
+
"rewards/accuracy_reward": 0.6718750149011612,
|
| 988 |
+
"rewards/format_reward": 0.0,
|
| 989 |
+
"step": 70
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"clip_ratio": 0.0,
|
| 993 |
+
"completion_length": 623.1432495117188,
|
| 994 |
+
"epoch": 0.4544,
|
| 995 |
+
"grad_norm": 0.10582008957862854,
|
| 996 |
+
"kl": 0.010004043579101562,
|
| 997 |
+
"learning_rate": 1.9954185929327507e-06,
|
| 998 |
+
"loss": 0.0483,
|
| 999 |
+
"reward": 0.6223958544433117,
|
| 1000 |
+
"reward_std": 0.24435177817940712,
|
| 1001 |
+
"rewards/accuracy_reward": 0.6223958544433117,
|
| 1002 |
+
"rewards/format_reward": 0.0,
|
| 1003 |
+
"step": 71
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"clip_ratio": 0.0,
|
| 1007 |
+
"completion_length": 568.3906402587891,
|
| 1008 |
+
"epoch": 0.4608,
|
| 1009 |
+
"grad_norm": 0.11884115636348724,
|
| 1010 |
+
"kl": 0.014675140380859375,
|
| 1011 |
+
"learning_rate": 1.963525491562421e-06,
|
| 1012 |
+
"loss": 0.0508,
|
| 1013 |
+
"reward": 0.6354166865348816,
|
| 1014 |
+
"reward_std": 0.26908400654792786,
|
| 1015 |
+
"rewards/accuracy_reward": 0.6354166865348816,
|
| 1016 |
+
"rewards/format_reward": 0.0,
|
| 1017 |
+
"step": 72
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"clip_ratio": 0.0,
|
| 1021 |
+
"completion_length": 585.3541946411133,
|
| 1022 |
+
"epoch": 0.4672,
|
| 1023 |
+
"grad_norm": 0.5041296482086182,
|
| 1024 |
+
"kl": 0.01261138916015625,
|
| 1025 |
+
"learning_rate": 1.931398991148026e-06,
|
| 1026 |
+
"loss": 0.0334,
|
| 1027 |
+
"reward": 0.6666666939854622,
|
| 1028 |
+
"reward_std": 0.25794812850654125,
|
| 1029 |
+
"rewards/accuracy_reward": 0.6666666939854622,
|
| 1030 |
+
"rewards/format_reward": 0.0,
|
| 1031 |
+
"step": 73
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"clip_ratio": 0.0,
|
| 1035 |
+
"completion_length": 571.6380348205566,
|
| 1036 |
+
"epoch": 0.4736,
|
| 1037 |
+
"grad_norm": 0.09192512929439545,
|
| 1038 |
+
"kl": 0.007915496826171875,
|
| 1039 |
+
"learning_rate": 1.8990552683500125e-06,
|
| 1040 |
+
"loss": 0.0433,
|
| 1041 |
+
"reward": 0.6796875223517418,
|
| 1042 |
+
"reward_std": 0.23698772117495537,
|
| 1043 |
+
"rewards/accuracy_reward": 0.6796875223517418,
|
| 1044 |
+
"rewards/format_reward": 0.0,
|
| 1045 |
+
"step": 74
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"clip_ratio": 0.0,
|
| 1049 |
+
"completion_length": 640.5026321411133,
|
| 1050 |
+
"epoch": 0.48,
|
| 1051 |
+
"grad_norm": 0.08046362549066544,
|
| 1052 |
+
"kl": 0.0056896209716796875,
|
| 1053 |
+
"learning_rate": 1.866510609206841e-06,
|
| 1054 |
+
"loss": 0.0414,
|
| 1055 |
+
"reward": 0.6145833507180214,
|
| 1056 |
+
"reward_std": 0.2606574771925807,
|
| 1057 |
+
"rewards/accuracy_reward": 0.6145833507180214,
|
| 1058 |
+
"rewards/format_reward": 0.0,
|
| 1059 |
+
"step": 75
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"clip_ratio": 0.0,
|
| 1063 |
+
"completion_length": 611.9713745117188,
|
| 1064 |
+
"epoch": 0.4864,
|
| 1065 |
+
"grad_norm": 0.06867582350969315,
|
| 1066 |
+
"kl": 0.0053558349609375,
|
| 1067 |
+
"learning_rate": 1.8337814009344715e-06,
|
| 1068 |
+
"loss": 0.0288,
|
| 1069 |
+
"reward": 0.6562500149011612,
|
| 1070 |
+
"reward_std": 0.24688384169712663,
|
| 1071 |
+
"rewards/accuracy_reward": 0.6562500149011612,
|
| 1072 |
+
"rewards/format_reward": 0.0,
|
| 1073 |
+
"step": 76
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"clip_ratio": 0.0,
|
| 1077 |
+
"completion_length": 630.6276168823242,
|
| 1078 |
+
"epoch": 0.4928,
|
| 1079 |
+
"grad_norm": 0.08496621996164322,
|
| 1080 |
+
"kl": 0.0078125,
|
| 1081 |
+
"learning_rate": 1.8008841236749094e-06,
|
| 1082 |
+
"loss": 0.0295,
|
| 1083 |
+
"reward": 0.6354166865348816,
|
| 1084 |
+
"reward_std": 0.24820272251963615,
|
| 1085 |
+
"rewards/accuracy_reward": 0.6354166865348816,
|
| 1086 |
+
"rewards/format_reward": 0.0,
|
| 1087 |
+
"step": 77
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"clip_ratio": 0.0,
|
| 1091 |
+
"completion_length": 613.4479293823242,
|
| 1092 |
+
"epoch": 0.4992,
|
| 1093 |
+
"grad_norm": 0.08193843811750412,
|
| 1094 |
+
"kl": 0.006076812744140625,
|
| 1095 |
+
"learning_rate": 1.7678353421979551e-06,
|
| 1096 |
+
"loss": 0.0364,
|
| 1097 |
+
"reward": 0.6536458507180214,
|
| 1098 |
+
"reward_std": 0.243922324385494,
|
| 1099 |
+
"rewards/accuracy_reward": 0.6536458507180214,
|
| 1100 |
+
"rewards/format_reward": 0.0,
|
| 1101 |
+
"step": 78
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"clip_ratio": 0.0,
|
| 1105 |
+
"completion_length": 661.4062805175781,
|
| 1106 |
+
"epoch": 0.5056,
|
| 1107 |
+
"grad_norm": 0.08683363348245621,
|
| 1108 |
+
"kl": 0.0048618316650390625,
|
| 1109 |
+
"learning_rate": 1.7346516975603465e-06,
|
| 1110 |
+
"loss": 0.0592,
|
| 1111 |
+
"reward": 0.5182291828095913,
|
| 1112 |
+
"reward_std": 0.24837740324437618,
|
| 1113 |
+
"rewards/accuracy_reward": 0.5182291828095913,
|
| 1114 |
+
"rewards/format_reward": 0.0,
|
| 1115 |
+
"step": 79
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"clip_ratio": 0.0,
|
| 1119 |
+
"completion_length": 628.0807342529297,
|
| 1120 |
+
"epoch": 0.512,
|
| 1121 |
+
"grad_norm": 0.08514772355556488,
|
| 1122 |
+
"kl": 0.008527755737304688,
|
| 1123 |
+
"learning_rate": 1.7013498987264833e-06,
|
| 1124 |
+
"loss": 0.0304,
|
| 1125 |
+
"reward": 0.6223958544433117,
|
| 1126 |
+
"reward_std": 0.21321814227849245,
|
| 1127 |
+
"rewards/accuracy_reward": 0.6223958544433117,
|
| 1128 |
+
"rewards/format_reward": 0.0,
|
| 1129 |
+
"step": 80
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"clip_ratio": 0.0,
|
| 1133 |
+
"completion_length": 585.869815826416,
|
| 1134 |
+
"epoch": 0.5184,
|
| 1135 |
+
"grad_norm": 0.08269178867340088,
|
| 1136 |
+
"kl": 0.0064334869384765625,
|
| 1137 |
+
"learning_rate": 1.6679467141549617e-06,
|
| 1138 |
+
"loss": 0.0344,
|
| 1139 |
+
"reward": 0.6119791828095913,
|
| 1140 |
+
"reward_std": 0.209374712780118,
|
| 1141 |
+
"rewards/accuracy_reward": 0.6119791828095913,
|
| 1142 |
+
"rewards/format_reward": 0.0,
|
| 1143 |
+
"step": 81
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"clip_ratio": 0.0,
|
| 1147 |
+
"completion_length": 611.2291870117188,
|
| 1148 |
+
"epoch": 0.5248,
|
| 1149 |
+
"grad_norm": 0.08860579878091812,
|
| 1150 |
+
"kl": 0.0071849822998046875,
|
| 1151 |
+
"learning_rate": 1.6344589633551504e-06,
|
| 1152 |
+
"loss": 0.0452,
|
| 1153 |
+
"reward": 0.6432291865348816,
|
| 1154 |
+
"reward_std": 0.2829308910295367,
|
| 1155 |
+
"rewards/accuracy_reward": 0.6432291865348816,
|
| 1156 |
+
"rewards/format_reward": 0.0,
|
| 1157 |
+
"step": 82
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"clip_ratio": 0.0,
|
| 1161 |
+
"completion_length": 577.3593978881836,
|
| 1162 |
+
"epoch": 0.5312,
|
| 1163 |
+
"grad_norm": 0.12256962805986404,
|
| 1164 |
+
"kl": 0.011285781860351562,
|
| 1165 |
+
"learning_rate": 1.6009035084180597e-06,
|
| 1166 |
+
"loss": 0.0362,
|
| 1167 |
+
"reward": 0.5989583618938923,
|
| 1168 |
+
"reward_std": 0.29459603875875473,
|
| 1169 |
+
"rewards/accuracy_reward": 0.5989583618938923,
|
| 1170 |
+
"rewards/format_reward": 0.0,
|
| 1171 |
+
"step": 83
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"clip_ratio": 0.0,
|
| 1175 |
+
"completion_length": 623.5781440734863,
|
| 1176 |
+
"epoch": 0.5376,
|
| 1177 |
+
"grad_norm": 0.09216293692588806,
|
| 1178 |
+
"kl": 0.006359100341796875,
|
| 1179 |
+
"learning_rate": 1.5672972455257725e-06,
|
| 1180 |
+
"loss": 0.0297,
|
| 1181 |
+
"reward": 0.6145833507180214,
|
| 1182 |
+
"reward_std": 0.2700417563319206,
|
| 1183 |
+
"rewards/accuracy_reward": 0.6145833507180214,
|
| 1184 |
+
"rewards/format_reward": 0.0,
|
| 1185 |
+
"step": 84
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"clip_ratio": 0.0,
|
| 1189 |
+
"completion_length": 594.4922103881836,
|
| 1190 |
+
"epoch": 0.544,
|
| 1191 |
+
"grad_norm": 0.074432373046875,
|
| 1192 |
+
"kl": 0.007892608642578125,
|
| 1193 |
+
"learning_rate": 1.5336570964437077e-06,
|
| 1194 |
+
"loss": 0.0215,
|
| 1195 |
+
"reward": 0.7057291865348816,
|
| 1196 |
+
"reward_std": 0.19550289679318666,
|
| 1197 |
+
"rewards/accuracy_reward": 0.7057291865348816,
|
| 1198 |
+
"rewards/format_reward": 0.0,
|
| 1199 |
+
"step": 85
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"clip_ratio": 0.0,
|
| 1203 |
+
"completion_length": 579.4895935058594,
|
| 1204 |
+
"epoch": 0.5504,
|
| 1205 |
+
"grad_norm": 0.10699887573719025,
|
| 1206 |
+
"kl": 0.006565093994140625,
|
| 1207 |
+
"learning_rate": 1.5e-06,
|
| 1208 |
+
"loss": 0.0462,
|
| 1209 |
+
"reward": 0.7109375223517418,
|
| 1210 |
+
"reward_std": 0.23190354462713003,
|
| 1211 |
+
"rewards/accuracy_reward": 0.7109375223517418,
|
| 1212 |
+
"rewards/format_reward": 0.0,
|
| 1213 |
+
"step": 86
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"clip_ratio": 0.0,
|
| 1217 |
+
"completion_length": 624.2838745117188,
|
| 1218 |
+
"epoch": 0.5568,
|
| 1219 |
+
"grad_norm": 0.10833382606506348,
|
| 1220 |
+
"kl": 0.008020401000976562,
|
| 1221 |
+
"learning_rate": 1.4663429035562928e-06,
|
| 1222 |
+
"loss": 0.0556,
|
| 1223 |
+
"reward": 0.5416666865348816,
|
| 1224 |
+
"reward_std": 0.26022475212812424,
|
| 1225 |
+
"rewards/accuracy_reward": 0.5416666865348816,
|
| 1226 |
+
"rewards/format_reward": 0.0,
|
| 1227 |
+
"step": 87
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"clip_ratio": 0.0,
|
| 1231 |
+
"completion_length": 558.2265777587891,
|
| 1232 |
+
"epoch": 0.5632,
|
| 1233 |
+
"grad_norm": 0.10274934023618698,
|
| 1234 |
+
"kl": 0.008214950561523438,
|
| 1235 |
+
"learning_rate": 1.4327027544742282e-06,
|
| 1236 |
+
"loss": 0.0489,
|
| 1237 |
+
"reward": 0.7473958507180214,
|
| 1238 |
+
"reward_std": 0.2477716295979917,
|
| 1239 |
+
"rewards/accuracy_reward": 0.7473958507180214,
|
| 1240 |
+
"rewards/format_reward": 0.0,
|
| 1241 |
+
"step": 88
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"clip_ratio": 0.0,
|
| 1245 |
+
"completion_length": 587.8672027587891,
|
| 1246 |
+
"epoch": 0.5696,
|
| 1247 |
+
"grad_norm": 0.08682374656200409,
|
| 1248 |
+
"kl": 0.0101776123046875,
|
| 1249 |
+
"learning_rate": 1.3990964915819406e-06,
|
| 1250 |
+
"loss": 0.02,
|
| 1251 |
+
"reward": 0.6432291828095913,
|
| 1252 |
+
"reward_std": 0.2609303444623947,
|
| 1253 |
+
"rewards/accuracy_reward": 0.6432291828095913,
|
| 1254 |
+
"rewards/format_reward": 0.0,
|
| 1255 |
+
"step": 89
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"clip_ratio": 0.0,
|
| 1259 |
+
"completion_length": 560.9869918823242,
|
| 1260 |
+
"epoch": 0.576,
|
| 1261 |
+
"grad_norm": 0.14457735419273376,
|
| 1262 |
+
"kl": 0.03089141845703125,
|
| 1263 |
+
"learning_rate": 1.3655410366448499e-06,
|
| 1264 |
+
"loss": 0.0723,
|
| 1265 |
+
"reward": 0.5729166846722364,
|
| 1266 |
+
"reward_std": 0.27257806807756424,
|
| 1267 |
+
"rewards/accuracy_reward": 0.5729166846722364,
|
| 1268 |
+
"rewards/format_reward": 0.0,
|
| 1269 |
+
"step": 90
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"clip_ratio": 0.0,
|
| 1273 |
+
"completion_length": 659.7422103881836,
|
| 1274 |
+
"epoch": 0.5824,
|
| 1275 |
+
"grad_norm": 0.09884804487228394,
|
| 1276 |
+
"kl": 0.007778167724609375,
|
| 1277 |
+
"learning_rate": 1.3320532858450384e-06,
|
| 1278 |
+
"loss": 0.0463,
|
| 1279 |
+
"reward": 0.6145833507180214,
|
| 1280 |
+
"reward_std": 0.28644399903714657,
|
| 1281 |
+
"rewards/accuracy_reward": 0.6145833507180214,
|
| 1282 |
+
"rewards/format_reward": 0.0,
|
| 1283 |
+
"step": 91
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"clip_ratio": 0.0,
|
| 1287 |
+
"completion_length": 699.0026245117188,
|
| 1288 |
+
"epoch": 0.5888,
|
| 1289 |
+
"grad_norm": 0.0860958844423294,
|
| 1290 |
+
"kl": 0.013418197631835938,
|
| 1291 |
+
"learning_rate": 1.2986501012735172e-06,
|
| 1292 |
+
"loss": 0.0167,
|
| 1293 |
+
"reward": 0.5078125149011612,
|
| 1294 |
+
"reward_std": 0.2878361213952303,
|
| 1295 |
+
"rewards/accuracy_reward": 0.5078125149011612,
|
| 1296 |
+
"rewards/format_reward": 0.0,
|
| 1297 |
+
"step": 92
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"clip_ratio": 0.0,
|
| 1301 |
+
"completion_length": 580.8489723205566,
|
| 1302 |
+
"epoch": 0.5952,
|
| 1303 |
+
"grad_norm": 0.07715660333633423,
|
| 1304 |
+
"kl": 0.007503509521484375,
|
| 1305 |
+
"learning_rate": 1.2653483024396534e-06,
|
| 1306 |
+
"loss": 0.0498,
|
| 1307 |
+
"reward": 0.7135416939854622,
|
| 1308 |
+
"reward_std": 0.20927490666508675,
|
| 1309 |
+
"rewards/accuracy_reward": 0.7135416939854622,
|
| 1310 |
+
"rewards/format_reward": 0.0,
|
| 1311 |
+
"step": 93
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"clip_ratio": 0.0,
|
| 1315 |
+
"completion_length": 597.4192810058594,
|
| 1316 |
+
"epoch": 0.6016,
|
| 1317 |
+
"grad_norm": 0.13049161434173584,
|
| 1318 |
+
"kl": 0.010112762451171875,
|
| 1319 |
+
"learning_rate": 1.2321646578020452e-06,
|
| 1320 |
+
"loss": 0.0603,
|
| 1321 |
+
"reward": 0.6302083507180214,
|
| 1322 |
+
"reward_std": 0.28328122943639755,
|
| 1323 |
+
"rewards/accuracy_reward": 0.6302083507180214,
|
| 1324 |
+
"rewards/format_reward": 0.0,
|
| 1325 |
+
"step": 94
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"clip_ratio": 0.0,
|
| 1329 |
+
"completion_length": 561.3333511352539,
|
| 1330 |
+
"epoch": 0.608,
|
| 1331 |
+
"grad_norm": 0.08519114553928375,
|
| 1332 |
+
"kl": 0.008464813232421875,
|
| 1333 |
+
"learning_rate": 1.199115876325091e-06,
|
| 1334 |
+
"loss": -0.005,
|
| 1335 |
+
"reward": 0.6223958507180214,
|
| 1336 |
+
"reward_std": 0.24969627987593412,
|
| 1337 |
+
"rewards/accuracy_reward": 0.6223958507180214,
|
| 1338 |
+
"rewards/format_reward": 0.0,
|
| 1339 |
+
"step": 95
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"clip_ratio": 0.0,
|
| 1343 |
+
"completion_length": 581.8073120117188,
|
| 1344 |
+
"epoch": 0.6144,
|
| 1345 |
+
"grad_norm": 0.1490948647260666,
|
| 1346 |
+
"kl": 0.0109405517578125,
|
| 1347 |
+
"learning_rate": 1.1662185990655286e-06,
|
| 1348 |
+
"loss": 0.0477,
|
| 1349 |
+
"reward": 0.6484375260770321,
|
| 1350 |
+
"reward_std": 0.24663167912513018,
|
| 1351 |
+
"rewards/accuracy_reward": 0.6484375260770321,
|
| 1352 |
+
"rewards/format_reward": 0.0,
|
| 1353 |
+
"step": 96
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"clip_ratio": 0.0,
|
| 1357 |
+
"completion_length": 606.9948043823242,
|
| 1358 |
+
"epoch": 0.6208,
|
| 1359 |
+
"grad_norm": 0.10985220223665237,
|
| 1360 |
+
"kl": 0.010274887084960938,
|
| 1361 |
+
"learning_rate": 1.1334893907931588e-06,
|
| 1362 |
+
"loss": 0.0505,
|
| 1363 |
+
"reward": 0.6614583432674408,
|
| 1364 |
+
"reward_std": 0.2865247614681721,
|
| 1365 |
+
"rewards/accuracy_reward": 0.6614583432674408,
|
| 1366 |
+
"rewards/format_reward": 0.0,
|
| 1367 |
+
"step": 97
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"clip_ratio": 0.0,
|
| 1371 |
+
"completion_length": 557.5833511352539,
|
| 1372 |
+
"epoch": 0.6272,
|
| 1373 |
+
"grad_norm": 0.10301777720451355,
|
| 1374 |
+
"kl": 0.014068603515625,
|
| 1375 |
+
"learning_rate": 1.1009447316499875e-06,
|
| 1376 |
+
"loss": 0.0434,
|
| 1377 |
+
"reward": 0.6979166865348816,
|
| 1378 |
+
"reward_std": 0.28450193256139755,
|
| 1379 |
+
"rewards/accuracy_reward": 0.6979166865348816,
|
| 1380 |
+
"rewards/format_reward": 0.0,
|
| 1381 |
+
"step": 98
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"clip_ratio": 0.0,
|
| 1385 |
+
"completion_length": 624.4713745117188,
|
| 1386 |
+
"epoch": 0.6336,
|
| 1387 |
+
"grad_norm": 0.10335499793291092,
|
| 1388 |
+
"kl": 0.0087432861328125,
|
| 1389 |
+
"learning_rate": 1.0686010088519742e-06,
|
| 1390 |
+
"loss": 0.0521,
|
| 1391 |
+
"reward": 0.609375013038516,
|
| 1392 |
+
"reward_std": 0.2595175374299288,
|
| 1393 |
+
"rewards/accuracy_reward": 0.609375013038516,
|
| 1394 |
+
"rewards/format_reward": 0.0,
|
| 1395 |
+
"step": 99
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"clip_ratio": 0.0,
|
| 1399 |
+
"completion_length": 607.7734489440918,
|
| 1400 |
+
"epoch": 0.64,
|
| 1401 |
+
"grad_norm": 0.11831183731555939,
|
| 1402 |
+
"kl": 0.010028839111328125,
|
| 1403 |
+
"learning_rate": 1.036474508437579e-06,
|
| 1404 |
+
"loss": 0.0517,
|
| 1405 |
+
"reward": 0.6796875223517418,
|
| 1406 |
+
"reward_std": 0.25354558415710926,
|
| 1407 |
+
"rewards/accuracy_reward": 0.6796875223517418,
|
| 1408 |
+
"rewards/format_reward": 0.0,
|
| 1409 |
+
"step": 100
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"clip_ratio": 0.0,
|
| 1413 |
+
"completion_length": 582.8567886352539,
|
| 1414 |
+
"epoch": 0.6464,
|
| 1415 |
+
"grad_norm": 0.08054033666849136,
|
| 1416 |
+
"kl": 0.007038116455078125,
|
| 1417 |
+
"learning_rate": 1.0045814070672498e-06,
|
| 1418 |
+
"loss": 0.0475,
|
| 1419 |
+
"reward": 0.6354166902601719,
|
| 1420 |
+
"reward_std": 0.2254233518615365,
|
| 1421 |
+
"rewards/accuracy_reward": 0.6354166902601719,
|
| 1422 |
+
"rewards/format_reward": 0.0,
|
| 1423 |
+
"step": 101
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"clip_ratio": 0.0,
|
| 1427 |
+
"completion_length": 568.1484527587891,
|
| 1428 |
+
"epoch": 0.6528,
|
| 1429 |
+
"grad_norm": 0.1000579297542572,
|
| 1430 |
+
"kl": 0.0096893310546875,
|
| 1431 |
+
"learning_rate": 9.729377638779858e-07,
|
| 1432 |
+
"loss": 0.0457,
|
| 1433 |
+
"reward": 0.6406250149011612,
|
| 1434 |
+
"reward_std": 0.3015099400654435,
|
| 1435 |
+
"rewards/accuracy_reward": 0.6406250149011612,
|
| 1436 |
+
"rewards/format_reward": 0.0,
|
| 1437 |
+
"step": 102
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"clip_ratio": 0.0,
|
| 1441 |
+
"completion_length": 532.4843978881836,
|
| 1442 |
+
"epoch": 0.6592,
|
| 1443 |
+
"grad_norm": 0.13967515528202057,
|
| 1444 |
+
"kl": 0.012454986572265625,
|
| 1445 |
+
"learning_rate": 9.415595123970823e-07,
|
| 1446 |
+
"loss": 0.0449,
|
| 1447 |
+
"reward": 0.6536458544433117,
|
| 1448 |
+
"reward_std": 0.19421059591695666,
|
| 1449 |
+
"rewards/accuracy_reward": 0.6536458544433117,
|
| 1450 |
+
"rewards/format_reward": 0.0,
|
| 1451 |
+
"step": 103
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"clip_ratio": 0.0,
|
| 1455 |
+
"completion_length": 615.4062728881836,
|
| 1456 |
+
"epoch": 0.6656,
|
| 1457 |
+
"grad_norm": 0.09052874892950058,
|
| 1458 |
+
"kl": 0.009624481201171875,
|
| 1459 |
+
"learning_rate": 9.104624525191147e-07,
|
| 1460 |
+
"loss": 0.0311,
|
| 1461 |
+
"reward": 0.6536458544433117,
|
| 1462 |
+
"reward_std": 0.24321836233139038,
|
| 1463 |
+
"rewards/accuracy_reward": 0.6536458544433117,
|
| 1464 |
+
"rewards/format_reward": 0.0,
|
| 1465 |
+
"step": 104
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"clip_ratio": 0.0,
|
| 1469 |
+
"completion_length": 620.0807495117188,
|
| 1470 |
+
"epoch": 0.672,
|
| 1471 |
+
"grad_norm": 0.10262376815080643,
|
| 1472 |
+
"kl": 0.009487152099609375,
|
| 1473 |
+
"learning_rate": 8.796622425502193e-07,
|
| 1474 |
+
"loss": 0.0438,
|
| 1475 |
+
"reward": 0.6510416865348816,
|
| 1476 |
+
"reward_std": 0.21620950940996408,
|
| 1477 |
+
"rewards/accuracy_reward": 0.6510416865348816,
|
| 1478 |
+
"rewards/format_reward": 0.0,
|
| 1479 |
+
"step": 105
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"clip_ratio": 0.0,
|
| 1483 |
+
"completion_length": 680.0390739440918,
|
| 1484 |
+
"epoch": 0.6784,
|
| 1485 |
+
"grad_norm": 0.07680626958608627,
|
| 1486 |
+
"kl": 0.007625579833984375,
|
| 1487 |
+
"learning_rate": 8.49174391323663e-07,
|
| 1488 |
+
"loss": 0.0273,
|
| 1489 |
+
"reward": 0.5182291828095913,
|
| 1490 |
+
"reward_std": 0.2917035147547722,
|
| 1491 |
+
"rewards/accuracy_reward": 0.5182291828095913,
|
| 1492 |
+
"rewards/format_reward": 0.0,
|
| 1493 |
+
"step": 106
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"clip_ratio": 0.0,
|
| 1497 |
+
"completion_length": 607.0390777587891,
|
| 1498 |
+
"epoch": 0.6848,
|
| 1499 |
+
"grad_norm": 0.13828518986701965,
|
| 1500 |
+
"kl": 0.008594512939453125,
|
| 1501 |
+
"learning_rate": 8.190142503906799e-07,
|
| 1502 |
+
"loss": 0.0475,
|
| 1503 |
+
"reward": 0.6250000149011612,
|
| 1504 |
+
"reward_std": 0.26452159881591797,
|
| 1505 |
+
"rewards/accuracy_reward": 0.6250000149011612,
|
| 1506 |
+
"rewards/format_reward": 0.0,
|
| 1507 |
+
"step": 107
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"clip_ratio": 0.0,
|
| 1511 |
+
"completion_length": 641.8073120117188,
|
| 1512 |
+
"epoch": 0.6912,
|
| 1513 |
+
"grad_norm": 0.13834954798221588,
|
| 1514 |
+
"kl": 0.0077571868896484375,
|
| 1515 |
+
"learning_rate": 7.891970062905021e-07,
|
| 1516 |
+
"loss": 0.0547,
|
| 1517 |
+
"reward": 0.5520833507180214,
|
| 1518 |
+
"reward_std": 0.29983909241855145,
|
| 1519 |
+
"rewards/accuracy_reward": 0.5520833507180214,
|
| 1520 |
+
"rewards/format_reward": 0.0,
|
| 1521 |
+
"step": 108
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"clip_ratio": 0.0,
|
| 1525 |
+
"completion_length": 613.0078315734863,
|
| 1526 |
+
"epoch": 0.6976,
|
| 1527 |
+
"grad_norm": 0.12272358685731888,
|
| 1528 |
+
"kl": 0.015625,
|
| 1529 |
+
"learning_rate": 7.597376729034797e-07,
|
| 1530 |
+
"loss": 0.053,
|
| 1531 |
+
"reward": 0.6041666828095913,
|
| 1532 |
+
"reward_std": 0.3014292009174824,
|
| 1533 |
+
"rewards/accuracy_reward": 0.6041666828095913,
|
| 1534 |
+
"rewards/format_reward": 0.0,
|
| 1535 |
+
"step": 109
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"clip_ratio": 0.0,
|
| 1539 |
+
"completion_length": 630.8359603881836,
|
| 1540 |
+
"epoch": 0.704,
|
| 1541 |
+
"grad_norm": 0.10720554739236832,
|
| 1542 |
+
"kl": 0.0121307373046875,
|
| 1543 |
+
"learning_rate": 7.30651083891141e-07,
|
| 1544 |
+
"loss": 0.0327,
|
| 1545 |
+
"reward": 0.630208358168602,
|
| 1546 |
+
"reward_std": 0.2655659755691886,
|
| 1547 |
+
"rewards/accuracy_reward": 0.630208358168602,
|
| 1548 |
+
"rewards/format_reward": 0.0,
|
| 1549 |
+
"step": 110
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"clip_ratio": 0.0,
|
| 1553 |
+
"completion_length": 585.4557495117188,
|
| 1554 |
+
"epoch": 0.7104,
|
| 1555 |
+
"grad_norm": 0.13796207308769226,
|
| 1556 |
+
"kl": 0.00902557373046875,
|
| 1557 |
+
"learning_rate": 7.019518852269954e-07,
|
| 1558 |
+
"loss": 0.0601,
|
| 1559 |
+
"reward": 0.6692708507180214,
|
| 1560 |
+
"reward_std": 0.296089599840343,
|
| 1561 |
+
"rewards/accuracy_reward": 0.6692708507180214,
|
| 1562 |
+
"rewards/format_reward": 0.0,
|
| 1563 |
+
"step": 111
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"clip_ratio": 0.0,
|
| 1567 |
+
"completion_length": 669.9349212646484,
|
| 1568 |
+
"epoch": 0.7168,
|
| 1569 |
+
"grad_norm": 0.09197106957435608,
|
| 1570 |
+
"kl": 0.008930206298828125,
|
| 1571 |
+
"learning_rate": 6.736545278218464e-07,
|
| 1572 |
+
"loss": 0.0422,
|
| 1573 |
+
"reward": 0.5364583507180214,
|
| 1574 |
+
"reward_std": 0.2704744841903448,
|
| 1575 |
+
"rewards/accuracy_reward": 0.5364583507180214,
|
| 1576 |
+
"rewards/format_reward": 0.0,
|
| 1577 |
+
"step": 112
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"clip_ratio": 0.0,
|
| 1581 |
+
"completion_length": 634.4739837646484,
|
| 1582 |
+
"epoch": 0.7232,
|
| 1583 |
+
"grad_norm": 0.0923609808087349,
|
| 1584 |
+
"kl": 0.010822296142578125,
|
| 1585 |
+
"learning_rate": 6.457732602473221e-07,
|
| 1586 |
+
"loss": 0.0331,
|
| 1587 |
+
"reward": 0.5677083618938923,
|
| 1588 |
+
"reward_std": 0.3102851826697588,
|
| 1589 |
+
"rewards/accuracy_reward": 0.5677083618938923,
|
| 1590 |
+
"rewards/format_reward": 0.0,
|
| 1591 |
+
"step": 113
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"clip_ratio": 0.0,
|
| 1595 |
+
"completion_length": 636.7994918823242,
|
| 1596 |
+
"epoch": 0.7296,
|
| 1597 |
+
"grad_norm": 0.081573486328125,
|
| 1598 |
+
"kl": 0.006977081298828125,
|
| 1599 |
+
"learning_rate": 6.183221215612905e-07,
|
| 1600 |
+
"loss": 0.0137,
|
| 1601 |
+
"reward": 0.7005208507180214,
|
| 1602 |
+
"reward_std": 0.2628549886867404,
|
| 1603 |
+
"rewards/accuracy_reward": 0.7005208507180214,
|
| 1604 |
+
"rewards/format_reward": 0.0,
|
| 1605 |
+
"step": 114
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"clip_ratio": 0.0,
|
| 1609 |
+
"completion_length": 666.0521049499512,
|
| 1610 |
+
"epoch": 0.736,
|
| 1611 |
+
"grad_norm": 0.07901042699813843,
|
| 1612 |
+
"kl": 0.010074615478515625,
|
| 1613 |
+
"learning_rate": 5.913149342387704e-07,
|
| 1614 |
+
"loss": -0.0007,
|
| 1615 |
+
"reward": 0.5442708507180214,
|
| 1616 |
+
"reward_std": 0.24364782264456153,
|
| 1617 |
+
"rewards/accuracy_reward": 0.5442708507180214,
|
| 1618 |
+
"rewards/format_reward": 0.0,
|
| 1619 |
+
"step": 115
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"clip_ratio": 0.0,
|
| 1623 |
+
"completion_length": 593.3750228881836,
|
| 1624 |
+
"epoch": 0.7424,
|
| 1625 |
+
"grad_norm": 2.1408307552337646,
|
| 1626 |
+
"kl": 0.11521148681640625,
|
| 1627 |
+
"learning_rate": 5.647652972118998e-07,
|
| 1628 |
+
"loss": 0.0378,
|
| 1629 |
+
"reward": 0.6328125149011612,
|
| 1630 |
+
"reward_std": 0.23286391608417034,
|
| 1631 |
+
"rewards/accuracy_reward": 0.6328125149011612,
|
| 1632 |
+
"rewards/format_reward": 0.0,
|
| 1633 |
+
"step": 116
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"clip_ratio": 0.0,
|
| 1637 |
+
"completion_length": 615.010440826416,
|
| 1638 |
+
"epoch": 0.7488,
|
| 1639 |
+
"grad_norm": 0.1459038108587265,
|
| 1640 |
+
"kl": 0.017993927001953125,
|
| 1641 |
+
"learning_rate": 5.386865790224638e-07,
|
| 1642 |
+
"loss": 0.0406,
|
| 1643 |
+
"reward": 0.6145833507180214,
|
| 1644 |
+
"reward_std": 0.27855165116488934,
|
| 1645 |
+
"rewards/accuracy_reward": 0.6145833507180214,
|
| 1646 |
+
"rewards/format_reward": 0.0,
|
| 1647 |
+
"step": 117
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"clip_ratio": 0.0,
|
| 1651 |
+
"completion_length": 598.5078353881836,
|
| 1652 |
+
"epoch": 0.7552,
|
| 1653 |
+
"grad_norm": 0.10137172043323517,
|
| 1654 |
+
"kl": 0.009675979614257812,
|
| 1655 |
+
"learning_rate": 5.130919110904311e-07,
|
| 1656 |
+
"loss": 0.0382,
|
| 1657 |
+
"reward": 0.6718750223517418,
|
| 1658 |
+
"reward_std": 0.2865214981138706,
|
| 1659 |
+
"rewards/accuracy_reward": 0.6718750223517418,
|
| 1660 |
+
"rewards/format_reward": 0.0,
|
| 1661 |
+
"step": 118
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"clip_ratio": 0.0,
|
| 1665 |
+
"completion_length": 589.893253326416,
|
| 1666 |
+
"epoch": 0.7616,
|
| 1667 |
+
"grad_norm": 0.12197765707969666,
|
| 1668 |
+
"kl": 0.011661529541015625,
|
| 1669 |
+
"learning_rate": 4.879941811018885e-07,
|
| 1670 |
+
"loss": 0.0467,
|
| 1671 |
+
"reward": 0.6041666865348816,
|
| 1672 |
+
"reward_std": 0.2229677839204669,
|
| 1673 |
+
"rewards/accuracy_reward": 0.6041666865348816,
|
| 1674 |
+
"rewards/format_reward": 0.0,
|
| 1675 |
+
"step": 119
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"clip_ratio": 0.0,
|
| 1679 |
+
"completion_length": 542.5468940734863,
|
| 1680 |
+
"epoch": 0.768,
|
| 1681 |
+
"grad_norm": 0.16627652943134308,
|
| 1682 |
+
"kl": 0.011745452880859375,
|
| 1683 |
+
"learning_rate": 4.63406026519703e-07,
|
| 1684 |
+
"loss": 0.0432,
|
| 1685 |
+
"reward": 0.6927083395421505,
|
| 1686 |
+
"reward_std": 0.22858351049944758,
|
| 1687 |
+
"rewards/accuracy_reward": 0.6927083395421505,
|
| 1688 |
+
"rewards/format_reward": 0.0,
|
| 1689 |
+
"step": 120
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"clip_ratio": 0.0,
|
| 1693 |
+
"completion_length": 621.4323120117188,
|
| 1694 |
+
"epoch": 0.7744,
|
| 1695 |
+
"grad_norm": 0.12212401628494263,
|
| 1696 |
+
"kl": 0.011493682861328125,
|
| 1697 |
+
"learning_rate": 4.3933982822017883e-07,
|
| 1698 |
+
"loss": 0.0686,
|
| 1699 |
+
"reward": 0.6406250186264515,
|
| 1700 |
+
"reward_std": 0.20165704423561692,
|
| 1701 |
+
"rewards/accuracy_reward": 0.6406250186264515,
|
| 1702 |
+
"rewards/format_reward": 0.0,
|
| 1703 |
+
"step": 121
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"clip_ratio": 0.0,
|
| 1707 |
+
"completion_length": 629.1536674499512,
|
| 1708 |
+
"epoch": 0.7808,
|
| 1709 |
+
"grad_norm": 0.19540025293827057,
|
| 1710 |
+
"kl": 0.013004302978515625,
|
| 1711 |
+
"learning_rate": 4.158077042589128e-07,
|
| 1712 |
+
"loss": 0.0758,
|
| 1713 |
+
"reward": 0.6067708507180214,
|
| 1714 |
+
"reward_std": 0.2859147386625409,
|
| 1715 |
+
"rewards/accuracy_reward": 0.6067708507180214,
|
| 1716 |
+
"rewards/format_reward": 0.0,
|
| 1717 |
+
"step": 122
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"clip_ratio": 0.0,
|
| 1721 |
+
"completion_length": 636.9297142028809,
|
| 1722 |
+
"epoch": 0.7872,
|
| 1723 |
+
"grad_norm": 0.12284351140260696,
|
| 1724 |
+
"kl": 0.012508392333984375,
|
| 1725 |
+
"learning_rate": 3.928215037689892e-07,
|
| 1726 |
+
"loss": 0.0419,
|
| 1727 |
+
"reward": 0.549479179084301,
|
| 1728 |
+
"reward_std": 0.30940655805170536,
|
| 1729 |
+
"rewards/accuracy_reward": 0.549479179084301,
|
| 1730 |
+
"rewards/format_reward": 0.0,
|
| 1731 |
+
"step": 123
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"clip_ratio": 0.0,
|
| 1735 |
+
"completion_length": 639.7604293823242,
|
| 1736 |
+
"epoch": 0.7936,
|
| 1737 |
+
"grad_norm": 0.11035679280757904,
|
| 1738 |
+
"kl": 0.009258270263671875,
|
| 1739 |
+
"learning_rate": 3.703928009945837e-07,
|
| 1740 |
+
"loss": 0.0234,
|
| 1741 |
+
"reward": 0.5651041902601719,
|
| 1742 |
+
"reward_std": 0.22610236518085003,
|
| 1743 |
+
"rewards/accuracy_reward": 0.5651041902601719,
|
| 1744 |
+
"rewards/format_reward": 0.0,
|
| 1745 |
+
"step": 124
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"clip_ratio": 0.0,
|
| 1749 |
+
"completion_length": 595.6744956970215,
|
| 1750 |
+
"epoch": 0.8,
|
| 1751 |
+
"grad_norm": 1.7241010665893555,
|
| 1752 |
+
"kl": 0.01811981201171875,
|
| 1753 |
+
"learning_rate": 3.4853288946298335e-07,
|
| 1754 |
+
"loss": 0.0198,
|
| 1755 |
+
"reward": 0.6536458507180214,
|
| 1756 |
+
"reward_std": 0.22094332706183195,
|
| 1757 |
+
"rewards/accuracy_reward": 0.6536458507180214,
|
| 1758 |
+
"rewards/format_reward": 0.0,
|
| 1759 |
+
"step": 125
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"clip_ratio": 0.0,
|
| 1763 |
+
"completion_length": 588.1692886352539,
|
| 1764 |
+
"epoch": 0.8064,
|
| 1765 |
+
"grad_norm": 0.15116219222545624,
|
| 1766 |
+
"kl": 0.01129913330078125,
|
| 1767 |
+
"learning_rate": 3.272527762979553e-07,
|
| 1768 |
+
"loss": 0.0436,
|
| 1769 |
+
"reward": 0.6848958507180214,
|
| 1770 |
+
"reward_std": 0.28214290738105774,
|
| 1771 |
+
"rewards/accuracy_reward": 0.6848958507180214,
|
| 1772 |
+
"rewards/format_reward": 0.0,
|
| 1773 |
+
"step": 126
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"clip_ratio": 0.0,
|
| 1777 |
+
"completion_length": 626.7161674499512,
|
| 1778 |
+
"epoch": 0.8128,
|
| 1779 |
+
"grad_norm": 0.11378590762615204,
|
| 1780 |
+
"kl": 0.01619720458984375,
|
| 1781 |
+
"learning_rate": 3.065631766773286e-07,
|
| 1782 |
+
"loss": 0.0478,
|
| 1783 |
+
"reward": 0.6250000223517418,
|
| 1784 |
+
"reward_std": 0.2649602070450783,
|
| 1785 |
+
"rewards/accuracy_reward": 0.6250000223517418,
|
| 1786 |
+
"rewards/format_reward": 0.0,
|
| 1787 |
+
"step": 127
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"clip_ratio": 0.0,
|
| 1791 |
+
"completion_length": 592.3489761352539,
|
| 1792 |
+
"epoch": 0.8192,
|
| 1793 |
+
"grad_norm": 0.09930500388145447,
|
| 1794 |
+
"kl": 0.015716552734375,
|
| 1795 |
+
"learning_rate": 2.86474508437579e-07,
|
| 1796 |
+
"loss": 0.0528,
|
| 1797 |
+
"reward": 0.6718750223517418,
|
| 1798 |
+
"reward_std": 0.28171182051301,
|
| 1799 |
+
"rewards/accuracy_reward": 0.6718750223517418,
|
| 1800 |
+
"rewards/format_reward": 0.0,
|
| 1801 |
+
"step": 128
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"clip_ratio": 0.0,
|
| 1805 |
+
"completion_length": 578.8854331970215,
|
| 1806 |
+
"epoch": 0.8256,
|
| 1807 |
+
"grad_norm": 0.10038118809461594,
|
| 1808 |
+
"kl": 0.00980377197265625,
|
| 1809 |
+
"learning_rate": 2.669968868281353e-07,
|
| 1810 |
+
"loss": 0.0554,
|
| 1811 |
+
"reward": 0.6666666865348816,
|
| 1812 |
+
"reward_std": 0.28783776238560677,
|
| 1813 |
+
"rewards/accuracy_reward": 0.6666666865348816,
|
| 1814 |
+
"rewards/format_reward": 0.0,
|
| 1815 |
+
"step": 129
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"clip_ratio": 0.0,
|
| 1819 |
+
"completion_length": 644.4713745117188,
|
| 1820 |
+
"epoch": 0.832,
|
| 1821 |
+
"grad_norm": 0.10138770937919617,
|
| 1822 |
+
"kl": 0.009368896484375,
|
| 1823 |
+
"learning_rate": 2.48140119418046e-07,
|
| 1824 |
+
"loss": 0.0495,
|
| 1825 |
+
"reward": 0.5963541828095913,
|
| 1826 |
+
"reward_std": 0.2717949990183115,
|
| 1827 |
+
"rewards/accuracy_reward": 0.5963541828095913,
|
| 1828 |
+
"rewards/format_reward": 0.0,
|
| 1829 |
+
"step": 130
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"clip_ratio": 0.0,
|
| 1833 |
+
"completion_length": 667.6406402587891,
|
| 1834 |
+
"epoch": 0.8384,
|
| 1835 |
+
"grad_norm": 0.10386987030506134,
|
| 1836 |
+
"kl": 0.010986328125,
|
| 1837 |
+
"learning_rate": 2.299137011575738e-07,
|
| 1838 |
+
"loss": 0.0519,
|
| 1839 |
+
"reward": 0.5677083469927311,
|
| 1840 |
+
"reward_std": 0.252684373408556,
|
| 1841 |
+
"rewards/accuracy_reward": 0.5677083469927311,
|
| 1842 |
+
"rewards/format_reward": 0.0,
|
| 1843 |
+
"step": 131
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"clip_ratio": 0.0,
|
| 1847 |
+
"completion_length": 647.0000228881836,
|
| 1848 |
+
"epoch": 0.8448,
|
| 1849 |
+
"grad_norm": 0.10453839600086212,
|
| 1850 |
+
"kl": 0.00896453857421875,
|
| 1851 |
+
"learning_rate": 2.1232680959720086e-07,
|
| 1852 |
+
"loss": 0.0571,
|
| 1853 |
+
"reward": 0.6197916865348816,
|
| 1854 |
+
"reward_std": 0.24627807177603245,
|
| 1855 |
+
"rewards/accuracy_reward": 0.6197916865348816,
|
| 1856 |
+
"rewards/format_reward": 0.0,
|
| 1857 |
+
"step": 132
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"clip_ratio": 0.0,
|
| 1861 |
+
"completion_length": 604.8047103881836,
|
| 1862 |
+
"epoch": 0.8512,
|
| 1863 |
+
"grad_norm": 0.12539909780025482,
|
| 1864 |
+
"kl": 0.01335906982421875,
|
| 1865 |
+
"learning_rate": 1.9538830026645698e-07,
|
| 1866 |
+
"loss": 0.0501,
|
| 1867 |
+
"reward": 0.6484375186264515,
|
| 1868 |
+
"reward_std": 0.25513568613678217,
|
| 1869 |
+
"rewards/accuracy_reward": 0.6484375186264515,
|
| 1870 |
+
"rewards/format_reward": 0.0,
|
| 1871 |
+
"step": 133
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"clip_ratio": 0.0,
|
| 1875 |
+
"completion_length": 599.2812728881836,
|
| 1876 |
+
"epoch": 0.8576,
|
| 1877 |
+
"grad_norm": 0.17024974524974823,
|
| 1878 |
+
"kl": 0.01447296142578125,
|
| 1879 |
+
"learning_rate": 1.79106702214893e-07,
|
| 1880 |
+
"loss": 0.0758,
|
| 1881 |
+
"reward": 0.6145833507180214,
|
| 1882 |
+
"reward_std": 0.2468089684844017,
|
| 1883 |
+
"rewards/accuracy_reward": 0.6145833507180214,
|
| 1884 |
+
"rewards/format_reward": 0.0,
|
| 1885 |
+
"step": 134
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"clip_ratio": 0.0,
|
| 1889 |
+
"completion_length": 627.3645935058594,
|
| 1890 |
+
"epoch": 0.864,
|
| 1891 |
+
"grad_norm": 0.10592558979988098,
|
| 1892 |
+
"kl": 0.00984954833984375,
|
| 1893 |
+
"learning_rate": 1.634902137174483e-07,
|
| 1894 |
+
"loss": 0.0531,
|
| 1895 |
+
"reward": 0.6223958544433117,
|
| 1896 |
+
"reward_std": 0.3007177282124758,
|
| 1897 |
+
"rewards/accuracy_reward": 0.6223958544433117,
|
| 1898 |
+
"rewards/format_reward": 0.0,
|
| 1899 |
+
"step": 135
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"clip_ratio": 0.0,
|
| 1903 |
+
"completion_length": 564.7031364440918,
|
| 1904 |
+
"epoch": 0.8704,
|
| 1905 |
+
"grad_norm": 0.2481531798839569,
|
| 1906 |
+
"kl": 0.01848602294921875,
|
| 1907 |
+
"learning_rate": 1.4854669814637145e-07,
|
| 1908 |
+
"loss": 0.0728,
|
| 1909 |
+
"reward": 0.6640625149011612,
|
| 1910 |
+
"reward_std": 0.3039671443402767,
|
| 1911 |
+
"rewards/accuracy_reward": 0.6640625149011612,
|
| 1912 |
+
"rewards/format_reward": 0.0,
|
| 1913 |
+
"step": 136
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"clip_ratio": 0.0,
|
| 1917 |
+
"completion_length": 666.7812652587891,
|
| 1918 |
+
"epoch": 0.8768,
|
| 1919 |
+
"grad_norm": 0.14774362742900848,
|
| 1920 |
+
"kl": 0.009685516357421875,
|
| 1921 |
+
"learning_rate": 1.342836800117762e-07,
|
| 1922 |
+
"loss": 0.05,
|
| 1923 |
+
"reward": 0.6015625149011612,
|
| 1924 |
+
"reward_std": 0.25645456463098526,
|
| 1925 |
+
"rewards/accuracy_reward": 0.6015625149011612,
|
| 1926 |
+
"rewards/format_reward": 0.0,
|
| 1927 |
+
"step": 137
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"clip_ratio": 0.0,
|
| 1931 |
+
"completion_length": 624.0573196411133,
|
| 1932 |
+
"epoch": 0.8832,
|
| 1933 |
+
"grad_norm": 0.12596814334392548,
|
| 1934 |
+
"kl": 0.010166168212890625,
|
| 1935 |
+
"learning_rate": 1.2070834117282415e-07,
|
| 1936 |
+
"loss": 0.0874,
|
| 1937 |
+
"reward": 0.6458333507180214,
|
| 1938 |
+
"reward_std": 0.2881148587912321,
|
| 1939 |
+
"rewards/accuracy_reward": 0.6458333507180214,
|
| 1940 |
+
"rewards/format_reward": 0.0,
|
| 1941 |
+
"step": 138
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"clip_ratio": 0.0,
|
| 1945 |
+
"completion_length": 633.1666831970215,
|
| 1946 |
+
"epoch": 0.8896,
|
| 1947 |
+
"grad_norm": 0.12767468392848969,
|
| 1948 |
+
"kl": 0.0145111083984375,
|
| 1949 |
+
"learning_rate": 1.0782751722144174e-07,
|
| 1950 |
+
"loss": 0.0462,
|
| 1951 |
+
"reward": 0.6145833469927311,
|
| 1952 |
+
"reward_std": 0.2834601551294327,
|
| 1953 |
+
"rewards/accuracy_reward": 0.6145833469927311,
|
| 1954 |
+
"rewards/format_reward": 0.0,
|
| 1955 |
+
"step": 139
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"clip_ratio": 0.0,
|
| 1959 |
+
"completion_length": 580.5312652587891,
|
| 1960 |
+
"epoch": 0.896,
|
| 1961 |
+
"grad_norm": 0.2081136852502823,
|
| 1962 |
+
"kl": 0.0196380615234375,
|
| 1963 |
+
"learning_rate": 9.564769404039419e-08,
|
| 1964 |
+
"loss": 0.0399,
|
| 1965 |
+
"reward": 0.6250000111758709,
|
| 1966 |
+
"reward_std": 0.26424711383879185,
|
| 1967 |
+
"rewards/accuracy_reward": 0.6250000111758709,
|
| 1968 |
+
"rewards/format_reward": 0.0,
|
| 1969 |
+
"step": 140
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"clip_ratio": 0.0,
|
| 1973 |
+
"completion_length": 676.6041870117188,
|
| 1974 |
+
"epoch": 0.9024,
|
| 1975 |
+
"grad_norm": 0.0916285365819931,
|
| 1976 |
+
"kl": 0.015132904052734375,
|
| 1977 |
+
"learning_rate": 8.417500453744864e-08,
|
| 1978 |
+
"loss": 0.0177,
|
| 1979 |
+
"reward": 0.5078125093132257,
|
| 1980 |
+
"reward_std": 0.23154503665864468,
|
| 1981 |
+
"rewards/accuracy_reward": 0.5078125093132257,
|
| 1982 |
+
"rewards/format_reward": 0.0,
|
| 1983 |
+
"step": 141
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"clip_ratio": 0.0,
|
| 1987 |
+
"completion_length": 582.924488067627,
|
| 1988 |
+
"epoch": 0.9088,
|
| 1989 |
+
"grad_norm": 0.15574556589126587,
|
| 1990 |
+
"kl": 0.01641082763671875,
|
| 1991 |
+
"learning_rate": 7.341522555726971e-08,
|
| 1992 |
+
"loss": 0.0592,
|
| 1993 |
+
"reward": 0.7005208507180214,
|
| 1994 |
+
"reward_std": 0.27556355856359005,
|
| 1995 |
+
"rewards/accuracy_reward": 0.7005208507180214,
|
| 1996 |
+
"rewards/format_reward": 0.0,
|
| 1997 |
+
"step": 142
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"clip_ratio": 0.0,
|
| 2001 |
+
"completion_length": 576.0130424499512,
|
| 2002 |
+
"epoch": 0.9152,
|
| 2003 |
+
"grad_norm": 0.11581158638000488,
|
| 2004 |
+
"kl": 0.01715087890625,
|
| 2005 |
+
"learning_rate": 6.337377497260432e-08,
|
| 2006 |
+
"loss": 0.055,
|
| 2007 |
+
"reward": 0.6770833432674408,
|
| 2008 |
+
"reward_std": 0.27591717429459095,
|
| 2009 |
+
"rewards/accuracy_reward": 0.6770833432674408,
|
| 2010 |
+
"rewards/format_reward": 0.0,
|
| 2011 |
+
"step": 143
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"clip_ratio": 0.0,
|
| 2015 |
+
"completion_length": 630.3619918823242,
|
| 2016 |
+
"epoch": 0.9216,
|
| 2017 |
+
"grad_norm": 0.22534584999084473,
|
| 2018 |
+
"kl": 0.019634246826171875,
|
| 2019 |
+
"learning_rate": 5.405570895622014e-08,
|
| 2020 |
+
"loss": 0.0865,
|
| 2021 |
+
"reward": 0.6406250260770321,
|
| 2022 |
+
"reward_std": 0.3224628400057554,
|
| 2023 |
+
"rewards/accuracy_reward": 0.6406250260770321,
|
| 2024 |
+
"rewards/format_reward": 0.0,
|
| 2025 |
+
"step": 144
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"clip_ratio": 0.0,
|
| 2029 |
+
"completion_length": 641.0911712646484,
|
| 2030 |
+
"epoch": 0.928,
|
| 2031 |
+
"grad_norm": 0.12959367036819458,
|
| 2032 |
+
"kl": 0.012157440185546875,
|
| 2033 |
+
"learning_rate": 4.546571943496969e-08,
|
| 2034 |
+
"loss": 0.0707,
|
| 2035 |
+
"reward": 0.6458333358168602,
|
| 2036 |
+
"reward_std": 0.31299781054258347,
|
| 2037 |
+
"rewards/accuracy_reward": 0.6458333358168602,
|
| 2038 |
+
"rewards/format_reward": 0.0,
|
| 2039 |
+
"step": 145
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"clip_ratio": 0.0,
|
| 2043 |
+
"completion_length": 622.8593902587891,
|
| 2044 |
+
"epoch": 0.9344,
|
| 2045 |
+
"grad_norm": 0.10907571762800217,
|
| 2046 |
+
"kl": 0.01004791259765625,
|
| 2047 |
+
"learning_rate": 3.7608131727264573e-08,
|
| 2048 |
+
"loss": 0.0394,
|
| 2049 |
+
"reward": 0.7005208656191826,
|
| 2050 |
+
"reward_std": 0.29626264050602913,
|
| 2051 |
+
"rewards/accuracy_reward": 0.7005208656191826,
|
| 2052 |
+
"rewards/format_reward": 0.0,
|
| 2053 |
+
"step": 146
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"clip_ratio": 0.0,
|
| 2057 |
+
"completion_length": 641.1328392028809,
|
| 2058 |
+
"epoch": 0.9408,
|
| 2059 |
+
"grad_norm": 0.10614655166864395,
|
| 2060 |
+
"kl": 0.01285552978515625,
|
| 2061 |
+
"learning_rate": 3.048690236514679e-08,
|
| 2062 |
+
"loss": 0.0218,
|
| 2063 |
+
"reward": 0.614583358168602,
|
| 2064 |
+
"reward_std": 0.28029411286115646,
|
| 2065 |
+
"rewards/accuracy_reward": 0.614583358168602,
|
| 2066 |
+
"rewards/format_reward": 0.0,
|
| 2067 |
+
"step": 147
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"clip_ratio": 0.0,
|
| 2071 |
+
"completion_length": 581.5442886352539,
|
| 2072 |
+
"epoch": 0.9472,
|
| 2073 |
+
"grad_norm": 0.11954043060541153,
|
| 2074 |
+
"kl": 0.017839431762695312,
|
| 2075 |
+
"learning_rate": 2.41056171020555e-08,
|
| 2076 |
+
"loss": 0.02,
|
| 2077 |
+
"reward": 0.6223958507180214,
|
| 2078 |
+
"reward_std": 0.27459665294736624,
|
| 2079 |
+
"rewards/accuracy_reward": 0.6223958507180214,
|
| 2080 |
+
"rewards/format_reward": 0.0,
|
| 2081 |
+
"step": 148
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"clip_ratio": 0.0,
|
| 2085 |
+
"completion_length": 623.4896049499512,
|
| 2086 |
+
"epoch": 0.9536,
|
| 2087 |
+
"grad_norm": 0.1879308968782425,
|
| 2088 |
+
"kl": 0.016384124755859375,
|
| 2089 |
+
"learning_rate": 1.846748910729351e-08,
|
| 2090 |
+
"loss": 0.0891,
|
| 2091 |
+
"reward": 0.6614583507180214,
|
| 2092 |
+
"reward_std": 0.3012502696365118,
|
| 2093 |
+
"rewards/accuracy_reward": 0.6614583507180214,
|
| 2094 |
+
"rewards/format_reward": 0.0,
|
| 2095 |
+
"step": 149
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"clip_ratio": 0.0,
|
| 2099 |
+
"completion_length": 587.9974098205566,
|
| 2100 |
+
"epoch": 0.96,
|
| 2101 |
+
"grad_norm": 0.15574875473976135,
|
| 2102 |
+
"kl": 0.010019302368164062,
|
| 2103 |
+
"learning_rate": 1.357535734809795e-08,
|
| 2104 |
+
"loss": 0.0974,
|
| 2105 |
+
"reward": 0.6927083432674408,
|
| 2106 |
+
"reward_std": 0.2541763046756387,
|
| 2107 |
+
"rewards/accuracy_reward": 0.6927083432674408,
|
| 2108 |
+
"rewards/format_reward": 0.0,
|
| 2109 |
+
"step": 150
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"clip_ratio": 0.0,
|
| 2113 |
+
"completion_length": 584.0442848205566,
|
| 2114 |
+
"epoch": 0.9664,
|
| 2115 |
+
"grad_norm": 0.13488668203353882,
|
| 2116 |
+
"kl": 0.02130889892578125,
|
| 2117 |
+
"learning_rate": 9.431685160136094e-09,
|
| 2118 |
+
"loss": 0.062,
|
| 2119 |
+
"reward": 0.6692708507180214,
|
| 2120 |
+
"reward_std": 0.2626819461584091,
|
| 2121 |
+
"rewards/accuracy_reward": 0.6692708507180214,
|
| 2122 |
+
"rewards/format_reward": 0.0,
|
| 2123 |
+
"step": 151
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"clip_ratio": 0.0,
|
| 2127 |
+
"completion_length": 618.4479293823242,
|
| 2128 |
+
"epoch": 0.9728,
|
| 2129 |
+
"grad_norm": 0.157786026597023,
|
| 2130 |
+
"kl": 0.017688751220703125,
|
| 2131 |
+
"learning_rate": 6.038559007141398e-09,
|
| 2132 |
+
"loss": 0.0492,
|
| 2133 |
+
"reward": 0.5546875149011612,
|
| 2134 |
+
"reward_std": 0.2780907116830349,
|
| 2135 |
+
"rewards/accuracy_reward": 0.5546875149011612,
|
| 2136 |
+
"rewards/format_reward": 0.0,
|
| 2137 |
+
"step": 152
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"clip_ratio": 0.0,
|
| 2141 |
+
"completion_length": 579.3203239440918,
|
| 2142 |
+
"epoch": 0.9792,
|
| 2143 |
+
"grad_norm": 0.09835051000118256,
|
| 2144 |
+
"kl": 0.011753082275390625,
|
| 2145 |
+
"learning_rate": 3.3976874303158767e-09,
|
| 2146 |
+
"loss": 0.0607,
|
| 2147 |
+
"reward": 0.6640625186264515,
|
| 2148 |
+
"reward_std": 0.19717375095933676,
|
| 2149 |
+
"rewards/accuracy_reward": 0.6640625186264515,
|
| 2150 |
+
"rewards/format_reward": 0.0,
|
| 2151 |
+
"step": 153
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"clip_ratio": 0.0,
|
| 2155 |
+
"completion_length": 643.7239723205566,
|
| 2156 |
+
"epoch": 0.9856,
|
| 2157 |
+
"grad_norm": 0.13841365277767181,
|
| 2158 |
+
"kl": 0.01361846923828125,
|
| 2159 |
+
"learning_rate": 1.5104001880281159e-09,
|
| 2160 |
+
"loss": 0.0547,
|
| 2161 |
+
"reward": 0.5494791828095913,
|
| 2162 |
+
"reward_std": 0.23084108112379909,
|
| 2163 |
+
"rewards/accuracy_reward": 0.5494791828095913,
|
| 2164 |
+
"rewards/format_reward": 0.0,
|
| 2165 |
+
"step": 154
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"clip_ratio": 0.0,
|
| 2169 |
+
"completion_length": 609.2422027587891,
|
| 2170 |
+
"epoch": 0.992,
|
| 2171 |
+
"grad_norm": 0.12696102261543274,
|
| 2172 |
+
"kl": 0.009654998779296875,
|
| 2173 |
+
"learning_rate": 3.77647586240204e-10,
|
| 2174 |
+
"loss": 0.0576,
|
| 2175 |
+
"reward": 0.6380208507180214,
|
| 2176 |
+
"reward_std": 0.3347429148852825,
|
| 2177 |
+
"rewards/accuracy_reward": 0.6380208507180214,
|
| 2178 |
+
"rewards/format_reward": 0.0,
|
| 2179 |
+
"step": 155
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"clip_ratio": 0.0,
|
| 2183 |
+
"completion_length": 572.3437652587891,
|
| 2184 |
+
"epoch": 0.9984,
|
| 2185 |
+
"grad_norm": 0.1173073798418045,
|
| 2186 |
+
"kl": 0.01114654541015625,
|
| 2187 |
+
"learning_rate": 0.0,
|
| 2188 |
+
"loss": 0.0226,
|
| 2189 |
+
"reward": 0.6302083507180214,
|
| 2190 |
+
"reward_std": 0.2554892925545573,
|
| 2191 |
+
"rewards/accuracy_reward": 0.6302083507180214,
|
| 2192 |
+
"rewards/format_reward": 0.0,
|
| 2193 |
+
"step": 156
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.9984,
|
| 2197 |
+
"step": 156,
|
| 2198 |
+
"total_flos": 0.0,
|
| 2199 |
+
"train_loss": 0.045506071436998,
|
| 2200 |
+
"train_runtime": 12746.7121,
|
| 2201 |
+
"train_samples_per_second": 0.588,
|
| 2202 |
+
"train_steps_per_second": 0.012
|
| 2203 |
+
}
|
| 2204 |
+
],
|
| 2205 |
+
"logging_steps": 1,
|
| 2206 |
+
"max_steps": 156,
|
| 2207 |
+
"num_input_tokens_seen": 0,
|
| 2208 |
+
"num_train_epochs": 1,
|
| 2209 |
+
"save_steps": 500,
|
| 2210 |
+
"stateful_callbacks": {
|
| 2211 |
+
"TrainerControl": {
|
| 2212 |
+
"args": {
|
| 2213 |
+
"should_epoch_stop": false,
|
| 2214 |
+
"should_evaluate": false,
|
| 2215 |
+
"should_log": false,
|
| 2216 |
+
"should_save": true,
|
| 2217 |
+
"should_training_stop": true
|
| 2218 |
+
},
|
| 2219 |
+
"attributes": {}
|
| 2220 |
+
}
|
| 2221 |
+
},
|
| 2222 |
+
"total_flos": 0.0,
|
| 2223 |
+
"train_batch_size": 16,
|
| 2224 |
+
"trial_name": null,
|
| 2225 |
+
"trial_params": null
|
| 2226 |
+
}
|