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 +2787 -0
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen2.5-7B
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen2.5-7B-Open-R1-GRPO-math-selected-default
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- grpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen2.5-7B-Open-R1-GRPO-math-selected-default
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B).
|
| 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="Lansechen/Qwen2.5-7B-Open-R1-GRPO-math-selected-default", 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/chenran1995-the-chinese-university-of-hong-kong/huggingface/runs/9qshrijc)
|
| 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
|
| 38 |
+
- Transformers: 4.50.0
|
| 39 |
+
- Pytorch: 2.5.1+cu121
|
| 40 |
+
- Datasets: 3.5.0
|
| 41 |
+
- Tokenizers: 0.21.1
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
Cite GRPO as:
|
| 46 |
+
|
| 47 |
+
```bibtex
|
| 48 |
+
@article{zhihong2024deepseekmath,
|
| 49 |
+
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
|
| 50 |
+
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
|
| 51 |
+
year = 2024,
|
| 52 |
+
eprint = {arXiv:2402.03300},
|
| 53 |
+
}
|
| 54 |
+
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
Cite TRL as:
|
| 58 |
+
|
| 59 |
+
```bibtex
|
| 60 |
+
@misc{vonwerra2022trl,
|
| 61 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 62 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
|
| 63 |
+
year = 2020,
|
| 64 |
+
journal = {GitHub repository},
|
| 65 |
+
publisher = {GitHub},
|
| 66 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 67 |
+
}
|
| 68 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.020176043500826808,
|
| 4 |
+
"train_runtime": 41639.173,
|
| 5 |
+
"train_samples": 11040,
|
| 6 |
+
"train_samples_per_second": 0.53,
|
| 7 |
+
"train_steps_per_second": 0.005
|
| 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": "4.50.0"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.020176043500826808,
|
| 4 |
+
"train_runtime": 41639.173,
|
| 5 |
+
"train_samples": 11040,
|
| 6 |
+
"train_samples_per_second": 0.53,
|
| 7 |
+
"train_steps_per_second": 0.005
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2787 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.9936628643852978,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 196,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"clip_ratio": 0.0,
|
| 14 |
+
"completion_length": 515.8303871154785,
|
| 15 |
+
"epoch": 0.010139416983523447,
|
| 16 |
+
"grad_norm": 0.4515054523944855,
|
| 17 |
+
"learning_rate": 5e-08,
|
| 18 |
+
"loss": 0.0123,
|
| 19 |
+
"num_tokens": 610184.0,
|
| 20 |
+
"reward": 0.26004465483129025,
|
| 21 |
+
"reward_std": 0.3379838764667511,
|
| 22 |
+
"rewards/accuracy_reward": 0.1986607126891613,
|
| 23 |
+
"rewards/format_reward": 0.06138392956927419,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"clip_ratio": 0.0,
|
| 28 |
+
"completion_length": 467.8660888671875,
|
| 29 |
+
"epoch": 0.020278833967046894,
|
| 30 |
+
"grad_norm": 0.6836522221565247,
|
| 31 |
+
"learning_rate": 1e-07,
|
| 32 |
+
"loss": 0.0058,
|
| 33 |
+
"num_tokens": 1164448.0,
|
| 34 |
+
"reward": 0.24776786752045155,
|
| 35 |
+
"reward_std": 0.3431246988475323,
|
| 36 |
+
"rewards/accuracy_reward": 0.171875,
|
| 37 |
+
"rewards/format_reward": 0.07589285867288709,
|
| 38 |
+
"step": 2
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"clip_ratio": 0.0,
|
| 42 |
+
"completion_length": 528.7500343322754,
|
| 43 |
+
"epoch": 0.030418250950570342,
|
| 44 |
+
"grad_norm": 0.5222043395042419,
|
| 45 |
+
"learning_rate": 1.5e-07,
|
| 46 |
+
"loss": 0.005,
|
| 47 |
+
"num_tokens": 1778328.0,
|
| 48 |
+
"reward": 0.26450893841683865,
|
| 49 |
+
"reward_std": 0.3584165945649147,
|
| 50 |
+
"rewards/accuracy_reward": 0.2187499962747097,
|
| 51 |
+
"rewards/format_reward": 0.045758930034935474,
|
| 52 |
+
"step": 3
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"clip_ratio": 0.0,
|
| 56 |
+
"completion_length": 508.4922103881836,
|
| 57 |
+
"epoch": 0.04055766793409379,
|
| 58 |
+
"grad_norm": 0.4442610442638397,
|
| 59 |
+
"learning_rate": 2e-07,
|
| 60 |
+
"loss": 0.0255,
|
| 61 |
+
"num_tokens": 2370889.0,
|
| 62 |
+
"reward": 0.26004465110599995,
|
| 63 |
+
"reward_std": 0.3489377610385418,
|
| 64 |
+
"rewards/accuracy_reward": 0.19642856996506453,
|
| 65 |
+
"rewards/format_reward": 0.06361607229337096,
|
| 66 |
+
"step": 4
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"clip_ratio": 0.0,
|
| 70 |
+
"completion_length": 519.3482360839844,
|
| 71 |
+
"epoch": 0.050697084917617236,
|
| 72 |
+
"grad_norm": 0.504645824432373,
|
| 73 |
+
"learning_rate": 2.5e-07,
|
| 74 |
+
"loss": 0.0363,
|
| 75 |
+
"num_tokens": 2963009.0,
|
| 76 |
+
"reward": 0.2700893022119999,
|
| 77 |
+
"reward_std": 0.3758334591984749,
|
| 78 |
+
"rewards/accuracy_reward": 0.1897321417927742,
|
| 79 |
+
"rewards/format_reward": 0.08035714202560484,
|
| 80 |
+
"step": 5
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"clip_ratio": 0.0,
|
| 84 |
+
"completion_length": 509.5201072692871,
|
| 85 |
+
"epoch": 0.060836501901140684,
|
| 86 |
+
"grad_norm": 0.4883345365524292,
|
| 87 |
+
"learning_rate": 3e-07,
|
| 88 |
+
"loss": 0.0146,
|
| 89 |
+
"num_tokens": 3548731.0,
|
| 90 |
+
"reward": 0.2924107275903225,
|
| 91 |
+
"reward_std": 0.3819224089384079,
|
| 92 |
+
"rewards/accuracy_reward": 0.21875,
|
| 93 |
+
"rewards/format_reward": 0.07366071455180645,
|
| 94 |
+
"step": 6
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"clip_ratio": 0.0,
|
| 98 |
+
"completion_length": 506.1562728881836,
|
| 99 |
+
"epoch": 0.07097591888466413,
|
| 100 |
+
"grad_norm": 0.7536768317222595,
|
| 101 |
+
"learning_rate": 3.5e-07,
|
| 102 |
+
"loss": -0.0085,
|
| 103 |
+
"num_tokens": 4126671.0,
|
| 104 |
+
"reward": 0.3169642947614193,
|
| 105 |
+
"reward_std": 0.37958647310733795,
|
| 106 |
+
"rewards/accuracy_reward": 0.2555803544819355,
|
| 107 |
+
"rewards/format_reward": 0.06138392980210483,
|
| 108 |
+
"step": 7
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"clip_ratio": 0.0,
|
| 112 |
+
"completion_length": 512.7143173217773,
|
| 113 |
+
"epoch": 0.08111533586818757,
|
| 114 |
+
"grad_norm": 0.5751662850379944,
|
| 115 |
+
"learning_rate": 4e-07,
|
| 116 |
+
"loss": -0.0186,
|
| 117 |
+
"num_tokens": 4713615.0,
|
| 118 |
+
"reward": 0.2879464440047741,
|
| 119 |
+
"reward_std": 0.3758578822016716,
|
| 120 |
+
"rewards/accuracy_reward": 0.2087053544819355,
|
| 121 |
+
"rewards/format_reward": 0.07924107136204839,
|
| 122 |
+
"step": 8
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"clip_ratio": 0.0,
|
| 126 |
+
"completion_length": 527.1506843566895,
|
| 127 |
+
"epoch": 0.09125475285171103,
|
| 128 |
+
"grad_norm": 0.727789044380188,
|
| 129 |
+
"learning_rate": 4.5e-07,
|
| 130 |
+
"loss": 0.0193,
|
| 131 |
+
"num_tokens": 5309270.0,
|
| 132 |
+
"reward": 0.28794644214212894,
|
| 133 |
+
"reward_std": 0.38178177177906036,
|
| 134 |
+
"rewards/accuracy_reward": 0.19866071455180645,
|
| 135 |
+
"rewards/format_reward": 0.08928571408614516,
|
| 136 |
+
"step": 9
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"clip_ratio": 0.0,
|
| 140 |
+
"completion_length": 462.6607437133789,
|
| 141 |
+
"epoch": 0.10139416983523447,
|
| 142 |
+
"grad_norm": 0.6003168821334839,
|
| 143 |
+
"learning_rate": 5e-07,
|
| 144 |
+
"loss": 0.0127,
|
| 145 |
+
"num_tokens": 5850542.0,
|
| 146 |
+
"reward": 0.3270089440047741,
|
| 147 |
+
"reward_std": 0.42656392976641655,
|
| 148 |
+
"rewards/accuracy_reward": 0.2176339253783226,
|
| 149 |
+
"rewards/format_reward": 0.109375,
|
| 150 |
+
"step": 10
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"clip_ratio": 0.0,
|
| 154 |
+
"completion_length": 500.8236846923828,
|
| 155 |
+
"epoch": 0.11153358681875793,
|
| 156 |
+
"grad_norm": 0.6242951154708862,
|
| 157 |
+
"learning_rate": 5.5e-07,
|
| 158 |
+
"loss": 0.0068,
|
| 159 |
+
"num_tokens": 6436816.0,
|
| 160 |
+
"reward": 0.3191964440047741,
|
| 161 |
+
"reward_std": 0.4202234297990799,
|
| 162 |
+
"rewards/accuracy_reward": 0.171875,
|
| 163 |
+
"rewards/format_reward": 0.14732142817229033,
|
| 164 |
+
"step": 11
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"clip_ratio": 0.0,
|
| 168 |
+
"completion_length": 517.0971221923828,
|
| 169 |
+
"epoch": 0.12167300380228137,
|
| 170 |
+
"grad_norm": 1.3958944082260132,
|
| 171 |
+
"learning_rate": 6e-07,
|
| 172 |
+
"loss": 0.0251,
|
| 173 |
+
"num_tokens": 7031447.0,
|
| 174 |
+
"reward": 0.3738839402794838,
|
| 175 |
+
"reward_std": 0.4328419789671898,
|
| 176 |
+
"rewards/accuracy_reward": 0.19531249720603228,
|
| 177 |
+
"rewards/format_reward": 0.17857142724096775,
|
| 178 |
+
"step": 12
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"clip_ratio": 0.0,
|
| 182 |
+
"completion_length": 540.4654197692871,
|
| 183 |
+
"epoch": 0.13181242078580482,
|
| 184 |
+
"grad_norm": 1.9689440727233887,
|
| 185 |
+
"learning_rate": 6.5e-07,
|
| 186 |
+
"loss": 0.0087,
|
| 187 |
+
"num_tokens": 7643888.0,
|
| 188 |
+
"reward": 0.4464285969734192,
|
| 189 |
+
"reward_std": 0.5018927529454231,
|
| 190 |
+
"rewards/accuracy_reward": 0.2310267835855484,
|
| 191 |
+
"rewards/format_reward": 0.2154017835855484,
|
| 192 |
+
"step": 13
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"clip_ratio": 0.0,
|
| 196 |
+
"completion_length": 489.197566986084,
|
| 197 |
+
"epoch": 0.14195183776932827,
|
| 198 |
+
"grad_norm": 0.798046350479126,
|
| 199 |
+
"learning_rate": 7e-07,
|
| 200 |
+
"loss": -0.0009,
|
| 201 |
+
"num_tokens": 8214385.0,
|
| 202 |
+
"reward": 0.4988839514553547,
|
| 203 |
+
"reward_std": 0.5417631343007088,
|
| 204 |
+
"rewards/accuracy_reward": 0.18973214458674192,
|
| 205 |
+
"rewards/format_reward": 0.3113839253783226,
|
| 206 |
+
"step": 14
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"clip_ratio": 0.0,
|
| 210 |
+
"completion_length": 507.252254486084,
|
| 211 |
+
"epoch": 0.1520912547528517,
|
| 212 |
+
"grad_norm": 0.4880025386810303,
|
| 213 |
+
"learning_rate": 7.5e-07,
|
| 214 |
+
"loss": 0.0114,
|
| 215 |
+
"num_tokens": 8797323.0,
|
| 216 |
+
"reward": 0.7120536044239998,
|
| 217 |
+
"reward_std": 0.5666098818182945,
|
| 218 |
+
"rewards/accuracy_reward": 0.19754464272409678,
|
| 219 |
+
"rewards/format_reward": 0.5145089402794838,
|
| 220 |
+
"step": 15
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"clip_ratio": 0.0,
|
| 224 |
+
"completion_length": 464.38841247558594,
|
| 225 |
+
"epoch": 0.16223067173637515,
|
| 226 |
+
"grad_norm": 0.3857059180736542,
|
| 227 |
+
"learning_rate": 8e-07,
|
| 228 |
+
"loss": -0.0014,
|
| 229 |
+
"num_tokens": 9340343.0,
|
| 230 |
+
"reward": 0.8526786044239998,
|
| 231 |
+
"reward_std": 0.5239133685827255,
|
| 232 |
+
"rewards/accuracy_reward": 0.21540178544819355,
|
| 233 |
+
"rewards/format_reward": 0.6372767761349678,
|
| 234 |
+
"step": 16
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"clip_ratio": 0.0,
|
| 238 |
+
"completion_length": 508.8448944091797,
|
| 239 |
+
"epoch": 0.17237008871989862,
|
| 240 |
+
"grad_norm": 0.44979313015937805,
|
| 241 |
+
"learning_rate": 8.499999999999999e-07,
|
| 242 |
+
"loss": 0.0215,
|
| 243 |
+
"num_tokens": 9930780.0,
|
| 244 |
+
"reward": 0.8381696939468384,
|
| 245 |
+
"reward_std": 0.540903877466917,
|
| 246 |
+
"rewards/accuracy_reward": 0.2020089291036129,
|
| 247 |
+
"rewards/format_reward": 0.6361607238650322,
|
| 248 |
+
"step": 17
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"clip_ratio": 0.0,
|
| 252 |
+
"completion_length": 480.3794860839844,
|
| 253 |
+
"epoch": 0.18250950570342206,
|
| 254 |
+
"grad_norm": 0.32462233304977417,
|
| 255 |
+
"learning_rate": 9e-07,
|
| 256 |
+
"loss": 0.03,
|
| 257 |
+
"num_tokens": 10484544.0,
|
| 258 |
+
"reward": 0.9810268357396126,
|
| 259 |
+
"reward_std": 0.5262674763798714,
|
| 260 |
+
"rewards/accuracy_reward": 0.2488839253783226,
|
| 261 |
+
"rewards/format_reward": 0.7321428582072258,
|
| 262 |
+
"step": 18
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"clip_ratio": 0.0,
|
| 266 |
+
"completion_length": 455.48104095458984,
|
| 267 |
+
"epoch": 0.1926489226869455,
|
| 268 |
+
"grad_norm": 0.34758785367012024,
|
| 269 |
+
"learning_rate": 9.499999999999999e-07,
|
| 270 |
+
"loss": 0.023,
|
| 271 |
+
"num_tokens": 11024639.0,
|
| 272 |
+
"reward": 1.0424107536673546,
|
| 273 |
+
"reward_std": 0.5019765421748161,
|
| 274 |
+
"rewards/accuracy_reward": 0.2578124962747097,
|
| 275 |
+
"rewards/format_reward": 0.784598208963871,
|
| 276 |
+
"step": 19
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"clip_ratio": 0.0,
|
| 280 |
+
"completion_length": 525.8482437133789,
|
| 281 |
+
"epoch": 0.20278833967046894,
|
| 282 |
+
"grad_norm": 0.44885051250457764,
|
| 283 |
+
"learning_rate": 1e-06,
|
| 284 |
+
"loss": 0.044,
|
| 285 |
+
"num_tokens": 11616223.0,
|
| 286 |
+
"reward": 1.1584821939468384,
|
| 287 |
+
"reward_std": 0.47795333340764046,
|
| 288 |
+
"rewards/accuracy_reward": 0.29352678544819355,
|
| 289 |
+
"rewards/format_reward": 0.8649553582072258,
|
| 290 |
+
"step": 20
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"clip_ratio": 0.0,
|
| 294 |
+
"completion_length": 486.7299346923828,
|
| 295 |
+
"epoch": 0.21292775665399238,
|
| 296 |
+
"grad_norm": 0.5567921996116638,
|
| 297 |
+
"learning_rate": 9.999203468625015e-07,
|
| 298 |
+
"loss": 0.0113,
|
| 299 |
+
"num_tokens": 12173605.0,
|
| 300 |
+
"reward": 1.2120536416769028,
|
| 301 |
+
"reward_std": 0.46485012769699097,
|
| 302 |
+
"rewards/accuracy_reward": 0.31473214365541935,
|
| 303 |
+
"rewards/format_reward": 0.8973214328289032,
|
| 304 |
+
"step": 21
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"clip_ratio": 0.0,
|
| 308 |
+
"completion_length": 542.1808204650879,
|
| 309 |
+
"epoch": 0.22306717363751585,
|
| 310 |
+
"grad_norm": 0.3349709212779999,
|
| 311 |
+
"learning_rate": 9.99681412828496e-07,
|
| 312 |
+
"loss": 0.0304,
|
| 313 |
+
"num_tokens": 12792559.0,
|
| 314 |
+
"reward": 1.1808036118745804,
|
| 315 |
+
"reward_std": 0.4690566882491112,
|
| 316 |
+
"rewards/accuracy_reward": 0.3024553544819355,
|
| 317 |
+
"rewards/format_reward": 0.878348208963871,
|
| 318 |
+
"step": 22
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"clip_ratio": 0.0,
|
| 322 |
+
"completion_length": 527.556941986084,
|
| 323 |
+
"epoch": 0.2332065906210393,
|
| 324 |
+
"grad_norm": 0.32437098026275635,
|
| 325 |
+
"learning_rate": 9.992832740253644e-07,
|
| 326 |
+
"loss": 0.0241,
|
| 327 |
+
"num_tokens": 13403498.0,
|
| 328 |
+
"reward": 1.2622768431901932,
|
| 329 |
+
"reward_std": 0.4536837413907051,
|
| 330 |
+
"rewards/accuracy_reward": 0.3504464253783226,
|
| 331 |
+
"rewards/format_reward": 0.9118303507566452,
|
| 332 |
+
"step": 23
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"clip_ratio": 0.0,
|
| 336 |
+
"completion_length": 549.2321624755859,
|
| 337 |
+
"epoch": 0.24334600760456274,
|
| 338 |
+
"grad_norm": 0.683305025100708,
|
| 339 |
+
"learning_rate": 9.987260573051267e-07,
|
| 340 |
+
"loss": 0.0239,
|
| 341 |
+
"num_tokens": 14025154.0,
|
| 342 |
+
"reward": 1.3013393580913544,
|
| 343 |
+
"reward_std": 0.4284757189452648,
|
| 344 |
+
"rewards/accuracy_reward": 0.3816964216530323,
|
| 345 |
+
"rewards/format_reward": 0.9196428656578064,
|
| 346 |
+
"step": 24
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"clip_ratio": 0.0,
|
| 350 |
+
"completion_length": 525.8181076049805,
|
| 351 |
+
"epoch": 0.2534854245880862,
|
| 352 |
+
"grad_norm": 0.34933602809906006,
|
| 353 |
+
"learning_rate": 9.98009940204023e-07,
|
| 354 |
+
"loss": 0.0268,
|
| 355 |
+
"num_tokens": 14627615.0,
|
| 356 |
+
"reward": 1.4040179401636124,
|
| 357 |
+
"reward_std": 0.41444528847932816,
|
| 358 |
+
"rewards/accuracy_reward": 0.4531249962747097,
|
| 359 |
+
"rewards/format_reward": 0.9508928656578064,
|
| 360 |
+
"step": 25
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"clip_ratio": 0.0,
|
| 364 |
+
"completion_length": 540.1384201049805,
|
| 365 |
+
"epoch": 0.26362484157160965,
|
| 366 |
+
"grad_norm": 0.3170299828052521,
|
| 367 |
+
"learning_rate": 9.971351508859486e-07,
|
| 368 |
+
"loss": 0.0192,
|
| 369 |
+
"num_tokens": 15230411.0,
|
| 370 |
+
"reward": 1.4040178954601288,
|
| 371 |
+
"reward_std": 0.45906292274594307,
|
| 372 |
+
"rewards/accuracy_reward": 0.4698660746216774,
|
| 373 |
+
"rewards/format_reward": 0.9341517835855484,
|
| 374 |
+
"step": 26
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"clip_ratio": 0.0,
|
| 378 |
+
"completion_length": 555.2734718322754,
|
| 379 |
+
"epoch": 0.2737642585551331,
|
| 380 |
+
"grad_norm": 0.2808490991592407,
|
| 381 |
+
"learning_rate": 9.961019680697591e-07,
|
| 382 |
+
"loss": 0.0915,
|
| 383 |
+
"num_tokens": 15856112.0,
|
| 384 |
+
"reward": 1.460937574505806,
|
| 385 |
+
"reward_std": 0.4059113450348377,
|
| 386 |
+
"rewards/accuracy_reward": 0.5189732201397419,
|
| 387 |
+
"rewards/format_reward": 0.9419642835855484,
|
| 388 |
+
"step": 27
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"clip_ratio": 0.0,
|
| 392 |
+
"completion_length": 596.7377471923828,
|
| 393 |
+
"epoch": 0.28390367553865653,
|
| 394 |
+
"grad_norm": 0.2290179580450058,
|
| 395 |
+
"learning_rate": 9.949107209404663e-07,
|
| 396 |
+
"loss": 0.0252,
|
| 397 |
+
"num_tokens": 16524429.0,
|
| 398 |
+
"reward": 1.5558036267757416,
|
| 399 |
+
"reward_std": 0.35689178109169006,
|
| 400 |
+
"rewards/accuracy_reward": 0.6015625037252903,
|
| 401 |
+
"rewards/format_reward": 0.9542410671710968,
|
| 402 |
+
"step": 28
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"clip_ratio": 0.0,
|
| 406 |
+
"completion_length": 586.0000267028809,
|
| 407 |
+
"epoch": 0.29404309252217997,
|
| 408 |
+
"grad_norm": 0.2592538595199585,
|
| 409 |
+
"learning_rate": 9.935617890443554e-07,
|
| 410 |
+
"loss": 0.0619,
|
| 411 |
+
"num_tokens": 17190221.0,
|
| 412 |
+
"reward": 1.5513393580913544,
|
| 413 |
+
"reward_std": 0.34823960438370705,
|
| 414 |
+
"rewards/accuracy_reward": 0.5937500074505806,
|
| 415 |
+
"rewards/format_reward": 0.9575892984867096,
|
| 416 |
+
"step": 29
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"clip_ratio": 0.0,
|
| 420 |
+
"completion_length": 659.8370819091797,
|
| 421 |
+
"epoch": 0.3041825095057034,
|
| 422 |
+
"grad_norm": 0.1614958941936493,
|
| 423 |
+
"learning_rate": 9.92055602168058e-07,
|
| 424 |
+
"loss": 0.0418,
|
| 425 |
+
"num_tokens": 17926923.0,
|
| 426 |
+
"reward": 1.5948661416769028,
|
| 427 |
+
"reward_std": 0.315753024071455,
|
| 428 |
+
"rewards/accuracy_reward": 0.6261160783469677,
|
| 429 |
+
"rewards/format_reward": 0.9687499925494194,
|
| 430 |
+
"step": 30
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"clip_ratio": 0.0,
|
| 434 |
+
"completion_length": 628.6830673217773,
|
| 435 |
+
"epoch": 0.31432192648922685,
|
| 436 |
+
"grad_norm": 0.13593189418315887,
|
| 437 |
+
"learning_rate": 9.90392640201615e-07,
|
| 438 |
+
"loss": 0.034,
|
| 439 |
+
"num_tokens": 18617831.0,
|
| 440 |
+
"reward": 1.6785715073347092,
|
| 441 |
+
"reward_std": 0.24619607254862785,
|
| 442 |
+
"rewards/accuracy_reward": 0.6975446492433548,
|
| 443 |
+
"rewards/format_reward": 0.9810267835855484,
|
| 444 |
+
"step": 31
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"clip_ratio": 0.0,
|
| 448 |
+
"completion_length": 635.0971374511719,
|
| 449 |
+
"epoch": 0.3244613434727503,
|
| 450 |
+
"grad_norm": 0.14025020599365234,
|
| 451 |
+
"learning_rate": 9.885734329855797e-07,
|
| 452 |
+
"loss": 0.0489,
|
| 453 |
+
"num_tokens": 19306622.0,
|
| 454 |
+
"reward": 1.6830357909202576,
|
| 455 |
+
"reward_std": 0.25594867765903473,
|
| 456 |
+
"rewards/accuracy_reward": 0.7087053582072258,
|
| 457 |
+
"rewards/format_reward": 0.9743303656578064,
|
| 458 |
+
"step": 32
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"clip_ratio": 0.0,
|
| 462 |
+
"completion_length": 662.7310485839844,
|
| 463 |
+
"epoch": 0.33460076045627374,
|
| 464 |
+
"grad_norm": 0.14114561676979065,
|
| 465 |
+
"learning_rate": 9.865985601422017e-07,
|
| 466 |
+
"loss": 0.0403,
|
| 467 |
+
"num_tokens": 20034485.0,
|
| 468 |
+
"reward": 1.5926340073347092,
|
| 469 |
+
"reward_std": 0.2743757776916027,
|
| 470 |
+
"rewards/accuracy_reward": 0.6274896934628487,
|
| 471 |
+
"rewards/format_reward": 0.9698660746216774,
|
| 472 |
+
"step": 33
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"clip_ratio": 0.0,
|
| 476 |
+
"completion_length": 674.9565048217773,
|
| 477 |
+
"epoch": 0.34474017743979724,
|
| 478 |
+
"grad_norm": 0.13052859902381897,
|
| 479 |
+
"learning_rate": 9.844686508907537e-07,
|
| 480 |
+
"loss": 0.036,
|
| 481 |
+
"num_tokens": 20766734.0,
|
| 482 |
+
"reward": 1.6595983058214188,
|
| 483 |
+
"reward_std": 0.22791905514895916,
|
| 484 |
+
"rewards/accuracy_reward": 0.676339291036129,
|
| 485 |
+
"rewards/format_reward": 0.983258917927742,
|
| 486 |
+
"step": 34
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"clip_ratio": 0.0,
|
| 490 |
+
"completion_length": 661.467658996582,
|
| 491 |
+
"epoch": 0.3548795944233207,
|
| 492 |
+
"grad_norm": 0.14716672897338867,
|
| 493 |
+
"learning_rate": 9.821843838470534e-07,
|
| 494 |
+
"loss": 0.0801,
|
| 495 |
+
"num_tokens": 21502137.0,
|
| 496 |
+
"reward": 1.658482238650322,
|
| 497 |
+
"reward_std": 0.2525596469640732,
|
| 498 |
+
"rewards/accuracy_reward": 0.6774553507566452,
|
| 499 |
+
"rewards/format_reward": 0.9810267835855484,
|
| 500 |
+
"step": 35
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"clip_ratio": 0.0,
|
| 504 |
+
"completion_length": 649.9118576049805,
|
| 505 |
+
"epoch": 0.3650190114068441,
|
| 506 |
+
"grad_norm": 0.1431976705789566,
|
| 507 |
+
"learning_rate": 9.797464868072486e-07,
|
| 508 |
+
"loss": 0.0625,
|
| 509 |
+
"num_tokens": 22212570.0,
|
| 510 |
+
"reward": 1.6852679550647736,
|
| 511 |
+
"reward_std": 0.2624010592699051,
|
| 512 |
+
"rewards/accuracy_reward": 0.7098214253783226,
|
| 513 |
+
"rewards/format_reward": 0.9754464253783226,
|
| 514 |
+
"step": 36
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"clip_ratio": 0.0,
|
| 518 |
+
"completion_length": 653.161865234375,
|
| 519 |
+
"epoch": 0.37515842839036756,
|
| 520 |
+
"grad_norm": 0.10097628086805344,
|
| 521 |
+
"learning_rate": 9.771557365159319e-07,
|
| 522 |
+
"loss": 0.0255,
|
| 523 |
+
"num_tokens": 22923331.0,
|
| 524 |
+
"reward": 1.6283482909202576,
|
| 525 |
+
"reward_std": 0.19372693821787834,
|
| 526 |
+
"rewards/accuracy_reward": 0.6428571417927742,
|
| 527 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 528 |
+
"step": 37
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"clip_ratio": 0.0,
|
| 532 |
+
"completion_length": 655.9073944091797,
|
| 533 |
+
"epoch": 0.385297845373891,
|
| 534 |
+
"grad_norm": 0.1087433397769928,
|
| 535 |
+
"learning_rate": 9.744129584186597e-07,
|
| 536 |
+
"loss": 0.0387,
|
| 537 |
+
"num_tokens": 23637632.0,
|
| 538 |
+
"reward": 1.6417411416769028,
|
| 539 |
+
"reward_std": 0.2093806378543377,
|
| 540 |
+
"rewards/accuracy_reward": 0.6629464328289032,
|
| 541 |
+
"rewards/format_reward": 0.9787946417927742,
|
| 542 |
+
"step": 38
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"clip_ratio": 0.0,
|
| 546 |
+
"completion_length": 703.9297103881836,
|
| 547 |
+
"epoch": 0.39543726235741444,
|
| 548 |
+
"grad_norm": 0.10768112540245056,
|
| 549 |
+
"learning_rate": 9.71519026398956e-07,
|
| 550 |
+
"loss": 0.0512,
|
| 551 |
+
"num_tokens": 24406137.0,
|
| 552 |
+
"reward": 1.6138393729925156,
|
| 553 |
+
"reward_std": 0.19195392541587353,
|
| 554 |
+
"rewards/accuracy_reward": 0.6428571455180645,
|
| 555 |
+
"rewards/format_reward": 0.9709821417927742,
|
| 556 |
+
"step": 39
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"clip_ratio": 0.0,
|
| 560 |
+
"completion_length": 626.8917617797852,
|
| 561 |
+
"epoch": 0.4055766793409379,
|
| 562 |
+
"grad_norm": 0.11539029330015182,
|
| 563 |
+
"learning_rate": 9.68474862499881e-07,
|
| 564 |
+
"loss": 0.0277,
|
| 565 |
+
"num_tokens": 25087096.0,
|
| 566 |
+
"reward": 1.7276786267757416,
|
| 567 |
+
"reward_std": 0.16980798542499542,
|
| 568 |
+
"rewards/accuracy_reward": 0.7366071343421936,
|
| 569 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 570 |
+
"step": 40
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"clip_ratio": 0.0,
|
| 574 |
+
"completion_length": 641.186408996582,
|
| 575 |
+
"epoch": 0.4157160963244613,
|
| 576 |
+
"grad_norm": 0.0916220024228096,
|
| 577 |
+
"learning_rate": 9.652814366302568e-07,
|
| 578 |
+
"loss": 0.0297,
|
| 579 |
+
"num_tokens": 25787575.0,
|
| 580 |
+
"reward": 1.7444197237491608,
|
| 581 |
+
"reward_std": 0.15002207644283772,
|
| 582 |
+
"rewards/accuracy_reward": 0.7566964328289032,
|
| 583 |
+
"rewards/format_reward": 0.9877232015132904,
|
| 584 |
+
"step": 41
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"clip_ratio": 0.0,
|
| 588 |
+
"completion_length": 635.1529388427734,
|
| 589 |
+
"epoch": 0.42585551330798477,
|
| 590 |
+
"grad_norm": 0.1067587211728096,
|
| 591 |
+
"learning_rate": 9.619397662556433e-07,
|
| 592 |
+
"loss": 0.0296,
|
| 593 |
+
"num_tokens": 26485640.0,
|
| 594 |
+
"reward": 1.6930804401636124,
|
| 595 |
+
"reward_std": 0.1523909503594041,
|
| 596 |
+
"rewards/accuracy_reward": 0.7064732164144516,
|
| 597 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 598 |
+
"step": 42
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"clip_ratio": 0.0,
|
| 602 |
+
"completion_length": 686.982177734375,
|
| 603 |
+
"epoch": 0.43599493029150826,
|
| 604 |
+
"grad_norm": 0.09716176986694336,
|
| 605 |
+
"learning_rate": 9.5845091607416e-07,
|
| 606 |
+
"loss": 0.0241,
|
| 607 |
+
"num_tokens": 27219880.0,
|
| 608 |
+
"reward": 1.7142858058214188,
|
| 609 |
+
"reward_std": 0.19766423292458057,
|
| 610 |
+
"rewards/accuracy_reward": 0.731026791036129,
|
| 611 |
+
"rewards/format_reward": 0.9832589253783226,
|
| 612 |
+
"step": 43
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"clip_ratio": 0.0,
|
| 616 |
+
"completion_length": 692.6172103881836,
|
| 617 |
+
"epoch": 0.4461343472750317,
|
| 618 |
+
"grad_norm": 0.0998157262802124,
|
| 619 |
+
"learning_rate": 9.548159976772592e-07,
|
| 620 |
+
"loss": 0.0304,
|
| 621 |
+
"num_tokens": 27970513.0,
|
| 622 |
+
"reward": 1.6975447237491608,
|
| 623 |
+
"reward_std": 0.2129265610128641,
|
| 624 |
+
"rewards/accuracy_reward": 0.7120535746216774,
|
| 625 |
+
"rewards/format_reward": 0.9854910597205162,
|
| 626 |
+
"step": 44
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"clip_ratio": 0.0,
|
| 630 |
+
"completion_length": 717.6428833007812,
|
| 631 |
+
"epoch": 0.45627376425855515,
|
| 632 |
+
"grad_norm": 0.110740527510643,
|
| 633 |
+
"learning_rate": 9.510361691955606e-07,
|
| 634 |
+
"loss": 0.0306,
|
| 635 |
+
"num_tokens": 28735969.0,
|
| 636 |
+
"reward": 1.6662947237491608,
|
| 637 |
+
"reward_std": 0.25920049473643303,
|
| 638 |
+
"rewards/accuracy_reward": 0.6796875,
|
| 639 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 640 |
+
"step": 45
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"clip_ratio": 0.0,
|
| 644 |
+
"completion_length": 651.1373062133789,
|
| 645 |
+
"epoch": 0.4664131812420786,
|
| 646 |
+
"grad_norm": 0.10963339358568192,
|
| 647 |
+
"learning_rate": 9.471126349298556e-07,
|
| 648 |
+
"loss": 0.031,
|
| 649 |
+
"num_tokens": 29449604.0,
|
| 650 |
+
"reward": 1.7377232760190964,
|
| 651 |
+
"reward_std": 0.19720236957073212,
|
| 652 |
+
"rewards/accuracy_reward": 0.75,
|
| 653 |
+
"rewards/format_reward": 0.987723208963871,
|
| 654 |
+
"step": 46
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"clip_ratio": 0.0,
|
| 658 |
+
"completion_length": 658.0803909301758,
|
| 659 |
+
"epoch": 0.47655259822560203,
|
| 660 |
+
"grad_norm": 0.11225859075784683,
|
| 661 |
+
"learning_rate": 9.430466449674013e-07,
|
| 662 |
+
"loss": 0.0419,
|
| 663 |
+
"num_tokens": 30187596.0,
|
| 664 |
+
"reward": 1.656250074505806,
|
| 665 |
+
"reward_std": 0.19186341715976596,
|
| 666 |
+
"rewards/accuracy_reward": 0.6662946417927742,
|
| 667 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 668 |
+
"step": 47
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"clip_ratio": 0.0,
|
| 672 |
+
"completion_length": 715.3538360595703,
|
| 673 |
+
"epoch": 0.4866920152091255,
|
| 674 |
+
"grad_norm": 0.09579791873693466,
|
| 675 |
+
"learning_rate": 9.388394947836278e-07,
|
| 676 |
+
"loss": 0.0133,
|
| 677 |
+
"num_tokens": 30955289.0,
|
| 678 |
+
"reward": 1.6250000596046448,
|
| 679 |
+
"reward_std": 0.1923097288236022,
|
| 680 |
+
"rewards/accuracy_reward": 0.6372767835855484,
|
| 681 |
+
"rewards/format_reward": 0.987723208963871,
|
| 682 |
+
"step": 48
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"clip_ratio": 0.0,
|
| 686 |
+
"completion_length": 694.5614242553711,
|
| 687 |
+
"epoch": 0.4968314321926489,
|
| 688 |
+
"grad_norm": 0.10001735389232635,
|
| 689 |
+
"learning_rate": 9.344925248293835e-07,
|
| 690 |
+
"loss": 0.0175,
|
| 691 |
+
"num_tokens": 31698624.0,
|
| 692 |
+
"reward": 1.6986608058214188,
|
| 693 |
+
"reward_std": 0.1780099142342806,
|
| 694 |
+
"rewards/accuracy_reward": 0.7087053582072258,
|
| 695 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 696 |
+
"step": 49
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"clip_ratio": 0.0,
|
| 700 |
+
"completion_length": 673.1484756469727,
|
| 701 |
+
"epoch": 0.5069708491761724,
|
| 702 |
+
"grad_norm": 0.10267457365989685,
|
| 703 |
+
"learning_rate": 9.300071201038501e-07,
|
| 704 |
+
"loss": 0.0246,
|
| 705 |
+
"num_tokens": 32427941.0,
|
| 706 |
+
"reward": 1.6919643878936768,
|
| 707 |
+
"reward_std": 0.203399364836514,
|
| 708 |
+
"rewards/accuracy_reward": 0.703125,
|
| 709 |
+
"rewards/format_reward": 0.9888392835855484,
|
| 710 |
+
"step": 50
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"clip_ratio": 0.0,
|
| 714 |
+
"completion_length": 617.7199020385742,
|
| 715 |
+
"epoch": 0.5171102661596958,
|
| 716 |
+
"grad_norm": 0.10793677717447281,
|
| 717 |
+
"learning_rate": 9.253847097132655e-07,
|
| 718 |
+
"loss": 0.0196,
|
| 719 |
+
"num_tokens": 33112242.0,
|
| 720 |
+
"reward": 1.7388393580913544,
|
| 721 |
+
"reward_std": 0.20277155563235283,
|
| 722 |
+
"rewards/accuracy_reward": 0.7477678582072258,
|
| 723 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 724 |
+
"step": 51
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"clip_ratio": 0.0,
|
| 728 |
+
"completion_length": 659.7522659301758,
|
| 729 |
+
"epoch": 0.5272496831432193,
|
| 730 |
+
"grad_norm": 0.09107255190610886,
|
| 731 |
+
"learning_rate": 9.206267664155906e-07,
|
| 732 |
+
"loss": 0.034,
|
| 733 |
+
"num_tokens": 33823756.0,
|
| 734 |
+
"reward": 1.7142857909202576,
|
| 735 |
+
"reward_std": 0.15979534853249788,
|
| 736 |
+
"rewards/accuracy_reward": 0.7243303582072258,
|
| 737 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 738 |
+
"step": 52
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"clip_ratio": 0.0,
|
| 742 |
+
"completion_length": 681.6015930175781,
|
| 743 |
+
"epoch": 0.5373891001267427,
|
| 744 |
+
"grad_norm": 0.10401283204555511,
|
| 745 |
+
"learning_rate": 9.157348061512726e-07,
|
| 746 |
+
"loss": 0.0255,
|
| 747 |
+
"num_tokens": 34570551.0,
|
| 748 |
+
"reward": 1.6975447088479996,
|
| 749 |
+
"reward_std": 0.19325901474803686,
|
| 750 |
+
"rewards/accuracy_reward": 0.7131696417927742,
|
| 751 |
+
"rewards/format_reward": 0.9843749925494194,
|
| 752 |
+
"step": 53
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"clip_ratio": 0.0,
|
| 756 |
+
"completion_length": 733.9632034301758,
|
| 757 |
+
"epoch": 0.5475285171102662,
|
| 758 |
+
"grad_norm": 0.09657347947359085,
|
| 759 |
+
"learning_rate": 9.107103875602458e-07,
|
| 760 |
+
"loss": 0.0269,
|
| 761 |
+
"num_tokens": 35346734.0,
|
| 762 |
+
"reward": 1.665178656578064,
|
| 763 |
+
"reward_std": 0.19283783435821533,
|
| 764 |
+
"rewards/accuracy_reward": 0.6808035671710968,
|
| 765 |
+
"rewards/format_reward": 0.9843749925494194,
|
| 766 |
+
"step": 54
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"clip_ratio": 0.0,
|
| 770 |
+
"completion_length": 695.3471298217773,
|
| 771 |
+
"epoch": 0.5576679340937896,
|
| 772 |
+
"grad_norm": 0.11091943085193634,
|
| 773 |
+
"learning_rate": 9.055551114853295e-07,
|
| 774 |
+
"loss": 0.0383,
|
| 775 |
+
"num_tokens": 36100157.0,
|
| 776 |
+
"reward": 1.6729911416769028,
|
| 777 |
+
"reward_std": 0.2295996304601431,
|
| 778 |
+
"rewards/accuracy_reward": 0.6874999925494194,
|
| 779 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 780 |
+
"step": 55
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"clip_ratio": 0.0,
|
| 784 |
+
"completion_length": 623.5301666259766,
|
| 785 |
+
"epoch": 0.5678073510773131,
|
| 786 |
+
"grad_norm": 0.1108093112707138,
|
| 787 |
+
"learning_rate": 9.002706204621802e-07,
|
| 788 |
+
"loss": 0.0293,
|
| 789 |
+
"num_tokens": 36782520.0,
|
| 790 |
+
"reward": 1.7477679401636124,
|
| 791 |
+
"reward_std": 0.20551220327615738,
|
| 792 |
+
"rewards/accuracy_reward": 0.7544642835855484,
|
| 793 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 794 |
+
"step": 56
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"clip_ratio": 0.0,
|
| 798 |
+
"completion_length": 666.6451263427734,
|
| 799 |
+
"epoch": 0.5779467680608364,
|
| 800 |
+
"grad_norm": 0.10158971697092056,
|
| 801 |
+
"learning_rate": 8.948585981959578e-07,
|
| 802 |
+
"loss": 0.0202,
|
| 803 |
+
"num_tokens": 37511418.0,
|
| 804 |
+
"reward": 1.7053572535514832,
|
| 805 |
+
"reward_std": 0.17258675582706928,
|
| 806 |
+
"rewards/accuracy_reward": 0.7154017761349678,
|
| 807 |
+
"rewards/format_reward": 0.9899553433060646,
|
| 808 |
+
"step": 57
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"clip_ratio": 0.0,
|
| 812 |
+
"completion_length": 651.3370819091797,
|
| 813 |
+
"epoch": 0.5880861850443599,
|
| 814 |
+
"grad_norm": 0.09192176908254623,
|
| 815 |
+
"learning_rate": 8.893207690248775e-07,
|
| 816 |
+
"loss": 0.0075,
|
| 817 |
+
"num_tokens": 38216080.0,
|
| 818 |
+
"reward": 1.7600447535514832,
|
| 819 |
+
"reward_std": 0.1495125712826848,
|
| 820 |
+
"rewards/accuracy_reward": 0.7633928582072258,
|
| 821 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 822 |
+
"step": 58
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"clip_ratio": 0.0,
|
| 826 |
+
"completion_length": 651.7176742553711,
|
| 827 |
+
"epoch": 0.5982256020278834,
|
| 828 |
+
"grad_norm": 0.11554893106222153,
|
| 829 |
+
"learning_rate": 8.836588973708128e-07,
|
| 830 |
+
"loss": 0.045,
|
| 831 |
+
"num_tokens": 38941611.0,
|
| 832 |
+
"reward": 1.688616156578064,
|
| 833 |
+
"reward_std": 0.20678117126226425,
|
| 834 |
+
"rewards/accuracy_reward": 0.7042410708963871,
|
| 835 |
+
"rewards/format_reward": 0.9843749925494194,
|
| 836 |
+
"step": 59
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"clip_ratio": 0.0,
|
| 840 |
+
"completion_length": 673.3739242553711,
|
| 841 |
+
"epoch": 0.6083650190114068,
|
| 842 |
+
"grad_norm": 0.11070550233125687,
|
| 843 |
+
"learning_rate": 8.778747871771291e-07,
|
| 844 |
+
"loss": 0.0224,
|
| 845 |
+
"num_tokens": 39682202.0,
|
| 846 |
+
"reward": 1.7388393729925156,
|
| 847 |
+
"reward_std": 0.18264728784561157,
|
| 848 |
+
"rewards/accuracy_reward": 0.7444196492433548,
|
| 849 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 850 |
+
"step": 60
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"clip_ratio": 0.0,
|
| 854 |
+
"completion_length": 649.6138610839844,
|
| 855 |
+
"epoch": 0.6185044359949303,
|
| 856 |
+
"grad_norm": 0.11238072067499161,
|
| 857 |
+
"learning_rate": 8.719702813339247e-07,
|
| 858 |
+
"loss": 0.0075,
|
| 859 |
+
"num_tokens": 40386600.0,
|
| 860 |
+
"reward": 1.6919643729925156,
|
| 861 |
+
"reward_std": 0.20359325036406517,
|
| 862 |
+
"rewards/accuracy_reward": 0.7020089328289032,
|
| 863 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 864 |
+
"step": 61
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"clip_ratio": 0.0,
|
| 868 |
+
"completion_length": 687.9553833007812,
|
| 869 |
+
"epoch": 0.6286438529784537,
|
| 870 |
+
"grad_norm": 0.09478152543306351,
|
| 871 |
+
"learning_rate": 8.659472610908627e-07,
|
| 872 |
+
"loss": 0.0211,
|
| 873 |
+
"num_tokens": 41132400.0,
|
| 874 |
+
"reward": 1.6729911416769028,
|
| 875 |
+
"reward_std": 0.17133308481425047,
|
| 876 |
+
"rewards/accuracy_reward": 0.6785714291036129,
|
| 877 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 878 |
+
"step": 62
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"clip_ratio": 0.0,
|
| 882 |
+
"completion_length": 615.4788284301758,
|
| 883 |
+
"epoch": 0.6387832699619772,
|
| 884 |
+
"grad_norm": 0.12343282252550125,
|
| 885 |
+
"learning_rate": 8.598076454577814e-07,
|
| 886 |
+
"loss": 0.0078,
|
| 887 |
+
"num_tokens": 41808165.0,
|
| 888 |
+
"reward": 1.7689732909202576,
|
| 889 |
+
"reward_std": 0.18264547362923622,
|
| 890 |
+
"rewards/accuracy_reward": 0.7734375,
|
| 891 |
+
"rewards/format_reward": 0.995535708963871,
|
| 892 |
+
"step": 63
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"clip_ratio": 0.0,
|
| 896 |
+
"completion_length": 636.3192367553711,
|
| 897 |
+
"epoch": 0.6489226869455006,
|
| 898 |
+
"grad_norm": 0.08991697430610657,
|
| 899 |
+
"learning_rate": 8.535533905932737e-07,
|
| 900 |
+
"loss": 0.0267,
|
| 901 |
+
"num_tokens": 42526435.0,
|
| 902 |
+
"reward": 1.7232143729925156,
|
| 903 |
+
"reward_std": 0.14434537151828408,
|
| 904 |
+
"rewards/accuracy_reward": 0.7287946492433548,
|
| 905 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 906 |
+
"step": 64
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"clip_ratio": 0.0,
|
| 910 |
+
"completion_length": 627.6741409301758,
|
| 911 |
+
"epoch": 0.6590621039290241,
|
| 912 |
+
"grad_norm": 0.091641865670681,
|
| 913 |
+
"learning_rate": 8.471864891814304e-07,
|
| 914 |
+
"loss": 0.0296,
|
| 915 |
+
"num_tokens": 43207935.0,
|
| 916 |
+
"reward": 1.7488840073347092,
|
| 917 |
+
"reward_std": 0.16924036061391234,
|
| 918 |
+
"rewards/accuracy_reward": 0.7589285671710968,
|
| 919 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 920 |
+
"step": 65
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"clip_ratio": 0.0,
|
| 924 |
+
"completion_length": 697.7109680175781,
|
| 925 |
+
"epoch": 0.6692015209125475,
|
| 926 |
+
"grad_norm": 0.09379775822162628,
|
| 927 |
+
"learning_rate": 8.407089697969456e-07,
|
| 928 |
+
"loss": 0.0402,
|
| 929 |
+
"num_tokens": 43957948.0,
|
| 930 |
+
"reward": 1.705357238650322,
|
| 931 |
+
"reward_std": 0.17076639458537102,
|
| 932 |
+
"rewards/accuracy_reward": 0.719866082072258,
|
| 933 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 934 |
+
"step": 66
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"clip_ratio": 0.0,
|
| 938 |
+
"completion_length": 719.4453353881836,
|
| 939 |
+
"epoch": 0.679340937896071,
|
| 940 |
+
"grad_norm": 0.09100989252328873,
|
| 941 |
+
"learning_rate": 8.341228962587881e-07,
|
| 942 |
+
"loss": 0.0285,
|
| 943 |
+
"num_tokens": 44738699.0,
|
| 944 |
+
"reward": 1.6875001192092896,
|
| 945 |
+
"reward_std": 0.18435446824878454,
|
| 946 |
+
"rewards/accuracy_reward": 0.7031250149011612,
|
| 947 |
+
"rewards/format_reward": 0.9843749850988388,
|
| 948 |
+
"step": 67
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"clip_ratio": 0.0,
|
| 952 |
+
"completion_length": 604.7611961364746,
|
| 953 |
+
"epoch": 0.6894803548795945,
|
| 954 |
+
"grad_norm": 0.1007450744509697,
|
| 955 |
+
"learning_rate": 8.274303669726426e-07,
|
| 956 |
+
"loss": 0.022,
|
| 957 |
+
"num_tokens": 45401885.0,
|
| 958 |
+
"reward": 1.7589286416769028,
|
| 959 |
+
"reward_std": 0.1566725349985063,
|
| 960 |
+
"rewards/accuracy_reward": 0.7678571455180645,
|
| 961 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 962 |
+
"step": 68
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"clip_ratio": 0.0,
|
| 966 |
+
"completion_length": 717.0123062133789,
|
| 967 |
+
"epoch": 0.6996197718631179,
|
| 968 |
+
"grad_norm": 0.10416340827941895,
|
| 969 |
+
"learning_rate": 8.206335142623304e-07,
|
| 970 |
+
"loss": 0.0253,
|
| 971 |
+
"num_tokens": 46174088.0,
|
| 972 |
+
"reward": 1.6651786267757416,
|
| 973 |
+
"reward_std": 0.19855336844921112,
|
| 974 |
+
"rewards/accuracy_reward": 0.671875,
|
| 975 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 976 |
+
"step": 69
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"clip_ratio": 0.0,
|
| 980 |
+
"completion_length": 669.8147583007812,
|
| 981 |
+
"epoch": 0.7097591888466414,
|
| 982 |
+
"grad_norm": 0.11590211093425751,
|
| 983 |
+
"learning_rate": 8.137345036904259e-07,
|
| 984 |
+
"loss": 0.0433,
|
| 985 |
+
"num_tokens": 46893618.0,
|
| 986 |
+
"reward": 1.7075893580913544,
|
| 987 |
+
"reward_std": 0.21949494630098343,
|
| 988 |
+
"rewards/accuracy_reward": 0.723214291036129,
|
| 989 |
+
"rewards/format_reward": 0.9843749925494194,
|
| 990 |
+
"step": 70
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"clip_ratio": 0.0,
|
| 994 |
+
"completion_length": 648.5803756713867,
|
| 995 |
+
"epoch": 0.7198986058301647,
|
| 996 |
+
"grad_norm": 0.10192694514989853,
|
| 997 |
+
"learning_rate": 8.067355333682797e-07,
|
| 998 |
+
"loss": 0.0238,
|
| 999 |
+
"num_tokens": 47613514.0,
|
| 1000 |
+
"reward": 1.66183041036129,
|
| 1001 |
+
"reward_std": 0.17979866359382868,
|
| 1002 |
+
"rewards/accuracy_reward": 0.6796875074505806,
|
| 1003 |
+
"rewards/format_reward": 0.9821428582072258,
|
| 1004 |
+
"step": 71
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"clip_ratio": 0.0,
|
| 1008 |
+
"completion_length": 658.716552734375,
|
| 1009 |
+
"epoch": 0.7300380228136882,
|
| 1010 |
+
"grad_norm": 0.10270233452320099,
|
| 1011 |
+
"learning_rate": 7.996388332556734e-07,
|
| 1012 |
+
"loss": 0.0461,
|
| 1013 |
+
"num_tokens": 48332444.0,
|
| 1014 |
+
"reward": 1.7109375894069672,
|
| 1015 |
+
"reward_std": 0.17445513978600502,
|
| 1016 |
+
"rewards/accuracy_reward": 0.7187500149011612,
|
| 1017 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1018 |
+
"step": 72
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"clip_ratio": 0.0,
|
| 1022 |
+
"completion_length": 727.381721496582,
|
| 1023 |
+
"epoch": 0.7401774397972116,
|
| 1024 |
+
"grad_norm": 0.09961652010679245,
|
| 1025 |
+
"learning_rate": 7.924466644503264e-07,
|
| 1026 |
+
"loss": 0.0144,
|
| 1027 |
+
"num_tokens": 49116674.0,
|
| 1028 |
+
"reward": 1.6529018580913544,
|
| 1029 |
+
"reward_std": 0.20062358025461435,
|
| 1030 |
+
"rewards/accuracy_reward": 0.6662946529686451,
|
| 1031 |
+
"rewards/format_reward": 0.9866071417927742,
|
| 1032 |
+
"step": 73
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"clip_ratio": 0.0,
|
| 1036 |
+
"completion_length": 648.5078430175781,
|
| 1037 |
+
"epoch": 0.7503168567807351,
|
| 1038 |
+
"grad_norm": 0.10154278576374054,
|
| 1039 |
+
"learning_rate": 7.85161318467482e-07,
|
| 1040 |
+
"loss": 0.0184,
|
| 1041 |
+
"num_tokens": 49833121.0,
|
| 1042 |
+
"reward": 1.765625074505806,
|
| 1043 |
+
"reward_std": 0.17119968868792057,
|
| 1044 |
+
"rewards/accuracy_reward": 0.7734374925494194,
|
| 1045 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1046 |
+
"step": 74
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"clip_ratio": 0.0,
|
| 1050 |
+
"completion_length": 657.3549385070801,
|
| 1051 |
+
"epoch": 0.7604562737642585,
|
| 1052 |
+
"grad_norm": 0.11170431971549988,
|
| 1053 |
+
"learning_rate": 7.777851165098011e-07,
|
| 1054 |
+
"loss": 0.0291,
|
| 1055 |
+
"num_tokens": 50552791.0,
|
| 1056 |
+
"reward": 1.7522321939468384,
|
| 1057 |
+
"reward_std": 0.18582619447261095,
|
| 1058 |
+
"rewards/accuracy_reward": 0.7589285671710968,
|
| 1059 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 1060 |
+
"step": 75
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"clip_ratio": 0.0,
|
| 1064 |
+
"completion_length": 618.2310562133789,
|
| 1065 |
+
"epoch": 0.770595690747782,
|
| 1066 |
+
"grad_norm": 0.09891778975725174,
|
| 1067 |
+
"learning_rate": 7.703204087277988e-07,
|
| 1068 |
+
"loss": 0.0087,
|
| 1069 |
+
"num_tokens": 51233966.0,
|
| 1070 |
+
"reward": 1.6707590073347092,
|
| 1071 |
+
"reward_std": 0.16447317646816373,
|
| 1072 |
+
"rewards/accuracy_reward": 0.6785714402794838,
|
| 1073 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1074 |
+
"step": 76
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"clip_ratio": 0.0,
|
| 1078 |
+
"completion_length": 691.1406631469727,
|
| 1079 |
+
"epoch": 0.7807351077313055,
|
| 1080 |
+
"grad_norm": 0.0960167646408081,
|
| 1081 |
+
"learning_rate": 7.627695734710564e-07,
|
| 1082 |
+
"loss": 0.0304,
|
| 1083 |
+
"num_tokens": 51975452.0,
|
| 1084 |
+
"reward": 1.7042411714792252,
|
| 1085 |
+
"reward_std": 0.1650120271369815,
|
| 1086 |
+
"rewards/accuracy_reward": 0.7220982164144516,
|
| 1087 |
+
"rewards/format_reward": 0.9821428507566452,
|
| 1088 |
+
"step": 77
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"clip_ratio": 0.0,
|
| 1092 |
+
"completion_length": 684.7500228881836,
|
| 1093 |
+
"epoch": 0.7908745247148289,
|
| 1094 |
+
"grad_norm": 0.09193353354930878,
|
| 1095 |
+
"learning_rate": 7.551350165304499e-07,
|
| 1096 |
+
"loss": 0.0281,
|
| 1097 |
+
"num_tokens": 52706764.0,
|
| 1098 |
+
"reward": 1.7488840222358704,
|
| 1099 |
+
"reward_std": 0.16149256937205791,
|
| 1100 |
+
"rewards/accuracy_reward": 0.7622767835855484,
|
| 1101 |
+
"rewards/format_reward": 0.9866071417927742,
|
| 1102 |
+
"step": 78
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"clip_ratio": 0.0,
|
| 1106 |
+
"completion_length": 661.9587326049805,
|
| 1107 |
+
"epoch": 0.8010139416983524,
|
| 1108 |
+
"grad_norm": 0.08813660591840744,
|
| 1109 |
+
"learning_rate": 7.474191703716338e-07,
|
| 1110 |
+
"loss": 0.0161,
|
| 1111 |
+
"num_tokens": 53430439.0,
|
| 1112 |
+
"reward": 1.7209822237491608,
|
| 1113 |
+
"reward_std": 0.14920138847082853,
|
| 1114 |
+
"rewards/accuracy_reward": 0.723214291036129,
|
| 1115 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1116 |
+
"step": 79
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"clip_ratio": 0.0,
|
| 1120 |
+
"completion_length": 657.2868576049805,
|
| 1121 |
+
"epoch": 0.8111533586818758,
|
| 1122 |
+
"grad_norm": 0.08803500235080719,
|
| 1123 |
+
"learning_rate": 7.396244933600284e-07,
|
| 1124 |
+
"loss": 0.0076,
|
| 1125 |
+
"num_tokens": 54143656.0,
|
| 1126 |
+
"reward": 1.7421875894069672,
|
| 1127 |
+
"reward_std": 0.15661878790706396,
|
| 1128 |
+
"rewards/accuracy_reward": 0.7433035746216774,
|
| 1129 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1130 |
+
"step": 80
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"clip_ratio": 0.0,
|
| 1134 |
+
"completion_length": 653.8337326049805,
|
| 1135 |
+
"epoch": 0.8212927756653993,
|
| 1136 |
+
"grad_norm": 0.11003640294075012,
|
| 1137 |
+
"learning_rate": 7.317534689775527e-07,
|
| 1138 |
+
"loss": 0.0296,
|
| 1139 |
+
"num_tokens": 54849195.0,
|
| 1140 |
+
"reward": 1.7332590073347092,
|
| 1141 |
+
"reward_std": 0.1878818403929472,
|
| 1142 |
+
"rewards/accuracy_reward": 0.737723208963871,
|
| 1143 |
+
"rewards/format_reward": 0.9955357015132904,
|
| 1144 |
+
"step": 81
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"clip_ratio": 0.0,
|
| 1148 |
+
"completion_length": 646.4810562133789,
|
| 1149 |
+
"epoch": 0.8314321926489227,
|
| 1150 |
+
"grad_norm": 0.10392062366008759,
|
| 1151 |
+
"learning_rate": 7.238086050313562e-07,
|
| 1152 |
+
"loss": 0.0264,
|
| 1153 |
+
"num_tokens": 55563698.0,
|
| 1154 |
+
"reward": 1.7165179401636124,
|
| 1155 |
+
"reward_std": 0.20238416641950607,
|
| 1156 |
+
"rewards/accuracy_reward": 0.7254464328289032,
|
| 1157 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1158 |
+
"step": 82
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"clip_ratio": 0.0,
|
| 1162 |
+
"completion_length": 661.6440048217773,
|
| 1163 |
+
"epoch": 0.8415716096324461,
|
| 1164 |
+
"grad_norm": 0.11124070733785629,
|
| 1165 |
+
"learning_rate": 7.157924328548002e-07,
|
| 1166 |
+
"loss": 0.0144,
|
| 1167 |
+
"num_tokens": 56290123.0,
|
| 1168 |
+
"reward": 1.7410715073347092,
|
| 1169 |
+
"reward_std": 0.16893591731786728,
|
| 1170 |
+
"rewards/accuracy_reward": 0.7477678582072258,
|
| 1171 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1172 |
+
"step": 83
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"clip_ratio": 0.0,
|
| 1176 |
+
"completion_length": 667.6004867553711,
|
| 1177 |
+
"epoch": 0.8517110266159695,
|
| 1178 |
+
"grad_norm": 0.1015985757112503,
|
| 1179 |
+
"learning_rate": 7.077075065009433e-07,
|
| 1180 |
+
"loss": 0.0289,
|
| 1181 |
+
"num_tokens": 57021605.0,
|
| 1182 |
+
"reward": 1.7712054401636124,
|
| 1183 |
+
"reward_std": 0.1773331630975008,
|
| 1184 |
+
"rewards/accuracy_reward": 0.7790178582072258,
|
| 1185 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1186 |
+
"step": 84
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"clip_ratio": 0.0,
|
| 1190 |
+
"completion_length": 654.0971298217773,
|
| 1191 |
+
"epoch": 0.861850443599493,
|
| 1192 |
+
"grad_norm": 0.10410565882921219,
|
| 1193 |
+
"learning_rate": 6.995564019287869e-07,
|
| 1194 |
+
"loss": 0.0127,
|
| 1195 |
+
"num_tokens": 57737492.0,
|
| 1196 |
+
"reward": 1.6863840073347092,
|
| 1197 |
+
"reward_std": 0.2015578467398882,
|
| 1198 |
+
"rewards/accuracy_reward": 0.6919642835855484,
|
| 1199 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 1200 |
+
"step": 85
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"clip_ratio": 0.0,
|
| 1204 |
+
"completion_length": 705.3315048217773,
|
| 1205 |
+
"epoch": 0.8719898605830165,
|
| 1206 |
+
"grad_norm": 0.12336766719818115,
|
| 1207 |
+
"learning_rate": 6.913417161825449e-07,
|
| 1208 |
+
"loss": 0.0313,
|
| 1209 |
+
"num_tokens": 58504381.0,
|
| 1210 |
+
"reward": 1.6819197237491608,
|
| 1211 |
+
"reward_std": 0.19116687029600143,
|
| 1212 |
+
"rewards/accuracy_reward": 0.6908482164144516,
|
| 1213 |
+
"rewards/format_reward": 0.991071417927742,
|
| 1214 |
+
"step": 86
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"clip_ratio": 0.0,
|
| 1218 |
+
"completion_length": 663.8895492553711,
|
| 1219 |
+
"epoch": 0.8821292775665399,
|
| 1220 |
+
"grad_norm": 0.09500221908092499,
|
| 1221 |
+
"learning_rate": 6.830660665641897e-07,
|
| 1222 |
+
"loss": 0.0007,
|
| 1223 |
+
"num_tokens": 59228946.0,
|
| 1224 |
+
"reward": 1.7243304252624512,
|
| 1225 |
+
"reward_std": 0.1611752938479185,
|
| 1226 |
+
"rewards/accuracy_reward": 0.731026791036129,
|
| 1227 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1228 |
+
"step": 87
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"clip_ratio": 0.0,
|
| 1232 |
+
"completion_length": 676.3370895385742,
|
| 1233 |
+
"epoch": 0.8922686945500634,
|
| 1234 |
+
"grad_norm": 0.10983681678771973,
|
| 1235 |
+
"learning_rate": 6.747320897995492e-07,
|
| 1236 |
+
"loss": 0.0173,
|
| 1237 |
+
"num_tokens": 59963936.0,
|
| 1238 |
+
"reward": 1.7120536118745804,
|
| 1239 |
+
"reward_std": 0.16041494347155094,
|
| 1240 |
+
"rewards/accuracy_reward": 0.7287946343421936,
|
| 1241 |
+
"rewards/format_reward": 0.983258917927742,
|
| 1242 |
+
"step": 88
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"clip_ratio": 0.0,
|
| 1246 |
+
"completion_length": 627.8236999511719,
|
| 1247 |
+
"epoch": 0.9024081115335868,
|
| 1248 |
+
"grad_norm": 0.10716623067855835,
|
| 1249 |
+
"learning_rate": 6.66342441198212e-07,
|
| 1250 |
+
"loss": 0.0255,
|
| 1251 |
+
"num_tokens": 60650370.0,
|
| 1252 |
+
"reward": 1.7276786416769028,
|
| 1253 |
+
"reward_std": 0.16107469610869884,
|
| 1254 |
+
"rewards/accuracy_reward": 0.7366071417927742,
|
| 1255 |
+
"rewards/format_reward": 0.991071417927742,
|
| 1256 |
+
"step": 89
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"clip_ratio": 0.0,
|
| 1260 |
+
"completion_length": 641.5011367797852,
|
| 1261 |
+
"epoch": 0.9125475285171103,
|
| 1262 |
+
"grad_norm": 0.12808987498283386,
|
| 1263 |
+
"learning_rate": 6.578997938075125e-07,
|
| 1264 |
+
"loss": 0.024,
|
| 1265 |
+
"num_tokens": 61349627.0,
|
| 1266 |
+
"reward": 1.6953125596046448,
|
| 1267 |
+
"reward_std": 0.18346151895821095,
|
| 1268 |
+
"rewards/accuracy_reward": 0.7109375,
|
| 1269 |
+
"rewards/format_reward": 0.9843749925494194,
|
| 1270 |
+
"step": 90
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"clip_ratio": 0.0,
|
| 1274 |
+
"completion_length": 647.2109603881836,
|
| 1275 |
+
"epoch": 0.9226869455006337,
|
| 1276 |
+
"grad_norm": 0.10632172971963882,
|
| 1277 |
+
"learning_rate": 6.494068375608646e-07,
|
| 1278 |
+
"loss": 0.0176,
|
| 1279 |
+
"num_tokens": 62064152.0,
|
| 1280 |
+
"reward": 1.7388393729925156,
|
| 1281 |
+
"reward_std": 0.17309935204684734,
|
| 1282 |
+
"rewards/accuracy_reward": 0.745535708963871,
|
| 1283 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 1284 |
+
"step": 91
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"clip_ratio": 0.0,
|
| 1288 |
+
"completion_length": 684.0725860595703,
|
| 1289 |
+
"epoch": 0.9328263624841572,
|
| 1290 |
+
"grad_norm": 0.10359478741884232,
|
| 1291 |
+
"learning_rate": 6.408662784207149e-07,
|
| 1292 |
+
"loss": 0.0327,
|
| 1293 |
+
"num_tokens": 62816457.0,
|
| 1294 |
+
"reward": 1.7265625596046448,
|
| 1295 |
+
"reward_std": 0.18876438401639462,
|
| 1296 |
+
"rewards/accuracy_reward": 0.7399553656578064,
|
| 1297 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 1298 |
+
"step": 92
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"clip_ratio": 0.0,
|
| 1302 |
+
"completion_length": 619.8136520385742,
|
| 1303 |
+
"epoch": 0.9429657794676806,
|
| 1304 |
+
"grad_norm": 0.110443115234375,
|
| 1305 |
+
"learning_rate": 6.322808375163895e-07,
|
| 1306 |
+
"loss": 0.0174,
|
| 1307 |
+
"num_tokens": 63498690.0,
|
| 1308 |
+
"reward": 1.7377233058214188,
|
| 1309 |
+
"reward_std": 0.16478892508894205,
|
| 1310 |
+
"rewards/accuracy_reward": 0.7399553582072258,
|
| 1311 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1312 |
+
"step": 93
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"clip_ratio": 0.0,
|
| 1316 |
+
"completion_length": 650.7254791259766,
|
| 1317 |
+
"epoch": 0.9531051964512041,
|
| 1318 |
+
"grad_norm": 0.10725186765193939,
|
| 1319 |
+
"learning_rate": 6.236532502771077e-07,
|
| 1320 |
+
"loss": 0.0106,
|
| 1321 |
+
"num_tokens": 64215340.0,
|
| 1322 |
+
"reward": 1.7031250894069672,
|
| 1323 |
+
"reward_std": 0.18828019313514233,
|
| 1324 |
+
"rewards/accuracy_reward": 0.7109375,
|
| 1325 |
+
"rewards/format_reward": 0.9921875,
|
| 1326 |
+
"step": 94
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"clip_ratio": 0.0,
|
| 1330 |
+
"completion_length": 643.1205749511719,
|
| 1331 |
+
"epoch": 0.9632446134347274,
|
| 1332 |
+
"grad_norm": 0.13609780371189117,
|
| 1333 |
+
"learning_rate": 6.149862655604403e-07,
|
| 1334 |
+
"loss": 0.0263,
|
| 1335 |
+
"num_tokens": 64917264.0,
|
| 1336 |
+
"reward": 1.7589286714792252,
|
| 1337 |
+
"reward_std": 0.19211565144360065,
|
| 1338 |
+
"rewards/accuracy_reward": 0.7734375074505806,
|
| 1339 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 1340 |
+
"step": 95
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"clip_ratio": 0.0,
|
| 1344 |
+
"completion_length": 640.569221496582,
|
| 1345 |
+
"epoch": 0.973384030418251,
|
| 1346 |
+
"grad_norm": 0.11930994689464569,
|
| 1347 |
+
"learning_rate": 6.062826447764883e-07,
|
| 1348 |
+
"loss": 0.0198,
|
| 1349 |
+
"num_tokens": 65613958.0,
|
| 1350 |
+
"reward": 1.757812574505806,
|
| 1351 |
+
"reward_std": 0.14032188896089792,
|
| 1352 |
+
"rewards/accuracy_reward": 0.7622767761349678,
|
| 1353 |
+
"rewards/format_reward": 0.9955357015132904,
|
| 1354 |
+
"step": 96
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"clip_ratio": 0.0,
|
| 1358 |
+
"completion_length": 654.4654388427734,
|
| 1359 |
+
"epoch": 0.9835234474017744,
|
| 1360 |
+
"grad_norm": 0.09683021903038025,
|
| 1361 |
+
"learning_rate": 5.975451610080642e-07,
|
| 1362 |
+
"loss": 0.0272,
|
| 1363 |
+
"num_tokens": 66328167.0,
|
| 1364 |
+
"reward": 1.6875000894069672,
|
| 1365 |
+
"reward_std": 0.1803465113043785,
|
| 1366 |
+
"rewards/accuracy_reward": 0.6953125074505806,
|
| 1367 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1368 |
+
"step": 97
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"clip_ratio": 0.0,
|
| 1372 |
+
"completion_length": 631.353515625,
|
| 1373 |
+
"epoch": 0.9936628643852978,
|
| 1374 |
+
"grad_norm": 0.1017894446849823,
|
| 1375 |
+
"learning_rate": 5.887765981271517e-07,
|
| 1376 |
+
"loss": 0.0292,
|
| 1377 |
+
"num_tokens": 67045777.0,
|
| 1378 |
+
"reward": 1.7265625894069672,
|
| 1379 |
+
"reward_std": 0.1662213634699583,
|
| 1380 |
+
"rewards/accuracy_reward": 0.7366071417927742,
|
| 1381 |
+
"rewards/format_reward": 0.9899553433060646,
|
| 1382 |
+
"step": 98
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"clip_ratio": 0.0,
|
| 1386 |
+
"completion_length": 662.9151916503906,
|
| 1387 |
+
"epoch": 1.0101394169835234,
|
| 1388 |
+
"grad_norm": 0.09788638353347778,
|
| 1389 |
+
"learning_rate": 5.7997974990793e-07,
|
| 1390 |
+
"loss": 0.0044,
|
| 1391 |
+
"num_tokens": 67768741.0,
|
| 1392 |
+
"reward": 1.7566965073347092,
|
| 1393 |
+
"reward_std": 0.1626696866005659,
|
| 1394 |
+
"rewards/accuracy_reward": 0.7578125,
|
| 1395 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1396 |
+
"step": 99
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"clip_ratio": 0.0,
|
| 1400 |
+
"completion_length": 659.5960006713867,
|
| 1401 |
+
"epoch": 1.020278833967047,
|
| 1402 |
+
"grad_norm": 0.09406567364931107,
|
| 1403 |
+
"learning_rate": 5.711574191366427e-07,
|
| 1404 |
+
"loss": 0.0125,
|
| 1405 |
+
"num_tokens": 68487115.0,
|
| 1406 |
+
"reward": 1.7321429550647736,
|
| 1407 |
+
"reward_std": 0.1440184023231268,
|
| 1408 |
+
"rewards/accuracy_reward": 0.7366071417927742,
|
| 1409 |
+
"rewards/format_reward": 0.995535708963871,
|
| 1410 |
+
"step": 100
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"clip_ratio": 0.0,
|
| 1414 |
+
"completion_length": 664.0189971923828,
|
| 1415 |
+
"epoch": 1.0304182509505704,
|
| 1416 |
+
"grad_norm": 0.10347539931535721,
|
| 1417 |
+
"learning_rate": 5.623124167185929e-07,
|
| 1418 |
+
"loss": 0.0152,
|
| 1419 |
+
"num_tokens": 69204604.0,
|
| 1420 |
+
"reward": 1.7410715222358704,
|
| 1421 |
+
"reward_std": 0.153701429720968,
|
| 1422 |
+
"rewards/accuracy_reward": 0.746651791036129,
|
| 1423 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1424 |
+
"step": 101
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"clip_ratio": 0.0,
|
| 1428 |
+
"completion_length": 623.9464569091797,
|
| 1429 |
+
"epoch": 1.0405576679340938,
|
| 1430 |
+
"grad_norm": 0.09895368665456772,
|
| 1431 |
+
"learning_rate": 5.534475607825565e-07,
|
| 1432 |
+
"loss": 0.033,
|
| 1433 |
+
"num_tokens": 69890948.0,
|
| 1434 |
+
"reward": 1.7165179401636124,
|
| 1435 |
+
"reward_std": 0.13192464038729668,
|
| 1436 |
+
"rewards/accuracy_reward": 0.7232142835855484,
|
| 1437 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1438 |
+
"step": 102
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"clip_ratio": 0.0,
|
| 1442 |
+
"completion_length": 683.225471496582,
|
| 1443 |
+
"epoch": 1.0506970849176172,
|
| 1444 |
+
"grad_norm": 0.11877602338790894,
|
| 1445 |
+
"learning_rate": 5.445656757828879e-07,
|
| 1446 |
+
"loss": 0.0322,
|
| 1447 |
+
"num_tokens": 70634422.0,
|
| 1448 |
+
"reward": 1.6361607909202576,
|
| 1449 |
+
"reward_std": 0.17785416916012764,
|
| 1450 |
+
"rewards/accuracy_reward": 0.645089291036129,
|
| 1451 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1452 |
+
"step": 103
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"clip_ratio": 0.0,
|
| 1456 |
+
"completion_length": 662.256721496582,
|
| 1457 |
+
"epoch": 1.0608365019011408,
|
| 1458 |
+
"grad_norm": 0.10182492434978485,
|
| 1459 |
+
"learning_rate": 5.356695915996161e-07,
|
| 1460 |
+
"loss": 0.0056,
|
| 1461 |
+
"num_tokens": 71362164.0,
|
| 1462 |
+
"reward": 1.7165179252624512,
|
| 1463 |
+
"reward_std": 0.15059774369001389,
|
| 1464 |
+
"rewards/accuracy_reward": 0.7187499925494194,
|
| 1465 |
+
"rewards/format_reward": 0.9977678582072258,
|
| 1466 |
+
"step": 104
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"clip_ratio": 0.0,
|
| 1470 |
+
"completion_length": 626.7701187133789,
|
| 1471 |
+
"epoch": 1.0709759188846641,
|
| 1472 |
+
"grad_norm": 0.1029505580663681,
|
| 1473 |
+
"learning_rate": 5.267621426368075e-07,
|
| 1474 |
+
"loss": 0.0062,
|
| 1475 |
+
"num_tokens": 72042966.0,
|
| 1476 |
+
"reward": 1.797991156578064,
|
| 1477 |
+
"reward_std": 0.14701888849958777,
|
| 1478 |
+
"rewards/accuracy_reward": 0.8013392835855484,
|
| 1479 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 1480 |
+
"step": 105
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"clip_ratio": 0.0,
|
| 1484 |
+
"completion_length": 577.3560562133789,
|
| 1485 |
+
"epoch": 1.0811153358681875,
|
| 1486 |
+
"grad_norm": 0.10245715826749802,
|
| 1487 |
+
"learning_rate": 5.178461669194903e-07,
|
| 1488 |
+
"loss": 0.013,
|
| 1489 |
+
"num_tokens": 72687485.0,
|
| 1490 |
+
"reward": 1.7946429401636124,
|
| 1491 |
+
"reward_std": 0.14055991917848587,
|
| 1492 |
+
"rewards/accuracy_reward": 0.7991071417927742,
|
| 1493 |
+
"rewards/format_reward": 0.9955357164144516,
|
| 1494 |
+
"step": 106
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"clip_ratio": 0.0,
|
| 1498 |
+
"completion_length": 634.2232437133789,
|
| 1499 |
+
"epoch": 1.091254752851711,
|
| 1500 |
+
"grad_norm": 0.09775260090827942,
|
| 1501 |
+
"learning_rate": 5.08924505189423e-07,
|
| 1502 |
+
"loss": 0.0073,
|
| 1503 |
+
"num_tokens": 73381709.0,
|
| 1504 |
+
"reward": 1.77120541036129,
|
| 1505 |
+
"reward_std": 0.13471420388668776,
|
| 1506 |
+
"rewards/accuracy_reward": 0.7745535597205162,
|
| 1507 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 1508 |
+
"step": 107
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"clip_ratio": 0.0,
|
| 1512 |
+
"completion_length": 671.7980308532715,
|
| 1513 |
+
"epoch": 1.1013941698352345,
|
| 1514 |
+
"grad_norm": 0.10289379954338074,
|
| 1515 |
+
"learning_rate": 5e-07,
|
| 1516 |
+
"loss": 0.0137,
|
| 1517 |
+
"num_tokens": 74107760.0,
|
| 1518 |
+
"reward": 1.7578125894069672,
|
| 1519 |
+
"reward_std": 0.16642380319535732,
|
| 1520 |
+
"rewards/accuracy_reward": 0.765625,
|
| 1521 |
+
"rewards/format_reward": 0.9921875,
|
| 1522 |
+
"step": 108
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"clip_ratio": 0.0,
|
| 1526 |
+
"completion_length": 700.1719055175781,
|
| 1527 |
+
"epoch": 1.111533586818758,
|
| 1528 |
+
"grad_norm": 0.08667897433042526,
|
| 1529 |
+
"learning_rate": 4.91075494810577e-07,
|
| 1530 |
+
"loss": 0.0214,
|
| 1531 |
+
"num_tokens": 74853106.0,
|
| 1532 |
+
"reward": 1.7477679252624512,
|
| 1533 |
+
"reward_std": 0.14406709745526314,
|
| 1534 |
+
"rewards/accuracy_reward": 0.7555803433060646,
|
| 1535 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1536 |
+
"step": 109
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"clip_ratio": 0.0,
|
| 1540 |
+
"completion_length": 654.7210083007812,
|
| 1541 |
+
"epoch": 1.1216730038022813,
|
| 1542 |
+
"grad_norm": 0.10183614492416382,
|
| 1543 |
+
"learning_rate": 4.821538330805098e-07,
|
| 1544 |
+
"loss": 0.0228,
|
| 1545 |
+
"num_tokens": 75566256.0,
|
| 1546 |
+
"reward": 1.7477679401636124,
|
| 1547 |
+
"reward_std": 0.16557483188807964,
|
| 1548 |
+
"rewards/accuracy_reward": 0.7600446417927742,
|
| 1549 |
+
"rewards/format_reward": 0.9877232015132904,
|
| 1550 |
+
"step": 110
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"clip_ratio": 0.0,
|
| 1554 |
+
"completion_length": 668.025691986084,
|
| 1555 |
+
"epoch": 1.131812420785805,
|
| 1556 |
+
"grad_norm": 0.11714053899049759,
|
| 1557 |
+
"learning_rate": 4.732378573631924e-07,
|
| 1558 |
+
"loss": 0.0246,
|
| 1559 |
+
"num_tokens": 76287535.0,
|
| 1560 |
+
"reward": 1.7421875894069672,
|
| 1561 |
+
"reward_std": 0.1860833875834942,
|
| 1562 |
+
"rewards/accuracy_reward": 0.7511160746216774,
|
| 1563 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1564 |
+
"step": 111
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"clip_ratio": 0.0,
|
| 1568 |
+
"completion_length": 669.5022659301758,
|
| 1569 |
+
"epoch": 1.1419518377693283,
|
| 1570 |
+
"grad_norm": 0.11465004831552505,
|
| 1571 |
+
"learning_rate": 4.643304084003838e-07,
|
| 1572 |
+
"loss": 0.0245,
|
| 1573 |
+
"num_tokens": 77014857.0,
|
| 1574 |
+
"reward": 1.7254465222358704,
|
| 1575 |
+
"reward_std": 0.19485904835164547,
|
| 1576 |
+
"rewards/accuracy_reward": 0.7321428656578064,
|
| 1577 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1578 |
+
"step": 112
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"clip_ratio": 0.0,
|
| 1582 |
+
"completion_length": 656.6093978881836,
|
| 1583 |
+
"epoch": 1.1520912547528517,
|
| 1584 |
+
"grad_norm": 0.160664901137352,
|
| 1585 |
+
"learning_rate": 4.55434324217112e-07,
|
| 1586 |
+
"loss": 0.0228,
|
| 1587 |
+
"num_tokens": 77732419.0,
|
| 1588 |
+
"reward": 1.6763393431901932,
|
| 1589 |
+
"reward_std": 0.1846153624355793,
|
| 1590 |
+
"rewards/accuracy_reward": 0.684151791036129,
|
| 1591 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1592 |
+
"step": 113
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"clip_ratio": 0.0,
|
| 1596 |
+
"completion_length": 655.0524826049805,
|
| 1597 |
+
"epoch": 1.162230671736375,
|
| 1598 |
+
"grad_norm": 0.10615188628435135,
|
| 1599 |
+
"learning_rate": 4.4655243921744367e-07,
|
| 1600 |
+
"loss": 0.0136,
|
| 1601 |
+
"num_tokens": 78444306.0,
|
| 1602 |
+
"reward": 1.695312574505806,
|
| 1603 |
+
"reward_std": 0.14989868365228176,
|
| 1604 |
+
"rewards/accuracy_reward": 0.7020089253783226,
|
| 1605 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1606 |
+
"step": 114
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"clip_ratio": 0.0,
|
| 1610 |
+
"completion_length": 655.928596496582,
|
| 1611 |
+
"epoch": 1.1723700887198987,
|
| 1612 |
+
"grad_norm": 0.1310528963804245,
|
| 1613 |
+
"learning_rate": 4.37687583281407e-07,
|
| 1614 |
+
"loss": 0.0203,
|
| 1615 |
+
"num_tokens": 79159234.0,
|
| 1616 |
+
"reward": 1.7176340222358704,
|
| 1617 |
+
"reward_std": 0.15918411500751972,
|
| 1618 |
+
"rewards/accuracy_reward": 0.7243303656578064,
|
| 1619 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1620 |
+
"step": 115
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"clip_ratio": 0.0,
|
| 1624 |
+
"completion_length": 618.7656478881836,
|
| 1625 |
+
"epoch": 1.182509505703422,
|
| 1626 |
+
"grad_norm": 0.11190109699964523,
|
| 1627 |
+
"learning_rate": 4.2884258086335745e-07,
|
| 1628 |
+
"loss": 0.0031,
|
| 1629 |
+
"num_tokens": 79852888.0,
|
| 1630 |
+
"reward": 1.774553656578064,
|
| 1631 |
+
"reward_std": 0.17635854333639145,
|
| 1632 |
+
"rewards/accuracy_reward": 0.7767857164144516,
|
| 1633 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1634 |
+
"step": 116
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"clip_ratio": 0.0,
|
| 1638 |
+
"completion_length": 629.7489166259766,
|
| 1639 |
+
"epoch": 1.1926489226869454,
|
| 1640 |
+
"grad_norm": 0.09713304787874222,
|
| 1641 |
+
"learning_rate": 4.2002025009206987e-07,
|
| 1642 |
+
"loss": 0.0098,
|
| 1643 |
+
"num_tokens": 80549095.0,
|
| 1644 |
+
"reward": 1.7287947237491608,
|
| 1645 |
+
"reward_std": 0.12099508009850979,
|
| 1646 |
+
"rewards/accuracy_reward": 0.7321428619325161,
|
| 1647 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 1648 |
+
"step": 117
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"clip_ratio": 0.0,
|
| 1652 |
+
"completion_length": 615.9765968322754,
|
| 1653 |
+
"epoch": 1.202788339670469,
|
| 1654 |
+
"grad_norm": 0.11283177137374878,
|
| 1655 |
+
"learning_rate": 4.1122340187284845e-07,
|
| 1656 |
+
"loss": 0.0315,
|
| 1657 |
+
"num_tokens": 81223298.0,
|
| 1658 |
+
"reward": 1.7388393580913544,
|
| 1659 |
+
"reward_std": 0.13626329507678747,
|
| 1660 |
+
"rewards/accuracy_reward": 0.745535708963871,
|
| 1661 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1662 |
+
"step": 118
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"clip_ratio": 0.0,
|
| 1666 |
+
"completion_length": 641.6629791259766,
|
| 1667 |
+
"epoch": 1.2129277566539924,
|
| 1668 |
+
"grad_norm": 0.0984925627708435,
|
| 1669 |
+
"learning_rate": 4.0245483899193586e-07,
|
| 1670 |
+
"loss": 0.0086,
|
| 1671 |
+
"num_tokens": 81925604.0,
|
| 1672 |
+
"reward": 1.7310268431901932,
|
| 1673 |
+
"reward_std": 0.16442857310175896,
|
| 1674 |
+
"rewards/accuracy_reward": 0.7354910671710968,
|
| 1675 |
+
"rewards/format_reward": 0.995535708963871,
|
| 1676 |
+
"step": 119
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"clip_ratio": 0.0,
|
| 1680 |
+
"completion_length": 628.5569534301758,
|
| 1681 |
+
"epoch": 1.2230671736375158,
|
| 1682 |
+
"grad_norm": 0.09760965406894684,
|
| 1683 |
+
"learning_rate": 3.937173552235116e-07,
|
| 1684 |
+
"loss": 0.0041,
|
| 1685 |
+
"num_tokens": 82621599.0,
|
| 1686 |
+
"reward": 1.7053572088479996,
|
| 1687 |
+
"reward_std": 0.11313999397680163,
|
| 1688 |
+
"rewards/accuracy_reward": 0.7098214253783226,
|
| 1689 |
+
"rewards/format_reward": 0.9955357164144516,
|
| 1690 |
+
"step": 120
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"clip_ratio": 0.0,
|
| 1694 |
+
"completion_length": 645.1350860595703,
|
| 1695 |
+
"epoch": 1.2332065906210392,
|
| 1696 |
+
"grad_norm": 0.08554679900407791,
|
| 1697 |
+
"learning_rate": 3.850137344395598e-07,
|
| 1698 |
+
"loss": 0.0151,
|
| 1699 |
+
"num_tokens": 83322856.0,
|
| 1700 |
+
"reward": 1.7354911714792252,
|
| 1701 |
+
"reward_std": 0.11471607023850083,
|
| 1702 |
+
"rewards/accuracy_reward": 0.7399553582072258,
|
| 1703 |
+
"rewards/format_reward": 0.995535708963871,
|
| 1704 |
+
"step": 121
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"clip_ratio": 0.0,
|
| 1708 |
+
"completion_length": 688.2924423217773,
|
| 1709 |
+
"epoch": 1.2433460076045628,
|
| 1710 |
+
"grad_norm": 0.11858882755041122,
|
| 1711 |
+
"learning_rate": 3.763467497228922e-07,
|
| 1712 |
+
"loss": 0.0248,
|
| 1713 |
+
"num_tokens": 84077598.0,
|
| 1714 |
+
"reward": 1.678571492433548,
|
| 1715 |
+
"reward_std": 0.15971866622567177,
|
| 1716 |
+
"rewards/accuracy_reward": 0.6930803582072258,
|
| 1717 |
+
"rewards/format_reward": 0.9854910597205162,
|
| 1718 |
+
"step": 122
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"clip_ratio": 0.0,
|
| 1722 |
+
"completion_length": 655.8861923217773,
|
| 1723 |
+
"epoch": 1.2534854245880862,
|
| 1724 |
+
"grad_norm": 0.09887714684009552,
|
| 1725 |
+
"learning_rate": 3.677191624836106e-07,
|
| 1726 |
+
"loss": 0.0222,
|
| 1727 |
+
"num_tokens": 84786984.0,
|
| 1728 |
+
"reward": 1.7790179252624512,
|
| 1729 |
+
"reward_std": 0.15774498507380486,
|
| 1730 |
+
"rewards/accuracy_reward": 0.787946417927742,
|
| 1731 |
+
"rewards/format_reward": 0.991071417927742,
|
| 1732 |
+
"step": 123
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"clip_ratio": 0.0,
|
| 1736 |
+
"completion_length": 663.8951187133789,
|
| 1737 |
+
"epoch": 1.2636248415716096,
|
| 1738 |
+
"grad_norm": 0.11492931097745895,
|
| 1739 |
+
"learning_rate": 3.591337215792851e-07,
|
| 1740 |
+
"loss": 0.019,
|
| 1741 |
+
"num_tokens": 85510002.0,
|
| 1742 |
+
"reward": 1.7433036416769028,
|
| 1743 |
+
"reward_std": 0.18467565812170506,
|
| 1744 |
+
"rewards/accuracy_reward": 0.7522321417927742,
|
| 1745 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1746 |
+
"step": 124
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"clip_ratio": 0.0,
|
| 1750 |
+
"completion_length": 623.6551513671875,
|
| 1751 |
+
"epoch": 1.2737642585551332,
|
| 1752 |
+
"grad_norm": 0.10936412215232849,
|
| 1753 |
+
"learning_rate": 3.505931624391355e-07,
|
| 1754 |
+
"loss": 0.0138,
|
| 1755 |
+
"num_tokens": 86187061.0,
|
| 1756 |
+
"reward": 1.7689732760190964,
|
| 1757 |
+
"reward_std": 0.15780174732208252,
|
| 1758 |
+
"rewards/accuracy_reward": 0.7712053433060646,
|
| 1759 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1760 |
+
"step": 125
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"clip_ratio": 0.0,
|
| 1764 |
+
"completion_length": 634.5870895385742,
|
| 1765 |
+
"epoch": 1.2839036755386566,
|
| 1766 |
+
"grad_norm": 0.11182162165641785,
|
| 1767 |
+
"learning_rate": 3.421002061924876e-07,
|
| 1768 |
+
"loss": 0.0187,
|
| 1769 |
+
"num_tokens": 86882515.0,
|
| 1770 |
+
"reward": 1.7410715073347092,
|
| 1771 |
+
"reward_std": 0.17043365351855755,
|
| 1772 |
+
"rewards/accuracy_reward": 0.746651791036129,
|
| 1773 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1774 |
+
"step": 126
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"clip_ratio": 0.0,
|
| 1778 |
+
"completion_length": 646.7031555175781,
|
| 1779 |
+
"epoch": 1.29404309252218,
|
| 1780 |
+
"grad_norm": 0.10373136401176453,
|
| 1781 |
+
"learning_rate": 3.3365755880178807e-07,
|
| 1782 |
+
"loss": 0.0117,
|
| 1783 |
+
"num_tokens": 87583665.0,
|
| 1784 |
+
"reward": 1.781250074505806,
|
| 1785 |
+
"reward_std": 0.1267631771042943,
|
| 1786 |
+
"rewards/accuracy_reward": 0.7845982238650322,
|
| 1787 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 1788 |
+
"step": 127
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"clip_ratio": 0.0,
|
| 1792 |
+
"completion_length": 675.5312728881836,
|
| 1793 |
+
"epoch": 1.3041825095057034,
|
| 1794 |
+
"grad_norm": 0.10996940732002258,
|
| 1795 |
+
"learning_rate": 3.2526791020045087e-07,
|
| 1796 |
+
"loss": 0.0199,
|
| 1797 |
+
"num_tokens": 88321541.0,
|
| 1798 |
+
"reward": 1.702008992433548,
|
| 1799 |
+
"reward_std": 0.18257617205381393,
|
| 1800 |
+
"rewards/accuracy_reward": 0.7120535671710968,
|
| 1801 |
+
"rewards/format_reward": 0.9899553582072258,
|
| 1802 |
+
"step": 128
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"clip_ratio": 0.0,
|
| 1806 |
+
"completion_length": 629.1652069091797,
|
| 1807 |
+
"epoch": 1.3143219264892267,
|
| 1808 |
+
"grad_norm": 0.1037248894572258,
|
| 1809 |
+
"learning_rate": 3.169339334358104e-07,
|
| 1810 |
+
"loss": 0.0005,
|
| 1811 |
+
"num_tokens": 89018617.0,
|
| 1812 |
+
"reward": 1.733258992433548,
|
| 1813 |
+
"reward_std": 0.15833040233701468,
|
| 1814 |
+
"rewards/accuracy_reward": 0.7366071417927742,
|
| 1815 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 1816 |
+
"step": 129
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"clip_ratio": 0.0,
|
| 1820 |
+
"completion_length": 661.2622985839844,
|
| 1821 |
+
"epoch": 1.3244613434727504,
|
| 1822 |
+
"grad_norm": 0.1321147084236145,
|
| 1823 |
+
"learning_rate": 3.086582838174551e-07,
|
| 1824 |
+
"loss": 0.0124,
|
| 1825 |
+
"num_tokens": 89743596.0,
|
| 1826 |
+
"reward": 1.7109375596046448,
|
| 1827 |
+
"reward_std": 0.19106812588870525,
|
| 1828 |
+
"rewards/accuracy_reward": 0.7198660671710968,
|
| 1829 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1830 |
+
"step": 130
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"clip_ratio": 0.0,
|
| 1834 |
+
"completion_length": 659.3091735839844,
|
| 1835 |
+
"epoch": 1.3346007604562737,
|
| 1836 |
+
"grad_norm": 0.09277363866567612,
|
| 1837 |
+
"learning_rate": 3.004435980712129e-07,
|
| 1838 |
+
"loss": 0.0114,
|
| 1839 |
+
"num_tokens": 90457801.0,
|
| 1840 |
+
"reward": 1.727678656578064,
|
| 1841 |
+
"reward_std": 0.13339263666421175,
|
| 1842 |
+
"rewards/accuracy_reward": 0.7366071492433548,
|
| 1843 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1844 |
+
"step": 131
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"clip_ratio": 0.0,
|
| 1848 |
+
"completion_length": 630.7388687133789,
|
| 1849 |
+
"epoch": 1.3447401774397973,
|
| 1850 |
+
"grad_norm": 0.09743496030569077,
|
| 1851 |
+
"learning_rate": 2.922924934990568e-07,
|
| 1852 |
+
"loss": 0.0247,
|
| 1853 |
+
"num_tokens": 91148095.0,
|
| 1854 |
+
"reward": 1.7700893729925156,
|
| 1855 |
+
"reward_std": 0.1285113412886858,
|
| 1856 |
+
"rewards/accuracy_reward": 0.7734375,
|
| 1857 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 1858 |
+
"step": 132
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"clip_ratio": 0.0,
|
| 1862 |
+
"completion_length": 611.0044898986816,
|
| 1863 |
+
"epoch": 1.3548795944233207,
|
| 1864 |
+
"grad_norm": 0.12098115682601929,
|
| 1865 |
+
"learning_rate": 2.8420756714519954e-07,
|
| 1866 |
+
"loss": 0.0017,
|
| 1867 |
+
"num_tokens": 91839859.0,
|
| 1868 |
+
"reward": 1.7444197237491608,
|
| 1869 |
+
"reward_std": 0.1756706228479743,
|
| 1870 |
+
"rewards/accuracy_reward": 0.745535708963871,
|
| 1871 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1872 |
+
"step": 133
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"clip_ratio": 0.0,
|
| 1876 |
+
"completion_length": 638.6317291259766,
|
| 1877 |
+
"epoch": 1.3650190114068441,
|
| 1878 |
+
"grad_norm": 0.10373256355524063,
|
| 1879 |
+
"learning_rate": 2.7619139496864376e-07,
|
| 1880 |
+
"loss": 0.0222,
|
| 1881 |
+
"num_tokens": 92548057.0,
|
| 1882 |
+
"reward": 1.6875000894069672,
|
| 1883 |
+
"reward_std": 0.15588384959846735,
|
| 1884 |
+
"rewards/accuracy_reward": 0.6975446492433548,
|
| 1885 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 1886 |
+
"step": 134
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"clip_ratio": 0.0,
|
| 1890 |
+
"completion_length": 606.5056076049805,
|
| 1891 |
+
"epoch": 1.3751584283903675,
|
| 1892 |
+
"grad_norm": 0.13831515610218048,
|
| 1893 |
+
"learning_rate": 2.6824653102244727e-07,
|
| 1894 |
+
"loss": 0.0172,
|
| 1895 |
+
"num_tokens": 93225558.0,
|
| 1896 |
+
"reward": 1.751116156578064,
|
| 1897 |
+
"reward_std": 0.16407952923327684,
|
| 1898 |
+
"rewards/accuracy_reward": 0.7578125,
|
| 1899 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 1900 |
+
"step": 135
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"clip_ratio": 0.0,
|
| 1904 |
+
"completion_length": 625.7645416259766,
|
| 1905 |
+
"epoch": 1.385297845373891,
|
| 1906 |
+
"grad_norm": 0.1232142373919487,
|
| 1907 |
+
"learning_rate": 2.603755066399718e-07,
|
| 1908 |
+
"loss": 0.01,
|
| 1909 |
+
"num_tokens": 93916323.0,
|
| 1910 |
+
"reward": 1.7801340371370316,
|
| 1911 |
+
"reward_std": 0.1467695259489119,
|
| 1912 |
+
"rewards/accuracy_reward": 0.7823660671710968,
|
| 1913 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1914 |
+
"step": 136
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"clip_ratio": 0.0,
|
| 1918 |
+
"completion_length": 686.4006958007812,
|
| 1919 |
+
"epoch": 1.3954372623574145,
|
| 1920 |
+
"grad_norm": 0.11197436600923538,
|
| 1921 |
+
"learning_rate": 2.5258082962836614e-07,
|
| 1922 |
+
"loss": 0.0197,
|
| 1923 |
+
"num_tokens": 94665898.0,
|
| 1924 |
+
"reward": 1.6082590073347092,
|
| 1925 |
+
"reward_std": 0.14906749688088894,
|
| 1926 |
+
"rewards/accuracy_reward": 0.613839291036129,
|
| 1927 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 1928 |
+
"step": 137
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"clip_ratio": 0.0,
|
| 1932 |
+
"completion_length": 634.619457244873,
|
| 1933 |
+
"epoch": 1.4055766793409379,
|
| 1934 |
+
"grad_norm": 0.09477974474430084,
|
| 1935 |
+
"learning_rate": 2.4486498346955023e-07,
|
| 1936 |
+
"loss": 0.0172,
|
| 1937 |
+
"num_tokens": 95357349.0,
|
| 1938 |
+
"reward": 1.7488840222358704,
|
| 1939 |
+
"reward_std": 0.14468519762158394,
|
| 1940 |
+
"rewards/accuracy_reward": 0.7555803582072258,
|
| 1941 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1942 |
+
"step": 138
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"clip_ratio": 0.0,
|
| 1946 |
+
"completion_length": 669.0736923217773,
|
| 1947 |
+
"epoch": 1.4157160963244613,
|
| 1948 |
+
"grad_norm": 0.11749354004859924,
|
| 1949 |
+
"learning_rate": 2.372304265289436e-07,
|
| 1950 |
+
"loss": 0.0018,
|
| 1951 |
+
"num_tokens": 96081135.0,
|
| 1952 |
+
"reward": 1.726562574505806,
|
| 1953 |
+
"reward_std": 0.17081871442496777,
|
| 1954 |
+
"rewards/accuracy_reward": 0.7276785746216774,
|
| 1955 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1956 |
+
"step": 139
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"clip_ratio": 0.0,
|
| 1960 |
+
"completion_length": 634.8705596923828,
|
| 1961 |
+
"epoch": 1.4258555133079849,
|
| 1962 |
+
"grad_norm": 0.10797657072544098,
|
| 1963 |
+
"learning_rate": 2.2967959127220137e-07,
|
| 1964 |
+
"loss": 0.0227,
|
| 1965 |
+
"num_tokens": 96771939.0,
|
| 1966 |
+
"reward": 1.7611608058214188,
|
| 1967 |
+
"reward_std": 0.13173524383455515,
|
| 1968 |
+
"rewards/accuracy_reward": 0.7689732015132904,
|
| 1969 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1970 |
+
"step": 140
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"clip_ratio": 0.0,
|
| 1974 |
+
"completion_length": 620.8694534301758,
|
| 1975 |
+
"epoch": 1.4359949302915083,
|
| 1976 |
+
"grad_norm": 0.12707878649234772,
|
| 1977 |
+
"learning_rate": 2.2221488349019902e-07,
|
| 1978 |
+
"loss": 0.0104,
|
| 1979 |
+
"num_tokens": 97476598.0,
|
| 1980 |
+
"reward": 1.7220982909202576,
|
| 1981 |
+
"reward_std": 0.17165770754218102,
|
| 1982 |
+
"rewards/accuracy_reward": 0.7265625,
|
| 1983 |
+
"rewards/format_reward": 0.995535708963871,
|
| 1984 |
+
"step": 141
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"clip_ratio": 0.0,
|
| 1988 |
+
"completion_length": 622.1797180175781,
|
| 1989 |
+
"epoch": 1.4461343472750317,
|
| 1990 |
+
"grad_norm": 0.1114574521780014,
|
| 1991 |
+
"learning_rate": 2.1483868153251788e-07,
|
| 1992 |
+
"loss": 0.0302,
|
| 1993 |
+
"num_tokens": 98165615.0,
|
| 1994 |
+
"reward": 1.7801340222358704,
|
| 1995 |
+
"reward_std": 0.15362949669361115,
|
| 1996 |
+
"rewards/accuracy_reward": 0.7924107164144516,
|
| 1997 |
+
"rewards/format_reward": 0.987723208963871,
|
| 1998 |
+
"step": 142
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"clip_ratio": 0.0,
|
| 2002 |
+
"completion_length": 687.9710159301758,
|
| 2003 |
+
"epoch": 1.456273764258555,
|
| 2004 |
+
"grad_norm": 0.10450585186481476,
|
| 2005 |
+
"learning_rate": 2.0755333554967346e-07,
|
| 2006 |
+
"loss": 0.0098,
|
| 2007 |
+
"num_tokens": 98919237.0,
|
| 2008 |
+
"reward": 1.679687574505806,
|
| 2009 |
+
"reward_std": 0.17744049057364464,
|
| 2010 |
+
"rewards/accuracy_reward": 0.6830357126891613,
|
| 2011 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 2012 |
+
"step": 143
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"clip_ratio": 0.0,
|
| 2016 |
+
"completion_length": 690.1897621154785,
|
| 2017 |
+
"epoch": 1.4664131812420786,
|
| 2018 |
+
"grad_norm": 0.10399246960878372,
|
| 2019 |
+
"learning_rate": 2.0036116674432652e-07,
|
| 2020 |
+
"loss": 0.0263,
|
| 2021 |
+
"num_tokens": 99674095.0,
|
| 2022 |
+
"reward": 1.6863840222358704,
|
| 2023 |
+
"reward_std": 0.15410742722451687,
|
| 2024 |
+
"rewards/accuracy_reward": 0.6941964402794838,
|
| 2025 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 2026 |
+
"step": 144
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"clip_ratio": 0.0,
|
| 2030 |
+
"completion_length": 599.2310600280762,
|
| 2031 |
+
"epoch": 1.476552598225602,
|
| 2032 |
+
"grad_norm": 0.09947122633457184,
|
| 2033 |
+
"learning_rate": 1.9326446663172035e-07,
|
| 2034 |
+
"loss": 0.0121,
|
| 2035 |
+
"num_tokens": 100342486.0,
|
| 2036 |
+
"reward": 1.7868304401636124,
|
| 2037 |
+
"reward_std": 0.13388300314545631,
|
| 2038 |
+
"rewards/accuracy_reward": 0.7879464253783226,
|
| 2039 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 2040 |
+
"step": 145
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"clip_ratio": 0.0,
|
| 2044 |
+
"completion_length": 595.1373062133789,
|
| 2045 |
+
"epoch": 1.4866920152091254,
|
| 2046 |
+
"grad_norm": 0.11993813514709473,
|
| 2047 |
+
"learning_rate": 1.8626549630957395e-07,
|
| 2048 |
+
"loss": 0.0045,
|
| 2049 |
+
"num_tokens": 101003873.0,
|
| 2050 |
+
"reward": 1.7544643580913544,
|
| 2051 |
+
"reward_std": 0.14560767449438572,
|
| 2052 |
+
"rewards/accuracy_reward": 0.7555803433060646,
|
| 2053 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 2054 |
+
"step": 146
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"clip_ratio": 0.0,
|
| 2058 |
+
"completion_length": 645.3449020385742,
|
| 2059 |
+
"epoch": 1.496831432192649,
|
| 2060 |
+
"grad_norm": 0.1037568524479866,
|
| 2061 |
+
"learning_rate": 1.7936648573766954e-07,
|
| 2062 |
+
"loss": 0.0197,
|
| 2063 |
+
"num_tokens": 101718814.0,
|
| 2064 |
+
"reward": 1.727678656578064,
|
| 2065 |
+
"reward_std": 0.1583123281598091,
|
| 2066 |
+
"rewards/accuracy_reward": 0.7366071492433548,
|
| 2067 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 2068 |
+
"step": 147
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"clip_ratio": 0.0,
|
| 2072 |
+
"completion_length": 665.5960121154785,
|
| 2073 |
+
"epoch": 1.5069708491761724,
|
| 2074 |
+
"grad_norm": 0.11654334515333176,
|
| 2075 |
+
"learning_rate": 1.725696330273575e-07,
|
| 2076 |
+
"loss": 0.0334,
|
| 2077 |
+
"num_tokens": 102445532.0,
|
| 2078 |
+
"reward": 1.6908482909202576,
|
| 2079 |
+
"reward_std": 0.15235634427517653,
|
| 2080 |
+
"rewards/accuracy_reward": 0.6997767873108387,
|
| 2081 |
+
"rewards/format_reward": 0.991071417927742,
|
| 2082 |
+
"step": 148
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"clip_ratio": 0.0,
|
| 2086 |
+
"completion_length": 615.9922103881836,
|
| 2087 |
+
"epoch": 1.5171102661596958,
|
| 2088 |
+
"grad_norm": 0.11088323593139648,
|
| 2089 |
+
"learning_rate": 1.65877103741212e-07,
|
| 2090 |
+
"loss": 0.0102,
|
| 2091 |
+
"num_tokens": 103122189.0,
|
| 2092 |
+
"reward": 1.7834822088479996,
|
| 2093 |
+
"reward_std": 0.15306678228080273,
|
| 2094 |
+
"rewards/accuracy_reward": 0.7901785671710968,
|
| 2095 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 2096 |
+
"step": 149
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"clip_ratio": 0.0,
|
| 2100 |
+
"completion_length": 649.7199020385742,
|
| 2101 |
+
"epoch": 1.5272496831432192,
|
| 2102 |
+
"grad_norm": 0.1219962015748024,
|
| 2103 |
+
"learning_rate": 1.592910302030544e-07,
|
| 2104 |
+
"loss": 0.0174,
|
| 2105 |
+
"num_tokens": 103823642.0,
|
| 2106 |
+
"reward": 1.7399554401636124,
|
| 2107 |
+
"reward_std": 0.18228058330714703,
|
| 2108 |
+
"rewards/accuracy_reward": 0.7488839402794838,
|
| 2109 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 2110 |
+
"step": 150
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"clip_ratio": 0.0,
|
| 2114 |
+
"completion_length": 668.3158836364746,
|
| 2115 |
+
"epoch": 1.5373891001267426,
|
| 2116 |
+
"grad_norm": 0.12323982268571854,
|
| 2117 |
+
"learning_rate": 1.5281351081856976e-07,
|
| 2118 |
+
"loss": 0.0218,
|
| 2119 |
+
"num_tokens": 104553709.0,
|
| 2120 |
+
"reward": 1.7120536416769028,
|
| 2121 |
+
"reward_std": 0.16315976157784462,
|
| 2122 |
+
"rewards/accuracy_reward": 0.723214291036129,
|
| 2123 |
+
"rewards/format_reward": 0.9888392761349678,
|
| 2124 |
+
"step": 151
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"clip_ratio": 0.0,
|
| 2128 |
+
"completion_length": 625.3861770629883,
|
| 2129 |
+
"epoch": 1.5475285171102662,
|
| 2130 |
+
"grad_norm": 0.1106390506029129,
|
| 2131 |
+
"learning_rate": 1.4644660940672627e-07,
|
| 2132 |
+
"loss": 0.0111,
|
| 2133 |
+
"num_tokens": 105239647.0,
|
| 2134 |
+
"reward": 1.781250074505806,
|
| 2135 |
+
"reward_std": 0.12235020846128464,
|
| 2136 |
+
"rewards/accuracy_reward": 0.7868303582072258,
|
| 2137 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 2138 |
+
"step": 152
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"clip_ratio": 0.0,
|
| 2142 |
+
"completion_length": 619.7009201049805,
|
| 2143 |
+
"epoch": 1.5576679340937896,
|
| 2144 |
+
"grad_norm": 0.10685606300830841,
|
| 2145 |
+
"learning_rate": 1.4019235454221856e-07,
|
| 2146 |
+
"loss": 0.012,
|
| 2147 |
+
"num_tokens": 105928971.0,
|
| 2148 |
+
"reward": 1.728794738650322,
|
| 2149 |
+
"reward_std": 0.13385361433029175,
|
| 2150 |
+
"rewards/accuracy_reward": 0.7321428656578064,
|
| 2151 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 2152 |
+
"step": 153
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"clip_ratio": 0.0,
|
| 2156 |
+
"completion_length": 624.036865234375,
|
| 2157 |
+
"epoch": 1.5678073510773132,
|
| 2158 |
+
"grad_norm": 0.11761969327926636,
|
| 2159 |
+
"learning_rate": 1.3405273890913737e-07,
|
| 2160 |
+
"loss": 0.0085,
|
| 2161 |
+
"num_tokens": 106610124.0,
|
| 2162 |
+
"reward": 1.7343751192092896,
|
| 2163 |
+
"reward_std": 0.17772767413407564,
|
| 2164 |
+
"rewards/accuracy_reward": 0.7354910746216774,
|
| 2165 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 2166 |
+
"step": 154
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"clip_ratio": 0.0,
|
| 2170 |
+
"completion_length": 635.014533996582,
|
| 2171 |
+
"epoch": 1.5779467680608366,
|
| 2172 |
+
"grad_norm": 0.11695584654808044,
|
| 2173 |
+
"learning_rate": 1.280297186660752e-07,
|
| 2174 |
+
"loss": 0.0378,
|
| 2175 |
+
"num_tokens": 107297473.0,
|
| 2176 |
+
"reward": 1.7935268580913544,
|
| 2177 |
+
"reward_std": 0.17546693049371243,
|
| 2178 |
+
"rewards/accuracy_reward": 0.8035714328289032,
|
| 2179 |
+
"rewards/format_reward": 0.9899553582072258,
|
| 2180 |
+
"step": 155
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"clip_ratio": 0.0,
|
| 2184 |
+
"completion_length": 708.3571853637695,
|
| 2185 |
+
"epoch": 1.58808618504436,
|
| 2186 |
+
"grad_norm": 0.11352672427892685,
|
| 2187 |
+
"learning_rate": 1.2212521282287093e-07,
|
| 2188 |
+
"loss": 0.0027,
|
| 2189 |
+
"num_tokens": 108052089.0,
|
| 2190 |
+
"reward": 1.7042411267757416,
|
| 2191 |
+
"reward_std": 0.17979852855205536,
|
| 2192 |
+
"rewards/accuracy_reward": 0.7053571417927742,
|
| 2193 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 2194 |
+
"step": 156
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"clip_ratio": 0.0,
|
| 2198 |
+
"completion_length": 633.8192291259766,
|
| 2199 |
+
"epoch": 1.5982256020278833,
|
| 2200 |
+
"grad_norm": 0.10707947611808777,
|
| 2201 |
+
"learning_rate": 1.1634110262918717e-07,
|
| 2202 |
+
"loss": 0.0118,
|
| 2203 |
+
"num_tokens": 108763999.0,
|
| 2204 |
+
"reward": 1.703125074505806,
|
| 2205 |
+
"reward_std": 0.14697580970823765,
|
| 2206 |
+
"rewards/accuracy_reward": 0.7109375,
|
| 2207 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 2208 |
+
"step": 157
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"clip_ratio": 0.0,
|
| 2212 |
+
"completion_length": 676.0625381469727,
|
| 2213 |
+
"epoch": 1.6083650190114067,
|
| 2214 |
+
"grad_norm": 0.11339569836854935,
|
| 2215 |
+
"learning_rate": 1.1067923097512255e-07,
|
| 2216 |
+
"loss": 0.0181,
|
| 2217 |
+
"num_tokens": 109500463.0,
|
| 2218 |
+
"reward": 1.7220983058214188,
|
| 2219 |
+
"reward_std": 0.18152274563908577,
|
| 2220 |
+
"rewards/accuracy_reward": 0.7287946492433548,
|
| 2221 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2222 |
+
"step": 158
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"clip_ratio": 0.0,
|
| 2226 |
+
"completion_length": 639.6294937133789,
|
| 2227 |
+
"epoch": 1.6185044359949303,
|
| 2228 |
+
"grad_norm": 0.10624842345714569,
|
| 2229 |
+
"learning_rate": 1.0514140180404202e-07,
|
| 2230 |
+
"loss": -0.0033,
|
| 2231 |
+
"num_tokens": 110202627.0,
|
| 2232 |
+
"reward": 1.7232143729925156,
|
| 2233 |
+
"reward_std": 0.16777998208999634,
|
| 2234 |
+
"rewards/accuracy_reward": 0.7265624925494194,
|
| 2235 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 2236 |
+
"step": 159
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"clip_ratio": 0.0,
|
| 2240 |
+
"completion_length": 681.1741485595703,
|
| 2241 |
+
"epoch": 1.6286438529784537,
|
| 2242 |
+
"grad_norm": 0.11243726313114166,
|
| 2243 |
+
"learning_rate": 9.972937953781984e-08,
|
| 2244 |
+
"loss": 0.0145,
|
| 2245 |
+
"num_tokens": 110959927.0,
|
| 2246 |
+
"reward": 1.678571492433548,
|
| 2247 |
+
"reward_std": 0.16837417893111706,
|
| 2248 |
+
"rewards/accuracy_reward": 0.6852678656578064,
|
| 2249 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2250 |
+
"step": 160
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"clip_ratio": 0.0,
|
| 2254 |
+
"completion_length": 605.4766006469727,
|
| 2255 |
+
"epoch": 1.6387832699619773,
|
| 2256 |
+
"grad_norm": 0.09180891513824463,
|
| 2257 |
+
"learning_rate": 9.444488851467041e-08,
|
| 2258 |
+
"loss": 0.019,
|
| 2259 |
+
"num_tokens": 111620106.0,
|
| 2260 |
+
"reward": 1.7968750894069672,
|
| 2261 |
+
"reward_std": 0.11058733332902193,
|
| 2262 |
+
"rewards/accuracy_reward": 0.8035714402794838,
|
| 2263 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2264 |
+
"step": 161
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"clip_ratio": 0.0,
|
| 2268 |
+
"completion_length": 650.7544937133789,
|
| 2269 |
+
"epoch": 1.6489226869455007,
|
| 2270 |
+
"grad_norm": 0.11670933663845062,
|
| 2271 |
+
"learning_rate": 8.928961243975436e-08,
|
| 2272 |
+
"loss": 0.0121,
|
| 2273 |
+
"num_tokens": 112336038.0,
|
| 2274 |
+
"reward": 1.7321429252624512,
|
| 2275 |
+
"reward_std": 0.1541335443034768,
|
| 2276 |
+
"rewards/accuracy_reward": 0.7402987703680992,
|
| 2277 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 2278 |
+
"step": 162
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"clip_ratio": 0.0,
|
| 2282 |
+
"completion_length": 630.4565010070801,
|
| 2283 |
+
"epoch": 1.659062103929024,
|
| 2284 |
+
"grad_norm": 0.10085251182317734,
|
| 2285 |
+
"learning_rate": 8.426519384872732e-08,
|
| 2286 |
+
"loss": 0.0036,
|
| 2287 |
+
"num_tokens": 113023063.0,
|
| 2288 |
+
"reward": 1.7455358058214188,
|
| 2289 |
+
"reward_std": 0.15896250680088997,
|
| 2290 |
+
"rewards/accuracy_reward": 0.7500000074505806,
|
| 2291 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2292 |
+
"step": 163
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"clip_ratio": 0.0,
|
| 2296 |
+
"completion_length": 636.2277069091797,
|
| 2297 |
+
"epoch": 1.6692015209125475,
|
| 2298 |
+
"grad_norm": 0.1013065055012703,
|
| 2299 |
+
"learning_rate": 7.937323358440934e-08,
|
| 2300 |
+
"loss": 0.0099,
|
| 2301 |
+
"num_tokens": 113724155.0,
|
| 2302 |
+
"reward": 1.7533483058214188,
|
| 2303 |
+
"reward_std": 0.13801472913473845,
|
| 2304 |
+
"rewards/accuracy_reward": 0.7555803582072258,
|
| 2305 |
+
"rewards/format_reward": 0.9977678582072258,
|
| 2306 |
+
"step": 164
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"clip_ratio": 0.0,
|
| 2310 |
+
"completion_length": 632.6138687133789,
|
| 2311 |
+
"epoch": 1.6793409378960709,
|
| 2312 |
+
"grad_norm": 0.09690902382135391,
|
| 2313 |
+
"learning_rate": 7.461529028673463e-08,
|
| 2314 |
+
"loss": 0.0219,
|
| 2315 |
+
"num_tokens": 114423785.0,
|
| 2316 |
+
"reward": 1.7834822237491608,
|
| 2317 |
+
"reward_std": 0.1157834529876709,
|
| 2318 |
+
"rewards/accuracy_reward": 0.7868303582072258,
|
| 2319 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 2320 |
+
"step": 165
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"clip_ratio": 0.0,
|
| 2324 |
+
"completion_length": 642.2444458007812,
|
| 2325 |
+
"epoch": 1.6894803548795945,
|
| 2326 |
+
"grad_norm": 0.111725352704525,
|
| 2327 |
+
"learning_rate": 6.999287989614971e-08,
|
| 2328 |
+
"loss": 0.0065,
|
| 2329 |
+
"num_tokens": 115131676.0,
|
| 2330 |
+
"reward": 1.7578126043081284,
|
| 2331 |
+
"reward_std": 0.15221323259174824,
|
| 2332 |
+
"rewards/accuracy_reward": 0.7600446417927742,
|
| 2333 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 2334 |
+
"step": 166
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"clip_ratio": 0.0,
|
| 2338 |
+
"completion_length": 625.8181190490723,
|
| 2339 |
+
"epoch": 1.6996197718631179,
|
| 2340 |
+
"grad_norm": 0.10648291558027267,
|
| 2341 |
+
"learning_rate": 6.550747517061656e-08,
|
| 2342 |
+
"loss": 0.0144,
|
| 2343 |
+
"num_tokens": 115818953.0,
|
| 2344 |
+
"reward": 1.7890625894069672,
|
| 2345 |
+
"reward_std": 0.1342666009441018,
|
| 2346 |
+
"rewards/accuracy_reward": 0.793526791036129,
|
| 2347 |
+
"rewards/format_reward": 0.9955357164144516,
|
| 2348 |
+
"step": 167
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"clip_ratio": 0.0,
|
| 2352 |
+
"completion_length": 649.0558166503906,
|
| 2353 |
+
"epoch": 1.7097591888466415,
|
| 2354 |
+
"grad_norm": 0.10185572504997253,
|
| 2355 |
+
"learning_rate": 6.116050521637218e-08,
|
| 2356 |
+
"loss": 0.0065,
|
| 2357 |
+
"num_tokens": 116528835.0,
|
| 2358 |
+
"reward": 1.7834822237491608,
|
| 2359 |
+
"reward_std": 0.1334951389580965,
|
| 2360 |
+
"rewards/accuracy_reward": 0.7868303507566452,
|
| 2361 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 2362 |
+
"step": 168
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"clip_ratio": 0.0,
|
| 2366 |
+
"completion_length": 623.458740234375,
|
| 2367 |
+
"epoch": 1.7198986058301649,
|
| 2368 |
+
"grad_norm": 0.10859635472297668,
|
| 2369 |
+
"learning_rate": 5.6953355032598795e-08,
|
| 2370 |
+
"loss": 0.0139,
|
| 2371 |
+
"num_tokens": 117224526.0,
|
| 2372 |
+
"reward": 1.7455358058214188,
|
| 2373 |
+
"reward_std": 0.15123825334012508,
|
| 2374 |
+
"rewards/accuracy_reward": 0.7522321417927742,
|
| 2375 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2376 |
+
"step": 169
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"clip_ratio": 0.0,
|
| 2380 |
+
"completion_length": 650.3861846923828,
|
| 2381 |
+
"epoch": 1.7300380228136882,
|
| 2382 |
+
"grad_norm": 0.11233941465616226,
|
| 2383 |
+
"learning_rate": 5.288736507014435e-08,
|
| 2384 |
+
"loss": 0.0094,
|
| 2385 |
+
"num_tokens": 117931272.0,
|
| 2386 |
+
"reward": 1.7790179401636124,
|
| 2387 |
+
"reward_std": 0.16706989519298077,
|
| 2388 |
+
"rewards/accuracy_reward": 0.7857142835855484,
|
| 2389 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2390 |
+
"step": 170
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"clip_ratio": 0.0,
|
| 2394 |
+
"completion_length": 621.7712326049805,
|
| 2395 |
+
"epoch": 1.7401774397972116,
|
| 2396 |
+
"grad_norm": 0.11732303351163864,
|
| 2397 |
+
"learning_rate": 4.896383080443933e-08,
|
| 2398 |
+
"loss": 0.0146,
|
| 2399 |
+
"num_tokens": 118613131.0,
|
| 2400 |
+
"reward": 1.7321429252624512,
|
| 2401 |
+
"reward_std": 0.16958792228251696,
|
| 2402 |
+
"rewards/accuracy_reward": 0.738839291036129,
|
| 2403 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2404 |
+
"step": 171
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"clip_ratio": 0.0,
|
| 2408 |
+
"completion_length": 619.6361885070801,
|
| 2409 |
+
"epoch": 1.750316856780735,
|
| 2410 |
+
"grad_norm": 0.11354614794254303,
|
| 2411 |
+
"learning_rate": 4.518400232274078e-08,
|
| 2412 |
+
"loss": 0.0137,
|
| 2413 |
+
"num_tokens": 119298829.0,
|
| 2414 |
+
"reward": 1.7466518580913544,
|
| 2415 |
+
"reward_std": 0.16586073767393827,
|
| 2416 |
+
"rewards/accuracy_reward": 0.7564388811588287,
|
| 2417 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 2418 |
+
"step": 172
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"clip_ratio": 0.0,
|
| 2422 |
+
"completion_length": 603.4085083007812,
|
| 2423 |
+
"epoch": 1.7604562737642584,
|
| 2424 |
+
"grad_norm": 0.11991310119628906,
|
| 2425 |
+
"learning_rate": 4.1549083925840165e-08,
|
| 2426 |
+
"loss": 0.0226,
|
| 2427 |
+
"num_tokens": 119956931.0,
|
| 2428 |
+
"reward": 1.8147322237491608,
|
| 2429 |
+
"reward_std": 0.12004590220749378,
|
| 2430 |
+
"rewards/accuracy_reward": 0.8191964328289032,
|
| 2431 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2432 |
+
"step": 173
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"clip_ratio": 0.0,
|
| 2436 |
+
"completion_length": 602.5100708007812,
|
| 2437 |
+
"epoch": 1.770595690747782,
|
| 2438 |
+
"grad_norm": 0.13672617077827454,
|
| 2439 |
+
"learning_rate": 3.806023374435663e-08,
|
| 2440 |
+
"loss": 0.0039,
|
| 2441 |
+
"num_tokens": 120625308.0,
|
| 2442 |
+
"reward": 1.7700893729925156,
|
| 2443 |
+
"reward_std": 0.17995007149875164,
|
| 2444 |
+
"rewards/accuracy_reward": 0.7712053582072258,
|
| 2445 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 2446 |
+
"step": 174
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"clip_ratio": 0.0,
|
| 2450 |
+
"completion_length": 647.302490234375,
|
| 2451 |
+
"epoch": 1.7807351077313056,
|
| 2452 |
+
"grad_norm": 0.10016340017318726,
|
| 2453 |
+
"learning_rate": 3.4718563369743213e-08,
|
| 2454 |
+
"loss": 0.0119,
|
| 2455 |
+
"num_tokens": 121338619.0,
|
| 2456 |
+
"reward": 1.7265626043081284,
|
| 2457 |
+
"reward_std": 0.1343973781913519,
|
| 2458 |
+
"rewards/accuracy_reward": 0.7310267984867096,
|
| 2459 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2460 |
+
"step": 175
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"clip_ratio": 0.0,
|
| 2464 |
+
"completion_length": 626.5011367797852,
|
| 2465 |
+
"epoch": 1.790874524714829,
|
| 2466 |
+
"grad_norm": 0.10784109681844711,
|
| 2467 |
+
"learning_rate": 3.15251375001192e-08,
|
| 2468 |
+
"loss": 0.0193,
|
| 2469 |
+
"num_tokens": 122040068.0,
|
| 2470 |
+
"reward": 1.7946429550647736,
|
| 2471 |
+
"reward_std": 0.13197333365678787,
|
| 2472 |
+
"rewards/accuracy_reward": 0.8013392835855484,
|
| 2473 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 2474 |
+
"step": 176
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"clip_ratio": 0.0,
|
| 2478 |
+
"completion_length": 627.9163284301758,
|
| 2479 |
+
"epoch": 1.8010139416983524,
|
| 2480 |
+
"grad_norm": 0.11255474388599396,
|
| 2481 |
+
"learning_rate": 2.8480973601043955e-08,
|
| 2482 |
+
"loss": 0.0134,
|
| 2483 |
+
"num_tokens": 122737489.0,
|
| 2484 |
+
"reward": 1.726562574505806,
|
| 2485 |
+
"reward_std": 0.10841276589781046,
|
| 2486 |
+
"rewards/accuracy_reward": 0.7299107126891613,
|
| 2487 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 2488 |
+
"step": 177
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"clip_ratio": 0.0,
|
| 2492 |
+
"completion_length": 641.2310562133789,
|
| 2493 |
+
"epoch": 1.8111533586818758,
|
| 2494 |
+
"grad_norm": 0.12718521058559418,
|
| 2495 |
+
"learning_rate": 2.558704158134023e-08,
|
| 2496 |
+
"loss": 0.0123,
|
| 2497 |
+
"num_tokens": 123452056.0,
|
| 2498 |
+
"reward": 1.712053656578064,
|
| 2499 |
+
"reward_std": 0.16341756749898195,
|
| 2500 |
+
"rewards/accuracy_reward": 0.7154017984867096,
|
| 2501 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 2502 |
+
"step": 178
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"clip_ratio": 0.0,
|
| 2506 |
+
"completion_length": 661.897346496582,
|
| 2507 |
+
"epoch": 1.8212927756653992,
|
| 2508 |
+
"grad_norm": 0.10139390826225281,
|
| 2509 |
+
"learning_rate": 2.2844263484068093e-08,
|
| 2510 |
+
"loss": 0.0179,
|
| 2511 |
+
"num_tokens": 124172116.0,
|
| 2512 |
+
"reward": 1.7165179401636124,
|
| 2513 |
+
"reward_std": 0.14442832861095667,
|
| 2514 |
+
"rewards/accuracy_reward": 0.7243303582072258,
|
| 2515 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 2516 |
+
"step": 179
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"clip_ratio": 0.0,
|
| 2520 |
+
"completion_length": 620.7277069091797,
|
| 2521 |
+
"epoch": 1.8314321926489225,
|
| 2522 |
+
"grad_norm": 0.11224541813135147,
|
| 2523 |
+
"learning_rate": 2.025351319275137e-08,
|
| 2524 |
+
"loss": 0.0239,
|
| 2525 |
+
"num_tokens": 124859256.0,
|
| 2526 |
+
"reward": 1.751116156578064,
|
| 2527 |
+
"reward_std": 0.14368766359984875,
|
| 2528 |
+
"rewards/accuracy_reward": 0.7578124925494194,
|
| 2529 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2530 |
+
"step": 180
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"clip_ratio": 0.0,
|
| 2534 |
+
"completion_length": 659.0647583007812,
|
| 2535 |
+
"epoch": 1.8415716096324461,
|
| 2536 |
+
"grad_norm": 0.11130577325820923,
|
| 2537 |
+
"learning_rate": 1.781561615294652e-08,
|
| 2538 |
+
"loss": 0.0061,
|
| 2539 |
+
"num_tokens": 125584242.0,
|
| 2540 |
+
"reward": 1.7154018580913544,
|
| 2541 |
+
"reward_std": 0.1235003275796771,
|
| 2542 |
+
"rewards/accuracy_reward": 0.7198660597205162,
|
| 2543 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2544 |
+
"step": 181
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"clip_ratio": 0.0,
|
| 2548 |
+
"completion_length": 663.6573867797852,
|
| 2549 |
+
"epoch": 1.8517110266159695,
|
| 2550 |
+
"grad_norm": 0.12259829789400101,
|
| 2551 |
+
"learning_rate": 1.553134910924636e-08,
|
| 2552 |
+
"loss": 0.0177,
|
| 2553 |
+
"num_tokens": 126314983.0,
|
| 2554 |
+
"reward": 1.671875074505806,
|
| 2555 |
+
"reward_std": 0.17521491926163435,
|
| 2556 |
+
"rewards/accuracy_reward": 0.6819196417927742,
|
| 2557 |
+
"rewards/format_reward": 0.9899553582072258,
|
| 2558 |
+
"step": 182
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"clip_ratio": 0.0,
|
| 2562 |
+
"completion_length": 610.7846221923828,
|
| 2563 |
+
"epoch": 1.8618504435994931,
|
| 2564 |
+
"grad_norm": 0.1277763843536377,
|
| 2565 |
+
"learning_rate": 1.340143985779829e-08,
|
| 2566 |
+
"loss": 0.0134,
|
| 2567 |
+
"num_tokens": 126989006.0,
|
| 2568 |
+
"reward": 1.7466518729925156,
|
| 2569 |
+
"reward_std": 0.13783067418262362,
|
| 2570 |
+
"rewards/accuracy_reward": 0.7511160671710968,
|
| 2571 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2572 |
+
"step": 183
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"clip_ratio": 0.0,
|
| 2576 |
+
"completion_length": 625.2879867553711,
|
| 2577 |
+
"epoch": 1.8719898605830165,
|
| 2578 |
+
"grad_norm": 0.10394810885190964,
|
| 2579 |
+
"learning_rate": 1.1426567014420297e-08,
|
| 2580 |
+
"loss": 0.0122,
|
| 2581 |
+
"num_tokens": 127675968.0,
|
| 2582 |
+
"reward": 1.8035715073347092,
|
| 2583 |
+
"reward_std": 0.12675560358911753,
|
| 2584 |
+
"rewards/accuracy_reward": 0.8102678507566452,
|
| 2585 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 2586 |
+
"step": 184
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"clip_ratio": 0.0,
|
| 2590 |
+
"completion_length": 661.4832916259766,
|
| 2591 |
+
"epoch": 1.88212927756654,
|
| 2592 |
+
"grad_norm": 0.10645025223493576,
|
| 2593 |
+
"learning_rate": 9.607359798384784e-09,
|
| 2594 |
+
"loss": 0.022,
|
| 2595 |
+
"num_tokens": 128384137.0,
|
| 2596 |
+
"reward": 1.7500000596046448,
|
| 2597 |
+
"reward_std": 0.1677253060042858,
|
| 2598 |
+
"rewards/accuracy_reward": 0.7555803582072258,
|
| 2599 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 2600 |
+
"step": 185
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"clip_ratio": 0.0,
|
| 2604 |
+
"completion_length": 628.4553909301758,
|
| 2605 |
+
"epoch": 1.8922686945500633,
|
| 2606 |
+
"grad_norm": 0.12001987546682358,
|
| 2607 |
+
"learning_rate": 7.944397831941951e-09,
|
| 2608 |
+
"loss": 0.0204,
|
| 2609 |
+
"num_tokens": 129093729.0,
|
| 2610 |
+
"reward": 1.710937574505806,
|
| 2611 |
+
"reward_std": 0.1775853894650936,
|
| 2612 |
+
"rewards/accuracy_reward": 0.7176339253783226,
|
| 2613 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 2614 |
+
"step": 186
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"clip_ratio": 0.0,
|
| 2618 |
+
"completion_length": 682.1506958007812,
|
| 2619 |
+
"epoch": 1.9024081115335867,
|
| 2620 |
+
"grad_norm": 0.11876320838928223,
|
| 2621 |
+
"learning_rate": 6.438210955644452e-09,
|
| 2622 |
+
"loss": 0.0227,
|
| 2623 |
+
"num_tokens": 129847360.0,
|
| 2624 |
+
"reward": 1.6707590073347092,
|
| 2625 |
+
"reward_std": 0.14890040084719658,
|
| 2626 |
+
"rewards/accuracy_reward": 0.6774553619325161,
|
| 2627 |
+
"rewards/format_reward": 0.9933035746216774,
|
| 2628 |
+
"step": 187
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"clip_ratio": 0.0,
|
| 2632 |
+
"completion_length": 632.9118576049805,
|
| 2633 |
+
"epoch": 1.9125475285171103,
|
| 2634 |
+
"grad_norm": 0.10546144843101501,
|
| 2635 |
+
"learning_rate": 5.0892790595336575e-09,
|
| 2636 |
+
"loss": 0.0116,
|
| 2637 |
+
"num_tokens": 130551465.0,
|
| 2638 |
+
"reward": 1.7399554401636124,
|
| 2639 |
+
"reward_std": 0.12469578720629215,
|
| 2640 |
+
"rewards/accuracy_reward": 0.7433035746216774,
|
| 2641 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 2642 |
+
"step": 188
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"clip_ratio": 0.0,
|
| 2646 |
+
"completion_length": 633.2812728881836,
|
| 2647 |
+
"epoch": 1.9226869455006337,
|
| 2648 |
+
"grad_norm": 0.11886601895093918,
|
| 2649 |
+
"learning_rate": 3.898031930240797e-09,
|
| 2650 |
+
"loss": 0.011,
|
| 2651 |
+
"num_tokens": 131243885.0,
|
| 2652 |
+
"reward": 1.7555804252624512,
|
| 2653 |
+
"reward_std": 0.14942110795527697,
|
| 2654 |
+
"rewards/accuracy_reward": 0.7578125,
|
| 2655 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 2656 |
+
"step": 189
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"clip_ratio": 0.0,
|
| 2660 |
+
"completion_length": 657.8850631713867,
|
| 2661 |
+
"epoch": 1.9328263624841573,
|
| 2662 |
+
"grad_norm": 0.14054881036281586,
|
| 2663 |
+
"learning_rate": 2.8648491140513264e-09,
|
| 2664 |
+
"loss": 0.0237,
|
| 2665 |
+
"num_tokens": 131959798.0,
|
| 2666 |
+
"reward": 1.7622768580913544,
|
| 2667 |
+
"reward_std": 0.1422654939815402,
|
| 2668 |
+
"rewards/accuracy_reward": 0.7656249925494194,
|
| 2669 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 2670 |
+
"step": 190
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"clip_ratio": 0.0,
|
| 2674 |
+
"completion_length": 580.6707878112793,
|
| 2675 |
+
"epoch": 1.9429657794676807,
|
| 2676 |
+
"grad_norm": 0.10732496529817581,
|
| 2677 |
+
"learning_rate": 1.9900597959770505e-09,
|
| 2678 |
+
"loss": 0.0063,
|
| 2679 |
+
"num_tokens": 132605207.0,
|
| 2680 |
+
"reward": 1.8348215222358704,
|
| 2681 |
+
"reward_std": 0.1142951613292098,
|
| 2682 |
+
"rewards/accuracy_reward": 0.8370535671710968,
|
| 2683 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 2684 |
+
"step": 191
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"clip_ratio": 0.0,
|
| 2688 |
+
"completion_length": 662.7555999755859,
|
| 2689 |
+
"epoch": 1.953105196451204,
|
| 2690 |
+
"grad_norm": 0.10711626708507538,
|
| 2691 |
+
"learning_rate": 1.2739426948732424e-09,
|
| 2692 |
+
"loss": 0.0112,
|
| 2693 |
+
"num_tokens": 133333004.0,
|
| 2694 |
+
"reward": 1.7488840073347092,
|
| 2695 |
+
"reward_std": 0.1441500186920166,
|
| 2696 |
+
"rewards/accuracy_reward": 0.7566964253783226,
|
| 2697 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 2698 |
+
"step": 192
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"clip_ratio": 0.0,
|
| 2702 |
+
"completion_length": 649.7835159301758,
|
| 2703 |
+
"epoch": 1.9632446134347274,
|
| 2704 |
+
"grad_norm": 0.12485576421022415,
|
| 2705 |
+
"learning_rate": 7.16725974635568e-10,
|
| 2706 |
+
"loss": 0.0127,
|
| 2707 |
+
"num_tokens": 134041050.0,
|
| 2708 |
+
"reward": 1.789062574505806,
|
| 2709 |
+
"reward_std": 0.15979287773370743,
|
| 2710 |
+
"rewards/accuracy_reward": 0.7946428582072258,
|
| 2711 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 2712 |
+
"step": 193
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"clip_ratio": 0.0,
|
| 2716 |
+
"completion_length": 666.5379867553711,
|
| 2717 |
+
"epoch": 1.9733840304182508,
|
| 2718 |
+
"grad_norm": 0.11115089803934097,
|
| 2719 |
+
"learning_rate": 3.185871715041255e-10,
|
| 2720 |
+
"loss": -0.0016,
|
| 2721 |
+
"num_tokens": 134771116.0,
|
| 2722 |
+
"reward": 1.7220982909202576,
|
| 2723 |
+
"reward_std": 0.14188571739941835,
|
| 2724 |
+
"rewards/accuracy_reward": 0.7243303507566452,
|
| 2725 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 2726 |
+
"step": 194
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"clip_ratio": 0.0,
|
| 2730 |
+
"completion_length": 599.9765777587891,
|
| 2731 |
+
"epoch": 1.9835234474017744,
|
| 2732 |
+
"grad_norm": 0.11773505806922913,
|
| 2733 |
+
"learning_rate": 7.96531374983589e-11,
|
| 2734 |
+
"loss": 0.0201,
|
| 2735 |
+
"num_tokens": 135429175.0,
|
| 2736 |
+
"reward": 1.7845983058214188,
|
| 2737 |
+
"reward_std": 0.16611923463642597,
|
| 2738 |
+
"rewards/accuracy_reward": 0.7890625,
|
| 2739 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2740 |
+
"step": 195
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"clip_ratio": 0.0,
|
| 2744 |
+
"completion_length": 647.62109375,
|
| 2745 |
+
"epoch": 1.9936628643852978,
|
| 2746 |
+
"grad_norm": 0.12847015261650085,
|
| 2747 |
+
"learning_rate": 0.0,
|
| 2748 |
+
"loss": 0.0274,
|
| 2749 |
+
"num_tokens": 136136408.0,
|
| 2750 |
+
"reward": 1.7332590222358704,
|
| 2751 |
+
"reward_std": 0.1759201493114233,
|
| 2752 |
+
"rewards/accuracy_reward": 0.7377232164144516,
|
| 2753 |
+
"rewards/format_reward": 0.995535708963871,
|
| 2754 |
+
"step": 196
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 1.9936628643852978,
|
| 2758 |
+
"step": 196,
|
| 2759 |
+
"total_flos": 0.0,
|
| 2760 |
+
"train_loss": 0.020176043500826808,
|
| 2761 |
+
"train_runtime": 41639.173,
|
| 2762 |
+
"train_samples_per_second": 0.53,
|
| 2763 |
+
"train_steps_per_second": 0.005
|
| 2764 |
+
}
|
| 2765 |
+
],
|
| 2766 |
+
"logging_steps": 1,
|
| 2767 |
+
"max_steps": 196,
|
| 2768 |
+
"num_input_tokens_seen": 0,
|
| 2769 |
+
"num_train_epochs": 2,
|
| 2770 |
+
"save_steps": 500,
|
| 2771 |
+
"stateful_callbacks": {
|
| 2772 |
+
"TrainerControl": {
|
| 2773 |
+
"args": {
|
| 2774 |
+
"should_epoch_stop": false,
|
| 2775 |
+
"should_evaluate": false,
|
| 2776 |
+
"should_log": false,
|
| 2777 |
+
"should_save": true,
|
| 2778 |
+
"should_training_stop": true
|
| 2779 |
+
},
|
| 2780 |
+
"attributes": {}
|
| 2781 |
+
}
|
| 2782 |
+
},
|
| 2783 |
+
"total_flos": 0.0,
|
| 2784 |
+
"train_batch_size": 16,
|
| 2785 |
+
"trial_name": null,
|
| 2786 |
+
"trial_params": null
|
| 2787 |
+
}
|