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 +1898 -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-lighteval
|
| 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-lighteval
|
| 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-lighteval", 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/q96trpqe)
|
| 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: 4.49.0
|
| 39 |
+
- Pytorch: 2.5.1+cu121
|
| 40 |
+
- Datasets: 3.3.1
|
| 41 |
+
- Tokenizers: 0.21.0
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
Cite GRPO as:
|
| 46 |
+
|
| 47 |
+
```bibtex
|
| 48 |
+
@article{zhihong2024deepseekmath,
|
| 49 |
+
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
|
| 50 |
+
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
|
| 51 |
+
year = 2024,
|
| 52 |
+
eprint = {arXiv:2402.03300},
|
| 53 |
+
}
|
| 54 |
+
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
Cite TRL as:
|
| 58 |
+
|
| 59 |
+
```bibtex
|
| 60 |
+
@misc{vonwerra2022trl,
|
| 61 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 62 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
|
| 63 |
+
year = 2020,
|
| 64 |
+
journal = {GitHub repository},
|
| 65 |
+
publisher = {GitHub},
|
| 66 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 67 |
+
}
|
| 68 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.023034477807496758,
|
| 4 |
+
"train_runtime": 35029.9588,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 0.428,
|
| 7 |
+
"train_steps_per_second": 0.004
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"eos_token_id": 151643,
|
| 4 |
+
"max_new_tokens": 2048,
|
| 5 |
+
"transformers_version": "4.49.0"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.023034477807496758,
|
| 4 |
+
"train_runtime": 35029.9588,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 0.428,
|
| 7 |
+
"train_steps_per_second": 0.004
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1898 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.9850746268656716,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 132,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"clip_ratio": 0.0,
|
| 13 |
+
"completion_length": 468.4821586608887,
|
| 14 |
+
"epoch": 0.014925373134328358,
|
| 15 |
+
"grad_norm": 0.6381782293319702,
|
| 16 |
+
"learning_rate": 7.142857142857142e-08,
|
| 17 |
+
"loss": 0.0029,
|
| 18 |
+
"num_tokens": 546936.0,
|
| 19 |
+
"reward": 0.27120537124574184,
|
| 20 |
+
"reward_std": 0.39265505224466324,
|
| 21 |
+
"rewards/accuracy_reward": 0.20647320989519358,
|
| 22 |
+
"rewards/format_reward": 0.06473214365541935,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"clip_ratio": 0.0,
|
| 27 |
+
"completion_length": 471.2355079650879,
|
| 28 |
+
"epoch": 0.029850746268656716,
|
| 29 |
+
"grad_norm": 0.47330012917518616,
|
| 30 |
+
"learning_rate": 1.4285714285714285e-07,
|
| 31 |
+
"loss": 0.0029,
|
| 32 |
+
"num_tokens": 1100635.0,
|
| 33 |
+
"reward": 0.29352679662406445,
|
| 34 |
+
"reward_std": 0.3840556889772415,
|
| 35 |
+
"rewards/accuracy_reward": 0.22767856996506453,
|
| 36 |
+
"rewards/format_reward": 0.06584821548312902,
|
| 37 |
+
"step": 2
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"clip_ratio": 0.0,
|
| 41 |
+
"completion_length": 503.05247497558594,
|
| 42 |
+
"epoch": 0.04477611940298507,
|
| 43 |
+
"grad_norm": 0.40551432967185974,
|
| 44 |
+
"learning_rate": 2.1428571428571426e-07,
|
| 45 |
+
"loss": -0.0092,
|
| 46 |
+
"num_tokens": 1698274.0,
|
| 47 |
+
"reward": 0.24888394214212894,
|
| 48 |
+
"reward_std": 0.36480507254600525,
|
| 49 |
+
"rewards/accuracy_reward": 0.20424107275903225,
|
| 50 |
+
"rewards/format_reward": 0.04464285809081048,
|
| 51 |
+
"step": 3
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"clip_ratio": 0.0,
|
| 55 |
+
"completion_length": 526.8214530944824,
|
| 56 |
+
"epoch": 0.05970149253731343,
|
| 57 |
+
"grad_norm": 0.4093160331249237,
|
| 58 |
+
"learning_rate": 2.857142857142857e-07,
|
| 59 |
+
"loss": 0.0128,
|
| 60 |
+
"num_tokens": 2299818.0,
|
| 61 |
+
"reward": 0.25781251303851604,
|
| 62 |
+
"reward_std": 0.35233646258711815,
|
| 63 |
+
"rewards/accuracy_reward": 0.1997767873108387,
|
| 64 |
+
"rewards/format_reward": 0.05803571571595967,
|
| 65 |
+
"step": 4
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"clip_ratio": 0.0,
|
| 69 |
+
"completion_length": 512.1830596923828,
|
| 70 |
+
"epoch": 0.07462686567164178,
|
| 71 |
+
"grad_norm": 0.5151273012161255,
|
| 72 |
+
"learning_rate": 3.5714285714285716e-07,
|
| 73 |
+
"loss": 0.0079,
|
| 74 |
+
"num_tokens": 2896238.0,
|
| 75 |
+
"reward": 0.2667410895228386,
|
| 76 |
+
"reward_std": 0.36962299421429634,
|
| 77 |
+
"rewards/accuracy_reward": 0.18750000279396772,
|
| 78 |
+
"rewards/format_reward": 0.07924107019789517,
|
| 79 |
+
"step": 5
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"clip_ratio": 0.0,
|
| 83 |
+
"completion_length": 459.7879638671875,
|
| 84 |
+
"epoch": 0.08955223880597014,
|
| 85 |
+
"grad_norm": 0.5247090458869934,
|
| 86 |
+
"learning_rate": 4.285714285714285e-07,
|
| 87 |
+
"loss": 0.0056,
|
| 88 |
+
"num_tokens": 3432696.0,
|
| 89 |
+
"reward": 0.2645089402794838,
|
| 90 |
+
"reward_std": 0.3864951953291893,
|
| 91 |
+
"rewards/accuracy_reward": 0.21316964365541935,
|
| 92 |
+
"rewards/format_reward": 0.0513392862631008,
|
| 93 |
+
"step": 6
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"clip_ratio": 0.0,
|
| 97 |
+
"completion_length": 485.81252670288086,
|
| 98 |
+
"epoch": 0.1044776119402985,
|
| 99 |
+
"grad_norm": 0.7675074338912964,
|
| 100 |
+
"learning_rate": 5e-07,
|
| 101 |
+
"loss": 0.0057,
|
| 102 |
+
"num_tokens": 3997032.0,
|
| 103 |
+
"reward": 0.2901785857975483,
|
| 104 |
+
"reward_std": 0.38608507812023163,
|
| 105 |
+
"rewards/accuracy_reward": 0.203125,
|
| 106 |
+
"rewards/format_reward": 0.08705357182770967,
|
| 107 |
+
"step": 7
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"clip_ratio": 0.0,
|
| 111 |
+
"completion_length": 503.35828399658203,
|
| 112 |
+
"epoch": 0.11940298507462686,
|
| 113 |
+
"grad_norm": 0.5886371731758118,
|
| 114 |
+
"learning_rate": 5.714285714285714e-07,
|
| 115 |
+
"loss": -0.0036,
|
| 116 |
+
"num_tokens": 4565689.0,
|
| 117 |
+
"reward": 0.3314732313156128,
|
| 118 |
+
"reward_std": 0.4302351512014866,
|
| 119 |
+
"rewards/accuracy_reward": 0.21205356996506453,
|
| 120 |
+
"rewards/format_reward": 0.11941964458674192,
|
| 121 |
+
"step": 8
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"clip_ratio": 0.0,
|
| 125 |
+
"completion_length": 454.2031440734863,
|
| 126 |
+
"epoch": 0.13432835820895522,
|
| 127 |
+
"grad_norm": 1.0013484954833984,
|
| 128 |
+
"learning_rate": 6.428571428571429e-07,
|
| 129 |
+
"loss": -0.0081,
|
| 130 |
+
"num_tokens": 5097991.0,
|
| 131 |
+
"reward": 0.3560268022119999,
|
| 132 |
+
"reward_std": 0.4213982783257961,
|
| 133 |
+
"rewards/accuracy_reward": 0.22321428451687098,
|
| 134 |
+
"rewards/format_reward": 0.1328125,
|
| 135 |
+
"step": 9
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"clip_ratio": 0.0,
|
| 139 |
+
"completion_length": 437.2924270629883,
|
| 140 |
+
"epoch": 0.14925373134328357,
|
| 141 |
+
"grad_norm": 1.295032262802124,
|
| 142 |
+
"learning_rate": 7.142857142857143e-07,
|
| 143 |
+
"loss": 0.0078,
|
| 144 |
+
"num_tokens": 5614445.0,
|
| 145 |
+
"reward": 0.4207589514553547,
|
| 146 |
+
"reward_std": 0.49341823533177376,
|
| 147 |
+
"rewards/accuracy_reward": 0.2254464291036129,
|
| 148 |
+
"rewards/format_reward": 0.19531249906867743,
|
| 149 |
+
"step": 10
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"clip_ratio": 0.0,
|
| 153 |
+
"completion_length": 522.3638687133789,
|
| 154 |
+
"epoch": 0.16417910447761194,
|
| 155 |
+
"grad_norm": 4.296813011169434,
|
| 156 |
+
"learning_rate": 7.857142857142856e-07,
|
| 157 |
+
"loss": 0.0091,
|
| 158 |
+
"num_tokens": 6212539.0,
|
| 159 |
+
"reward": 0.4508928842842579,
|
| 160 |
+
"reward_std": 0.48767242580652237,
|
| 161 |
+
"rewards/accuracy_reward": 0.203125,
|
| 162 |
+
"rewards/format_reward": 0.2477678582072258,
|
| 163 |
+
"step": 11
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"clip_ratio": 0.0,
|
| 167 |
+
"completion_length": 484.03127670288086,
|
| 168 |
+
"epoch": 0.1791044776119403,
|
| 169 |
+
"grad_norm": 0.9652361869812012,
|
| 170 |
+
"learning_rate": 8.57142857142857e-07,
|
| 171 |
+
"loss": 0.0202,
|
| 172 |
+
"num_tokens": 6776311.0,
|
| 173 |
+
"reward": 0.525669664144516,
|
| 174 |
+
"reward_std": 0.5405469685792923,
|
| 175 |
+
"rewards/accuracy_reward": 0.2198660708963871,
|
| 176 |
+
"rewards/format_reward": 0.3058035746216774,
|
| 177 |
+
"step": 12
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"clip_ratio": 0.0,
|
| 181 |
+
"completion_length": 491.166316986084,
|
| 182 |
+
"epoch": 0.19402985074626866,
|
| 183 |
+
"grad_norm": 0.43946775794029236,
|
| 184 |
+
"learning_rate": 9.285714285714285e-07,
|
| 185 |
+
"loss": 0.0234,
|
| 186 |
+
"num_tokens": 7339580.0,
|
| 187 |
+
"reward": 0.7756696790456772,
|
| 188 |
+
"reward_std": 0.5723891109228134,
|
| 189 |
+
"rewards/accuracy_reward": 0.2388392873108387,
|
| 190 |
+
"rewards/format_reward": 0.5368303470313549,
|
| 191 |
+
"step": 13
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"clip_ratio": 0.0,
|
| 195 |
+
"completion_length": 453.806941986084,
|
| 196 |
+
"epoch": 0.208955223880597,
|
| 197 |
+
"grad_norm": 1.8365284204483032,
|
| 198 |
+
"learning_rate": 1e-06,
|
| 199 |
+
"loss": 0.0215,
|
| 200 |
+
"num_tokens": 7870063.0,
|
| 201 |
+
"reward": 0.8973214775323868,
|
| 202 |
+
"reward_std": 0.5517464429140091,
|
| 203 |
+
"rewards/accuracy_reward": 0.2645089253783226,
|
| 204 |
+
"rewards/format_reward": 0.6328125074505806,
|
| 205 |
+
"step": 14
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"clip_ratio": 0.0,
|
| 209 |
+
"completion_length": 442.7131881713867,
|
| 210 |
+
"epoch": 0.22388059701492538,
|
| 211 |
+
"grad_norm": 0.4182904362678528,
|
| 212 |
+
"learning_rate": 9.998286624877785e-07,
|
| 213 |
+
"loss": 0.0323,
|
| 214 |
+
"num_tokens": 8381238.0,
|
| 215 |
+
"reward": 0.9241071864962578,
|
| 216 |
+
"reward_std": 0.5212078019976616,
|
| 217 |
+
"rewards/accuracy_reward": 0.2198660708963871,
|
| 218 |
+
"rewards/format_reward": 0.7042410671710968,
|
| 219 |
+
"step": 15
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"clip_ratio": 0.0,
|
| 223 |
+
"completion_length": 489.45203399658203,
|
| 224 |
+
"epoch": 0.23880597014925373,
|
| 225 |
+
"grad_norm": 0.4200661778450012,
|
| 226 |
+
"learning_rate": 9.99314767377287e-07,
|
| 227 |
+
"loss": 0.0269,
|
| 228 |
+
"num_tokens": 8943155.0,
|
| 229 |
+
"reward": 1.010044701397419,
|
| 230 |
+
"reward_std": 0.5311430767178535,
|
| 231 |
+
"rewards/accuracy_reward": 0.2633928544819355,
|
| 232 |
+
"rewards/format_reward": 0.746651791036129,
|
| 233 |
+
"step": 16
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"clip_ratio": 0.0,
|
| 237 |
+
"completion_length": 437.2343864440918,
|
| 238 |
+
"epoch": 0.2537313432835821,
|
| 239 |
+
"grad_norm": 0.35031601786613464,
|
| 240 |
+
"learning_rate": 9.98458666866564e-07,
|
| 241 |
+
"loss": -0.0231,
|
| 242 |
+
"num_tokens": 9484653.0,
|
| 243 |
+
"reward": 1.1908482760190964,
|
| 244 |
+
"reward_std": 0.4788440987467766,
|
| 245 |
+
"rewards/accuracy_reward": 0.30691963993012905,
|
| 246 |
+
"rewards/format_reward": 0.8839285671710968,
|
| 247 |
+
"step": 17
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"clip_ratio": 0.0,
|
| 251 |
+
"completion_length": 466.06028747558594,
|
| 252 |
+
"epoch": 0.26865671641791045,
|
| 253 |
+
"grad_norm": 0.31605255603790283,
|
| 254 |
+
"learning_rate": 9.972609476841365e-07,
|
| 255 |
+
"loss": 0.031,
|
| 256 |
+
"num_tokens": 10038427.0,
|
| 257 |
+
"reward": 1.2421875596046448,
|
| 258 |
+
"reward_std": 0.4797636419534683,
|
| 259 |
+
"rewards/accuracy_reward": 0.34263392724096775,
|
| 260 |
+
"rewards/format_reward": 0.8995535746216774,
|
| 261 |
+
"step": 18
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"clip_ratio": 0.0,
|
| 265 |
+
"completion_length": 492.533504486084,
|
| 266 |
+
"epoch": 0.2835820895522388,
|
| 267 |
+
"grad_norm": 0.3311372399330139,
|
| 268 |
+
"learning_rate": 9.957224306869053e-07,
|
| 269 |
+
"loss": 0.0183,
|
| 270 |
+
"num_tokens": 10604529.0,
|
| 271 |
+
"reward": 1.3191964626312256,
|
| 272 |
+
"reward_std": 0.4729856923222542,
|
| 273 |
+
"rewards/accuracy_reward": 0.4151785634458065,
|
| 274 |
+
"rewards/format_reward": 0.9040178507566452,
|
| 275 |
+
"step": 19
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"clip_ratio": 0.0,
|
| 279 |
+
"completion_length": 479.89734649658203,
|
| 280 |
+
"epoch": 0.29850746268656714,
|
| 281 |
+
"grad_norm": 0.339732825756073,
|
| 282 |
+
"learning_rate": 9.938441702975689e-07,
|
| 283 |
+
"loss": 0.0107,
|
| 284 |
+
"num_tokens": 11161797.0,
|
| 285 |
+
"reward": 1.3560268729925156,
|
| 286 |
+
"reward_std": 0.47978325933218,
|
| 287 |
+
"rewards/accuracy_reward": 0.4374999962747097,
|
| 288 |
+
"rewards/format_reward": 0.918526791036129,
|
| 289 |
+
"step": 20
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"clip_ratio": 0.0,
|
| 293 |
+
"completion_length": 552.010066986084,
|
| 294 |
+
"epoch": 0.31343283582089554,
|
| 295 |
+
"grad_norm": 0.23753176629543304,
|
| 296 |
+
"learning_rate": 9.916274537819773e-07,
|
| 297 |
+
"loss": 0.0283,
|
| 298 |
+
"num_tokens": 11780870.0,
|
| 299 |
+
"reward": 1.4386161267757416,
|
| 300 |
+
"reward_std": 0.46573930606245995,
|
| 301 |
+
"rewards/accuracy_reward": 0.5189732164144516,
|
| 302 |
+
"rewards/format_reward": 0.9196428656578064,
|
| 303 |
+
"step": 21
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"clip_ratio": 0.0,
|
| 307 |
+
"completion_length": 487.6372947692871,
|
| 308 |
+
"epoch": 0.3283582089552239,
|
| 309 |
+
"grad_norm": 0.2946343719959259,
|
| 310 |
+
"learning_rate": 9.890738003669027e-07,
|
| 311 |
+
"loss": 0.0361,
|
| 312 |
+
"num_tokens": 12347257.0,
|
| 313 |
+
"reward": 1.491071492433548,
|
| 314 |
+
"reward_std": 0.42494936659932137,
|
| 315 |
+
"rewards/accuracy_reward": 0.5357142835855484,
|
| 316 |
+
"rewards/format_reward": 0.9553571343421936,
|
| 317 |
+
"step": 22
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"clip_ratio": 0.0,
|
| 321 |
+
"completion_length": 479.8527069091797,
|
| 322 |
+
"epoch": 0.34328358208955223,
|
| 323 |
+
"grad_norm": 0.2775673568248749,
|
| 324 |
+
"learning_rate": 9.861849601988383e-07,
|
| 325 |
+
"loss": 0.013,
|
| 326 |
+
"num_tokens": 12907781.0,
|
| 327 |
+
"reward": 1.545758992433548,
|
| 328 |
+
"reward_std": 0.39185454323887825,
|
| 329 |
+
"rewards/accuracy_reward": 0.6004464291036129,
|
| 330 |
+
"rewards/format_reward": 0.9453125074505806,
|
| 331 |
+
"step": 23
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"clip_ratio": 0.0,
|
| 335 |
+
"completion_length": 570.8906555175781,
|
| 336 |
+
"epoch": 0.3582089552238806,
|
| 337 |
+
"grad_norm": 0.2118677943944931,
|
| 338 |
+
"learning_rate": 9.82962913144534e-07,
|
| 339 |
+
"loss": 0.0149,
|
| 340 |
+
"num_tokens": 13555419.0,
|
| 341 |
+
"reward": 1.547991156578064,
|
| 342 |
+
"reward_std": 0.369963688775897,
|
| 343 |
+
"rewards/accuracy_reward": 0.6071428544819355,
|
| 344 |
+
"rewards/format_reward": 0.940848208963871,
|
| 345 |
+
"step": 24
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"clip_ratio": 0.0,
|
| 349 |
+
"completion_length": 533.5781478881836,
|
| 350 |
+
"epoch": 0.373134328358209,
|
| 351 |
+
"grad_norm": 0.18608489632606506,
|
| 352 |
+
"learning_rate": 9.794098674340966e-07,
|
| 353 |
+
"loss": 0.011,
|
| 354 |
+
"num_tokens": 14148665.0,
|
| 355 |
+
"reward": 1.6908482909202576,
|
| 356 |
+
"reward_std": 0.2971882149577141,
|
| 357 |
+
"rewards/accuracy_reward": 0.7209821492433548,
|
| 358 |
+
"rewards/format_reward": 0.9698660746216774,
|
| 359 |
+
"step": 25
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"clip_ratio": 0.0,
|
| 363 |
+
"completion_length": 579.241096496582,
|
| 364 |
+
"epoch": 0.3880597014925373,
|
| 365 |
+
"grad_norm": 0.19908326864242554,
|
| 366 |
+
"learning_rate": 9.755282581475767e-07,
|
| 367 |
+
"loss": 0.0489,
|
| 368 |
+
"num_tokens": 14805265.0,
|
| 369 |
+
"reward": 1.5781250596046448,
|
| 370 |
+
"reward_std": 0.3244300540536642,
|
| 371 |
+
"rewards/accuracy_reward": 0.6238839328289032,
|
| 372 |
+
"rewards/format_reward": 0.9542410746216774,
|
| 373 |
+
"step": 26
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"clip_ratio": 0.0,
|
| 377 |
+
"completion_length": 601.6830520629883,
|
| 378 |
+
"epoch": 0.40298507462686567,
|
| 379 |
+
"grad_norm": 0.16485629975795746,
|
| 380 |
+
"learning_rate": 9.713207455460892e-07,
|
| 381 |
+
"loss": 0.0507,
|
| 382 |
+
"num_tokens": 15471061.0,
|
| 383 |
+
"reward": 1.6294643580913544,
|
| 384 |
+
"reward_std": 0.259660467505455,
|
| 385 |
+
"rewards/accuracy_reward": 0.6629464402794838,
|
| 386 |
+
"rewards/format_reward": 0.9665178656578064,
|
| 387 |
+
"step": 27
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"clip_ratio": 0.0,
|
| 391 |
+
"completion_length": 574.0223426818848,
|
| 392 |
+
"epoch": 0.417910447761194,
|
| 393 |
+
"grad_norm": 0.1541140377521515,
|
| 394 |
+
"learning_rate": 9.667902132486008e-07,
|
| 395 |
+
"loss": 0.0453,
|
| 396 |
+
"num_tokens": 16104777.0,
|
| 397 |
+
"reward": 1.6495536267757416,
|
| 398 |
+
"reward_std": 0.25614376924932003,
|
| 399 |
+
"rewards/accuracy_reward": 0.684151791036129,
|
| 400 |
+
"rewards/format_reward": 0.9654017835855484,
|
| 401 |
+
"step": 28
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"clip_ratio": 0.0,
|
| 405 |
+
"completion_length": 663.6596221923828,
|
| 406 |
+
"epoch": 0.43283582089552236,
|
| 407 |
+
"grad_norm": 0.14992094039916992,
|
| 408 |
+
"learning_rate": 9.619397662556433e-07,
|
| 409 |
+
"loss": 0.0387,
|
| 410 |
+
"num_tokens": 16838656.0,
|
| 411 |
+
"reward": 1.617187574505806,
|
| 412 |
+
"reward_std": 0.25877264328300953,
|
| 413 |
+
"rewards/accuracy_reward": 0.6529017873108387,
|
| 414 |
+
"rewards/format_reward": 0.9642857164144516,
|
| 415 |
+
"step": 29
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"clip_ratio": 0.0,
|
| 419 |
+
"completion_length": 632.5580749511719,
|
| 420 |
+
"epoch": 0.44776119402985076,
|
| 421 |
+
"grad_norm": 0.1843111515045166,
|
| 422 |
+
"learning_rate": 9.567727288213004e-07,
|
| 423 |
+
"loss": 0.0511,
|
| 424 |
+
"num_tokens": 17528148.0,
|
| 425 |
+
"reward": 1.7154018729925156,
|
| 426 |
+
"reward_std": 0.23728702031075954,
|
| 427 |
+
"rewards/accuracy_reward": 0.7410714328289032,
|
| 428 |
+
"rewards/format_reward": 0.9743303582072258,
|
| 429 |
+
"step": 30
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"clip_ratio": 0.0,
|
| 433 |
+
"completion_length": 628.217658996582,
|
| 434 |
+
"epoch": 0.4626865671641791,
|
| 435 |
+
"grad_norm": 0.11407707631587982,
|
| 436 |
+
"learning_rate": 9.512926421749303e-07,
|
| 437 |
+
"loss": 0.0391,
|
| 438 |
+
"num_tokens": 18214823.0,
|
| 439 |
+
"reward": 1.6651786416769028,
|
| 440 |
+
"reward_std": 0.18615373829379678,
|
| 441 |
+
"rewards/accuracy_reward": 0.6841517835855484,
|
| 442 |
+
"rewards/format_reward": 0.9810267835855484,
|
| 443 |
+
"step": 31
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"clip_ratio": 0.0,
|
| 447 |
+
"completion_length": 595.1272659301758,
|
| 448 |
+
"epoch": 0.47761194029850745,
|
| 449 |
+
"grad_norm": 0.12041105329990387,
|
| 450 |
+
"learning_rate": 9.455032620941839e-07,
|
| 451 |
+
"loss": 0.0291,
|
| 452 |
+
"num_tokens": 18876793.0,
|
| 453 |
+
"reward": 1.7310268729925156,
|
| 454 |
+
"reward_std": 0.1854203026741743,
|
| 455 |
+
"rewards/accuracy_reward": 0.7500000074505806,
|
| 456 |
+
"rewards/format_reward": 0.9810267835855484,
|
| 457 |
+
"step": 32
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"clip_ratio": 0.0,
|
| 461 |
+
"completion_length": 664.4553909301758,
|
| 462 |
+
"epoch": 0.4925373134328358,
|
| 463 |
+
"grad_norm": 0.15564288198947906,
|
| 464 |
+
"learning_rate": 9.394085563309826e-07,
|
| 465 |
+
"loss": 0.0552,
|
| 466 |
+
"num_tokens": 19597657.0,
|
| 467 |
+
"reward": 1.6562500596046448,
|
| 468 |
+
"reward_std": 0.2201189510524273,
|
| 469 |
+
"rewards/accuracy_reward": 0.680803582072258,
|
| 470 |
+
"rewards/format_reward": 0.9754464253783226,
|
| 471 |
+
"step": 33
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"clip_ratio": 0.0,
|
| 475 |
+
"completion_length": 683.2623062133789,
|
| 476 |
+
"epoch": 0.5074626865671642,
|
| 477 |
+
"grad_norm": 0.11131128668785095,
|
| 478 |
+
"learning_rate": 9.330127018922193e-07,
|
| 479 |
+
"loss": 0.044,
|
| 480 |
+
"num_tokens": 20340292.0,
|
| 481 |
+
"reward": 1.672991156578064,
|
| 482 |
+
"reward_std": 0.2345678173005581,
|
| 483 |
+
"rewards/accuracy_reward": 0.6941964291036129,
|
| 484 |
+
"rewards/format_reward": 0.978794626891613,
|
| 485 |
+
"step": 34
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"clip_ratio": 0.0,
|
| 489 |
+
"completion_length": 643.5792694091797,
|
| 490 |
+
"epoch": 0.5223880597014925,
|
| 491 |
+
"grad_norm": 0.10805458575487137,
|
| 492 |
+
"learning_rate": 9.26320082177046e-07,
|
| 493 |
+
"loss": 0.0302,
|
| 494 |
+
"num_tokens": 21056931.0,
|
| 495 |
+
"reward": 1.6930804401636124,
|
| 496 |
+
"reward_std": 0.2090724390000105,
|
| 497 |
+
"rewards/accuracy_reward": 0.7064732126891613,
|
| 498 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 499 |
+
"step": 35
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"clip_ratio": 0.0,
|
| 503 |
+
"completion_length": 671.2388725280762,
|
| 504 |
+
"epoch": 0.5373134328358209,
|
| 505 |
+
"grad_norm": 0.11671151965856552,
|
| 506 |
+
"learning_rate": 9.19335283972712e-07,
|
| 507 |
+
"loss": 0.0517,
|
| 508 |
+
"num_tokens": 21801897.0,
|
| 509 |
+
"reward": 1.6685268580913544,
|
| 510 |
+
"reward_std": 0.23604967631399632,
|
| 511 |
+
"rewards/accuracy_reward": 0.6908482164144516,
|
| 512 |
+
"rewards/format_reward": 0.9776785671710968,
|
| 513 |
+
"step": 36
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"clip_ratio": 0.0,
|
| 517 |
+
"completion_length": 627.1127471923828,
|
| 518 |
+
"epoch": 0.5522388059701493,
|
| 519 |
+
"grad_norm": 0.11113366484642029,
|
| 520 |
+
"learning_rate": 9.120630943110077e-07,
|
| 521 |
+
"loss": 0.0203,
|
| 522 |
+
"num_tokens": 22491214.0,
|
| 523 |
+
"reward": 1.7321429252624512,
|
| 524 |
+
"reward_std": 0.17728407308459282,
|
| 525 |
+
"rewards/accuracy_reward": 0.7477678507566452,
|
| 526 |
+
"rewards/format_reward": 0.984375,
|
| 527 |
+
"step": 37
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"clip_ratio": 0.0,
|
| 531 |
+
"completion_length": 648.8895416259766,
|
| 532 |
+
"epoch": 0.5671641791044776,
|
| 533 |
+
"grad_norm": 0.09676128625869751,
|
| 534 |
+
"learning_rate": 9.045084971874737e-07,
|
| 535 |
+
"loss": 0.0296,
|
| 536 |
+
"num_tokens": 23218403.0,
|
| 537 |
+
"reward": 1.7366072237491608,
|
| 538 |
+
"reward_std": 0.16557986475527287,
|
| 539 |
+
"rewards/accuracy_reward": 0.745535708963871,
|
| 540 |
+
"rewards/format_reward": 0.9910714104771614,
|
| 541 |
+
"step": 38
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"clip_ratio": 0.0,
|
| 545 |
+
"completion_length": 621.256721496582,
|
| 546 |
+
"epoch": 0.582089552238806,
|
| 547 |
+
"grad_norm": 0.10444407165050507,
|
| 548 |
+
"learning_rate": 8.966766701456176e-07,
|
| 549 |
+
"loss": 0.0293,
|
| 550 |
+
"num_tokens": 23901713.0,
|
| 551 |
+
"reward": 1.6908483058214188,
|
| 552 |
+
"reward_std": 0.18806083872914314,
|
| 553 |
+
"rewards/accuracy_reward": 0.7109375,
|
| 554 |
+
"rewards/format_reward": 0.979910708963871,
|
| 555 |
+
"step": 39
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"clip_ratio": 0.0,
|
| 559 |
+
"completion_length": 713.9565048217773,
|
| 560 |
+
"epoch": 0.5970149253731343,
|
| 561 |
+
"grad_norm": 0.09341216832399368,
|
| 562 |
+
"learning_rate": 8.885729807284854e-07,
|
| 563 |
+
"loss": 0.0245,
|
| 564 |
+
"num_tokens": 24665114.0,
|
| 565 |
+
"reward": 1.6953125596046448,
|
| 566 |
+
"reward_std": 0.1814118754118681,
|
| 567 |
+
"rewards/accuracy_reward": 0.7131696492433548,
|
| 568 |
+
"rewards/format_reward": 0.9821428507566452,
|
| 569 |
+
"step": 40
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"clip_ratio": 0.0,
|
| 573 |
+
"completion_length": 651.0513610839844,
|
| 574 |
+
"epoch": 0.6119402985074627,
|
| 575 |
+
"grad_norm": 0.1323300451040268,
|
| 576 |
+
"learning_rate": 8.802029828000155e-07,
|
| 577 |
+
"loss": 0.0425,
|
| 578 |
+
"num_tokens": 25389064.0,
|
| 579 |
+
"reward": 1.6674107909202576,
|
| 580 |
+
"reward_std": 0.2174726091325283,
|
| 581 |
+
"rewards/accuracy_reward": 0.6841517835855484,
|
| 582 |
+
"rewards/format_reward": 0.9832589253783226,
|
| 583 |
+
"step": 41
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"clip_ratio": 0.0,
|
| 587 |
+
"completion_length": 607.2377471923828,
|
| 588 |
+
"epoch": 0.6268656716417911,
|
| 589 |
+
"grad_norm": 0.11196275800466537,
|
| 590 |
+
"learning_rate": 8.71572412738697e-07,
|
| 591 |
+
"loss": 0.0262,
|
| 592 |
+
"num_tokens": 26055629.0,
|
| 593 |
+
"reward": 1.8258929252624512,
|
| 594 |
+
"reward_std": 0.15602111723273993,
|
| 595 |
+
"rewards/accuracy_reward": 0.834821417927742,
|
| 596 |
+
"rewards/format_reward": 0.991071417927742,
|
| 597 |
+
"step": 42
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"clip_ratio": 0.0,
|
| 601 |
+
"completion_length": 627.4364242553711,
|
| 602 |
+
"epoch": 0.6417910447761194,
|
| 603 |
+
"grad_norm": 0.10694889724254608,
|
| 604 |
+
"learning_rate": 8.626871855061437e-07,
|
| 605 |
+
"loss": 0.0058,
|
| 606 |
+
"num_tokens": 26753428.0,
|
| 607 |
+
"reward": 1.7589286416769028,
|
| 608 |
+
"reward_std": 0.15414638444781303,
|
| 609 |
+
"rewards/accuracy_reward": 0.7645089328289032,
|
| 610 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 611 |
+
"step": 43
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"clip_ratio": 0.0,
|
| 615 |
+
"completion_length": 601.2935638427734,
|
| 616 |
+
"epoch": 0.6567164179104478,
|
| 617 |
+
"grad_norm": 0.09472611546516418,
|
| 618 |
+
"learning_rate": 8.535533905932737e-07,
|
| 619 |
+
"loss": 0.0191,
|
| 620 |
+
"num_tokens": 27419099.0,
|
| 621 |
+
"reward": 1.7455357760190964,
|
| 622 |
+
"reward_std": 0.15037838742136955,
|
| 623 |
+
"rewards/accuracy_reward": 0.7544642835855484,
|
| 624 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 625 |
+
"step": 44
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"clip_ratio": 0.0,
|
| 629 |
+
"completion_length": 621.0245742797852,
|
| 630 |
+
"epoch": 0.6716417910447762,
|
| 631 |
+
"grad_norm": 0.10363104939460754,
|
| 632 |
+
"learning_rate": 8.441772878468769e-07,
|
| 633 |
+
"loss": 0.034,
|
| 634 |
+
"num_tokens": 28101745.0,
|
| 635 |
+
"reward": 1.743303656578064,
|
| 636 |
+
"reward_std": 0.1812288984656334,
|
| 637 |
+
"rewards/accuracy_reward": 0.756696417927742,
|
| 638 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 639 |
+
"step": 45
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"clip_ratio": 0.0,
|
| 643 |
+
"completion_length": 713.8482437133789,
|
| 644 |
+
"epoch": 0.6865671641791045,
|
| 645 |
+
"grad_norm": 0.10190044343471527,
|
| 646 |
+
"learning_rate": 8.34565303179429e-07,
|
| 647 |
+
"loss": 0.0563,
|
| 648 |
+
"num_tokens": 28875145.0,
|
| 649 |
+
"reward": 1.6640625894069672,
|
| 650 |
+
"reward_std": 0.2248408328741789,
|
| 651 |
+
"rewards/accuracy_reward": 0.6886160746216774,
|
| 652 |
+
"rewards/format_reward": 0.9754464253783226,
|
| 653 |
+
"step": 46
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"clip_ratio": 0.0,
|
| 657 |
+
"completion_length": 670.1551666259766,
|
| 658 |
+
"epoch": 0.7014925373134329,
|
| 659 |
+
"grad_norm": 0.11430803686380386,
|
| 660 |
+
"learning_rate": 8.247240241650917e-07,
|
| 661 |
+
"loss": 0.0096,
|
| 662 |
+
"num_tokens": 29591660.0,
|
| 663 |
+
"reward": 1.7377232760190964,
|
| 664 |
+
"reward_std": 0.1653135661035776,
|
| 665 |
+
"rewards/accuracy_reward": 0.7455357052385807,
|
| 666 |
+
"rewards/format_reward": 0.9921874850988388,
|
| 667 |
+
"step": 47
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"clip_ratio": 0.0,
|
| 671 |
+
"completion_length": 653.9408874511719,
|
| 672 |
+
"epoch": 0.7164179104477612,
|
| 673 |
+
"grad_norm": 0.12349370867013931,
|
| 674 |
+
"learning_rate": 8.146601955249187e-07,
|
| 675 |
+
"loss": 0.0196,
|
| 676 |
+
"num_tokens": 30304407.0,
|
| 677 |
+
"reward": 1.7377232909202576,
|
| 678 |
+
"reward_std": 0.19286326505243778,
|
| 679 |
+
"rewards/accuracy_reward": 0.7511160746216774,
|
| 680 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 681 |
+
"step": 48
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"clip_ratio": 0.0,
|
| 685 |
+
"completion_length": 667.8393173217773,
|
| 686 |
+
"epoch": 0.7313432835820896,
|
| 687 |
+
"grad_norm": 0.10884106904268265,
|
| 688 |
+
"learning_rate": 8.043807145043603e-07,
|
| 689 |
+
"loss": 0.0286,
|
| 690 |
+
"num_tokens": 31039191.0,
|
| 691 |
+
"reward": 1.7377233058214188,
|
| 692 |
+
"reward_std": 0.1800588108599186,
|
| 693 |
+
"rewards/accuracy_reward": 0.746651791036129,
|
| 694 |
+
"rewards/format_reward": 0.991071417927742,
|
| 695 |
+
"step": 49
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"clip_ratio": 0.0,
|
| 699 |
+
"completion_length": 632.0323944091797,
|
| 700 |
+
"epoch": 0.746268656716418,
|
| 701 |
+
"grad_norm": 0.10950858145952225,
|
| 702 |
+
"learning_rate": 7.938926261462365e-07,
|
| 703 |
+
"loss": 0.0221,
|
| 704 |
+
"num_tokens": 31743644.0,
|
| 705 |
+
"reward": 1.7098215073347092,
|
| 706 |
+
"reward_std": 0.15341119468212128,
|
| 707 |
+
"rewards/accuracy_reward": 0.71875,
|
| 708 |
+
"rewards/format_reward": 0.9910714104771614,
|
| 709 |
+
"step": 50
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"clip_ratio": 0.0,
|
| 713 |
+
"completion_length": 633.6294937133789,
|
| 714 |
+
"epoch": 0.7611940298507462,
|
| 715 |
+
"grad_norm": 0.12094567716121674,
|
| 716 |
+
"learning_rate": 7.832031184624164e-07,
|
| 717 |
+
"loss": 0.0344,
|
| 718 |
+
"num_tokens": 32449160.0,
|
| 719 |
+
"reward": 1.7287947237491608,
|
| 720 |
+
"reward_std": 0.184023879468441,
|
| 721 |
+
"rewards/accuracy_reward": 0.7366071343421936,
|
| 722 |
+
"rewards/format_reward": 0.9921874850988388,
|
| 723 |
+
"step": 51
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"clip_ratio": 0.0,
|
| 727 |
+
"completion_length": 659.0848617553711,
|
| 728 |
+
"epoch": 0.7761194029850746,
|
| 729 |
+
"grad_norm": 0.10622338950634003,
|
| 730 |
+
"learning_rate": 7.723195175075135e-07,
|
| 731 |
+
"loss": 0.0386,
|
| 732 |
+
"num_tokens": 33168348.0,
|
| 733 |
+
"reward": 1.6964286416769028,
|
| 734 |
+
"reward_std": 0.17664530966430902,
|
| 735 |
+
"rewards/accuracy_reward": 0.717633917927742,
|
| 736 |
+
"rewards/format_reward": 0.9787946417927742,
|
| 737 |
+
"step": 52
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"clip_ratio": 0.0,
|
| 741 |
+
"completion_length": 671.2433319091797,
|
| 742 |
+
"epoch": 0.7910447761194029,
|
| 743 |
+
"grad_norm": 0.10682238638401031,
|
| 744 |
+
"learning_rate": 7.612492823579744e-07,
|
| 745 |
+
"loss": 0.0376,
|
| 746 |
+
"num_tokens": 33897454.0,
|
| 747 |
+
"reward": 1.671875074505806,
|
| 748 |
+
"reward_std": 0.1950401533395052,
|
| 749 |
+
"rewards/accuracy_reward": 0.6886160746216774,
|
| 750 |
+
"rewards/format_reward": 0.983258917927742,
|
| 751 |
+
"step": 53
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"clip_ratio": 0.0,
|
| 755 |
+
"completion_length": 606.1585159301758,
|
| 756 |
+
"epoch": 0.8059701492537313,
|
| 757 |
+
"grad_norm": 0.10302180051803589,
|
| 758 |
+
"learning_rate": 7.5e-07,
|
| 759 |
+
"loss": 0.0104,
|
| 760 |
+
"num_tokens": 34575404.0,
|
| 761 |
+
"reward": 1.7410715073347092,
|
| 762 |
+
"reward_std": 0.16222931072115898,
|
| 763 |
+
"rewards/accuracy_reward": 0.745535708963871,
|
| 764 |
+
"rewards/format_reward": 0.9955357015132904,
|
| 765 |
+
"step": 54
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"clip_ratio": 0.0,
|
| 769 |
+
"completion_length": 625.7221260070801,
|
| 770 |
+
"epoch": 0.8208955223880597,
|
| 771 |
+
"grad_norm": 0.10820939391851425,
|
| 772 |
+
"learning_rate": 7.385793801298042e-07,
|
| 773 |
+
"loss": 0.0312,
|
| 774 |
+
"num_tokens": 35275171.0,
|
| 775 |
+
"reward": 1.7243304252624512,
|
| 776 |
+
"reward_std": 0.20950445905327797,
|
| 777 |
+
"rewards/accuracy_reward": 0.7366071417927742,
|
| 778 |
+
"rewards/format_reward": 0.9877232015132904,
|
| 779 |
+
"step": 55
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"clip_ratio": 0.0,
|
| 783 |
+
"completion_length": 633.1205596923828,
|
| 784 |
+
"epoch": 0.835820895522388,
|
| 785 |
+
"grad_norm": 0.13289578258991241,
|
| 786 |
+
"learning_rate": 7.269952498697734e-07,
|
| 787 |
+
"loss": 0.0142,
|
| 788 |
+
"num_tokens": 35973503.0,
|
| 789 |
+
"reward": 1.6941965073347092,
|
| 790 |
+
"reward_std": 0.176253211684525,
|
| 791 |
+
"rewards/accuracy_reward": 0.703125,
|
| 792 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 793 |
+
"step": 56
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"clip_ratio": 0.0,
|
| 797 |
+
"completion_length": 625.6227951049805,
|
| 798 |
+
"epoch": 0.8507462686567164,
|
| 799 |
+
"grad_norm": 0.13948650658130646,
|
| 800 |
+
"learning_rate": 7.152555484041475e-07,
|
| 801 |
+
"loss": 0.0413,
|
| 802 |
+
"num_tokens": 36657061.0,
|
| 803 |
+
"reward": 1.7879465222358704,
|
| 804 |
+
"reward_std": 0.15797736030071974,
|
| 805 |
+
"rewards/accuracy_reward": 0.8035714253783226,
|
| 806 |
+
"rewards/format_reward": 0.9843749925494194,
|
| 807 |
+
"step": 57
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"clip_ratio": 0.0,
|
| 811 |
+
"completion_length": 650.3381958007812,
|
| 812 |
+
"epoch": 0.8656716417910447,
|
| 813 |
+
"grad_norm": 0.2045215368270874,
|
| 814 |
+
"learning_rate": 7.033683215379002e-07,
|
| 815 |
+
"loss": 0.0374,
|
| 816 |
+
"num_tokens": 37365732.0,
|
| 817 |
+
"reward": 1.695312574505806,
|
| 818 |
+
"reward_std": 0.17142421007156372,
|
| 819 |
+
"rewards/accuracy_reward": 0.7120535671710968,
|
| 820 |
+
"rewards/format_reward": 0.983258917927742,
|
| 821 |
+
"step": 58
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"clip_ratio": 0.0,
|
| 825 |
+
"completion_length": 641.5770378112793,
|
| 826 |
+
"epoch": 0.8805970149253731,
|
| 827 |
+
"grad_norm": 0.13572581112384796,
|
| 828 |
+
"learning_rate": 6.913417161825449e-07,
|
| 829 |
+
"loss": 0.0067,
|
| 830 |
+
"num_tokens": 38065993.0,
|
| 831 |
+
"reward": 1.7890625894069672,
|
| 832 |
+
"reward_std": 0.16594232060015202,
|
| 833 |
+
"rewards/accuracy_reward": 0.7946428656578064,
|
| 834 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 835 |
+
"step": 59
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"clip_ratio": 0.0,
|
| 839 |
+
"completion_length": 598.4498100280762,
|
| 840 |
+
"epoch": 0.8955223880597015,
|
| 841 |
+
"grad_norm": 0.118675097823143,
|
| 842 |
+
"learning_rate": 6.7918397477265e-07,
|
| 843 |
+
"loss": 0.0253,
|
| 844 |
+
"num_tokens": 38737236.0,
|
| 845 |
+
"reward": 1.7466518580913544,
|
| 846 |
+
"reward_std": 0.17820186354219913,
|
| 847 |
+
"rewards/accuracy_reward": 0.7522321492433548,
|
| 848 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 849 |
+
"step": 60
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"clip_ratio": 0.0,
|
| 853 |
+
"completion_length": 641.1663208007812,
|
| 854 |
+
"epoch": 0.9104477611940298,
|
| 855 |
+
"grad_norm": 0.08891814202070236,
|
| 856 |
+
"learning_rate": 6.669034296168854e-07,
|
| 857 |
+
"loss": 0.0319,
|
| 858 |
+
"num_tokens": 39453897.0,
|
| 859 |
+
"reward": 1.709821492433548,
|
| 860 |
+
"reward_std": 0.14528072997927666,
|
| 861 |
+
"rewards/accuracy_reward": 0.71875,
|
| 862 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 863 |
+
"step": 61
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"clip_ratio": 0.0,
|
| 867 |
+
"completion_length": 612.7120742797852,
|
| 868 |
+
"epoch": 0.9253731343283582,
|
| 869 |
+
"grad_norm": 0.18538488447666168,
|
| 870 |
+
"learning_rate": 6.545084971874736e-07,
|
| 871 |
+
"loss": 0.0169,
|
| 872 |
+
"num_tokens": 40128303.0,
|
| 873 |
+
"reward": 1.7712054252624512,
|
| 874 |
+
"reward_std": 0.16406728280708194,
|
| 875 |
+
"rewards/accuracy_reward": 0.7812499925494194,
|
| 876 |
+
"rewards/format_reward": 0.9899553433060646,
|
| 877 |
+
"step": 62
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"clip_ratio": 0.0,
|
| 881 |
+
"completion_length": 678.2556076049805,
|
| 882 |
+
"epoch": 0.9402985074626866,
|
| 883 |
+
"grad_norm": 0.1315467208623886,
|
| 884 |
+
"learning_rate": 6.420076723519614e-07,
|
| 885 |
+
"loss": 0.0448,
|
| 886 |
+
"num_tokens": 40865532.0,
|
| 887 |
+
"reward": 1.7254465073347092,
|
| 888 |
+
"reward_std": 0.2296902798116207,
|
| 889 |
+
"rewards/accuracy_reward": 0.7544642984867096,
|
| 890 |
+
"rewards/format_reward": 0.9709821417927742,
|
| 891 |
+
"step": 63
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"clip_ratio": 0.0,
|
| 895 |
+
"completion_length": 583.9475708007812,
|
| 896 |
+
"epoch": 0.9552238805970149,
|
| 897 |
+
"grad_norm": 0.1169319748878479,
|
| 898 |
+
"learning_rate": 6.294095225512604e-07,
|
| 899 |
+
"loss": 0.0223,
|
| 900 |
+
"num_tokens": 41525429.0,
|
| 901 |
+
"reward": 1.7522322088479996,
|
| 902 |
+
"reward_std": 0.18113385140895844,
|
| 903 |
+
"rewards/accuracy_reward": 0.7622767761349678,
|
| 904 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 905 |
+
"step": 64
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"clip_ratio": 0.0,
|
| 909 |
+
"completion_length": 596.311408996582,
|
| 910 |
+
"epoch": 0.9701492537313433,
|
| 911 |
+
"grad_norm": 0.12513795495033264,
|
| 912 |
+
"learning_rate": 6.167226819279527e-07,
|
| 913 |
+
"loss": 0.0178,
|
| 914 |
+
"num_tokens": 42192916.0,
|
| 915 |
+
"reward": 1.7477679252624512,
|
| 916 |
+
"reward_std": 0.14999555610120296,
|
| 917 |
+
"rewards/accuracy_reward": 0.753348208963871,
|
| 918 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 919 |
+
"step": 65
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"clip_ratio": 0.0,
|
| 923 |
+
"completion_length": 649.2395858764648,
|
| 924 |
+
"epoch": 0.9850746268656716,
|
| 925 |
+
"grad_norm": 0.6537142992019653,
|
| 926 |
+
"learning_rate": 6.039558454088795e-07,
|
| 927 |
+
"loss": 0.0283,
|
| 928 |
+
"num_tokens": 42898514.0,
|
| 929 |
+
"reward": 1.733258992433548,
|
| 930 |
+
"reward_std": 0.15077468007802963,
|
| 931 |
+
"rewards/accuracy_reward": 0.7455357164144516,
|
| 932 |
+
"rewards/format_reward": 0.9877232164144516,
|
| 933 |
+
"step": 66
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"clip_ratio": 0.0,
|
| 937 |
+
"completion_length": 617.8482437133789,
|
| 938 |
+
"epoch": 1.0149253731343284,
|
| 939 |
+
"grad_norm": 0.09364171326160431,
|
| 940 |
+
"learning_rate": 5.911177627460738e-07,
|
| 941 |
+
"loss": 0.0312,
|
| 942 |
+
"num_tokens": 43570490.0,
|
| 943 |
+
"reward": 1.7712054550647736,
|
| 944 |
+
"reward_std": 0.16040645446628332,
|
| 945 |
+
"rewards/accuracy_reward": 0.7924107164144516,
|
| 946 |
+
"rewards/format_reward": 0.9787946343421936,
|
| 947 |
+
"step": 67
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"clip_ratio": 0.0,
|
| 951 |
+
"completion_length": 664.5725784301758,
|
| 952 |
+
"epoch": 1.0298507462686568,
|
| 953 |
+
"grad_norm": 0.09059074521064758,
|
| 954 |
+
"learning_rate": 5.782172325201155e-07,
|
| 955 |
+
"loss": 0.04,
|
| 956 |
+
"num_tokens": 44298899.0,
|
| 957 |
+
"reward": 1.7410715073347092,
|
| 958 |
+
"reward_std": 0.1766328364610672,
|
| 959 |
+
"rewards/accuracy_reward": 0.7555803656578064,
|
| 960 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 961 |
+
"step": 68
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"clip_ratio": 0.0,
|
| 965 |
+
"completion_length": 597.5100708007812,
|
| 966 |
+
"epoch": 1.044776119402985,
|
| 967 |
+
"grad_norm": 0.10916374623775482,
|
| 968 |
+
"learning_rate": 5.652630961100258e-07,
|
| 969 |
+
"loss": 0.0065,
|
| 970 |
+
"num_tokens": 44959604.0,
|
| 971 |
+
"reward": 1.7511161714792252,
|
| 972 |
+
"reward_std": 0.14327580528333783,
|
| 973 |
+
"rewards/accuracy_reward": 0.754464291036129,
|
| 974 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 975 |
+
"step": 69
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"clip_ratio": 0.0,
|
| 979 |
+
"completion_length": 612.0524826049805,
|
| 980 |
+
"epoch": 1.0597014925373134,
|
| 981 |
+
"grad_norm": 0.11315930634737015,
|
| 982 |
+
"learning_rate": 5.522642316338268e-07,
|
| 983 |
+
"loss": 0.0129,
|
| 984 |
+
"num_tokens": 45635299.0,
|
| 985 |
+
"reward": 1.7700893580913544,
|
| 986 |
+
"reward_std": 0.17250201851129532,
|
| 987 |
+
"rewards/accuracy_reward": 0.776785708963871,
|
| 988 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 989 |
+
"step": 70
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"clip_ratio": 0.0,
|
| 993 |
+
"completion_length": 597.2611808776855,
|
| 994 |
+
"epoch": 1.0746268656716418,
|
| 995 |
+
"grad_norm": 0.12196239084005356,
|
| 996 |
+
"learning_rate": 5.392295478639225e-07,
|
| 997 |
+
"loss": 0.0234,
|
| 998 |
+
"num_tokens": 46305437.0,
|
| 999 |
+
"reward": 1.7700893729925156,
|
| 1000 |
+
"reward_std": 0.15816278103739023,
|
| 1001 |
+
"rewards/accuracy_reward": 0.7756696417927742,
|
| 1002 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1003 |
+
"step": 71
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"clip_ratio": 0.0,
|
| 1007 |
+
"completion_length": 642.2678756713867,
|
| 1008 |
+
"epoch": 1.0895522388059702,
|
| 1009 |
+
"grad_norm": 0.11447075009346008,
|
| 1010 |
+
"learning_rate": 5.26167978121472e-07,
|
| 1011 |
+
"loss": 0.0379,
|
| 1012 |
+
"num_tokens": 47003701.0,
|
| 1013 |
+
"reward": 1.75334832072258,
|
| 1014 |
+
"reward_std": 0.17948786355555058,
|
| 1015 |
+
"rewards/accuracy_reward": 0.7667410746216774,
|
| 1016 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 1017 |
+
"step": 72
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"clip_ratio": 0.0,
|
| 1021 |
+
"completion_length": 640.4107437133789,
|
| 1022 |
+
"epoch": 1.1044776119402986,
|
| 1023 |
+
"grad_norm": 0.11286026239395142,
|
| 1024 |
+
"learning_rate": 5.130884741539366e-07,
|
| 1025 |
+
"loss": 0.029,
|
| 1026 |
+
"num_tokens": 47710189.0,
|
| 1027 |
+
"reward": 1.6953125894069672,
|
| 1028 |
+
"reward_std": 0.20528794033452868,
|
| 1029 |
+
"rewards/accuracy_reward": 0.7042410671710968,
|
| 1030 |
+
"rewards/format_reward": 0.991071417927742,
|
| 1031 |
+
"step": 73
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"clip_ratio": 0.0,
|
| 1035 |
+
"completion_length": 616.1919898986816,
|
| 1036 |
+
"epoch": 1.1194029850746268,
|
| 1037 |
+
"grad_norm": 0.11942121386528015,
|
| 1038 |
+
"learning_rate": 5e-07,
|
| 1039 |
+
"loss": 0.019,
|
| 1040 |
+
"num_tokens": 48375529.0,
|
| 1041 |
+
"reward": 1.7991072237491608,
|
| 1042 |
+
"reward_std": 0.17037044698372483,
|
| 1043 |
+
"rewards/accuracy_reward": 0.8058035746216774,
|
| 1044 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1045 |
+
"step": 74
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"clip_ratio": 0.0,
|
| 1049 |
+
"completion_length": 626.8359680175781,
|
| 1050 |
+
"epoch": 1.1343283582089552,
|
| 1051 |
+
"grad_norm": 0.12714813649654388,
|
| 1052 |
+
"learning_rate": 4.869115258460634e-07,
|
| 1053 |
+
"loss": 0.0344,
|
| 1054 |
+
"num_tokens": 49061558.0,
|
| 1055 |
+
"reward": 1.7343750894069672,
|
| 1056 |
+
"reward_std": 0.17231870535761118,
|
| 1057 |
+
"rewards/accuracy_reward": 0.7488839253783226,
|
| 1058 |
+
"rewards/format_reward": 0.9854910597205162,
|
| 1059 |
+
"step": 75
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"clip_ratio": 0.0,
|
| 1063 |
+
"completion_length": 593.3370780944824,
|
| 1064 |
+
"epoch": 1.1492537313432836,
|
| 1065 |
+
"grad_norm": 0.09834135323762894,
|
| 1066 |
+
"learning_rate": 4.7383202187852804e-07,
|
| 1067 |
+
"loss": 0.0158,
|
| 1068 |
+
"num_tokens": 49720004.0,
|
| 1069 |
+
"reward": 1.8203125894069672,
|
| 1070 |
+
"reward_std": 0.12710504699498415,
|
| 1071 |
+
"rewards/accuracy_reward": 0.8325892984867096,
|
| 1072 |
+
"rewards/format_reward": 0.987723208963871,
|
| 1073 |
+
"step": 76
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"clip_ratio": 0.0,
|
| 1077 |
+
"completion_length": 532.5000267028809,
|
| 1078 |
+
"epoch": 1.164179104477612,
|
| 1079 |
+
"grad_norm": 0.10094507038593292,
|
| 1080 |
+
"learning_rate": 4.6077045213607755e-07,
|
| 1081 |
+
"loss": 0.0103,
|
| 1082 |
+
"num_tokens": 50322556.0,
|
| 1083 |
+
"reward": 1.845982238650322,
|
| 1084 |
+
"reward_std": 0.10092606954276562,
|
| 1085 |
+
"rewards/accuracy_reward": 0.8459821417927742,
|
| 1086 |
+
"rewards/format_reward": 1.0,
|
| 1087 |
+
"step": 77
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"clip_ratio": 0.0,
|
| 1091 |
+
"completion_length": 619.5044860839844,
|
| 1092 |
+
"epoch": 1.1791044776119404,
|
| 1093 |
+
"grad_norm": 0.09271937608718872,
|
| 1094 |
+
"learning_rate": 4.477357683661733e-07,
|
| 1095 |
+
"loss": 0.0121,
|
| 1096 |
+
"num_tokens": 51004440.0,
|
| 1097 |
+
"reward": 1.712053656578064,
|
| 1098 |
+
"reward_std": 0.13478131592273712,
|
| 1099 |
+
"rewards/accuracy_reward": 0.7243303582072258,
|
| 1100 |
+
"rewards/format_reward": 0.987723208963871,
|
| 1101 |
+
"step": 78
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"clip_ratio": 0.0,
|
| 1105 |
+
"completion_length": 661.3080673217773,
|
| 1106 |
+
"epoch": 1.1940298507462686,
|
| 1107 |
+
"grad_norm": 0.1302761286497116,
|
| 1108 |
+
"learning_rate": 4.347369038899743e-07,
|
| 1109 |
+
"loss": 0.0196,
|
| 1110 |
+
"num_tokens": 51728116.0,
|
| 1111 |
+
"reward": 1.7321429401636124,
|
| 1112 |
+
"reward_std": 0.16369346249848604,
|
| 1113 |
+
"rewards/accuracy_reward": 0.7399553507566452,
|
| 1114 |
+
"rewards/format_reward": 0.9921875,
|
| 1115 |
+
"step": 79
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"clip_ratio": 0.0,
|
| 1119 |
+
"completion_length": 673.4252510070801,
|
| 1120 |
+
"epoch": 1.208955223880597,
|
| 1121 |
+
"grad_norm": 0.10427302122116089,
|
| 1122 |
+
"learning_rate": 4.2178276747988444e-07,
|
| 1123 |
+
"loss": 0.0256,
|
| 1124 |
+
"num_tokens": 52457921.0,
|
| 1125 |
+
"reward": 1.6930804401636124,
|
| 1126 |
+
"reward_std": 0.13809803500771523,
|
| 1127 |
+
"rewards/accuracy_reward": 0.7008928582072258,
|
| 1128 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1129 |
+
"step": 80
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"clip_ratio": 0.0,
|
| 1133 |
+
"completion_length": 632.3080596923828,
|
| 1134 |
+
"epoch": 1.2238805970149254,
|
| 1135 |
+
"grad_norm": 0.09508039057254791,
|
| 1136 |
+
"learning_rate": 4.0888223725392624e-07,
|
| 1137 |
+
"loss": 0.0227,
|
| 1138 |
+
"num_tokens": 53149381.0,
|
| 1139 |
+
"reward": 1.742187574505806,
|
| 1140 |
+
"reward_std": 0.147241884842515,
|
| 1141 |
+
"rewards/accuracy_reward": 0.75,
|
| 1142 |
+
"rewards/format_reward": 0.9921874850988388,
|
| 1143 |
+
"step": 81
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"clip_ratio": 0.0,
|
| 1147 |
+
"completion_length": 602.2265930175781,
|
| 1148 |
+
"epoch": 1.2388059701492538,
|
| 1149 |
+
"grad_norm": 0.09547509253025055,
|
| 1150 |
+
"learning_rate": 3.960441545911204e-07,
|
| 1151 |
+
"loss": 0.0134,
|
| 1152 |
+
"num_tokens": 53826080.0,
|
| 1153 |
+
"reward": 1.7991072088479996,
|
| 1154 |
+
"reward_std": 0.1459453795105219,
|
| 1155 |
+
"rewards/accuracy_reward": 0.8002232015132904,
|
| 1156 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1157 |
+
"step": 82
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"clip_ratio": 0.0,
|
| 1161 |
+
"completion_length": 634.0826187133789,
|
| 1162 |
+
"epoch": 1.2537313432835822,
|
| 1163 |
+
"grad_norm": 0.0878840833902359,
|
| 1164 |
+
"learning_rate": 3.8327731807204744e-07,
|
| 1165 |
+
"loss": 0.0183,
|
| 1166 |
+
"num_tokens": 54522986.0,
|
| 1167 |
+
"reward": 1.7377232760190964,
|
| 1168 |
+
"reward_std": 0.17751624016091228,
|
| 1169 |
+
"rewards/accuracy_reward": 0.745535708963871,
|
| 1170 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1171 |
+
"step": 83
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"clip_ratio": 0.0,
|
| 1175 |
+
"completion_length": 627.9698944091797,
|
| 1176 |
+
"epoch": 1.2686567164179103,
|
| 1177 |
+
"grad_norm": 0.1188497319817543,
|
| 1178 |
+
"learning_rate": 3.7059047744873955e-07,
|
| 1179 |
+
"loss": 0.0205,
|
| 1180 |
+
"num_tokens": 55224271.0,
|
| 1181 |
+
"reward": 1.7555804401636124,
|
| 1182 |
+
"reward_std": 0.16282380744814873,
|
| 1183 |
+
"rewards/accuracy_reward": 0.761160708963871,
|
| 1184 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1185 |
+
"step": 84
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"clip_ratio": 0.0,
|
| 1189 |
+
"completion_length": 610.5402069091797,
|
| 1190 |
+
"epoch": 1.2835820895522387,
|
| 1191 |
+
"grad_norm": 0.0854121670126915,
|
| 1192 |
+
"learning_rate": 3.5799232764803867e-07,
|
| 1193 |
+
"loss": 0.0124,
|
| 1194 |
+
"num_tokens": 55889819.0,
|
| 1195 |
+
"reward": 1.8024554699659348,
|
| 1196 |
+
"reward_std": 0.12497595604509115,
|
| 1197 |
+
"rewards/accuracy_reward": 0.8035714328289032,
|
| 1198 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1199 |
+
"step": 85
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"clip_ratio": 0.0,
|
| 1203 |
+
"completion_length": 610.6161003112793,
|
| 1204 |
+
"epoch": 1.2985074626865671,
|
| 1205 |
+
"grad_norm": 0.1087036058306694,
|
| 1206 |
+
"learning_rate": 3.454915028125263e-07,
|
| 1207 |
+
"loss": 0.0278,
|
| 1208 |
+
"num_tokens": 56559627.0,
|
| 1209 |
+
"reward": 1.710937574505806,
|
| 1210 |
+
"reward_std": 0.15050204377621412,
|
| 1211 |
+
"rewards/accuracy_reward": 0.71875,
|
| 1212 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 1213 |
+
"step": 86
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"clip_ratio": 0.0,
|
| 1217 |
+
"completion_length": 659.1116409301758,
|
| 1218 |
+
"epoch": 1.3134328358208955,
|
| 1219 |
+
"grad_norm": 0.09039624035358429,
|
| 1220 |
+
"learning_rate": 3.330965703831146e-07,
|
| 1221 |
+
"loss": 0.0385,
|
| 1222 |
+
"num_tokens": 57282351.0,
|
| 1223 |
+
"reward": 1.7176340222358704,
|
| 1224 |
+
"reward_std": 0.16426212899386883,
|
| 1225 |
+
"rewards/accuracy_reward": 0.7321428582072258,
|
| 1226 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 1227 |
+
"step": 87
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"clip_ratio": 0.0,
|
| 1231 |
+
"completion_length": 578.8861999511719,
|
| 1232 |
+
"epoch": 1.328358208955224,
|
| 1233 |
+
"grad_norm": 0.09018764644861221,
|
| 1234 |
+
"learning_rate": 3.2081602522734985e-07,
|
| 1235 |
+
"loss": 0.0131,
|
| 1236 |
+
"num_tokens": 57939417.0,
|
| 1237 |
+
"reward": 1.85491082072258,
|
| 1238 |
+
"reward_std": 0.11039695050567389,
|
| 1239 |
+
"rewards/accuracy_reward": 0.8560267835855484,
|
| 1240 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1241 |
+
"step": 88
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"clip_ratio": 0.0,
|
| 1245 |
+
"completion_length": 578.7924346923828,
|
| 1246 |
+
"epoch": 1.3432835820895521,
|
| 1247 |
+
"grad_norm": 0.12793904542922974,
|
| 1248 |
+
"learning_rate": 3.086582838174551e-07,
|
| 1249 |
+
"loss": 0.0431,
|
| 1250 |
+
"num_tokens": 58583903.0,
|
| 1251 |
+
"reward": 1.77678582072258,
|
| 1252 |
+
"reward_std": 0.1774543970823288,
|
| 1253 |
+
"rewards/accuracy_reward": 0.7834821417927742,
|
| 1254 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 1255 |
+
"step": 89
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"clip_ratio": 0.0,
|
| 1259 |
+
"completion_length": 635.9888610839844,
|
| 1260 |
+
"epoch": 1.3582089552238805,
|
| 1261 |
+
"grad_norm": 0.10590506345033646,
|
| 1262 |
+
"learning_rate": 2.9663167846209996e-07,
|
| 1263 |
+
"loss": 0.0308,
|
| 1264 |
+
"num_tokens": 59286877.0,
|
| 1265 |
+
"reward": 1.705357238650322,
|
| 1266 |
+
"reward_std": 0.16925656888633966,
|
| 1267 |
+
"rewards/accuracy_reward": 0.71875,
|
| 1268 |
+
"rewards/format_reward": 0.9866071417927742,
|
| 1269 |
+
"step": 90
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"clip_ratio": 0.0,
|
| 1273 |
+
"completion_length": 559.2232398986816,
|
| 1274 |
+
"epoch": 1.373134328358209,
|
| 1275 |
+
"grad_norm": 0.11253120750188828,
|
| 1276 |
+
"learning_rate": 2.847444515958523e-07,
|
| 1277 |
+
"loss": 0.0138,
|
| 1278 |
+
"num_tokens": 59920069.0,
|
| 1279 |
+
"reward": 1.8337054401636124,
|
| 1280 |
+
"reward_std": 0.1393702062778175,
|
| 1281 |
+
"rewards/accuracy_reward": 0.8370535671710968,
|
| 1282 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 1283 |
+
"step": 91
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"clip_ratio": 0.0,
|
| 1287 |
+
"completion_length": 627.5044937133789,
|
| 1288 |
+
"epoch": 1.3880597014925373,
|
| 1289 |
+
"grad_norm": 0.10160024464130402,
|
| 1290 |
+
"learning_rate": 2.730047501302266e-07,
|
| 1291 |
+
"loss": 0.0299,
|
| 1292 |
+
"num_tokens": 60608377.0,
|
| 1293 |
+
"reward": 1.7879465222358704,
|
| 1294 |
+
"reward_std": 0.1563611626625061,
|
| 1295 |
+
"rewards/accuracy_reward": 0.8024553507566452,
|
| 1296 |
+
"rewards/format_reward": 0.9854910671710968,
|
| 1297 |
+
"step": 92
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"clip_ratio": 0.0,
|
| 1301 |
+
"completion_length": 625.7399826049805,
|
| 1302 |
+
"epoch": 1.4029850746268657,
|
| 1303 |
+
"grad_norm": 0.09554547071456909,
|
| 1304 |
+
"learning_rate": 2.6142061987019574e-07,
|
| 1305 |
+
"loss": 0.0153,
|
| 1306 |
+
"num_tokens": 61288944.0,
|
| 1307 |
+
"reward": 1.783482238650322,
|
| 1308 |
+
"reward_std": 0.1566515825688839,
|
| 1309 |
+
"rewards/accuracy_reward": 0.7890625,
|
| 1310 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1311 |
+
"step": 93
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"clip_ratio": 0.0,
|
| 1315 |
+
"completion_length": 581.5256996154785,
|
| 1316 |
+
"epoch": 1.417910447761194,
|
| 1317 |
+
"grad_norm": 0.08792009949684143,
|
| 1318 |
+
"learning_rate": 2.500000000000001e-07,
|
| 1319 |
+
"loss": 0.0273,
|
| 1320 |
+
"num_tokens": 61952575.0,
|
| 1321 |
+
"reward": 1.7310268878936768,
|
| 1322 |
+
"reward_std": 0.14925330318510532,
|
| 1323 |
+
"rewards/accuracy_reward": 0.7377232164144516,
|
| 1324 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1325 |
+
"step": 94
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"clip_ratio": 0.0,
|
| 1329 |
+
"completion_length": 628.5502471923828,
|
| 1330 |
+
"epoch": 1.4328358208955223,
|
| 1331 |
+
"grad_norm": 0.12849754095077515,
|
| 1332 |
+
"learning_rate": 2.387507176420256e-07,
|
| 1333 |
+
"loss": 0.0301,
|
| 1334 |
+
"num_tokens": 62647396.0,
|
| 1335 |
+
"reward": 1.7087054401636124,
|
| 1336 |
+
"reward_std": 0.1948286723345518,
|
| 1337 |
+
"rewards/accuracy_reward": 0.7310267835855484,
|
| 1338 |
+
"rewards/format_reward": 0.9776785597205162,
|
| 1339 |
+
"step": 95
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"clip_ratio": 0.0,
|
| 1343 |
+
"completion_length": 645.8973579406738,
|
| 1344 |
+
"epoch": 1.4477611940298507,
|
| 1345 |
+
"grad_norm": 0.30540189146995544,
|
| 1346 |
+
"learning_rate": 2.2768048249248644e-07,
|
| 1347 |
+
"loss": 0.0235,
|
| 1348 |
+
"num_tokens": 63355344.0,
|
| 1349 |
+
"reward": 1.7477679252624512,
|
| 1350 |
+
"reward_std": 0.16450455971062183,
|
| 1351 |
+
"rewards/accuracy_reward": 0.7533482164144516,
|
| 1352 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 1353 |
+
"step": 96
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"clip_ratio": 0.0,
|
| 1357 |
+
"completion_length": 581.8024864196777,
|
| 1358 |
+
"epoch": 1.462686567164179,
|
| 1359 |
+
"grad_norm": 0.09905046969652176,
|
| 1360 |
+
"learning_rate": 2.167968815375837e-07,
|
| 1361 |
+
"loss": 0.0151,
|
| 1362 |
+
"num_tokens": 64012431.0,
|
| 1363 |
+
"reward": 1.7756697088479996,
|
| 1364 |
+
"reward_std": 0.13652249239385128,
|
| 1365 |
+
"rewards/accuracy_reward": 0.78125,
|
| 1366 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1367 |
+
"step": 97
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"clip_ratio": 0.0,
|
| 1371 |
+
"completion_length": 658.9587326049805,
|
| 1372 |
+
"epoch": 1.4776119402985075,
|
| 1373 |
+
"grad_norm": 0.10181548446416855,
|
| 1374 |
+
"learning_rate": 2.0610737385376348e-07,
|
| 1375 |
+
"loss": 0.0381,
|
| 1376 |
+
"num_tokens": 64744754.0,
|
| 1377 |
+
"reward": 1.699776828289032,
|
| 1378 |
+
"reward_std": 0.19275180995464325,
|
| 1379 |
+
"rewards/accuracy_reward": 0.7098214253783226,
|
| 1380 |
+
"rewards/format_reward": 0.9899553507566452,
|
| 1381 |
+
"step": 98
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"clip_ratio": 0.0,
|
| 1385 |
+
"completion_length": 672.4922180175781,
|
| 1386 |
+
"epoch": 1.4925373134328357,
|
| 1387 |
+
"grad_norm": 0.10780028253793716,
|
| 1388 |
+
"learning_rate": 1.9561928549563966e-07,
|
| 1389 |
+
"loss": 0.0107,
|
| 1390 |
+
"num_tokens": 65485395.0,
|
| 1391 |
+
"reward": 1.633928656578064,
|
| 1392 |
+
"reward_std": 0.1676093488931656,
|
| 1393 |
+
"rewards/accuracy_reward": 0.6350446492433548,
|
| 1394 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1395 |
+
"step": 99
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.5074626865671643,
|
| 1399 |
+
"grad_norm": 0.09826894104480743,
|
| 1400 |
+
"learning_rate": 1.8533980447508135e-07,
|
| 1401 |
+
"loss": 0.0277,
|
| 1402 |
+
"step": 100
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.5074626865671643,
|
| 1406 |
+
"eval_clip_ratio": 0.0,
|
| 1407 |
+
"eval_completion_length": 608.394744617313,
|
| 1408 |
+
"eval_loss": 0.02166515402495861,
|
| 1409 |
+
"eval_num_tokens": 66141067.0,
|
| 1410 |
+
"eval_reward": 1.7128642008291277,
|
| 1411 |
+
"eval_reward_std": 0.17956038917201525,
|
| 1412 |
+
"eval_rewards/accuracy_reward": 0.7203461689323021,
|
| 1413 |
+
"eval_rewards/format_reward": 0.99251795181349,
|
| 1414 |
+
"eval_runtime": 7500.3981,
|
| 1415 |
+
"eval_samples_per_second": 0.667,
|
| 1416 |
+
"eval_steps_per_second": 0.006,
|
| 1417 |
+
"step": 100
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"clip_ratio": 0.0,
|
| 1421 |
+
"completion_length": 610.5853996276855,
|
| 1422 |
+
"epoch": 1.5223880597014925,
|
| 1423 |
+
"grad_norm": 0.09803326427936554,
|
| 1424 |
+
"learning_rate": 1.7527597583490823e-07,
|
| 1425 |
+
"loss": 0.0144,
|
| 1426 |
+
"num_tokens": 66825204.0,
|
| 1427 |
+
"reward": 1.758370615541935,
|
| 1428 |
+
"reward_std": 0.14415120193734765,
|
| 1429 |
+
"rewards/accuracy_reward": 0.7678571417927742,
|
| 1430 |
+
"rewards/format_reward": 0.9905133880674839,
|
| 1431 |
+
"step": 101
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"clip_ratio": 0.0,
|
| 1435 |
+
"completion_length": 589.2846298217773,
|
| 1436 |
+
"epoch": 1.537313432835821,
|
| 1437 |
+
"grad_norm": 0.09236069023609161,
|
| 1438 |
+
"learning_rate": 1.6543469682057104e-07,
|
| 1439 |
+
"loss": 0.023,
|
| 1440 |
+
"num_tokens": 67489387.0,
|
| 1441 |
+
"reward": 1.7488840073347092,
|
| 1442 |
+
"reward_std": 0.12999783549457788,
|
| 1443 |
+
"rewards/accuracy_reward": 0.7578124925494194,
|
| 1444 |
+
"rewards/format_reward": 0.9910714253783226,
|
| 1445 |
+
"step": 102
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"clip_ratio": 0.0,
|
| 1449 |
+
"completion_length": 631.8917617797852,
|
| 1450 |
+
"epoch": 1.5522388059701493,
|
| 1451 |
+
"grad_norm": 0.09945357590913773,
|
| 1452 |
+
"learning_rate": 1.5582271215312293e-07,
|
| 1453 |
+
"loss": 0.0348,
|
| 1454 |
+
"num_tokens": 68182122.0,
|
| 1455 |
+
"reward": 1.7488840073347092,
|
| 1456 |
+
"reward_std": 0.14755096472799778,
|
| 1457 |
+
"rewards/accuracy_reward": 0.7589285746216774,
|
| 1458 |
+
"rewards/format_reward": 0.9899553433060646,
|
| 1459 |
+
"step": 103
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"clip_ratio": 0.0,
|
| 1463 |
+
"completion_length": 585.780158996582,
|
| 1464 |
+
"epoch": 1.5671641791044775,
|
| 1465 |
+
"grad_norm": 0.1044122576713562,
|
| 1466 |
+
"learning_rate": 1.4644660940672627e-07,
|
| 1467 |
+
"loss": 0.0259,
|
| 1468 |
+
"num_tokens": 68843053.0,
|
| 1469 |
+
"reward": 1.765625074505806,
|
| 1470 |
+
"reward_std": 0.15429440513253212,
|
| 1471 |
+
"rewards/accuracy_reward": 0.7834821417927742,
|
| 1472 |
+
"rewards/format_reward": 0.9821428582072258,
|
| 1473 |
+
"step": 104
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"clip_ratio": 0.0,
|
| 1477 |
+
"completion_length": 621.1797103881836,
|
| 1478 |
+
"epoch": 1.582089552238806,
|
| 1479 |
+
"grad_norm": 0.11087001115083694,
|
| 1480 |
+
"learning_rate": 1.3731281449385628e-07,
|
| 1481 |
+
"loss": 0.0144,
|
| 1482 |
+
"num_tokens": 69536934.0,
|
| 1483 |
+
"reward": 1.7779018729925156,
|
| 1484 |
+
"reward_std": 0.1481210682541132,
|
| 1485 |
+
"rewards/accuracy_reward": 0.7823660671710968,
|
| 1486 |
+
"rewards/format_reward": 0.995535708963871,
|
| 1487 |
+
"step": 105
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"clip_ratio": 0.0,
|
| 1491 |
+
"completion_length": 660.2511444091797,
|
| 1492 |
+
"epoch": 1.5970149253731343,
|
| 1493 |
+
"grad_norm": 0.09636900573968887,
|
| 1494 |
+
"learning_rate": 1.284275872613028e-07,
|
| 1495 |
+
"loss": 0.0152,
|
| 1496 |
+
"num_tokens": 70259407.0,
|
| 1497 |
+
"reward": 1.7109375894069672,
|
| 1498 |
+
"reward_std": 0.17005854472517967,
|
| 1499 |
+
"rewards/accuracy_reward": 0.7220982164144516,
|
| 1500 |
+
"rewards/format_reward": 0.9888392761349678,
|
| 1501 |
+
"step": 106
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"clip_ratio": 0.0,
|
| 1505 |
+
"completion_length": 674.9152069091797,
|
| 1506 |
+
"epoch": 1.6119402985074627,
|
| 1507 |
+
"grad_norm": 0.10703866928815842,
|
| 1508 |
+
"learning_rate": 1.1979701719998454e-07,
|
| 1509 |
+
"loss": 0.0272,
|
| 1510 |
+
"num_tokens": 70998891.0,
|
| 1511 |
+
"reward": 1.7343750894069672,
|
| 1512 |
+
"reward_std": 0.18828168138861656,
|
| 1513 |
+
"rewards/accuracy_reward": 0.7444196343421936,
|
| 1514 |
+
"rewards/format_reward": 0.9899553433060646,
|
| 1515 |
+
"step": 107
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"clip_ratio": 0.0,
|
| 1519 |
+
"completion_length": 604.8225708007812,
|
| 1520 |
+
"epoch": 1.626865671641791,
|
| 1521 |
+
"grad_norm": 0.107986681163311,
|
| 1522 |
+
"learning_rate": 1.1142701927151454e-07,
|
| 1523 |
+
"loss": 0.0242,
|
| 1524 |
+
"num_tokens": 71660740.0,
|
| 1525 |
+
"reward": 1.805803656578064,
|
| 1526 |
+
"reward_std": 0.1751266662031412,
|
| 1527 |
+
"rewards/accuracy_reward": 0.8125000074505806,
|
| 1528 |
+
"rewards/format_reward": 0.9933035522699356,
|
| 1529 |
+
"step": 108
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"clip_ratio": 0.0,
|
| 1533 |
+
"completion_length": 639.0335083007812,
|
| 1534 |
+
"epoch": 1.6417910447761193,
|
| 1535 |
+
"grad_norm": 0.09717655181884766,
|
| 1536 |
+
"learning_rate": 1.0332332985438247e-07,
|
| 1537 |
+
"loss": 0.0221,
|
| 1538 |
+
"num_tokens": 72367578.0,
|
| 1539 |
+
"reward": 1.7700893580913544,
|
| 1540 |
+
"reward_std": 0.16510999109596014,
|
| 1541 |
+
"rewards/accuracy_reward": 0.7779017835855484,
|
| 1542 |
+
"rewards/format_reward": 0.9921874850988388,
|
| 1543 |
+
"step": 109
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"clip_ratio": 0.0,
|
| 1547 |
+
"completion_length": 596.9654197692871,
|
| 1548 |
+
"epoch": 1.6567164179104479,
|
| 1549 |
+
"grad_norm": 0.09363789856433868,
|
| 1550 |
+
"learning_rate": 9.549150281252632e-08,
|
| 1551 |
+
"loss": 0.0057,
|
| 1552 |
+
"num_tokens": 73028371.0,
|
| 1553 |
+
"reward": 1.7689733058214188,
|
| 1554 |
+
"reward_std": 0.1158353891223669,
|
| 1555 |
+
"rewards/accuracy_reward": 0.7745535671710968,
|
| 1556 |
+
"rewards/format_reward": 0.9944196417927742,
|
| 1557 |
+
"step": 110
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"clip_ratio": 0.0,
|
| 1561 |
+
"completion_length": 646.8337478637695,
|
| 1562 |
+
"epoch": 1.671641791044776,
|
| 1563 |
+
"grad_norm": 0.10138875991106033,
|
| 1564 |
+
"learning_rate": 8.793690568899215e-08,
|
| 1565 |
+
"loss": 0.0239,
|
| 1566 |
+
"num_tokens": 73731510.0,
|
| 1567 |
+
"reward": 1.7578125894069672,
|
| 1568 |
+
"reward_std": 0.15553954057395458,
|
| 1569 |
+
"rewards/accuracy_reward": 0.762276791036129,
|
| 1570 |
+
"rewards/format_reward": 0.995535708963871,
|
| 1571 |
+
"step": 111
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"clip_ratio": 0.0,
|
| 1575 |
+
"completion_length": 651.131721496582,
|
| 1576 |
+
"epoch": 1.6865671641791045,
|
| 1577 |
+
"grad_norm": 0.08721350878477097,
|
| 1578 |
+
"learning_rate": 8.066471602728803e-08,
|
| 1579 |
+
"loss": 0.0261,
|
| 1580 |
+
"num_tokens": 74432588.0,
|
| 1581 |
+
"reward": 1.7589286416769028,
|
| 1582 |
+
"reward_std": 0.14256859384477139,
|
| 1583 |
+
"rewards/accuracy_reward": 0.7712053507566452,
|
| 1584 |
+
"rewards/format_reward": 0.9877232164144516,
|
| 1585 |
+
"step": 112
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"clip_ratio": 0.0,
|
| 1589 |
+
"completion_length": 607.5703277587891,
|
| 1590 |
+
"epoch": 1.7014925373134329,
|
| 1591 |
+
"grad_norm": 0.10357489436864853,
|
| 1592 |
+
"learning_rate": 7.36799178229539e-08,
|
| 1593 |
+
"loss": 0.0318,
|
| 1594 |
+
"num_tokens": 75104779.0,
|
| 1595 |
+
"reward": 1.7220982760190964,
|
| 1596 |
+
"reward_std": 0.16208719182759523,
|
| 1597 |
+
"rewards/accuracy_reward": 0.7377232164144516,
|
| 1598 |
+
"rewards/format_reward": 0.984375,
|
| 1599 |
+
"step": 113
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"clip_ratio": 0.0,
|
| 1603 |
+
"completion_length": 582.2377395629883,
|
| 1604 |
+
"epoch": 1.716417910447761,
|
| 1605 |
+
"grad_norm": 0.09729992598295212,
|
| 1606 |
+
"learning_rate": 6.698729810778064e-08,
|
| 1607 |
+
"loss": 0.015,
|
| 1608 |
+
"num_tokens": 75758304.0,
|
| 1609 |
+
"reward": 1.7935268580913544,
|
| 1610 |
+
"reward_std": 0.1367027312517166,
|
| 1611 |
+
"rewards/accuracy_reward": 0.7957589328289032,
|
| 1612 |
+
"rewards/format_reward": 0.9977678582072258,
|
| 1613 |
+
"step": 114
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"clip_ratio": 0.0,
|
| 1617 |
+
"completion_length": 644.0134124755859,
|
| 1618 |
+
"epoch": 1.7313432835820897,
|
| 1619 |
+
"grad_norm": 0.10120779275894165,
|
| 1620 |
+
"learning_rate": 6.059144366901736e-08,
|
| 1621 |
+
"loss": 0.0228,
|
| 1622 |
+
"num_tokens": 76463884.0,
|
| 1623 |
+
"reward": 1.7354911714792252,
|
| 1624 |
+
"reward_std": 0.14026323426514864,
|
| 1625 |
+
"rewards/accuracy_reward": 0.7488839253783226,
|
| 1626 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 1627 |
+
"step": 115
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"clip_ratio": 0.0,
|
| 1631 |
+
"completion_length": 577.9464569091797,
|
| 1632 |
+
"epoch": 1.7462686567164178,
|
| 1633 |
+
"grad_norm": 0.08987674117088318,
|
| 1634 |
+
"learning_rate": 5.44967379058161e-08,
|
| 1635 |
+
"loss": 0.0337,
|
| 1636 |
+
"num_tokens": 77102308.0,
|
| 1637 |
+
"reward": 1.7555804550647736,
|
| 1638 |
+
"reward_std": 0.1271761180832982,
|
| 1639 |
+
"rewards/accuracy_reward": 0.762276791036129,
|
| 1640 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 1641 |
+
"step": 116
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"clip_ratio": 0.0,
|
| 1645 |
+
"completion_length": 644.6161041259766,
|
| 1646 |
+
"epoch": 1.7611940298507462,
|
| 1647 |
+
"grad_norm": 0.11021538078784943,
|
| 1648 |
+
"learning_rate": 4.870735782506979e-08,
|
| 1649 |
+
"loss": 0.0105,
|
| 1650 |
+
"num_tokens": 77835332.0,
|
| 1651 |
+
"reward": 1.7142857909202576,
|
| 1652 |
+
"reward_std": 0.18193747848272324,
|
| 1653 |
+
"rewards/accuracy_reward": 0.7165178582072258,
|
| 1654 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1655 |
+
"step": 117
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"clip_ratio": 0.0,
|
| 1659 |
+
"completion_length": 640.1942176818848,
|
| 1660 |
+
"epoch": 1.7761194029850746,
|
| 1661 |
+
"grad_norm": 0.12281131744384766,
|
| 1662 |
+
"learning_rate": 4.322727117869951e-08,
|
| 1663 |
+
"loss": 0.0103,
|
| 1664 |
+
"num_tokens": 78535354.0,
|
| 1665 |
+
"reward": 1.7522322237491608,
|
| 1666 |
+
"reward_std": 0.15158831980079412,
|
| 1667 |
+
"rewards/accuracy_reward": 0.7555803507566452,
|
| 1668 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 1669 |
+
"step": 118
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"clip_ratio": 0.0,
|
| 1673 |
+
"completion_length": 620.7723617553711,
|
| 1674 |
+
"epoch": 1.7910447761194028,
|
| 1675 |
+
"grad_norm": 0.09940478950738907,
|
| 1676 |
+
"learning_rate": 3.806023374435663e-08,
|
| 1677 |
+
"loss": 0.0351,
|
| 1678 |
+
"num_tokens": 79216974.0,
|
| 1679 |
+
"reward": 1.7254465073347092,
|
| 1680 |
+
"reward_std": 0.16859996505081654,
|
| 1681 |
+
"rewards/accuracy_reward": 0.7366071492433548,
|
| 1682 |
+
"rewards/format_reward": 0.9888392835855484,
|
| 1683 |
+
"step": 119
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"clip_ratio": 0.0,
|
| 1687 |
+
"completion_length": 641.8460083007812,
|
| 1688 |
+
"epoch": 1.8059701492537314,
|
| 1689 |
+
"grad_norm": 0.09752647578716278,
|
| 1690 |
+
"learning_rate": 3.3209786751399184e-08,
|
| 1691 |
+
"loss": 0.0242,
|
| 1692 |
+
"num_tokens": 79931012.0,
|
| 1693 |
+
"reward": 1.7723215073347092,
|
| 1694 |
+
"reward_std": 0.17135184817016125,
|
| 1695 |
+
"rewards/accuracy_reward": 0.7845982164144516,
|
| 1696 |
+
"rewards/format_reward": 0.9877232164144516,
|
| 1697 |
+
"step": 120
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"clip_ratio": 0.0,
|
| 1701 |
+
"completion_length": 605.100471496582,
|
| 1702 |
+
"epoch": 1.8208955223880596,
|
| 1703 |
+
"grad_norm": 0.09651315957307816,
|
| 1704 |
+
"learning_rate": 2.8679254453910785e-08,
|
| 1705 |
+
"loss": 0.025,
|
| 1706 |
+
"num_tokens": 80603486.0,
|
| 1707 |
+
"reward": 1.7366072237491608,
|
| 1708 |
+
"reward_std": 0.14862325228750706,
|
| 1709 |
+
"rewards/accuracy_reward": 0.7410714328289032,
|
| 1710 |
+
"rewards/format_reward": 0.9955357164144516,
|
| 1711 |
+
"step": 121
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"clip_ratio": 0.0,
|
| 1715 |
+
"completion_length": 651.4765853881836,
|
| 1716 |
+
"epoch": 1.835820895522388,
|
| 1717 |
+
"grad_norm": 0.09396825730800629,
|
| 1718 |
+
"learning_rate": 2.4471741852423233e-08,
|
| 1719 |
+
"loss": 0.0172,
|
| 1720 |
+
"num_tokens": 81334001.0,
|
| 1721 |
+
"reward": 1.7354911714792252,
|
| 1722 |
+
"reward_std": 0.16738921124488115,
|
| 1723 |
+
"rewards/accuracy_reward": 0.7511160746216774,
|
| 1724 |
+
"rewards/format_reward": 0.9843749850988388,
|
| 1725 |
+
"step": 122
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"clip_ratio": 0.0,
|
| 1729 |
+
"completion_length": 638.3326187133789,
|
| 1730 |
+
"epoch": 1.8507462686567164,
|
| 1731 |
+
"grad_norm": 0.1153474748134613,
|
| 1732 |
+
"learning_rate": 2.0590132565903473e-08,
|
| 1733 |
+
"loss": 0.0411,
|
| 1734 |
+
"num_tokens": 82038987.0,
|
| 1735 |
+
"reward": 1.7734375894069672,
|
| 1736 |
+
"reward_std": 0.2178222257643938,
|
| 1737 |
+
"rewards/accuracy_reward": 0.7868303507566452,
|
| 1738 |
+
"rewards/format_reward": 0.9866071417927742,
|
| 1739 |
+
"step": 123
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"clip_ratio": 0.0,
|
| 1743 |
+
"completion_length": 599.5000228881836,
|
| 1744 |
+
"epoch": 1.8656716417910446,
|
| 1745 |
+
"grad_norm": 0.10678599029779434,
|
| 1746 |
+
"learning_rate": 1.7037086855465898e-08,
|
| 1747 |
+
"loss": 0.0201,
|
| 1748 |
+
"num_tokens": 82700603.0,
|
| 1749 |
+
"reward": 1.7354911714792252,
|
| 1750 |
+
"reward_std": 0.1632389174774289,
|
| 1751 |
+
"rewards/accuracy_reward": 0.7421874925494194,
|
| 1752 |
+
"rewards/format_reward": 0.9933035597205162,
|
| 1753 |
+
"step": 124
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"clip_ratio": 0.0,
|
| 1757 |
+
"completion_length": 644.1205596923828,
|
| 1758 |
+
"epoch": 1.8805970149253732,
|
| 1759 |
+
"grad_norm": 0.08962120860815048,
|
| 1760 |
+
"learning_rate": 1.3815039801161722e-08,
|
| 1761 |
+
"loss": 0.0185,
|
| 1762 |
+
"num_tokens": 83404503.0,
|
| 1763 |
+
"reward": 1.750000074505806,
|
| 1764 |
+
"reward_std": 0.15682824235409498,
|
| 1765 |
+
"rewards/accuracy_reward": 0.7566964253783226,
|
| 1766 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1767 |
+
"step": 125
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"clip_ratio": 0.0,
|
| 1771 |
+
"completion_length": 565.2109603881836,
|
| 1772 |
+
"epoch": 1.8955223880597014,
|
| 1773 |
+
"grad_norm": 0.10807648301124573,
|
| 1774 |
+
"learning_rate": 1.0926199633097154e-08,
|
| 1775 |
+
"loss": 0.0042,
|
| 1776 |
+
"num_tokens": 84037108.0,
|
| 1777 |
+
"reward": 1.8046875894069672,
|
| 1778 |
+
"reward_std": 0.13849271647632122,
|
| 1779 |
+
"rewards/accuracy_reward": 0.8080357164144516,
|
| 1780 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 1781 |
+
"step": 126
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"clip_ratio": 0.0,
|
| 1785 |
+
"completion_length": 633.560302734375,
|
| 1786 |
+
"epoch": 1.9104477611940298,
|
| 1787 |
+
"grad_norm": 0.10242209583520889,
|
| 1788 |
+
"learning_rate": 8.372546218022746e-09,
|
| 1789 |
+
"loss": 0.0202,
|
| 1790 |
+
"num_tokens": 84741714.0,
|
| 1791 |
+
"reward": 1.7120536714792252,
|
| 1792 |
+
"reward_std": 0.158757739700377,
|
| 1793 |
+
"rewards/accuracy_reward": 0.7187500074505806,
|
| 1794 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 1795 |
+
"step": 127
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"clip_ratio": 0.0,
|
| 1799 |
+
"completion_length": 597.8248062133789,
|
| 1800 |
+
"epoch": 1.9253731343283582,
|
| 1801 |
+
"grad_norm": 0.09526298940181732,
|
| 1802 |
+
"learning_rate": 6.15582970243117e-09,
|
| 1803 |
+
"loss": 0.0091,
|
| 1804 |
+
"num_tokens": 85407733.0,
|
| 1805 |
+
"reward": 1.8069197237491608,
|
| 1806 |
+
"reward_std": 0.11288107186555862,
|
| 1807 |
+
"rewards/accuracy_reward": 0.8069196492433548,
|
| 1808 |
+
"rewards/format_reward": 1.0,
|
| 1809 |
+
"step": 128
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"clip_ratio": 0.0,
|
| 1813 |
+
"completion_length": 644.9765853881836,
|
| 1814 |
+
"epoch": 1.9402985074626866,
|
| 1815 |
+
"grad_norm": 0.09915687888860703,
|
| 1816 |
+
"learning_rate": 4.277569313094809e-09,
|
| 1817 |
+
"loss": 0.0311,
|
| 1818 |
+
"num_tokens": 86108960.0,
|
| 1819 |
+
"reward": 1.7622768729925156,
|
| 1820 |
+
"reward_std": 0.1561026033014059,
|
| 1821 |
+
"rewards/accuracy_reward": 0.7756696343421936,
|
| 1822 |
+
"rewards/format_reward": 0.9866071343421936,
|
| 1823 |
+
"step": 129
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"clip_ratio": 0.0,
|
| 1827 |
+
"completion_length": 646.7611846923828,
|
| 1828 |
+
"epoch": 1.955223880597015,
|
| 1829 |
+
"grad_norm": 0.09854050725698471,
|
| 1830 |
+
"learning_rate": 2.739052315863355e-09,
|
| 1831 |
+
"loss": 0.007,
|
| 1832 |
+
"num_tokens": 86830202.0,
|
| 1833 |
+
"reward": 1.6919643580913544,
|
| 1834 |
+
"reward_std": 0.14856339246034622,
|
| 1835 |
+
"rewards/accuracy_reward": 0.6953125074505806,
|
| 1836 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 1837 |
+
"step": 130
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"clip_ratio": 0.0,
|
| 1841 |
+
"completion_length": 597.3638725280762,
|
| 1842 |
+
"epoch": 1.9701492537313432,
|
| 1843 |
+
"grad_norm": 0.11233126372098923,
|
| 1844 |
+
"learning_rate": 1.541333133436018e-09,
|
| 1845 |
+
"loss": -0.002,
|
| 1846 |
+
"num_tokens": 87492608.0,
|
| 1847 |
+
"reward": 1.7845982909202576,
|
| 1848 |
+
"reward_std": 0.14027062244713306,
|
| 1849 |
+
"rewards/accuracy_reward": 0.7868303582072258,
|
| 1850 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1851 |
+
"step": 131
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"clip_ratio": 0.0,
|
| 1855 |
+
"completion_length": 627.9294013977051,
|
| 1856 |
+
"epoch": 1.9850746268656716,
|
| 1857 |
+
"grad_norm": 0.10069043189287186,
|
| 1858 |
+
"learning_rate": 6.852326227130833e-10,
|
| 1859 |
+
"loss": 0.0303,
|
| 1860 |
+
"num_tokens": 88193234.0,
|
| 1861 |
+
"reward": 1.7500000894069672,
|
| 1862 |
+
"reward_std": 0.14365093689411879,
|
| 1863 |
+
"rewards/accuracy_reward": 0.7555803582072258,
|
| 1864 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 1865 |
+
"step": 132
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.9850746268656716,
|
| 1869 |
+
"step": 132,
|
| 1870 |
+
"total_flos": 0.0,
|
| 1871 |
+
"train_loss": 0.023034477807496758,
|
| 1872 |
+
"train_runtime": 35029.9588,
|
| 1873 |
+
"train_samples_per_second": 0.428,
|
| 1874 |
+
"train_steps_per_second": 0.004
|
| 1875 |
+
}
|
| 1876 |
+
],
|
| 1877 |
+
"logging_steps": 1,
|
| 1878 |
+
"max_steps": 134,
|
| 1879 |
+
"num_input_tokens_seen": 0,
|
| 1880 |
+
"num_train_epochs": 2,
|
| 1881 |
+
"save_steps": 500,
|
| 1882 |
+
"stateful_callbacks": {
|
| 1883 |
+
"TrainerControl": {
|
| 1884 |
+
"args": {
|
| 1885 |
+
"should_epoch_stop": false,
|
| 1886 |
+
"should_evaluate": false,
|
| 1887 |
+
"should_log": false,
|
| 1888 |
+
"should_save": true,
|
| 1889 |
+
"should_training_stop": false
|
| 1890 |
+
},
|
| 1891 |
+
"attributes": {}
|
| 1892 |
+
}
|
| 1893 |
+
},
|
| 1894 |
+
"total_flos": 0.0,
|
| 1895 |
+
"train_batch_size": 16,
|
| 1896 |
+
"trial_name": null,
|
| 1897 |
+
"trial_params": null
|
| 1898 |
+
}
|