Model save
Browse files- README.md +67 -0
- all_results.json +8 -0
- generation_config.json +6 -0
- train_results.json +8 -0
- trainer_state.json +1199 -0
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
model_name: Qwen2.5-7B-EMPO-NM-COT-20K-2epoch
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
- trl
|
| 7 |
+
- grpo
|
| 8 |
+
licence: license
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Qwen2.5-7B-EMPO-NM-COT-20K-2epoch
|
| 12 |
+
|
| 13 |
+
This model is a fine-tuned version of [None](https://huggingface.co/None).
|
| 14 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 15 |
+
|
| 16 |
+
## Quick start
|
| 17 |
+
|
| 18 |
+
```python
|
| 19 |
+
from transformers import pipeline
|
| 20 |
+
|
| 21 |
+
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?"
|
| 22 |
+
generator = pipeline("text-generation", model="Haitao999/Qwen2.5-7B-EMPO-NM-COT-20K-2epoch", device="cuda")
|
| 23 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 24 |
+
print(output["generated_text"])
|
| 25 |
+
```
|
| 26 |
+
|
| 27 |
+
## Training procedure
|
| 28 |
+
|
| 29 |
+
[<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/tjucsailab/huggingface/runs/9tiixwvm)
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
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).
|
| 33 |
+
|
| 34 |
+
### Framework versions
|
| 35 |
+
|
| 36 |
+
- TRL: 0.14.0
|
| 37 |
+
- Transformers: 4.48.3
|
| 38 |
+
- Pytorch: 2.5.1+cu124
|
| 39 |
+
- Datasets: 3.1.0
|
| 40 |
+
- Tokenizers: 0.21.0
|
| 41 |
+
|
| 42 |
+
## Citations
|
| 43 |
+
|
| 44 |
+
Cite GRPO as:
|
| 45 |
+
|
| 46 |
+
```bibtex
|
| 47 |
+
@article{zhihong2024deepseekmath,
|
| 48 |
+
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
|
| 49 |
+
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},
|
| 50 |
+
year = 2024,
|
| 51 |
+
eprint = {arXiv:2402.03300},
|
| 52 |
+
}
|
| 53 |
+
|
| 54 |
+
```
|
| 55 |
+
|
| 56 |
+
Cite TRL as:
|
| 57 |
+
|
| 58 |
+
```bibtex
|
| 59 |
+
@misc{vonwerra2022trl,
|
| 60 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 61 |
+
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},
|
| 62 |
+
year = 2020,
|
| 63 |
+
journal = {GitHub repository},
|
| 64 |
+
publisher = {GitHub},
|
| 65 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 66 |
+
}
|
| 67 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 2.8008765204773544e-08,
|
| 4 |
+
"train_runtime": 57265.8095,
|
| 5 |
+
"train_samples": 20000,
|
| 6 |
+
"train_samples_per_second": 0.349,
|
| 7 |
+
"train_steps_per_second": 0.002
|
| 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": "4.48.3"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 2.8008765204773544e-08,
|
| 4 |
+
"train_runtime": 57265.8095,
|
| 5 |
+
"train_samples": 20000,
|
| 6 |
+
"train_samples_per_second": 0.349,
|
| 7 |
+
"train_steps_per_second": 0.002
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1199 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9965010496850945,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 89,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"completion_length": 719.0248603820801,
|
| 13 |
+
"epoch": 0.01119664100769769,
|
| 14 |
+
"grad_norm": 5.308396339416504,
|
| 15 |
+
"learning_rate": 3e-06,
|
| 16 |
+
"loss": 0.0,
|
| 17 |
+
"reward": 0.7022594679147005,
|
| 18 |
+
"reward_std": 0.11167733068577945,
|
| 19 |
+
"rewards/accuracy_reward": 0.4343112222850323,
|
| 20 |
+
"rewards/semantic_entropy_math_reward": 0.7022594679147005,
|
| 21 |
+
"rewards/total_entropy_reward": 1.2448364309966564,
|
| 22 |
+
"step": 1
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"completion_length": 697.6454010009766,
|
| 26 |
+
"epoch": 0.02239328201539538,
|
| 27 |
+
"grad_norm": 3.310378313064575,
|
| 28 |
+
"learning_rate": 3e-06,
|
| 29 |
+
"loss": 0.0,
|
| 30 |
+
"reward": 0.6712827906012535,
|
| 31 |
+
"reward_std": 0.1011071486864239,
|
| 32 |
+
"rewards/accuracy_reward": 0.352040808647871,
|
| 33 |
+
"rewards/semantic_entropy_math_reward": 0.6712827868759632,
|
| 34 |
+
"rewards/total_entropy_reward": 1.305861696600914,
|
| 35 |
+
"step": 2
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"completion_length": 701.3826370239258,
|
| 39 |
+
"epoch": 0.03358992302309307,
|
| 40 |
+
"grad_norm": 4.717713356018066,
|
| 41 |
+
"learning_rate": 3e-06,
|
| 42 |
+
"loss": 0.0,
|
| 43 |
+
"reward": 0.7075437195599079,
|
| 44 |
+
"reward_std": 0.07735519809648395,
|
| 45 |
+
"rewards/accuracy_reward": 0.4126275437884033,
|
| 46 |
+
"rewards/semantic_entropy_math_reward": 0.7075437270104885,
|
| 47 |
+
"rewards/total_entropy_reward": 1.2551036067306995,
|
| 48 |
+
"step": 3
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"completion_length": 692.5803489685059,
|
| 52 |
+
"epoch": 0.04478656403079076,
|
| 53 |
+
"grad_norm": 4.699649810791016,
|
| 54 |
+
"learning_rate": 3e-06,
|
| 55 |
+
"loss": 0.0,
|
| 56 |
+
"reward": 0.6926020309329033,
|
| 57 |
+
"reward_std": 0.09249642631039023,
|
| 58 |
+
"rewards/accuracy_reward": 0.4196428433060646,
|
| 59 |
+
"rewards/semantic_entropy_math_reward": 0.6926020495593548,
|
| 60 |
+
"rewards/total_entropy_reward": 1.2794943004846573,
|
| 61 |
+
"step": 4
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"completion_length": 643.8430976867676,
|
| 65 |
+
"epoch": 0.05598320503848846,
|
| 66 |
+
"grad_norm": 7.118838310241699,
|
| 67 |
+
"learning_rate": 3e-06,
|
| 68 |
+
"loss": 0.0,
|
| 69 |
+
"reward": 0.7232142686843872,
|
| 70 |
+
"reward_std": 0.09167159674689174,
|
| 71 |
+
"rewards/accuracy_reward": 0.41326529532670975,
|
| 72 |
+
"rewards/semantic_entropy_math_reward": 0.7232142575085163,
|
| 73 |
+
"rewards/total_entropy_reward": 1.2130939476191998,
|
| 74 |
+
"step": 5
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"completion_length": 711.6734580993652,
|
| 78 |
+
"epoch": 0.06717984604618614,
|
| 79 |
+
"grad_norm": 4.829713344573975,
|
| 80 |
+
"learning_rate": 3e-06,
|
| 81 |
+
"loss": 0.0,
|
| 82 |
+
"reward": 0.6625364199280739,
|
| 83 |
+
"reward_std": 0.1073409270029515,
|
| 84 |
+
"rewards/accuracy_reward": 0.37436223588883877,
|
| 85 |
+
"rewards/semantic_entropy_math_reward": 0.6625364273786545,
|
| 86 |
+
"rewards/total_entropy_reward": 1.320713147521019,
|
| 87 |
+
"step": 6
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"completion_length": 676.0541915893555,
|
| 91 |
+
"epoch": 0.07837648705388384,
|
| 92 |
+
"grad_norm": 4.374125003814697,
|
| 93 |
+
"learning_rate": 3e-06,
|
| 94 |
+
"loss": 0.0,
|
| 95 |
+
"reward": 0.7006195187568665,
|
| 96 |
+
"reward_std": 0.09755392652004957,
|
| 97 |
+
"rewards/accuracy_reward": 0.4247448882088065,
|
| 98 |
+
"rewards/semantic_entropy_math_reward": 0.7006195187568665,
|
| 99 |
+
"rewards/total_entropy_reward": 1.2528423443436623,
|
| 100 |
+
"step": 7
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"completion_length": 712.5401573181152,
|
| 104 |
+
"epoch": 0.08957312806158152,
|
| 105 |
+
"grad_norm": 6.558979034423828,
|
| 106 |
+
"learning_rate": 3e-06,
|
| 107 |
+
"loss": 0.0,
|
| 108 |
+
"reward": 0.633381923660636,
|
| 109 |
+
"reward_std": 0.10544528882019222,
|
| 110 |
+
"rewards/accuracy_reward": 0.3463010173290968,
|
| 111 |
+
"rewards/semantic_entropy_math_reward": 0.6333819553256035,
|
| 112 |
+
"rewards/total_entropy_reward": 1.3834920637309551,
|
| 113 |
+
"step": 8
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"completion_length": 711.6345539093018,
|
| 117 |
+
"epoch": 0.10076976906927922,
|
| 118 |
+
"grad_norm": 7.621776103973389,
|
| 119 |
+
"learning_rate": 3e-06,
|
| 120 |
+
"loss": 0.0,
|
| 121 |
+
"reward": 0.6800291426479816,
|
| 122 |
+
"reward_std": 0.10248321201652288,
|
| 123 |
+
"rewards/accuracy_reward": 0.39604590833187103,
|
| 124 |
+
"rewards/semantic_entropy_math_reward": 0.6800291538238525,
|
| 125 |
+
"rewards/total_entropy_reward": 1.2968212738633156,
|
| 126 |
+
"step": 9
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"completion_length": 674.1779174804688,
|
| 130 |
+
"epoch": 0.11196641007697691,
|
| 131 |
+
"grad_norm": 7.523331165313721,
|
| 132 |
+
"learning_rate": 3e-06,
|
| 133 |
+
"loss": 0.0,
|
| 134 |
+
"reward": 0.6800291538238525,
|
| 135 |
+
"reward_std": 0.11024063220247626,
|
| 136 |
+
"rewards/accuracy_reward": 0.411352033726871,
|
| 137 |
+
"rewards/semantic_entropy_math_reward": 0.6800291500985622,
|
| 138 |
+
"rewards/total_entropy_reward": 1.2904038280248642,
|
| 139 |
+
"step": 10
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"completion_length": 741.4017696380615,
|
| 143 |
+
"epoch": 0.1231630510846746,
|
| 144 |
+
"grad_norm": 6.336475372314453,
|
| 145 |
+
"learning_rate": 3e-06,
|
| 146 |
+
"loss": 0.0,
|
| 147 |
+
"reward": 0.6415816266089678,
|
| 148 |
+
"reward_std": 0.10356245189905167,
|
| 149 |
+
"rewards/accuracy_reward": 0.3858418306335807,
|
| 150 |
+
"rewards/semantic_entropy_math_reward": 0.6415816303342581,
|
| 151 |
+
"rewards/total_entropy_reward": 1.363376997411251,
|
| 152 |
+
"step": 11
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"completion_length": 692.8437347412109,
|
| 156 |
+
"epoch": 0.13435969209237228,
|
| 157 |
+
"grad_norm": 7.185462474822998,
|
| 158 |
+
"learning_rate": 3e-06,
|
| 159 |
+
"loss": 0.0,
|
| 160 |
+
"reward": 0.7002550885081291,
|
| 161 |
+
"reward_std": 0.10777880600653589,
|
| 162 |
+
"rewards/accuracy_reward": 0.43048468697816133,
|
| 163 |
+
"rewards/semantic_entropy_math_reward": 0.7002551108598709,
|
| 164 |
+
"rewards/total_entropy_reward": 1.2493381686508656,
|
| 165 |
+
"step": 12
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"completion_length": 704.6638870239258,
|
| 169 |
+
"epoch": 0.14555633310007,
|
| 170 |
+
"grad_norm": 17.740726470947266,
|
| 171 |
+
"learning_rate": 3e-06,
|
| 172 |
+
"loss": 0.0,
|
| 173 |
+
"reward": 0.673833817243576,
|
| 174 |
+
"reward_std": 0.09210015833377838,
|
| 175 |
+
"rewards/accuracy_reward": 0.43558672815561295,
|
| 176 |
+
"rewards/semantic_entropy_math_reward": 0.6738338358700275,
|
| 177 |
+
"rewards/total_entropy_reward": 1.3185552880167961,
|
| 178 |
+
"step": 13
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"completion_length": 679.9374847412109,
|
| 182 |
+
"epoch": 0.15675297410776767,
|
| 183 |
+
"grad_norm": 9.5872802734375,
|
| 184 |
+
"learning_rate": 3e-06,
|
| 185 |
+
"loss": 0.0,
|
| 186 |
+
"reward": 0.7075437121093273,
|
| 187 |
+
"reward_std": 0.09575178450904787,
|
| 188 |
+
"rewards/accuracy_reward": 0.40624999068677425,
|
| 189 |
+
"rewards/semantic_entropy_math_reward": 0.7075437419116497,
|
| 190 |
+
"rewards/total_entropy_reward": 1.237540539354086,
|
| 191 |
+
"step": 14
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"completion_length": 731.2404136657715,
|
| 195 |
+
"epoch": 0.16794961511546536,
|
| 196 |
+
"grad_norm": 4.687215805053711,
|
| 197 |
+
"learning_rate": 3e-06,
|
| 198 |
+
"loss": 0.0,
|
| 199 |
+
"reward": 0.6375728864222765,
|
| 200 |
+
"reward_std": 0.108485147356987,
|
| 201 |
+
"rewards/accuracy_reward": 0.3762755021452904,
|
| 202 |
+
"rewards/semantic_entropy_math_reward": 0.6375729013234377,
|
| 203 |
+
"rewards/total_entropy_reward": 1.3838917911052704,
|
| 204 |
+
"step": 15
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"completion_length": 700.0522842407227,
|
| 208 |
+
"epoch": 0.17914625612316304,
|
| 209 |
+
"grad_norm": 8.418819427490234,
|
| 210 |
+
"learning_rate": 3e-06,
|
| 211 |
+
"loss": 0.0,
|
| 212 |
+
"reward": 0.6547011584043503,
|
| 213 |
+
"reward_std": 0.10533164534717798,
|
| 214 |
+
"rewards/accuracy_reward": 0.40624999161809683,
|
| 215 |
+
"rewards/semantic_entropy_math_reward": 0.6547011733055115,
|
| 216 |
+
"rewards/total_entropy_reward": 1.3433180004358292,
|
| 217 |
+
"step": 16
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"completion_length": 684.034423828125,
|
| 221 |
+
"epoch": 0.19034289713086075,
|
| 222 |
+
"grad_norm": 7.032946586608887,
|
| 223 |
+
"learning_rate": 3e-06,
|
| 224 |
+
"loss": 0.0,
|
| 225 |
+
"reward": 0.664905235171318,
|
| 226 |
+
"reward_std": 0.0974241562653333,
|
| 227 |
+
"rewards/accuracy_reward": 0.3832908095791936,
|
| 228 |
+
"rewards/semantic_entropy_math_reward": 0.6649052500724792,
|
| 229 |
+
"rewards/total_entropy_reward": 1.3396263718605042,
|
| 230 |
+
"step": 17
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"completion_length": 677.9177227020264,
|
| 234 |
+
"epoch": 0.20153953813855843,
|
| 235 |
+
"grad_norm": 15.37868595123291,
|
| 236 |
+
"learning_rate": 3e-06,
|
| 237 |
+
"loss": 0.0,
|
| 238 |
+
"reward": 0.6975218504667282,
|
| 239 |
+
"reward_std": 0.09150373586453497,
|
| 240 |
+
"rewards/accuracy_reward": 0.45153060369193554,
|
| 241 |
+
"rewards/semantic_entropy_math_reward": 0.69752187281847,
|
| 242 |
+
"rewards/total_entropy_reward": 1.262941613793373,
|
| 243 |
+
"step": 18
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"completion_length": 683.5771484375,
|
| 247 |
+
"epoch": 0.21273617914625612,
|
| 248 |
+
"grad_norm": 36.41266632080078,
|
| 249 |
+
"learning_rate": 3e-06,
|
| 250 |
+
"loss": 0.0,
|
| 251 |
+
"reward": 0.7057215645909309,
|
| 252 |
+
"reward_std": 0.09910683194175363,
|
| 253 |
+
"rewards/accuracy_reward": 0.4368622377514839,
|
| 254 |
+
"rewards/semantic_entropy_math_reward": 0.7057215794920921,
|
| 255 |
+
"rewards/total_entropy_reward": 1.2459207847714424,
|
| 256 |
+
"step": 19
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"completion_length": 688.8073768615723,
|
| 260 |
+
"epoch": 0.22393282015395383,
|
| 261 |
+
"grad_norm": 27.405065536499023,
|
| 262 |
+
"learning_rate": 3e-06,
|
| 263 |
+
"loss": 0.0,
|
| 264 |
+
"reward": 0.7097303122282028,
|
| 265 |
+
"reward_std": 0.09203298413194716,
|
| 266 |
+
"rewards/accuracy_reward": 0.4598214225843549,
|
| 267 |
+
"rewards/semantic_entropy_math_reward": 0.7097303308546543,
|
| 268 |
+
"rewards/total_entropy_reward": 1.2417229264974594,
|
| 269 |
+
"step": 20
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"completion_length": 736.3092956542969,
|
| 273 |
+
"epoch": 0.2351294611616515,
|
| 274 |
+
"grad_norm": 19.091941833496094,
|
| 275 |
+
"learning_rate": 3e-06,
|
| 276 |
+
"loss": 0.0,
|
| 277 |
+
"reward": 0.6541544962674379,
|
| 278 |
+
"reward_std": 0.0986680502537638,
|
| 279 |
+
"rewards/accuracy_reward": 0.4228316266089678,
|
| 280 |
+
"rewards/semantic_entropy_math_reward": 0.6541545186191797,
|
| 281 |
+
"rewards/total_entropy_reward": 1.3527532257139683,
|
| 282 |
+
"step": 21
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"completion_length": 709.4795722961426,
|
| 286 |
+
"epoch": 0.2463261021693492,
|
| 287 |
+
"grad_norm": 14.640216827392578,
|
| 288 |
+
"learning_rate": 3e-06,
|
| 289 |
+
"loss": 0.0,
|
| 290 |
+
"reward": 0.6836734637618065,
|
| 291 |
+
"reward_std": 0.10867427196353674,
|
| 292 |
+
"rewards/accuracy_reward": 0.46301019564270973,
|
| 293 |
+
"rewards/semantic_entropy_math_reward": 0.6836734749376774,
|
| 294 |
+
"rewards/total_entropy_reward": 1.2770788073539734,
|
| 295 |
+
"step": 22
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"completion_length": 681.6218032836914,
|
| 299 |
+
"epoch": 0.2575227431770469,
|
| 300 |
+
"grad_norm": 13.511483192443848,
|
| 301 |
+
"learning_rate": 3e-06,
|
| 302 |
+
"loss": 0.0,
|
| 303 |
+
"reward": 0.6967929936945438,
|
| 304 |
+
"reward_std": 0.09672949556261301,
|
| 305 |
+
"rewards/accuracy_reward": 0.44579080305993557,
|
| 306 |
+
"rewards/semantic_entropy_math_reward": 0.6967929899692535,
|
| 307 |
+
"rewards/total_entropy_reward": 1.2772413976490498,
|
| 308 |
+
"step": 23
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"completion_length": 705.6811141967773,
|
| 312 |
+
"epoch": 0.26871938418474456,
|
| 313 |
+
"grad_norm": 15.547083854675293,
|
| 314 |
+
"learning_rate": 3e-06,
|
| 315 |
+
"loss": 0.0,
|
| 316 |
+
"reward": 0.6461370252072811,
|
| 317 |
+
"reward_std": 0.1031385101377964,
|
| 318 |
+
"rewards/accuracy_reward": 0.3985969265922904,
|
| 319 |
+
"rewards/semantic_entropy_math_reward": 0.6461370065808296,
|
| 320 |
+
"rewards/total_entropy_reward": 1.3622385039925575,
|
| 321 |
+
"step": 24
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"completion_length": 731.4566116333008,
|
| 325 |
+
"epoch": 0.27991602519244224,
|
| 326 |
+
"grad_norm": 8.434446334838867,
|
| 327 |
+
"learning_rate": 3e-06,
|
| 328 |
+
"loss": 0.0,
|
| 329 |
+
"reward": 0.6233600564301014,
|
| 330 |
+
"reward_std": 0.08827707398450002,
|
| 331 |
+
"rewards/accuracy_reward": 0.3494897885248065,
|
| 332 |
+
"rewards/semantic_entropy_math_reward": 0.6233600713312626,
|
| 333 |
+
"rewards/total_entropy_reward": 1.4261119738221169,
|
| 334 |
+
"step": 25
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"completion_length": 685.5197486877441,
|
| 338 |
+
"epoch": 0.29111266620014,
|
| 339 |
+
"grad_norm": 12.586360931396484,
|
| 340 |
+
"learning_rate": 3e-06,
|
| 341 |
+
"loss": 0.0,
|
| 342 |
+
"reward": 0.6765670366585255,
|
| 343 |
+
"reward_std": 0.1049469755962491,
|
| 344 |
+
"rewards/accuracy_reward": 0.412627543322742,
|
| 345 |
+
"rewards/semantic_entropy_math_reward": 0.6765670329332352,
|
| 346 |
+
"rewards/total_entropy_reward": 1.3074346259236336,
|
| 347 |
+
"step": 26
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"completion_length": 711.061840057373,
|
| 351 |
+
"epoch": 0.30230930720783766,
|
| 352 |
+
"grad_norm": 9.39128303527832,
|
| 353 |
+
"learning_rate": 3e-06,
|
| 354 |
+
"loss": 0.0,
|
| 355 |
+
"reward": 0.6521501429378986,
|
| 356 |
+
"reward_std": 0.1099298931658268,
|
| 357 |
+
"rewards/accuracy_reward": 0.40624999068677425,
|
| 358 |
+
"rewards/semantic_entropy_math_reward": 0.6521501652896404,
|
| 359 |
+
"rewards/total_entropy_reward": 1.3500806987285614,
|
| 360 |
+
"step": 27
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"completion_length": 707.5612106323242,
|
| 364 |
+
"epoch": 0.31350594821553535,
|
| 365 |
+
"grad_norm": 8.279614448547363,
|
| 366 |
+
"learning_rate": 3e-06,
|
| 367 |
+
"loss": 0.0,
|
| 368 |
+
"reward": 0.6692784316837788,
|
| 369 |
+
"reward_std": 0.10041180578991771,
|
| 370 |
+
"rewards/accuracy_reward": 0.36096938233822584,
|
| 371 |
+
"rewards/semantic_entropy_math_reward": 0.6692784205079079,
|
| 372 |
+
"rewards/total_entropy_reward": 1.3163355849683285,
|
| 373 |
+
"step": 28
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"completion_length": 675.4157962799072,
|
| 377 |
+
"epoch": 0.32470258922323303,
|
| 378 |
+
"grad_norm": 13.983796119689941,
|
| 379 |
+
"learning_rate": 3e-06,
|
| 380 |
+
"loss": 0.0,
|
| 381 |
+
"reward": 0.7091836705803871,
|
| 382 |
+
"reward_std": 0.09643913782201707,
|
| 383 |
+
"rewards/accuracy_reward": 0.478316318243742,
|
| 384 |
+
"rewards/semantic_entropy_math_reward": 0.7091836743056774,
|
| 385 |
+
"rewards/total_entropy_reward": 1.2364819720387459,
|
| 386 |
+
"step": 29
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"completion_length": 723.9189872741699,
|
| 390 |
+
"epoch": 0.3358992302309307,
|
| 391 |
+
"grad_norm": 13.94961166381836,
|
| 392 |
+
"learning_rate": 3e-06,
|
| 393 |
+
"loss": 0.0,
|
| 394 |
+
"reward": 0.614431481808424,
|
| 395 |
+
"reward_std": 0.12222768180072308,
|
| 396 |
+
"rewards/accuracy_reward": 0.3533163210377097,
|
| 397 |
+
"rewards/semantic_entropy_math_reward": 0.614431481808424,
|
| 398 |
+
"rewards/total_entropy_reward": 1.4096959978342056,
|
| 399 |
+
"step": 30
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"completion_length": 721.4604339599609,
|
| 403 |
+
"epoch": 0.3470958712386284,
|
| 404 |
+
"grad_norm": 7.932522773742676,
|
| 405 |
+
"learning_rate": 3e-06,
|
| 406 |
+
"loss": 0.0,
|
| 407 |
+
"reward": 0.6603498533368111,
|
| 408 |
+
"reward_std": 0.09537219302728772,
|
| 409 |
+
"rewards/accuracy_reward": 0.3992346851155162,
|
| 410 |
+
"rewards/semantic_entropy_math_reward": 0.6603498607873917,
|
| 411 |
+
"rewards/total_entropy_reward": 1.3325160779058933,
|
| 412 |
+
"step": 31
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"completion_length": 702.3788185119629,
|
| 416 |
+
"epoch": 0.3582925122463261,
|
| 417 |
+
"grad_norm": 9.096883773803711,
|
| 418 |
+
"learning_rate": 3e-06,
|
| 419 |
+
"loss": 0.0,
|
| 420 |
+
"reward": 0.6829446069896221,
|
| 421 |
+
"reward_std": 0.09760210802778602,
|
| 422 |
+
"rewards/accuracy_reward": 0.40816325694322586,
|
| 423 |
+
"rewards/semantic_entropy_math_reward": 0.682944618165493,
|
| 424 |
+
"rewards/total_entropy_reward": 1.303523451089859,
|
| 425 |
+
"step": 32
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"completion_length": 700.5924644470215,
|
| 429 |
+
"epoch": 0.3694891532540238,
|
| 430 |
+
"grad_norm": 3.891953229904175,
|
| 431 |
+
"learning_rate": 3e-06,
|
| 432 |
+
"loss": 0.0,
|
| 433 |
+
"reward": 0.7055393569171429,
|
| 434 |
+
"reward_std": 0.10629667551256716,
|
| 435 |
+
"rewards/accuracy_reward": 0.45982141979038715,
|
| 436 |
+
"rewards/semantic_entropy_math_reward": 0.7055393569171429,
|
| 437 |
+
"rewards/total_entropy_reward": 1.2391385585069656,
|
| 438 |
+
"step": 33
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"completion_length": 716.7027950286865,
|
| 442 |
+
"epoch": 0.3806857942617215,
|
| 443 |
+
"grad_norm": 12.71116828918457,
|
| 444 |
+
"learning_rate": 3e-06,
|
| 445 |
+
"loss": 0.0,
|
| 446 |
+
"reward": 0.6698250789195299,
|
| 447 |
+
"reward_std": 0.09906701685395092,
|
| 448 |
+
"rewards/accuracy_reward": 0.40369897056370974,
|
| 449 |
+
"rewards/semantic_entropy_math_reward": 0.6698250807821751,
|
| 450 |
+
"rewards/total_entropy_reward": 1.3271235637366772,
|
| 451 |
+
"step": 34
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"completion_length": 769.0414352416992,
|
| 455 |
+
"epoch": 0.3918824352694192,
|
| 456 |
+
"grad_norm": 9.855749130249023,
|
| 457 |
+
"learning_rate": 3e-06,
|
| 458 |
+
"loss": 0.0,
|
| 459 |
+
"reward": 0.6481414064764977,
|
| 460 |
+
"reward_std": 0.10145364608615637,
|
| 461 |
+
"rewards/accuracy_reward": 0.3743622349575162,
|
| 462 |
+
"rewards/semantic_entropy_math_reward": 0.6481414288282394,
|
| 463 |
+
"rewards/total_entropy_reward": 1.3565044924616814,
|
| 464 |
+
"step": 35
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"completion_length": 722.349479675293,
|
| 468 |
+
"epoch": 0.40307907627711687,
|
| 469 |
+
"grad_norm": 5.131447792053223,
|
| 470 |
+
"learning_rate": 3e-06,
|
| 471 |
+
"loss": 0.0,
|
| 472 |
+
"reward": 0.6638119556009769,
|
| 473 |
+
"reward_std": 0.09085186710581183,
|
| 474 |
+
"rewards/accuracy_reward": 0.4170918297022581,
|
| 475 |
+
"rewards/semantic_entropy_math_reward": 0.6638119481503963,
|
| 476 |
+
"rewards/total_entropy_reward": 1.350273534655571,
|
| 477 |
+
"step": 36
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"completion_length": 703.8832778930664,
|
| 481 |
+
"epoch": 0.41427571728481455,
|
| 482 |
+
"grad_norm": 7.057519435882568,
|
| 483 |
+
"learning_rate": 3e-06,
|
| 484 |
+
"loss": 0.0,
|
| 485 |
+
"reward": 0.690597664564848,
|
| 486 |
+
"reward_std": 0.09433076065033674,
|
| 487 |
+
"rewards/accuracy_reward": 0.43686223309487104,
|
| 488 |
+
"rewards/semantic_entropy_math_reward": 0.6905976496636868,
|
| 489 |
+
"rewards/total_entropy_reward": 1.2801040560007095,
|
| 490 |
+
"step": 37
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"completion_length": 725.2799606323242,
|
| 494 |
+
"epoch": 0.42547235829251223,
|
| 495 |
+
"grad_norm": 7.460778713226318,
|
| 496 |
+
"learning_rate": 3e-06,
|
| 497 |
+
"loss": 0.0,
|
| 498 |
+
"reward": 0.6669096164405346,
|
| 499 |
+
"reward_std": 0.10552376857958734,
|
| 500 |
+
"rewards/accuracy_reward": 0.41326529532670975,
|
| 501 |
+
"rewards/semantic_entropy_math_reward": 0.6669096313416958,
|
| 502 |
+
"rewards/total_entropy_reward": 1.3323625773191452,
|
| 503 |
+
"step": 38
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"completion_length": 725.7671966552734,
|
| 507 |
+
"epoch": 0.4366689993002099,
|
| 508 |
+
"grad_norm": 7.124528408050537,
|
| 509 |
+
"learning_rate": 3e-06,
|
| 510 |
+
"loss": 0.0,
|
| 511 |
+
"reward": 0.6472303047776222,
|
| 512 |
+
"reward_std": 0.10429512546397746,
|
| 513 |
+
"rewards/accuracy_reward": 0.3909438718110323,
|
| 514 |
+
"rewards/semantic_entropy_math_reward": 0.6472303159534931,
|
| 515 |
+
"rewards/total_entropy_reward": 1.3546394035220146,
|
| 516 |
+
"step": 39
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"completion_length": 697.7423439025879,
|
| 520 |
+
"epoch": 0.44786564030790765,
|
| 521 |
+
"grad_norm": 7.065954208374023,
|
| 522 |
+
"learning_rate": 3e-06,
|
| 523 |
+
"loss": 0.0,
|
| 524 |
+
"reward": 0.7193877585232258,
|
| 525 |
+
"reward_std": 0.09428266366012394,
|
| 526 |
+
"rewards/accuracy_reward": 0.4336734600365162,
|
| 527 |
+
"rewards/semantic_entropy_math_reward": 0.7193877547979355,
|
| 528 |
+
"rewards/total_entropy_reward": 1.2143822945654392,
|
| 529 |
+
"step": 40
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"completion_length": 703.3673400878906,
|
| 533 |
+
"epoch": 0.45906228131560534,
|
| 534 |
+
"grad_norm": 6.764792442321777,
|
| 535 |
+
"learning_rate": 3e-06,
|
| 536 |
+
"loss": 0.0,
|
| 537 |
+
"reward": 0.706632649526,
|
| 538 |
+
"reward_std": 0.0942279752343893,
|
| 539 |
+
"rewards/accuracy_reward": 0.4649234591051936,
|
| 540 |
+
"rewards/semantic_entropy_math_reward": 0.7066326681524515,
|
| 541 |
+
"rewards/total_entropy_reward": 1.238373503088951,
|
| 542 |
+
"step": 41
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"completion_length": 651.4183578491211,
|
| 546 |
+
"epoch": 0.470258922323303,
|
| 547 |
+
"grad_norm": 16.47541618347168,
|
| 548 |
+
"learning_rate": 3e-06,
|
| 549 |
+
"loss": 0.0,
|
| 550 |
+
"reward": 0.6813046522438526,
|
| 551 |
+
"reward_std": 0.10622056131251156,
|
| 552 |
+
"rewards/accuracy_reward": 0.43239795230329037,
|
| 553 |
+
"rewards/semantic_entropy_math_reward": 0.681304682046175,
|
| 554 |
+
"rewards/total_entropy_reward": 1.284136950969696,
|
| 555 |
+
"step": 42
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"completion_length": 743.2646636962891,
|
| 559 |
+
"epoch": 0.4814555633310007,
|
| 560 |
+
"grad_norm": 15.174323081970215,
|
| 561 |
+
"learning_rate": 3e-06,
|
| 562 |
+
"loss": 0.0,
|
| 563 |
+
"reward": 0.6525145824998617,
|
| 564 |
+
"reward_std": 0.10325121926143765,
|
| 565 |
+
"rewards/accuracy_reward": 0.3947704015299678,
|
| 566 |
+
"rewards/semantic_entropy_math_reward": 0.652514586225152,
|
| 567 |
+
"rewards/total_entropy_reward": 1.358573641628027,
|
| 568 |
+
"step": 43
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"completion_length": 725.5962905883789,
|
| 572 |
+
"epoch": 0.4926522043386984,
|
| 573 |
+
"grad_norm": 7.421966075897217,
|
| 574 |
+
"learning_rate": 3e-06,
|
| 575 |
+
"loss": 0.0,
|
| 576 |
+
"reward": 0.6867711264640093,
|
| 577 |
+
"reward_std": 0.09842351987026632,
|
| 578 |
+
"rewards/accuracy_reward": 0.4323979504406452,
|
| 579 |
+
"rewards/semantic_entropy_math_reward": 0.6867711190134287,
|
| 580 |
+
"rewards/total_entropy_reward": 1.2839920222759247,
|
| 581 |
+
"step": 44
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"completion_length": 698.1390113830566,
|
| 585 |
+
"epoch": 0.5038488453463961,
|
| 586 |
+
"grad_norm": 9.836834907531738,
|
| 587 |
+
"learning_rate": 3e-06,
|
| 588 |
+
"loss": 0.0,
|
| 589 |
+
"reward": 0.6851311810314655,
|
| 590 |
+
"reward_std": 0.10096711455844343,
|
| 591 |
+
"rewards/accuracy_reward": 0.44196427427232265,
|
| 592 |
+
"rewards/semantic_entropy_math_reward": 0.6851312182843685,
|
| 593 |
+
"rewards/total_entropy_reward": 1.290800966322422,
|
| 594 |
+
"step": 45
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"completion_length": 737.8379898071289,
|
| 598 |
+
"epoch": 0.5150454863540938,
|
| 599 |
+
"grad_norm": 13.060182571411133,
|
| 600 |
+
"learning_rate": 3e-06,
|
| 601 |
+
"loss": 0.0,
|
| 602 |
+
"reward": 0.6867711432278156,
|
| 603 |
+
"reward_std": 0.10890168650075793,
|
| 604 |
+
"rewards/accuracy_reward": 0.403061218559742,
|
| 605 |
+
"rewards/semantic_entropy_math_reward": 0.6867711097002029,
|
| 606 |
+
"rewards/total_entropy_reward": 1.2737670093774796,
|
| 607 |
+
"step": 46
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"completion_length": 720.0937309265137,
|
| 611 |
+
"epoch": 0.5262421273617914,
|
| 612 |
+
"grad_norm": 17.424467086791992,
|
| 613 |
+
"learning_rate": 3e-06,
|
| 614 |
+
"loss": 0.0,
|
| 615 |
+
"reward": 0.6268221419304609,
|
| 616 |
+
"reward_std": 0.11281977966427803,
|
| 617 |
+
"rewards/accuracy_reward": 0.39668366219848394,
|
| 618 |
+
"rewards/semantic_entropy_math_reward": 0.6268221419304609,
|
| 619 |
+
"rewards/total_entropy_reward": 1.396425575017929,
|
| 620 |
+
"step": 47
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"completion_length": 730.9495983123779,
|
| 624 |
+
"epoch": 0.5374387683694891,
|
| 625 |
+
"grad_norm": 7.613661766052246,
|
| 626 |
+
"learning_rate": 3e-06,
|
| 627 |
+
"loss": 0.0,
|
| 628 |
+
"reward": 0.6333818975836039,
|
| 629 |
+
"reward_std": 0.09289166564121842,
|
| 630 |
+
"rewards/accuracy_reward": 0.38201529905200005,
|
| 631 |
+
"rewards/semantic_entropy_math_reward": 0.6333819199353456,
|
| 632 |
+
"rewards/total_entropy_reward": 1.3891723416745663,
|
| 633 |
+
"step": 48
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"completion_length": 730.4712867736816,
|
| 637 |
+
"epoch": 0.5486354093771868,
|
| 638 |
+
"grad_norm": 11.16511344909668,
|
| 639 |
+
"learning_rate": 3e-06,
|
| 640 |
+
"loss": 0.0,
|
| 641 |
+
"reward": 0.628097664564848,
|
| 642 |
+
"reward_std": 0.10562112857587636,
|
| 643 |
+
"rewards/accuracy_reward": 0.3871173355728388,
|
| 644 |
+
"rewards/semantic_entropy_math_reward": 0.6280976608395576,
|
| 645 |
+
"rewards/total_entropy_reward": 1.3984056264162064,
|
| 646 |
+
"step": 49
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"completion_length": 700.4126129150391,
|
| 650 |
+
"epoch": 0.5598320503848845,
|
| 651 |
+
"grad_norm": 7.883020401000977,
|
| 652 |
+
"learning_rate": 3e-06,
|
| 653 |
+
"loss": 0.0,
|
| 654 |
+
"reward": 0.6543367393314838,
|
| 655 |
+
"reward_std": 0.08848634106107056,
|
| 656 |
+
"rewards/accuracy_reward": 0.43494897056370974,
|
| 657 |
+
"rewards/semantic_entropy_math_reward": 0.6543367449194193,
|
| 658 |
+
"rewards/total_entropy_reward": 1.368666134774685,
|
| 659 |
+
"step": 50
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"completion_length": 726.035701751709,
|
| 663 |
+
"epoch": 0.5710286913925823,
|
| 664 |
+
"grad_norm": 16.26348114013672,
|
| 665 |
+
"learning_rate": 3e-06,
|
| 666 |
+
"loss": 0.0,
|
| 667 |
+
"reward": 0.6665451787412167,
|
| 668 |
+
"reward_std": 0.09329186473041773,
|
| 669 |
+
"rewards/accuracy_reward": 0.40114795323461294,
|
| 670 |
+
"rewards/semantic_entropy_math_reward": 0.6665451973676682,
|
| 671 |
+
"rewards/total_entropy_reward": 1.3389556668698788,
|
| 672 |
+
"step": 51
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"completion_length": 719.875617980957,
|
| 676 |
+
"epoch": 0.58222533240028,
|
| 677 |
+
"grad_norm": 10.366020202636719,
|
| 678 |
+
"learning_rate": 3e-06,
|
| 679 |
+
"loss": 0.0,
|
| 680 |
+
"reward": 0.6474125292152166,
|
| 681 |
+
"reward_std": 0.10870802192948759,
|
| 682 |
+
"rewards/accuracy_reward": 0.3934948956593871,
|
| 683 |
+
"rewards/semantic_entropy_math_reward": 0.6474125385284424,
|
| 684 |
+
"rewards/total_entropy_reward": 1.3537504002451897,
|
| 685 |
+
"step": 52
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"completion_length": 691.538890838623,
|
| 689 |
+
"epoch": 0.5934219734079776,
|
| 690 |
+
"grad_norm": 10.94919204711914,
|
| 691 |
+
"learning_rate": 3e-06,
|
| 692 |
+
"loss": 0.0,
|
| 693 |
+
"reward": 0.6860422715544701,
|
| 694 |
+
"reward_std": 0.0946744061075151,
|
| 695 |
+
"rewards/accuracy_reward": 0.489158159121871,
|
| 696 |
+
"rewards/semantic_entropy_math_reward": 0.6860422790050507,
|
| 697 |
+
"rewards/total_entropy_reward": 1.3007621616125107,
|
| 698 |
+
"step": 53
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"completion_length": 705.672176361084,
|
| 702 |
+
"epoch": 0.6046186144156753,
|
| 703 |
+
"grad_norm": 23.8171329498291,
|
| 704 |
+
"learning_rate": 3e-06,
|
| 705 |
+
"loss": 0.0,
|
| 706 |
+
"reward": 0.6762026119977236,
|
| 707 |
+
"reward_std": 0.09615424065850675,
|
| 708 |
+
"rewards/accuracy_reward": 0.39668366592377424,
|
| 709 |
+
"rewards/semantic_entropy_math_reward": 0.6762026362121105,
|
| 710 |
+
"rewards/total_entropy_reward": 1.3121707029640675,
|
| 711 |
+
"step": 54
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"completion_length": 677.1664428710938,
|
| 715 |
+
"epoch": 0.615815255423373,
|
| 716 |
+
"grad_norm": 9.031086921691895,
|
| 717 |
+
"learning_rate": 3e-06,
|
| 718 |
+
"loss": 0.0,
|
| 719 |
+
"reward": 0.6896865852177143,
|
| 720 |
+
"reward_std": 0.1015757208224386,
|
| 721 |
+
"rewards/accuracy_reward": 0.4317601965740323,
|
| 722 |
+
"rewards/semantic_entropy_math_reward": 0.6896865479648113,
|
| 723 |
+
"rewards/total_entropy_reward": 1.287308655679226,
|
| 724 |
+
"step": 55
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"completion_length": 713.0114631652832,
|
| 728 |
+
"epoch": 0.6270118964310707,
|
| 729 |
+
"grad_norm": 10.871789932250977,
|
| 730 |
+
"learning_rate": 3e-06,
|
| 731 |
+
"loss": 0.0,
|
| 732 |
+
"reward": 0.6656341031193733,
|
| 733 |
+
"reward_std": 0.11424232507124543,
|
| 734 |
+
"rewards/accuracy_reward": 0.4221938643604517,
|
| 735 |
+
"rewards/semantic_entropy_math_reward": 0.6656341068446636,
|
| 736 |
+
"rewards/total_entropy_reward": 1.3229641020298004,
|
| 737 |
+
"step": 56
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"completion_length": 677.5694923400879,
|
| 741 |
+
"epoch": 0.6382085374387684,
|
| 742 |
+
"grad_norm": 8.904267311096191,
|
| 743 |
+
"learning_rate": 3e-06,
|
| 744 |
+
"loss": 0.0,
|
| 745 |
+
"reward": 0.6729227248579264,
|
| 746 |
+
"reward_std": 0.10780132608488202,
|
| 747 |
+
"rewards/accuracy_reward": 0.4381377436220646,
|
| 748 |
+
"rewards/semantic_entropy_math_reward": 0.6729227565228939,
|
| 749 |
+
"rewards/total_entropy_reward": 1.3122014850378036,
|
| 750 |
+
"step": 57
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"completion_length": 687.1345500946045,
|
| 754 |
+
"epoch": 0.6494051784464661,
|
| 755 |
+
"grad_norm": 19.945981979370117,
|
| 756 |
+
"learning_rate": 3e-06,
|
| 757 |
+
"loss": 0.0,
|
| 758 |
+
"reward": 0.6621720213443041,
|
| 759 |
+
"reward_std": 0.10787450219504535,
|
| 760 |
+
"rewards/accuracy_reward": 0.4183673388324678,
|
| 761 |
+
"rewards/semantic_entropy_math_reward": 0.6621720027178526,
|
| 762 |
+
"rewards/total_entropy_reward": 1.3265771567821503,
|
| 763 |
+
"step": 58
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"completion_length": 673.6224384307861,
|
| 767 |
+
"epoch": 0.6606018194541637,
|
| 768 |
+
"grad_norm": 25.872291564941406,
|
| 769 |
+
"learning_rate": 3e-06,
|
| 770 |
+
"loss": 0.0,
|
| 771 |
+
"reward": 0.6900510117411613,
|
| 772 |
+
"reward_std": 0.09328114939853549,
|
| 773 |
+
"rewards/accuracy_reward": 0.3998724389821291,
|
| 774 |
+
"rewards/semantic_entropy_math_reward": 0.6900510005652905,
|
| 775 |
+
"rewards/total_entropy_reward": 1.2932880148291588,
|
| 776 |
+
"step": 59
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"completion_length": 712.2027912139893,
|
| 780 |
+
"epoch": 0.6717984604618614,
|
| 781 |
+
"grad_norm": 12.222248077392578,
|
| 782 |
+
"learning_rate": 3e-06,
|
| 783 |
+
"loss": 0.0,
|
| 784 |
+
"reward": 0.6426749210804701,
|
| 785 |
+
"reward_std": 0.09614149574190378,
|
| 786 |
+
"rewards/accuracy_reward": 0.41007652156986296,
|
| 787 |
+
"rewards/semantic_entropy_math_reward": 0.6426749173551798,
|
| 788 |
+
"rewards/total_entropy_reward": 1.383428543806076,
|
| 789 |
+
"step": 60
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"completion_length": 739.6472969055176,
|
| 793 |
+
"epoch": 0.6829951014695591,
|
| 794 |
+
"grad_norm": 9.738767623901367,
|
| 795 |
+
"learning_rate": 3e-06,
|
| 796 |
+
"loss": 0.0,
|
| 797 |
+
"reward": 0.6273687966167927,
|
| 798 |
+
"reward_std": 0.11152111599221826,
|
| 799 |
+
"rewards/accuracy_reward": 0.34885203186422586,
|
| 800 |
+
"rewards/semantic_entropy_math_reward": 0.6273688077926636,
|
| 801 |
+
"rewards/total_entropy_reward": 1.3971327617764473,
|
| 802 |
+
"step": 61
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"completion_length": 685.2659378051758,
|
| 806 |
+
"epoch": 0.6941917424772568,
|
| 807 |
+
"grad_norm": 20.52920913696289,
|
| 808 |
+
"learning_rate": 3e-06,
|
| 809 |
+
"loss": 0.0,
|
| 810 |
+
"reward": 0.6596209742128849,
|
| 811 |
+
"reward_std": 0.10549976932816207,
|
| 812 |
+
"rewards/accuracy_reward": 0.43622447922825813,
|
| 813 |
+
"rewards/semantic_entropy_math_reward": 0.6596209909766912,
|
| 814 |
+
"rewards/total_entropy_reward": 1.3266954682767391,
|
| 815 |
+
"step": 62
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"completion_length": 684.1128730773926,
|
| 819 |
+
"epoch": 0.7053883834849545,
|
| 820 |
+
"grad_norm": 13.253375053405762,
|
| 821 |
+
"learning_rate": 3e-06,
|
| 822 |
+
"loss": 0.0,
|
| 823 |
+
"reward": 0.6740160267800093,
|
| 824 |
+
"reward_std": 0.11020976235158741,
|
| 825 |
+
"rewards/accuracy_reward": 0.4419642761349678,
|
| 826 |
+
"rewards/semantic_entropy_math_reward": 0.6740160342305899,
|
| 827 |
+
"rewards/total_entropy_reward": 1.295092262327671,
|
| 828 |
+
"step": 63
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"completion_length": 702.8647766113281,
|
| 832 |
+
"epoch": 0.7165850244926522,
|
| 833 |
+
"grad_norm": 11.436138153076172,
|
| 834 |
+
"learning_rate": 3e-06,
|
| 835 |
+
"loss": 0.0,
|
| 836 |
+
"reward": 0.6612609177827835,
|
| 837 |
+
"reward_std": 0.0923373675905168,
|
| 838 |
+
"rewards/accuracy_reward": 0.43239795230329037,
|
| 839 |
+
"rewards/semantic_entropy_math_reward": 0.661260936409235,
|
| 840 |
+
"rewards/total_entropy_reward": 1.3441792502999306,
|
| 841 |
+
"step": 64
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"completion_length": 713.3341636657715,
|
| 845 |
+
"epoch": 0.72778166550035,
|
| 846 |
+
"grad_norm": 15.771661758422852,
|
| 847 |
+
"learning_rate": 3e-06,
|
| 848 |
+
"loss": 0.0,
|
| 849 |
+
"reward": 0.6313775349408388,
|
| 850 |
+
"reward_std": 0.1141419094055891,
|
| 851 |
+
"rewards/accuracy_reward": 0.4139030510559678,
|
| 852 |
+
"rewards/semantic_entropy_math_reward": 0.6313775237649679,
|
| 853 |
+
"rewards/total_entropy_reward": 1.3911906033754349,
|
| 854 |
+
"step": 65
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"completion_length": 691.0420761108398,
|
| 858 |
+
"epoch": 0.7389783065080476,
|
| 859 |
+
"grad_norm": 9.556193351745605,
|
| 860 |
+
"learning_rate": 3e-06,
|
| 861 |
+
"loss": 0.0,
|
| 862 |
+
"reward": 0.6576165966689587,
|
| 863 |
+
"reward_std": 0.10697318986058235,
|
| 864 |
+
"rewards/accuracy_reward": 0.3386479513719678,
|
| 865 |
+
"rewards/semantic_entropy_math_reward": 0.657616626471281,
|
| 866 |
+
"rewards/total_entropy_reward": 1.3421761691570282,
|
| 867 |
+
"step": 66
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"completion_length": 688.1179618835449,
|
| 871 |
+
"epoch": 0.7501749475157453,
|
| 872 |
+
"grad_norm": 18.490732192993164,
|
| 873 |
+
"learning_rate": 3e-06,
|
| 874 |
+
"loss": 0.0,
|
| 875 |
+
"reward": 0.6712827831506729,
|
| 876 |
+
"reward_std": 0.10525703080929816,
|
| 877 |
+
"rewards/accuracy_reward": 0.445790808647871,
|
| 878 |
+
"rewards/semantic_entropy_math_reward": 0.6712828055024147,
|
| 879 |
+
"rewards/total_entropy_reward": 1.3154872134327888,
|
| 880 |
+
"step": 67
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"completion_length": 647.132007598877,
|
| 884 |
+
"epoch": 0.761371588523443,
|
| 885 |
+
"grad_norm": 7.94115686416626,
|
| 886 |
+
"learning_rate": 3e-06,
|
| 887 |
+
"loss": 0.0,
|
| 888 |
+
"reward": 0.6920553781092167,
|
| 889 |
+
"reward_std": 0.09936971426941454,
|
| 890 |
+
"rewards/accuracy_reward": 0.43176019564270973,
|
| 891 |
+
"rewards/semantic_entropy_math_reward": 0.6920554004609585,
|
| 892 |
+
"rewards/total_entropy_reward": 1.2725396156311035,
|
| 893 |
+
"step": 68
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"completion_length": 699.4005012512207,
|
| 897 |
+
"epoch": 0.7725682295311407,
|
| 898 |
+
"grad_norm": 12.146166801452637,
|
| 899 |
+
"learning_rate": 3e-06,
|
| 900 |
+
"loss": 0.0,
|
| 901 |
+
"reward": 0.6652696616947651,
|
| 902 |
+
"reward_std": 0.11806112038902938,
|
| 903 |
+
"rewards/accuracy_reward": 0.409438768401742,
|
| 904 |
+
"rewards/semantic_entropy_math_reward": 0.6652696691453457,
|
| 905 |
+
"rewards/total_entropy_reward": 1.3024558648467064,
|
| 906 |
+
"step": 69
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"completion_length": 667.9285659790039,
|
| 910 |
+
"epoch": 0.7837648705388384,
|
| 911 |
+
"grad_norm": 15.953444480895996,
|
| 912 |
+
"learning_rate": 3e-06,
|
| 913 |
+
"loss": 0.0,
|
| 914 |
+
"reward": 0.679846927523613,
|
| 915 |
+
"reward_std": 0.09083303948864341,
|
| 916 |
+
"rewards/accuracy_reward": 0.4311224427074194,
|
| 917 |
+
"rewards/semantic_entropy_math_reward": 0.6798469312489033,
|
| 918 |
+
"rewards/total_entropy_reward": 1.3062404170632362,
|
| 919 |
+
"step": 70
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"completion_length": 705.4400367736816,
|
| 923 |
+
"epoch": 0.794961511546536,
|
| 924 |
+
"grad_norm": 12.320276260375977,
|
| 925 |
+
"learning_rate": 3e-06,
|
| 926 |
+
"loss": 0.0,
|
| 927 |
+
"reward": 0.6556122340261936,
|
| 928 |
+
"reward_std": 0.09324299031868577,
|
| 929 |
+
"rewards/accuracy_reward": 0.37946427892893553,
|
| 930 |
+
"rewards/semantic_entropy_math_reward": 0.6556122545152903,
|
| 931 |
+
"rewards/total_entropy_reward": 1.3581575378775597,
|
| 932 |
+
"step": 71
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"completion_length": 654.1511325836182,
|
| 936 |
+
"epoch": 0.8061581525542337,
|
| 937 |
+
"grad_norm": 22.0692138671875,
|
| 938 |
+
"learning_rate": 3e-06,
|
| 939 |
+
"loss": 0.0,
|
| 940 |
+
"reward": 0.6794824972748756,
|
| 941 |
+
"reward_std": 0.10328507493250072,
|
| 942 |
+
"rewards/accuracy_reward": 0.4209183603525162,
|
| 943 |
+
"rewards/semantic_entropy_math_reward": 0.6794825121760368,
|
| 944 |
+
"rewards/total_entropy_reward": 1.2847715727984905,
|
| 945 |
+
"step": 72
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"completion_length": 690.308666229248,
|
| 949 |
+
"epoch": 0.8173547935619314,
|
| 950 |
+
"grad_norm": 10.22040843963623,
|
| 951 |
+
"learning_rate": 3e-06,
|
| 952 |
+
"loss": 0.0,
|
| 953 |
+
"reward": 0.6780247762799263,
|
| 954 |
+
"reward_std": 0.105113809928298,
|
| 955 |
+
"rewards/accuracy_reward": 0.45918366499245167,
|
| 956 |
+
"rewards/semantic_entropy_math_reward": 0.678024772554636,
|
| 957 |
+
"rewards/total_entropy_reward": 1.2880272567272186,
|
| 958 |
+
"step": 73
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"completion_length": 696.0803337097168,
|
| 962 |
+
"epoch": 0.8285514345696291,
|
| 963 |
+
"grad_norm": 20.899675369262695,
|
| 964 |
+
"learning_rate": 3e-06,
|
| 965 |
+
"loss": 0.0,
|
| 966 |
+
"reward": 0.6516034882515669,
|
| 967 |
+
"reward_std": 0.10480827221181244,
|
| 968 |
+
"rewards/accuracy_reward": 0.3852040730416775,
|
| 969 |
+
"rewards/semantic_entropy_math_reward": 0.6516034882515669,
|
| 970 |
+
"rewards/total_entropy_reward": 1.3282338082790375,
|
| 971 |
+
"step": 74
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"completion_length": 649.8654270172119,
|
| 975 |
+
"epoch": 0.8397480755773268,
|
| 976 |
+
"grad_norm": 9.469245910644531,
|
| 977 |
+
"learning_rate": 3e-06,
|
| 978 |
+
"loss": 0.0,
|
| 979 |
+
"reward": 0.7246719934046268,
|
| 980 |
+
"reward_std": 0.08447014342527837,
|
| 981 |
+
"rewards/accuracy_reward": 0.433035708963871,
|
| 982 |
+
"rewards/semantic_entropy_math_reward": 0.7246720045804977,
|
| 983 |
+
"rewards/total_entropy_reward": 1.220706269145012,
|
| 984 |
+
"step": 75
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"completion_length": 746.2142791748047,
|
| 988 |
+
"epoch": 0.8509447165850245,
|
| 989 |
+
"grad_norm": 10.530217170715332,
|
| 990 |
+
"learning_rate": 3e-06,
|
| 991 |
+
"loss": 0.0,
|
| 992 |
+
"reward": 0.639395035803318,
|
| 993 |
+
"reward_std": 0.10488821647595614,
|
| 994 |
+
"rewards/accuracy_reward": 0.40816325321793556,
|
| 995 |
+
"rewards/semantic_entropy_math_reward": 0.6393950544297695,
|
| 996 |
+
"rewards/total_entropy_reward": 1.3760143592953682,
|
| 997 |
+
"step": 76
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"completion_length": 655.6836585998535,
|
| 1001 |
+
"epoch": 0.8621413575927221,
|
| 1002 |
+
"grad_norm": 14.1907958984375,
|
| 1003 |
+
"learning_rate": 3e-06,
|
| 1004 |
+
"loss": 0.0,
|
| 1005 |
+
"reward": 0.7228498421609402,
|
| 1006 |
+
"reward_std": 0.09376341942697763,
|
| 1007 |
+
"rewards/accuracy_reward": 0.4674744727090001,
|
| 1008 |
+
"rewards/semantic_entropy_math_reward": 0.7228498421609402,
|
| 1009 |
+
"rewards/total_entropy_reward": 1.2132117934525013,
|
| 1010 |
+
"step": 77
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"completion_length": 672.8137626647949,
|
| 1014 |
+
"epoch": 0.8733379986004198,
|
| 1015 |
+
"grad_norm": 13.5209379196167,
|
| 1016 |
+
"learning_rate": 3e-06,
|
| 1017 |
+
"loss": 0.0,
|
| 1018 |
+
"reward": 0.652332354336977,
|
| 1019 |
+
"reward_std": 0.09460025187581778,
|
| 1020 |
+
"rewards/accuracy_reward": 0.3858418334275484,
|
| 1021 |
+
"rewards/semantic_entropy_math_reward": 0.6523323617875576,
|
| 1022 |
+
"rewards/total_entropy_reward": 1.351367250084877,
|
| 1023 |
+
"step": 78
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"completion_length": 696.9342880249023,
|
| 1027 |
+
"epoch": 0.8845346396081175,
|
| 1028 |
+
"grad_norm": 17.302839279174805,
|
| 1029 |
+
"learning_rate": 3e-06,
|
| 1030 |
+
"loss": 0.0,
|
| 1031 |
+
"reward": 0.6514212656766176,
|
| 1032 |
+
"reward_std": 0.09626807877793908,
|
| 1033 |
+
"rewards/accuracy_reward": 0.4062499897554517,
|
| 1034 |
+
"rewards/semantic_entropy_math_reward": 0.651421282440424,
|
| 1035 |
+
"rewards/total_entropy_reward": 1.366665042936802,
|
| 1036 |
+
"step": 79
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"completion_length": 682.3073806762695,
|
| 1040 |
+
"epoch": 0.8957312806158153,
|
| 1041 |
+
"grad_norm": 9.640515327453613,
|
| 1042 |
+
"learning_rate": 3e-06,
|
| 1043 |
+
"loss": 0.0,
|
| 1044 |
+
"reward": 0.6767492536455393,
|
| 1045 |
+
"reward_std": 0.10075846454128623,
|
| 1046 |
+
"rewards/accuracy_reward": 0.42028060276061296,
|
| 1047 |
+
"rewards/semantic_entropy_math_reward": 0.6767492648214102,
|
| 1048 |
+
"rewards/total_entropy_reward": 1.3104709684848785,
|
| 1049 |
+
"step": 80
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"completion_length": 722.4725646972656,
|
| 1053 |
+
"epoch": 0.906927921623513,
|
| 1054 |
+
"grad_norm": 6.948408603668213,
|
| 1055 |
+
"learning_rate": 3e-06,
|
| 1056 |
+
"loss": 0.0,
|
| 1057 |
+
"reward": 0.6521501280367374,
|
| 1058 |
+
"reward_std": 0.08798706252127886,
|
| 1059 |
+
"rewards/accuracy_reward": 0.4534438718110323,
|
| 1060 |
+
"rewards/semantic_entropy_math_reward": 0.6521501541137695,
|
| 1061 |
+
"rewards/total_entropy_reward": 1.3641655445098877,
|
| 1062 |
+
"step": 81
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"completion_length": 686.7238388061523,
|
| 1066 |
+
"epoch": 0.9181245626312107,
|
| 1067 |
+
"grad_norm": 47.02455139160156,
|
| 1068 |
+
"learning_rate": 3e-06,
|
| 1069 |
+
"loss": 0.0,
|
| 1070 |
+
"reward": 0.6811224482953548,
|
| 1071 |
+
"reward_std": 0.11016466980800033,
|
| 1072 |
+
"rewards/accuracy_reward": 0.42283162754029036,
|
| 1073 |
+
"rewards/semantic_entropy_math_reward": 0.6811224408447742,
|
| 1074 |
+
"rewards/total_entropy_reward": 1.2916913330554962,
|
| 1075 |
+
"step": 82
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"completion_length": 709.4598007202148,
|
| 1079 |
+
"epoch": 0.9293212036389084,
|
| 1080 |
+
"grad_norm": 31.659669876098633,
|
| 1081 |
+
"learning_rate": 3e-06,
|
| 1082 |
+
"loss": 0.0,
|
| 1083 |
+
"reward": 0.6086005661636591,
|
| 1084 |
+
"reward_std": 0.10625140275806189,
|
| 1085 |
+
"rewards/accuracy_reward": 0.37882652692496777,
|
| 1086 |
+
"rewards/semantic_entropy_math_reward": 0.60860057733953,
|
| 1087 |
+
"rewards/total_entropy_reward": 1.4536337479948997,
|
| 1088 |
+
"step": 83
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"completion_length": 661.408784866333,
|
| 1092 |
+
"epoch": 0.940517844646606,
|
| 1093 |
+
"grad_norm": 8.08792781829834,
|
| 1094 |
+
"learning_rate": 3e-06,
|
| 1095 |
+
"loss": 0.0,
|
| 1096 |
+
"reward": 0.6858600489795208,
|
| 1097 |
+
"reward_std": 0.09285477036610246,
|
| 1098 |
+
"rewards/accuracy_reward": 0.4221938652917743,
|
| 1099 |
+
"rewards/semantic_entropy_math_reward": 0.6858600452542305,
|
| 1100 |
+
"rewards/total_entropy_reward": 1.283115666359663,
|
| 1101 |
+
"step": 84
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"completion_length": 657.9400405883789,
|
| 1105 |
+
"epoch": 0.9517144856543037,
|
| 1106 |
+
"grad_norm": 21.801326751708984,
|
| 1107 |
+
"learning_rate": 3e-06,
|
| 1108 |
+
"loss": 0.0,
|
| 1109 |
+
"reward": 0.6805757954716682,
|
| 1110 |
+
"reward_std": 0.09728616522625089,
|
| 1111 |
+
"rewards/accuracy_reward": 0.43749999441206455,
|
| 1112 |
+
"rewards/semantic_entropy_math_reward": 0.6805758327245712,
|
| 1113 |
+
"rewards/total_entropy_reward": 1.3078671097755432,
|
| 1114 |
+
"step": 85
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"completion_length": 674.7142734527588,
|
| 1118 |
+
"epoch": 0.9629111266620014,
|
| 1119 |
+
"grad_norm": 11.559253692626953,
|
| 1120 |
+
"learning_rate": 3e-06,
|
| 1121 |
+
"loss": 0.0,
|
| 1122 |
+
"reward": 0.6862244829535484,
|
| 1123 |
+
"reward_std": 0.10050268471240997,
|
| 1124 |
+
"rewards/accuracy_reward": 0.4304846851155162,
|
| 1125 |
+
"rewards/semantic_entropy_math_reward": 0.686224490404129,
|
| 1126 |
+
"rewards/total_entropy_reward": 1.2905268669128418,
|
| 1127 |
+
"step": 86
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"completion_length": 636.0388870239258,
|
| 1131 |
+
"epoch": 0.9741077676696991,
|
| 1132 |
+
"grad_norm": 11.289789199829102,
|
| 1133 |
+
"learning_rate": 3e-06,
|
| 1134 |
+
"loss": 0.0,
|
| 1135 |
+
"reward": 0.6986151486635208,
|
| 1136 |
+
"reward_std": 0.10155197884887457,
|
| 1137 |
+
"rewards/accuracy_reward": 0.43686223216354847,
|
| 1138 |
+
"rewards/semantic_entropy_math_reward": 0.6986151374876499,
|
| 1139 |
+
"rewards/total_entropy_reward": 1.2592380531132221,
|
| 1140 |
+
"step": 87
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"completion_length": 683.5822525024414,
|
| 1144 |
+
"epoch": 0.9853044086773968,
|
| 1145 |
+
"grad_norm": 10.526566505432129,
|
| 1146 |
+
"learning_rate": 3e-06,
|
| 1147 |
+
"loss": 0.0,
|
| 1148 |
+
"reward": 0.6585276797413826,
|
| 1149 |
+
"reward_std": 0.10897616110742092,
|
| 1150 |
+
"rewards/accuracy_reward": 0.4017857098951936,
|
| 1151 |
+
"rewards/semantic_entropy_math_reward": 0.6585277020931244,
|
| 1152 |
+
"rewards/total_entropy_reward": 1.324688896536827,
|
| 1153 |
+
"step": 88
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"completion_length": 683.3411960601807,
|
| 1157 |
+
"epoch": 0.9965010496850945,
|
| 1158 |
+
"grad_norm": 19.354259490966797,
|
| 1159 |
+
"learning_rate": 3e-06,
|
| 1160 |
+
"loss": 0.0,
|
| 1161 |
+
"reward": 0.6687317825853825,
|
| 1162 |
+
"reward_std": 0.09195416304282844,
|
| 1163 |
+
"rewards/accuracy_reward": 0.380102027207613,
|
| 1164 |
+
"rewards/semantic_entropy_math_reward": 0.6687317807227373,
|
| 1165 |
+
"rewards/total_entropy_reward": 1.3135743215680122,
|
| 1166 |
+
"step": 89
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.9965010496850945,
|
| 1170 |
+
"step": 89,
|
| 1171 |
+
"total_flos": 0.0,
|
| 1172 |
+
"train_loss": 2.8008765204773544e-08,
|
| 1173 |
+
"train_runtime": 57265.8095,
|
| 1174 |
+
"train_samples_per_second": 0.349,
|
| 1175 |
+
"train_steps_per_second": 0.002
|
| 1176 |
+
}
|
| 1177 |
+
],
|
| 1178 |
+
"logging_steps": 1,
|
| 1179 |
+
"max_steps": 89,
|
| 1180 |
+
"num_input_tokens_seen": 0,
|
| 1181 |
+
"num_train_epochs": 1,
|
| 1182 |
+
"save_steps": 20,
|
| 1183 |
+
"stateful_callbacks": {
|
| 1184 |
+
"TrainerControl": {
|
| 1185 |
+
"args": {
|
| 1186 |
+
"should_epoch_stop": false,
|
| 1187 |
+
"should_evaluate": false,
|
| 1188 |
+
"should_log": false,
|
| 1189 |
+
"should_save": true,
|
| 1190 |
+
"should_training_stop": true
|
| 1191 |
+
},
|
| 1192 |
+
"attributes": {}
|
| 1193 |
+
}
|
| 1194 |
+
},
|
| 1195 |
+
"total_flos": 0.0,
|
| 1196 |
+
"train_batch_size": 2,
|
| 1197 |
+
"trial_name": null,
|
| 1198 |
+
"trial_params": null
|
| 1199 |
+
}
|