44 lines
1.6 KiB
Markdown
44 lines
1.6 KiB
Markdown
---
|
|
license: apache-2.0
|
|
base_model:
|
|
- Qwen/Qwen3-1.7B
|
|
library_name: transformers
|
|
pipeline_tag: text-generation
|
|
tags:
|
|
- mathematical-reasoning
|
|
- reinforcement-learning
|
|
- grpo
|
|
---
|
|
|
|
# Model Card for PrAg-PO Qwen3-1.7B step720
|
|
|
|
**If you are using the model, a star to our [github repo](https://github.com/wenquanlu/PrAg-PO) would be really appreciated! 😊**
|
|
|
|
This is the step 720 checkpoint when training Qwen3-1.7B on MATH Level-3-to-5 Dataset using PrAg-PO. The training procedure is outlined in the paper [PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning](https://arxiv.org/abs/2602.03190).
|
|
|
|
### Model Sources
|
|
|
|
- **Repository 🤖:** [https://github.com/wenquanlu/PrAg-PO](https://github.com/wenquanlu/PrAg-PO)
|
|
- **Paper 📝:** [PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning](https://arxiv.org/abs/2602.03190)
|
|
|
|
## Uses
|
|
|
|
This model is intended for mathematical reasoning tasks. It leverages prompt augmentation to generate reasoning traces under diverse templates, increasing rollout diversity and stability during RL training.
|
|
|
|
### Results
|
|
|
|
<p align="center">
|
|
<img src="https://raw.githubusercontent.com/wenquanlu/prompt-augmentation-GRPO/master/imgs/table_result.png">
|
|
</p>
|
|
|
|
|
|
## Citation
|
|
```bibtex
|
|
@misc{lu2026pragpopromptaugmentedpolicy,
|
|
title={PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning},
|
|
author={Wenquan Lu and Hai Huang and Enqi Liu and Randall Balestriero},
|
|
journal={arXiv preprint arXiv:2602.03190},
|
|
url={https://arxiv.org/abs/2602.03190},
|
|
year={2026},
|
|
}
|
|
``` |