Yicheng Zou issues

Repositories
Issues
Comments

Results 3 issues of


                                            Yicheng Zou

Can I use this toolkit for other languages?

The toolkit is really amazing and convenient. I wonder how it can work for non-English languages.

添加集录祈愿

4.5版本更新新增了集录祈愿

RL Reward Instability and Reproducibility Issues Tied to vLLM Version

Hello OpenRLHF Team, We have encountered a significant issue with reward instability and a lack of reproducibility during RL sampling (PPO) when using **vLLM v0.8.3**. Our experiments show that **vLLM...