Wei-Peng

University of Chinese Academy of Sciences Beijing Ph.D, NLP and CyberSecurity in AI

Results 5 comments of


                                            Wei-Peng

Pytorch版本的网盘失效了

+1

[Chatllama] Add multiple sources for generating synthetic data

hi did you add support for HF models in dataset generation? It seems only OpenAI’s davinci-003 in line 21 in generate_rewards.py.

[Chatllama] it seems you don't support flan_t5_xl to generate rewards training data

I've faced the same issue, have you fixed it?

有没有在tensorflow上训练出很好的效果？

我的效果也比论文里的低不少，不知道为啥

tensorflow版本的baseline结果性能指标是多少？

我实现的效果很差，b和rouge只有25和30，在验证集