search

jingyaogong / minimind

「大模型」3小时完全从0训练26M的小参数GPT，个人显卡即可推理训练！
https://jingyaogong.github.io/minimind
Apache License 2.0
2.7k stars 329 forks source link

5-dpo_train运行报错，是版本问题吗？ #85

Closed srconly closed 1 week ago

srconly commented 1 week ago

D:\Anaconda3\envs\minni\python.exe D:/work/qyQwen/minimind-master/5-dpo_train.py Traceback (most recent call last): File "D:\work\qyQwen\minimind-master\5-dpo_train.py", line 49, in dpo_trainer = DPOTrainer( File "D:\Anaconda3\envs\minni\lib\site-packages**transformers\utils\deprecation.py*", line 165, in wrapped_func return func(args, kwargs) TypeError: init() got an unexpected keyword argument 'beta'**

Process finished with exit code 1 transformer=4.46.0和4.44.0都试过，trl=0.11.3

jingyaogong commented 1 week ago

版本问题，你需要确认 所谓"transformer=4.46.0和4.44.0都试过，trl=0.11.3" 是否来自 minni 环境？

召集多人测试了所有不同平台的机器，不可复现，暂时close

srconly commented 1 week ago

感谢感谢，我的trl确实不来自minni环境，我本地clone了最新的trl项目，它和mind同级目录，造成了这个问题，已解决