ChatGLM-6B icon indicating copy to clipboard operation
ChatGLM-6B copied to clipboard

[BUG/Help]如何对chatglm-6b进行RLHF,有相关的代码实现吗?

Open derrickcyt opened this issue 1 year ago • 5 comments

Is there an existing issue for this?

  • [X] I have searched the existing issues

Current Behavior

没找到相关的实现代码

Expected Behavior

No response

Steps To Reproduce

None.

Environment

OS: Ubuntu 20.04
Python: 3.8
Transformers: 4.26.1
PyTorch: 1.12
CUDA Support: True

Anything else?

No response

derrickcyt avatar Apr 12 '23 03:04 derrickcyt

同问 可以出一个教程吗

OceannTwT avatar Apr 12 '23 03:04 OceannTwT

同问

dragononly avatar Apr 12 '23 14:04 dragononly

参考trl和trlx这两个项目,GPT2+PPO 、GPT2+ILQL

valkryhx avatar Apr 13 '23 17:04 valkryhx

微软的deepspeed-chat,改改代码应该可以支持吧

white-wolf-tech avatar Apr 14 '23 01:04 white-wolf-tech

同问

netwolf712 avatar Apr 18 '23 08:04 netwolf712

参考trl和trlx这两个项目,GPT2+PPO 、GPT2+ILQL

你试过可行吗

dayL-W avatar May 09 '23 08:05 dayL-W

Duplicate of #3

zhangch9 avatar Aug 16 '23 07:08 zhangch9