Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

SimPO如何设置beta参数 #4414

Closed
1 task done
zhangfan-algo opened this issue Jun 21, 2024 · 6 comments
Closed
1 task done

SimPO如何设置beta参数 #4414

zhangfan-algo opened this issue Jun 21, 2024 · 6 comments
Labels
solved This problem has been already solved

Comments

@zhangfan-algo
Copy link

Reminder

  • I have read the README and searched the existing issues.

System Info

目前看到命令行参数里面只能设置gamma

Reproduction

llamafactory-cli train qwen_full_simpo.yaml

Expected behavior

No response

Others

No response
@github-actions github-actions bot added the pending This problem is yet to be addressed label Jun 21, 2024
@zhangfan-algo
Copy link
Author

主要是想解决在进行simpo训练的过程中'rewards/chosen','rewards/rejected' 这俩都为负 而且都是在变小

@exceedzhang
Copy link

同样的问题SimPO论文提到2个超参数：如何进行设置？对应参数是哪2个？Beta 参数和sim_gamma吗？
image

@hiyouga
Copy link
Owner

hiyouga commented Jun 23, 2024

pref_beta

@hiyouga hiyouga added solved This problem has been already solved and removed pending This problem is yet to be addressed labels Jun 23, 2024
@hiyouga hiyouga closed this as completed Jun 23, 2024
@exceedzhang
Copy link

@hiyouga pref_beta 值设置可以超过1吗？ SimPO论文使用2 和2.5

@hiyouga
Copy link
Owner

hiyouga commented Jun 23, 2024

可以

@exceedzhang
Copy link

可以

感谢！

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
solved This problem has been already solved
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@exceedzhang @hiyouga @zhangfan-algo