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.

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

Can you support fast resume with streaming option? #6352

Open
1 task done
JonghwanMun opened this issue Dec 16, 2024 · 0 comments
Open
1 task done

Can you support fast resume with streaming option? #6352

JonghwanMun opened this issue Dec 16, 2024 · 0 comments
Labels
pending This problem is yet to be addressed

Comments

@JonghwanMun
Copy link

JonghwanMun commented Dec 16, 2024

Reminder

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

System Info

No response

Reproduction

The problem is too slow resume when using data with streaming option.

  • This is mainly because streaming option makes the data as iterable dataset.
  • Then, resume with iterable dataset requires the data skip process (until the saved last iterations) that takes a lot of times.
  • Actually, this is an known issue and there are some efforts to address this issue:

Can you add this feature into LLaMA-Factory?

Expected behavior

Add a feature of supporting fast resume with streaming data.

Others

No response

@github-actions github-actions bot added the pending This problem is yet to be addressed label Dec 16, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
pending This problem is yet to be addressed
Projects
None yet
Development

No branches or pull requests

1 participant