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

resuming training from LoRA checkpoint fails when base model is quantised #659

Closed
bghira opened this issue Aug 7, 2024 · 3 comments
Closed

resuming training from LoRA checkpoint fails when base model is quantised #659

bghira opened this issue Aug 7, 2024 · 3 comments
Assignees
@bghira
Labels
help wanted Extra attention is needed regression This bug has regressed behaviour that previously worked. upstream-bug We can't do anything but wait.

Comments

@bghira
Copy link
Owner

bghira commented Aug 7, 2024

workaround is to continue training it without the base being quantised but obviously that's difficult-to-impossible.

the bug is seemingly upstream in PEFT.
@bghira bghira added bug Something isn't working help wanted Extra attention is needed regression This bug has regressed behaviour that previously worked. upstream-bug We can't do anything but wait. labels Aug 7, 2024
@bghira bghira self-assigned this Aug 7, 2024
@bghira
Copy link
Owner Author

bghira commented Aug 7, 2024

huggingface/diffusers#9108

@bghira bghira removed the bug Something isn't working label Aug 8, 2024
This was referenced Aug 13, 2024
Closed
Closed
@bghira
Copy link
Owner Author

bghira commented Aug 18, 2024

for anyone hitting this, see --init_lora as a workaround

@bghira
Copy link
Owner Author

bghira commented Aug 21, 2024

fixed in main & release branches

@bghira bghira closed this as completed Aug 21, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@bghira bghira

Labels
help wanted Extra attention is needed regression This bug has regressed behaviour that previously worked. upstream-bug We can't do anything but wait.
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@bghira