-
Notifications
You must be signed in to change notification settings - Fork 1.4k
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
Fix unwrapping peft models #948
Fix unwrapping peft models #948
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Makes sense, thanks!
Hi @kkteru |
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. |
Interesting, somehow the test_sft_trainer.py tests failed on latest main branch (without any changes from this branch) in my local. Investigating further. |
@kkteru what are the failed tests? might be unrelated to your changes I think - you can just merge your branch with upstream main and see here if the tests pass |
b9c7662
to
7c12444
Compare
These ones fail both on the main branch and this branch (merged with main) on my system. I just rebased on top of latest main and pushed. I dont see the tests running here?
|
The tests were failing because you had an old version of transformers, installing the 4.35.0 will fix your isses! As you can see, CI is green now :D |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Awesome clean up !
* First unwrap the model and then process the input embeddings * Changed base_model to base_model.model to stay consistent with peft model abstractions
Fixes issues discussed here.
cc @younesbelkada