-
Notifications
You must be signed in to change notification settings - Fork 31.7k
Fix no_split_modules for Llama4 pretrained models #37673
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 no_split_modules for Llama4 pretrained models #37673
Conversation
|
Hi 👋, thank you for opening this pull request! The pull request is converted to draft by default. The CI will be paused while the PR is in draft mode. When it is ready for review, please click the |
771dd68 to
e5976aa
Compare
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.
Thanks! Can you merge main or push an empty commit to trigger CI?
bb8718f to
14483b5
Compare
|
@zucchini-nlp thanks, I've just rebased and re-pushed. |
|
The CI error seems related to #37675. |
|
Hey @astefanutti, should be fixed now ! you can just update the branch |
14483b5 to
0f7432a
Compare
|
@MekkCyber thanks for the update, I've just rebased on it. |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
What does this PR do?
This PR moves the definition of
_no_split_modulesfor Llama4 pre-trained models from the baseLlama4PreTrainedModelclass to the subclasses so each model has the correct set of modules.Otherwise loading the
Llama4ForCausalLMcurrently fails because it doesn't have theLlama4VisionEncoderLayermodule.Fixes #37672
Who can review?
@ArthurZucker @amyeroberts @qubvel