Skip to content

Torch compile first call of prepare_cos_sin method #1395

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

Open
wants to merge 13 commits into
base: habana_main
Choose a base branch
from

Conversation

Kacper-Pietkun
Copy link

Currently prepare_cos_sin method calls inside of LlamaDecoderLayer are torch.compiled. However, first call to that method inside of HpuModelAdapter's forward runs in eager mode. This PR changes that and torch.compiles that call and fixes recompilations that were happening because of that compilation.

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

@Kacper-Pietkun
Copy link
Author

/run-gaudi-tests

1 similar comment
@anko-intel
Copy link

/run-gaudi-tests

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants