Skip to content

Conversation

@snowclipsed
Copy link
Contributor

No description provided.

@snowclipsed snowclipsed reopened this May 19, 2025
self._decode_one_tok = torch.compile(
self._decode_one_tok, fullgraph=True, mode="reduce-overhead"
)
self._vis_enc = torch.compile(self._vis_enc, fullgraph=False, mode="reduce-overhead")
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@EthanReid what do you think about this? The prefill/decode won't compile for me (even at fp16)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It works for me. Since compile is optional, it should be kept.

@EthanReid
Copy link
Contributor

Screenshot 2025-05-19 at 10 41 57 PM Please clean up print statements

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