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.

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

LLM: add esimd sdp for pvc #10543

Merged
merged 4 commits into from
Mar 26, 2024
Merged

LLM: add esimd sdp for pvc #10543

merged 4 commits into from
Mar 26, 2024

Conversation

rnwang04
Copy link
Contributor

Description

1. Why the change?

https://github.com/analytics-zoo/nano/issues/1176#issuecomment-2005541797

2. User API changes

No change.

3. Summary of the change

  • add esimd sdp for pvc , only update llama now

4. How to test?

  • Unit test

@rnwang04 rnwang04 mentioned this pull request Mar 26, 2024
1 task
Copy link
Contributor

@lalalapotter lalalapotter left a comment

Choose a reason for hiding this comment

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

LGTM.

@rnwang04 rnwang04 merged commit ea4bc45 into intel-analytics:main Mar 26, 2024
13 checks passed
@rnwang04 rnwang04 deleted the esimd_sdp branch March 26, 2024 11:04
elif query_states.shape[0] > 1 and attention_mask is not None:
# for batched input, can't accept attention_mask
# TODO: this check needs some time
if not torch.all(attention_mask.eq(0)):
Copy link
Contributor

Choose a reason for hiding this comment

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

Copy link
Contributor Author

Choose a reason for hiding this comment

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

attention_mask.any() is slightly faster than not torch.all(attention_mask.eq(0)).
Have raised a PR to update it: #10549

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.

3 participants