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.

Sign up for GitHub

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

Jump to bottom

Are fuseQKV masked attention and Flash Attention the same? #786

Open
likejazz opened this issue Feb 4, 2024 · 0 comments
Open

Are fuseQKV masked attention and Flash Attention the same? #786

likejazz opened this issue Feb 4, 2024 · 0 comments

Comments

@likejazz
Copy link

likejazz commented Feb 4, 2024 •
edited
Loading

In the GPT guide(https://github.com/NVIDIA/FasterTransformer/blob/main/docs/gpt_guide.md#workflow),
Fig 2 shows fuseQKV masked attention, which looks very similar to Flash Attention. However, there's no longer any mention of fuseQKV masked attention or 'Flash Attention' in the text, so I'm wondering if it's the same technology as Flash Attention.

Am I understanding it correctly?
@likejazz likejazz changed the title fuseQKV masked attention same with Flash Attention? Are fuseQKV masked attention and Flash Attention the same? Feb 4, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@likejazz