Skip to content

feat: Add support for flash attention converter #2560

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
Merged
merged 1 commit into from
Jan 9, 2024
Merged

feat: Add support for flash attention converter #2560

merged 1 commit into from
Jan 9, 2024

Conversation

gs-olive
Copy link
Collaborator

Description

  • Add new subgraph-matching variants to align with flash attention paradigm in SD + SDXL models
  • Add support for scale kwarg specification in both attention variants
  • Add testing for flash attention ATen operator

Fixes #2427

Type of change

  • New converter

Checklist:

  • [ x ] My code follows the style guidelines of this project (You can use the linters)
  • [ x ] I have performed a self-review of my own code
  • [ x ] I have commented my code, particularly in hard-to-understand areas and hacks
  • [ x ] I have made corresponding changes to the documentation
  • [ x ] I have added tests to verify my fix or my feature
  • [ x ] New and existing unit tests pass locally with my changes
  • [ x ] I have added the relevant labels to my PR in so that relevant reviewers are notified

@gs-olive
feat: Add support for flash attention converter
e819202 
- Add new subgraph-matching variants to align with flash attention
paradigm in SD + SDXL models
- Add support for `scale` kwarg specification in both attention variants
- Add testing for flash attention ATen operator
@gs-olive gs-olive requested review from zewenli98 and apbose December 27, 2023 18:35
@gs-olive gs-olive self-assigned this Dec 27, 2023
@github-actions github-actions bot added component: api [Python] Issues re: Python API component: conversion Issues re: Conversion stage component: converters Issues re: Specific op converters component: dynamo Issues relating to the `torch.compile` or `torch._dynamo.export` paths component: lowering Issues re: The lowering / preprocessing passes component: tests Issues re: Tests labels Dec 27, 2023
@github-actions github-actions bot requested a review from narendasan December 27, 2023 18:36
zewenli98
zewenli98 approved these changes Dec 28, 2023
View reviewed changes
Copy link
Collaborator

@zewenli98 zewenli98 left a comment

Choose a reason for hiding this comment

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

Looks good to me!

@gs-olive gs-olive merged commit de49d62 into pytorch:main Jan 9, 2024
@gs-olive gs-olive deleted the scaled_dot_product_attention_converter branch January 9, 2024 23:22
gs-olive added a commit that referenced this pull request Feb 5, 2024
@gs-olive
feat: Add support for flash attention converter (#2560)
e7fe504
@gs-olive gs-olive mentioned this pull request Feb 5, 2024
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@zewenli98 zewenli98 zewenli98 approved these changes

@apbose apbose Awaiting requested review from apbose

@narendasan narendasan Awaiting requested review from narendasan

Assignees

@gs-olive gs-olive

Labels
cla signed component: api [Python] Issues re: Python API component: conversion Issues re: Conversion stage component: converters Issues re: Specific op converters component: dynamo Issues relating to the `torch.compile` or `torch._dynamo.export` paths component: lowering Issues re: The lowering / preprocessing passes component: tests Issues re: Tests
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Support aten._scaled_dot_product_flash_attention
3 participants
@gs-olive @zewenli98 @facebook-github-bot