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

[release] wheel should be published with local version available in wheel METADATA #730

Open
yundai424 opened this issue Dec 18, 2023 · 0 comments · May be fixed by #856
Open

[release] wheel should be published with local version available in wheel METADATA #730

yundai424 opened this issue Dec 18, 2023 · 0 comments · May be fixed by #856

Comments

@yundai424
Copy link

yundai424 commented Dec 18, 2023

Hi team,

This is an issue following up this PR: #229

Right now the released wheels, though with local versions available in the wheel's filename, doesn't encode the proper local version in the METADATA:

wget https://github.com/Dao-AILab/flash-attention/releases/download/v2.3.6/flash_attn-2.3.6+cu118torch2.0cxx11abiFALSE-cp38-cp38-linux_x86_64.whl && unzip flash_attn-2.3.6+cu118torch2.0cxx11abiFALSE-cp38-cp38-linux_x86_64.whl && cat flash_attn-2.3.6.dist-info/METADATA | head -n 10

This will give

Metadata-Version: 2.1
Name: flash-attn
Version: 2.3.6
Summary: Flash Attention: Fast and Memory-Efficient Exact Attention
Home-page: https://github.com/Dao-AILab/flash-attention
Author: Tri Dao
Author-email: [email protected]
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: BSD License
Classifier: Operating System :: Unix

it would be useful to encode the local version in wheel metadata as well and I believe the only thing needed is to add FLASH_ATTN_LOCAL_VERSION env var into GH action flow.

Would be happy to raise a tentative PR if this is something you agree upon. Thanks!

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 a pull request may close this issue.

1 participant