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

Return old versions to conda #646

Closed
TsykunovDmitriy opened this issue Jan 16, 2023 · 2 comments
Closed

Return old versions to conda #646

TsykunovDmitriy opened this issue Jan 16, 2023 · 2 comments

Comments

@TsykunovDmitriy
Copy link

TsykunovDmitriy commented Jan 16, 2023

Hi! Thanks for your work. Your package is incredibly useful for research tasks.

I noticed that some builds started disappearing from conda. I worked with this xformers=0.0.15.dev343+git.1b1fd8a=py39_cu11.6_pyt1.13 version and everything was fine. After updating to dev395 I get errors like this. Could you bring back previous builds?

Maybe you have the possibility to share the tar.gz archive that we could continue to work on?

@danthe3rd
Copy link
Contributor

danthe3rd commented Jan 16, 2023

Hi,
Thanks for opening this issue. Unfortunately, each xFormers build is quite big (~100MB) and we build for multiple combinations of pytorch version/python/CUDA etc... which adds up pretty fast.
Pypi and conda both limit the storage we can use to ~10GB, so we have to remove older builds to be able to upload new ones...
We plan to leave some builds for longer (for stable releases for instance).

You can re-install the version you mention (or any), but you will have to build manually from source. Follow instructions from there, and replace main with the commit hash you want (1b1fd8a for you I think). In you case, it should be something like

# (Optional) Makes the build much faster
pip install ninja
# Set TORCH_CUDA_ARCH_LIST if running and building on different GPU types
pip install -v -U git+https://github.com/facebookresearch/xformers.git@1b1fd8a#egg=xformers
# (this can take dozens of minutes)

@TsykunovDmitriy
Copy link
Author

Thanks for the answer. I solved my problem by installing this version of xformers and update pytorch version to 1.13.1+cu117.

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

No branches or pull requests

2 participants