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

Tensorflow for mac arm64 #208

Open
7 tasks
jwokaty opened this issue Oct 3, 2022 · 1 comment
Open
7 tasks

Tensorflow for mac arm64 #208

jwokaty opened this issue Oct 3, 2022 · 1 comment

Comments

@jwokaty
Copy link
Collaborator

jwokaty commented Oct 3, 2022

AFAIK pip install tensorflow doesn't work for Mac ARM64. It's suggested to use pip install Tensorflow-macos Tensorflow-metal. See https://stackoverflow.com/questions/68515561/installing-tensorflow-on-macos-on-an-arm-mbp.

  • Assess impact on packages
  • Search on code.bioconductor.org for tensorflow file:DESCRIPTION: scAlign, ttgsea, infinityFlow, pareg (note: scAlign also needs to update python, some of these may need to include tensorflow as a SystemRequirements)
  • Determine if they should add INSTALL file
  • Update Mac ARM64 Doc
  • Make pip_arm64.txt (what path?)
  • Make issue in github.com/Bioconductor/bioconductor_salt
  • Update kjohnson

For posterity, also adding from Slack discussion, quoting @hpages :

If a Bioc package installs a CRAN package that has a SystemRequirement, it should explicitly declare it in the Bioc package's SystemRequirements

Generally speaking I would say yes. For Python modules like tensorflow: definitely! With the exception of packages that manage their Python deps via basilisk. All their Python needs are taken care of automatically by basilisk so they don't need to declare them in SystemRequirements.

@hpages
Copy link
Contributor

hpages commented Oct 3, 2022

A bummer that pip3 install tensorflow does not work on Mac ARM64 😞

I would be cautious about the workaround you link to on SO. Seems tricky/hacky, has not been approved yet, has only one upvote, and it's not clear that it will play well with other Python modules that depend on the tensorflow module (see Side-Note: Package Install Issues). Also I suspect that most M1 users are not aware of this workaround so will probably struggle trying to install the tensorflow module on their machine. If you hit too many snags with this, it's ok to mark Bioconductor packages that depend on this module as unsupported on arm64, until pip3 install tensorflow is actually supported.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants