-
Notifications
You must be signed in to change notification settings - Fork 9
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 to PyTorch migration #189
Conversation
…to PyTorch instead of tensorflow
I had one question about these lines, after we remove TensorFlow are we still going to recommend a Python 3.10 dev environment? If so, what's our official reason? |
Nope, once we move to torch, the plan is to update everything to recommend 3.11. Once all dependencies also support it, we can then move to 3.12. tensorflow is/was the only thing keeping us on 3.10 |
IIUC as we're following NEP 29, we will recommend 3.11 but still support 3.10 until April 2025 |
Yeah we haven't updated supported versions while we've been stuck with 3.10 as the latest. I suggest once the torch version of cf is released, we go through and NEP29-ify everything. |
There was a problem hiding this 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 - we should wait until brainglobe/cellfinder#418 (comment) is merged before merging this.
*is released. |
Build Sphinx Docs is failing due to:
The link works when I try to access it via browser. |
I've re-run the build. I've sometimes noticed that biorxiv does this, but it works itself out |
Before submitting a pull request (PR), please read the contributing guide.
Please fill out as much of this template as you can, but if you have any problems or questions, just leave a comment and we will help out :)
Description
What is this PR
Why is this PR needed?
We're changing the backend for
cellfinder
to use PyTorch instead of Tensorflow.What does this PR do?
Removes most mentions of TensorFlow as it relates to
cellfinder
and replaces them with the relevant PyTorch instruction.I've made these changes with the understanding that
cellfinder
is the only package that currently uses TensorFlow so migrating to PyTorch allows us to not instruct users on setting it up. I didn't change the mentions in error-messages.md as those remain relevant for people using the older version.References
Closes #183
How has this PR been tested?
Built locally.
Is this a breaking change?
No.
Checklist: