-
Notifications
You must be signed in to change notification settings - Fork 5.7k
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
simplify requirements.txt #28824
simplify requirements.txt #28824
Conversation
This reverts commit 5009eaa.
@hmahmood24 @YushaArif99 am I ok to merge this? seems to be fine with local testing, but I think the only way to test it out properly will be as part of the 'full ivy test from fresh dev perspective' which we can do soon. |
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.
lgtm at a first glance! But i guess we'll need to do some e2e testing to verify everything works smoothly from an end user's perspective.
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.
@Sam-Armstrong Thanks for working on this. The first question/observation that comes to mind is that requirements.txt
is supposed to represent the minimal requirements in order to use ivy
out of the box without minimal effort and since we are launching ivy
with the purpose of using it as a transpiler with support for going from torch
to tensorflow
, do you think it makes sense to have these deps in the regular requirements.txt
file because if someone does pip install ivy
, we'd only install numpy
and that's no good (it was good when it was to be used as a framework but is useless if we are marketing ivy
as a transpiler with no support for transpiling to/fro numpy
at launch). If you look at transformers for example, they have tensorflow
, torch
, jax
and numpy
all listed down directly in their setup.py
rather than have optional
requirements files that the users themselves have to install regardless of whether they're installing from source or from pip
Also would be good to pin the versions as part of this PR @Sam-Armstrong |
I'll make a new PR for pinning the frameworks, so we can check we agree on the pinned versions. |
Remove unnecessary dependencies from the requirements.txt file.