-
Notifications
You must be signed in to change notification settings - Fork 63
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
Remove weak dependencies #294
Comments
Yes, CUDA is going to be optional starting v0.5 (no planned timeline for it yet).
Correct. However, for the Project.toml to be backwards compatible, we need to list all dependencies in |
Thanks for the clarification 👍 You are right, I found out where I triggered precompilation of Zygote myself. In DiffEqBase, most of the weakdeps do not appear in the If removing the weak dependencies from |
Ah I see. I think we can get away with just putting them under |
If the tests pass for the linked PR this should be fixed |
Big fan of this project, thanks a lot!
I noticed that installing/precompiling takes a while and that there are quite some heavy dependencies.
By looking at LuxCUDA.jl, it seems that CUDA is going to be optional?
However, there is more:
From how I interpret the
__init__
function, it appears that Flux, Zygote etc. are meant to be optional dependencies.But they appear in the
[deps]
section of the Project.toml and are automatically installed in Julia 1.8.I did not test Julia 1.9 yet, but as they are also listed in
[weakdeps]
, it might not be the case for Julia 1.9+Is it mandatory to have Flux, ... as hard dependencies? Or could they be removed from the Project.toml dependency section?
This issue also mentions Flux as a dependency, so might be related.
The text was updated successfully, but these errors were encountered: