Support for CUDA in devcontainer

[gschramm]

@KrisThielemans @johnstairs would it be possible to install CUDA in the devcontainer such that we could run recons using cuda (e.g. via using cupy or or the pytorch build with cuda)?

[KrisThielemans]

GitHub Codespaces doesn't currently support CUDA AFAIK (see e.g.here.
Even if this would be possible, you would have to pay for it likely.

These restrictions might not apply for locally run codespaces via VS Code.

If it would be possible, you'd need 2 different configs (no/with GPU). the only way I know that that can be done is by having 2 different branches.

@johnstairs will know more.

[gschramm]

A local solution would be sufficient for me.
As an alternative, I could also build yardl from the sources in a conda environment.

[KrisThielemans]

As an alternative, I could also build yardl from the sources in a conda environment.

Easier to download binaries from https://github.com/microsoft/yardl/releases