You shouldn't invoke setup.py directly

Re:You shouldn't invoke setup.py directly by ampdepolymerase on Tuesday October 19, 2021 @11:43AM (#28919393)

Poetry doesn't play nice with PyTorch and many other ML libraries.

[ Parent | on HN ]

Re:You shouldn't invoke setup.py directly by bigbillheck on Tuesday October 19, 2021 @11:51AM (#28919496)

We use it for both pytorch and tensorflow (sometimes in the same package) and it works fine.

[ Parent | on HN ]

Re:You shouldn't invoke setup.py directly by camjw on Tuesday October 19, 2021 @12:03PM (#28919668)

How do you manage cuda versions of pytorch? At the moment I'm using poethepoet to install them separately but it feels very suboptimal.

[ Parent | on HN ]

Re:You shouldn't invoke setup.py directly by bigbillheck on Tuesday October 19, 2021 @03:25PM (#28922474)

So I personally didn't do the sysadminning, but it's a docker container in which CUDA has been loaded, and after that it 'just works'.

[ Parent | on HN ]

Re:You shouldn't invoke setup.py directly by disgruntledphd2 on Wednesday October 20, 2021 @08:19AM (#28929779)

One of the nice things about Conda is that it will handle these kinds of dependencies for you within the environment it sets up.

While it doesn't really matter for web/scripting use cases, it's an absolute godsend for DS/ML workflows.

But the resolver, oh why does the resolver take so long?

[ Parent | on HN ]

Re:You shouldn't invoke setup.py directly by lmeyerov on Wednesday October 20, 2021 @11:12AM (#28931634)

we do conda -> mamba for that exact scenario :)

[ Parent | on HN ]