Add azure windows runner for pytorch

[aktech]

This adds a new configuration for windows runner for pytorch-cpu-feedstock.

Relevant issues:

• External runners for Windows conda-forge.github.io#2169
• Windows VMs Quansight/open-gpu-server#31

cc @jaimergp @isuruf

@wolfv @baszalmstra can you confirm the following:

• This is only for pytorch feedstock (as in only sponsoring windows runners for pytorch?)
• Runner machine size (Standard_D2s_v3) is sufficient, i.e. 2 cores 8 GB RAM or need a larger runner (update: 8 cores and 32GB RAM now)

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• Reset the build number to 0 (if the version changed)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

[wolfv]

Hey @aktech - we are happy with sponsoring a larger Windows runner also for other feedstocks but the immediate need was for pytorch, indeed. E.g. Tensorflow or other feedstocks should ideally also benefit 🚀

I think something with 4- or 8 cores would be more fun, no? :)

Thanks for pushing this through!

[aktech]

I think something with 4- or 8 cores would be more fun, no? :)

Sure, bumped to Standard_D8s_v3 8 cores and 32GB RAM. (I wasn't sure of the quota, but looks like the account has enough quota to support this)

[jaimergp]

LGTM! Thanks for the work @aktech, and the generosity @wolfv.

I guess we would just need to leave in writing what the process is for other feedstocks to request access to the Windows resource. Pinging you here? Something in admin-requests?

[wolfv]

@aktech any steps that I need to do to get it running in this PR? conda-forge/pytorch-cpu-feedstock#231

[jaimergp]

You probably need changing these sections:

• https://github.com/conda-forge/pytorch-cpu-feedstock/blob/437d51a1c09199b8a050a3a1b11c97625995cac6/conda-forge.yml#L25-L27
• https://github.com/conda-forge/pytorch-cpu-feedstock/blob/437d51a1c09199b8a050a3a1b11c97625995cac6/recipe/conda_build_config.yaml#L13