Pytorch cuda in registry nightly images

[bhack]

Are you testing if the nightly image is usable with cuda?

torch.cuda._is_compiled() is false inside the last nightly image

We should add docker images to validation framework.

Workflow:
https://github.com/pytorch/pytorch/blob/main/.github/workflows/docker-release.yml

Docker containers are located here:
https://github.com/orgs/pytorch/packages/container/package/pytorch

Simple install command:
docker pull ghcr.io/pytorch/pytorch:2.2.1-cuda11.8-cudnn8-devel

Build workflow:
https://github.com/pytorch/pytorch/actions/runs/8200189724/job/22426518545

[atalman]

We should add automation around validation of docker images for both nightly and releases. Release workflow: https://github.com/pytorch/pytorch/actions/runs/8393526521/job/22988732918

Onboard to validation framework:
https://github.com/pytorch/builder/actions/workflows/validate-binaries.yml

[atalman]

cc @juliagmt-google

[juliagmt-google]

Thanks for sharing the task and details. Here are my questions:

1. torch.cuda._is_compiled() is false inside the last nightly image: where can I see the output?
2. What exactly is the validation?
3. I saw docker pull ghcr.io/pytorch/pytorch:2.2.2-cuda11.8-cudnn8-devel in https://github.com/orgs/pytorch/packages/container/package/pytorch where Docker containers are located, but the instruction says installing docker pull ghcr.io/pytorch/pytorch:2.2.1-cuda11.8-cudnn8-devel, which has a different PyTorch version. Which command should I use?
4. Which files do we need to change to add automation and validation?

[atalman]

Link to the images to validate: https://github.com/orgs/pytorch/packages/container/package/pytorch-nightly
Nova workflows for reference: https://github.com/pytorch/test-infra/wiki/Using-Nova-Reusable-Build-Workflows

Try to call: https://github.com/pytorch/test-infra/blob/main/.github/workflows/linux_job.yml

[juliagmt-google]

For gpu runners we need to use pytorch/test-infra/.github/workflows/linux_job.yml@main

[bhack]

It would be nice to check in the CI job the layers invalidation as currently tracking nightly day by day is going to fill quite soon artifact registry space and local build cache. See pytorch/pytorch#125862