Pin fairscale version

[carmocca]

What does this PR do?

The 0.4.1 release breaks our tests.

Even though this release was more than a month ago. The nightly job that generates the CUDA docker image used by Azure CI has been failing for a long time.

That was until yesterday, when #10088 fixed it. At midnight that day, the image got rebuilt, now with the new fairscale version (0.4.1). This explains why the failure did not appear before.

Fixes #10196

Proof:

Fixed image build (with 0.4.1, 21h ago): https://github.com/PyTorchLightning/pytorch-lightning/runs/4016407989?check_suite_focus=true#step:5:1957

Previous failing build (2 days ago): https://github.com/PyTorchLightning/pytorch-lightning/runs/4003726347?check_suite_focus=true

Haven't checked when was the last successful build.

This is a good example why #10060 is important.

Does your PR introduce any breaking changes? If yes, please list them.

None

Before submitting

• [n/a] Was this discussed/approved via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• [n/a] Did you make sure to update the documentation with your changes? (if necessary)
• [n/a] Did you write any new necessary tests? (not for typos and docs)
• [n/a] Did you verify new and existing tests pass locally with your changes?
• [n/a] Did you list all the breaking changes introduced by this pull request?
• [n/a] Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

[daniellepintz]

Thanks for the fix!!