update ngc for 1.3

[Borda]

What does this PR do?

update base and ...

Before submitting

• 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?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

• 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

Did you have fun?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #7414 (d48131d) into master (fbc8b20) will decrease coverage by 0%.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #7414   +/-   ##
======================================
- Coverage      92%     92%   -0%     
======================================
  Files         200     200           
  Lines       12983   12983           
======================================
- Hits        11937   11897   -40     
- Misses       1046    1086   +40     

[s-rog]

CI saying the build ran out of storage?

BTW we recently started using 21.04 (from 21.02) as well and it required a driver update (460.x to 465.x) even for data center GPUs (V100) despite nvidia docs stating otherwise (ddp fails from nccl errors). I think this breaks 21.04 on some cloud providers (AWS) that currently have an older driver version.

[Borda]

CI saying the build ran out of storage?

yes we need to move it to another CI, probably Azure
the issue comes that the base PyTorch container is too large about 14GB

BTW we recently started using 21.04 (from 21.02) as well and it required a driver update (460.x to 465.x) even for data center GPUs (V100) despite nvidia docs stating otherwise (ddp fails from nccl errors). I think this breaks 21.04 on some cloud providers (AWS) that currently have an older driver version.

this is directly for NGC only so it shall be fine as we based it on the latest PyTorch container