Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Raise exceptions when torch distributed is not available #10418

Merged
merged 3 commits into from
Nov 9, 2021

Conversation

four4fish
Copy link
Contributor

@four4fish four4fish commented Nov 8, 2021

What does this PR do?

Avoids init_dist_connection failing silently

If torch.distributed is not available, the process_group cannot be created. Later, the runtime will fail when wrapping model or after training steps (when call distributed_c10d). So we should fail faster at init_dist_connection step

partially fixes #10348

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

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 list all the breaking changes introduced by this pull request?
  • Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is welcome to review the PR.
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 🙃

@four4fish four4fish marked this pull request as draft November 8, 2021 18:36
@four4fish four4fish added the distributed Generic distributed-related topic label Nov 8, 2021
@four4fish four4fish added this to the 1.6 milestone Nov 8, 2021
@codecov
Copy link

codecov bot commented Nov 8, 2021

Codecov Report

Merging #10418 (ecb1f24) into master (e61a98a) will decrease coverage by 0%.
The diff coverage is n/a.

❗ Current head ecb1f24 differs from pull request most recent head bcf0679. Consider uploading reports for the commit bcf0679 to get more accurate results

@@           Coverage Diff           @@
##           master   #10418   +/-   ##
=======================================
- Coverage      89%      89%   -0%     
=======================================
  Files         181      181           
  Lines       16386    16319   -67     
=======================================
- Hits        14523    14455   -68     
- Misses       1863     1864    +1     

@four4fish four4fish marked this pull request as ready for review November 8, 2021 21:34
@ananthsub ananthsub changed the title Raise exceptions when torch distributed is not avalible Raise exceptions when torch distributed is not available Nov 8, 2021
@mergify mergify bot added the ready PRs ready to be merged label Nov 9, 2021
@ananthsub
Copy link
Contributor

@four4fish could you update the changelog as well?

@justusschock justusschock enabled auto-merge (squash) November 9, 2021 08:27
@justusschock justusschock merged commit 0ed5e3d into Lightning-AI:master Nov 9, 2021
Raalsky pushed a commit to neptune-ai/pytorch-lightning that referenced this pull request Nov 23, 2021
…I#10418)

* Raise exceptions when torch distributed is not avalible

* add changelog
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
distributed Generic distributed-related topic ready PRs ready to be merged
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Module 'torch.distributed' has no attribute 'ProcessGroup' when importing PyTorch Lightning
6 participants