Classification references does not work without distributed setup
Open
Nobody has claimed this yet.
bug
help wanted
module: reference scripts
- Dominant language
- Python
- Stars
- 17.9k
- Forks
- 7.3k
- Avg merge
- 1d 15h
- Merged PRs (30d)
- 13
Description
If you don't set the respective env vars
training will not be distributed and in turn the backend will not be initialized. However, during evaluation we check
unguarded, which then fails with
RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.
cc @datumbox
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with references/classification/utils.py at the environment-variable handling and references/classification/train.py at the evaluation check linked in the issue. Reproduce classification training without the distributed environment variables and confirm the evaluation path no longer raises the uninitialized process-group error.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision, distributed-systems
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 50/100