mlfoundations / mlfoundations/evalchemy
Race condition when running alpaca eval?
Nobody has claimed this yet.
- Dominant language
- HTML
- Stars
- 610
- Forks
- 87
- PR merge metrics
- No merged PRs in 30d
Description
Alpaca eval seems to save annotations under json files in eval/chat_benchmarks/alpaca_eval/src/alpaca_eval/evaluators_configs - this leads to a race condition if multiple processes try to use the same annotator at once.
This issue can be circumvented by serializing evaluation, but posting this here in case there is a better way.
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 in eval/chat_benchmarks/alpaca_eval/src/alpaca_eval/evaluators_configs and trace where annotation JSON files are written. Reproduce evaluation with multiple processes using the same annotator, then determine whether concurrent runs can complete without conflicting writes or corrupted annotations.
Written by the indexing model from the issue text.
Assessment
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100