Add recommendation for how to run distributed eval
Open
- Dominant language
- Python
- Stars
- 64
- Forks
- 18
- Avg merge
- 4h 46m
- Merged PRs (30d)
- 1
Description
- Add example/demo notebook for how to parallelize eval
- Mention gotchas in metric computation (i.e. avoiding only taking results from the first shard)
Followup from #17
Contributor guide
Assessment
This issue has not been assessed yet.