verifiers
verifiers copied to clipboard
Support env mixing with interleave_datasets
Description
This PR adds support for mixing the underlying datasets in an EnvGroup using interleave_datasets or concatenate_datasets. The former allows for easy customization of the env mixes (e.g. sampling probabilities to deal with uneven dataset sizes). The PR does not change any default behavior, ie. its fully backwards compatible. Added tests for new interleaving functionality.
Type of Change
- [ ] Bug fix (non-breaking change which fixes an issue)
- [x] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
- [ ] Documentation update
- [ ] Test improvement
Testing
- [ ] All existing tests pass when running
uv run pytestlocally. - [x] New tests have been added to cover the changes
Checklist
- [ ] My code follows the style guidelines of this project as outlined in AGENTS.md
- [ ] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] Any dependent changes have been merged and published
Additional Notes
@mikasenghaas can we remove the probabilities field? then should be good to merge.
@mikasenghaas can we remove the
probabilitiesfield? then should be good to merge.
done!