MLflow
Store evaluation inputs, outputs, scores, and traces alongside versioned machine learning experiments.
Published by MLflow Project
How to use it
Create an evaluation dataset or reuse fixed examples, then run your model or agent and apply the appropriate scorers. Store inputs, run parameters, and metrics with the experiment so a later comparison can be reproduced.
Review changed scores and their underlying examples in the local UI. For a model-based judge, keep its configuration and rubric with the run.
- Compatibility
- Python, Model APIs
- Evaluation areas
- Test datasets, Run logs, Comparative reporting
- Access
- Local Python evaluation and tracking. Optional integrations have separate requirements.