v3.1.1
📦 mlflowView on GitHub →
✨ 2 features🐛 6 fixes🔧 6 symbols
Summary
MLflow 3.1.1 adds a larger prompt limit, pagination for history retrieval, and several bug fixes including artifact download support and security hardening.
✨ New Features
- Increase prompt text limit from 5K to 100K in Model Registry / Sqlalchemy.
- Add pagination support to get-history API for FileStore and SqlAlchemyStore in Tracking.
🐛 Bug Fixes
- Add support for downloading logged model artifacts in Artifacts.
- Fix Bedrock provider and configured inference profile compatibility in Models.
- Ensure attribute.run_id is specified when search_traces filters by run_id in Tracking.
- Mitigate GraphQL batching attacks in Tracking.
- Make chunk size configurable in DatabricksSDKModelsArtifactRepository for Model Registry.
- Various small bug fixes and documentation updates (see PRs #16261, #16411, #16352, #16327, #16324, #16279, #16193, #16197, #16409, #16348, #16347, #16290, #16286, #16283, #16271, #16223, #16326, #16364, #16308, #16218, #16262, #16191, #16212, #16208, #16205, #16200, #16198).