MLflow 2.20.0

[Harutaka Kawamura]

We are excited to announce the release of MLflow 2.20.0! This release includes a number of significant features, enhancements, and bug fixes.

Major New Features

• 💡Type Hint-Based Model Signature: Define your model's signature in the most Pythonic way. MLflow now supports defining a model signature based on the type hints in your PythonModel's predict function, and validating input data payloads against it. (#14182, #14168, #14130, #14100, #14099, @serena-ruan)

• 🧠 Bedrock / Groq Tracing Support: MLflow Tracing now offers a one-line auto-tracing experience for Amazon Bedrock and Groq LLMs. Track LLM invocation within your model by simply adding mlflow.bedrock.tracing or mlflow.groq.tracing call to the code. (#14018, @B-Step62, #14006, @anumita0203)

• 🗒️ Inline Trace Rendering in Jupyter Notebook: MLflow now supports rendering a trace UI within the notebook where you are running models. This eliminates the need to frequently switch between the notebook and browser, creating a seamless local model debugging experience. (#13955, @daniellok-db)

• ⚡️Faster Model Validation with uv Package Manager: MLflow has adopted uv, a new Rust-based, super-fast Python package manager. This release adds support for the new package manager in the mlflow.models.predict API, enabling faster model environment validation. Stay tuned for more updates! (#13824, @serena-ruan)

• 🖥️ New Chat Panel in Trace UI: THe MLflow Trace UI now shows a unified chat panel for LLM invocations. The update allows you to view chat messages and function calls in a rich and consistent UI across LLM providers, as well as inspect the raw input and output payloads. (#14211, @TomuHirata)

See the release note for more details!