Skip to content

Latest commit

 

History

History
19 lines (15 loc) · 1.01 KB

EXTRA_DEPENDENCIES.rst

File metadata and controls

19 lines (15 loc) · 1.01 KB

Extra MLflow Dependencies

When you install the MLflow Python package, a set of core dependencies needed to use most MLflow functionality (tracking, projects, models APIs) is also installed.

However, in order to use certain framework-specific MLflow APIs or configuration options, you need to install additional, "extra" dependencies. For example, the model persistence APIs under the mlflow.sklearn module require scikit-learn to be installed. Some of the most common MLflow extra dependencies can be installed via pip install mlflow[extras].

The full set of extra dependencies are documented, along with the modules that depend on them, in the following files:

  • extra-ml-requirements.txt: ML libraries needed to use model persistence and inference APIs
  • small-requirements.txt, large-requirements.txt: Libraries required to use non-default artifact-logging and tracking server configurations