Skip to content

Proposal for mlflow integration #1029

@ablaom

Description

@ablaom

mlflow is an API and web-based UI for logging parameters, code
versions, metrics, and output files when running machine learning experiments, for later
visualizing the results. Integration of mlflow already exists for these other ML
platforms: Scikit-learn, Keras, Gluon, XGBoost, LightGBM, Statsmodels, Spark, Fastai,
Pytorch.

Further to this short project
outline
, and after preliminary discussions with @pebeto and @deyandyankov,
I give below a tentative design proposal for integration of
mlflow with MLJ, using
MLFlowClient.jl, which already provides a
Julia interface to mlflow.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    Status

    tracking/discussion/metaissues/misc

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions