Skip to content

InfinStor Mlflow Plugin Installation

InfinStor Machine Learning Platform is a tool for Data Scientists and Data Engineers to manage data. InfinStor's core capabilities include fine grained snapshots of S3 buckets at any point in time in the past, slices of S3 bucket data that was ingested in a specified period of time, snapshots of tensors, etc.

There are three pip packages that provide functionality in the local system. These are:

  • infinstor
  • infinstor-mlflow-plugin
  • jupyterlab-infinstor

the 'infinstor' pip package is a client library that provides various functionality for things related to the InfinStor service. 'infinstor-mlflow-plugin' package provides the mlflow plugins for tracking, projects, artifacts, etc. Finally, the 'jupyterlab-infinstor' package provides a jupyterlab server extension for the InfinStor jupyterlab sidebar UI component. The 'jupyterlab-infinstor' component is optional for those who do not use jupyterlab.

The plugins are installed using pip. They must be installed on all python environments where the machine learning activity is to take place. This includes the jupyterlab server python env, the ipython env, the InfinStor single VM Cloud instance and any cluster instances (e.g. EMR)

Prerequisites

The python libraries that InfinStor requires as prerequisites must be installed using the following commands

pip install mlflow

InfinStor pip packages

The following commands uninstall older versions and install new versions of the infinstor pip packages

pip uninstall -y infinstor
pip install infinstor
pip uninstall -y infinstor-mlflow-plugin
pip install infinstor-mlflow-plugin
pip uninstall -y jupyterlab_infinstor
pip install jupyterlab_infinstor