subudhidebadash / mlflowdocker

Docker for mlflow framework

mlflowdocker

Docker container for mlflow 0.8 framework with azure backend.

Docker build: howto

If you are new to docker you will first need to install docker on your machine. If the installation is finished, you will need to clone this repo to your local machine. Once it has been cloned you can build the docker with the following command:

• docker build -t mlflowserver -f Dockerfile . --no-cache

Docker run: howto

If the build was successful you can start the container with the following command:

• docker run -p 5000:5000 --env MLFLOW_SERVER_DEFAULT_ARTIFACT_ROOT=<wasb> --env AZURE_STORAGE_ACCESS_KEY=<access_key> -it mlflowserver:latest

Please note that the port is by default on 5000

Configuring the azure backend

The current container was configured to have the storage backend in Microsoft Azure. Please do the following actions:

• Create a blobstorage account
• Create a container in the blobstorage account
• Write down the connection string and replace wasb with the connection string (eg: "wasbs://[email protected]")
• Write down the authentication key and replace access_key with that value

Access the website

If you run the docker local the website will be available from https://localhost:5000

MlFlow documentation

All documentation with regards to mlflow can be found here: https://mlflow.org/docs/latest/index.html