pranaychandekar / mlflow-server Goto Github PK

MLflow is an open source platform to manage the ML lifecycle, including experimentation, reproducibility, deployment, and a central model registry.

MLFlow Tracking Server

MLflow currently offers four components:

To track ML experiments and version models, we need to host the MLFlow Tracking Server or MLFlow Server. To make this task simple, this repository offers the containerized solution for hosting the MLFlow Server in minutes. The only thing that you need to host it on your local system is the Docker Engine.

The documentation contains:

1. Descriptions of the files
2. Instructions to host the MLFlow Server
3. Instructions to use the MLFlow Server

Step 01: Build the docker image of the MLFlow Server

Run the following command in the directory with the Dockerfile

 docker build --network=host -t mlflow-server .

Check the container with the following command

docker images

You should see the mlflow-server image in the output.

Step 02: Run the lambda function container

Run the following command.

docker run -d --network=host --name=mlflow-server mlflow-server

This will create the lambda function container. Now your container is up and running to process the invocation.

Step 03: Verify the running container

Please execute the command below to verify whether the container is running.

docker ps

You should see the mlflow-server container running.

Step 04: Launch the MLFlow Server UI

Go to your browser and enter the below url

http://localhost:5000

Your MLFlow Server is now ready to track your ML experiments and model versions.

MLFlow - Model Registry

Author: Pranay Chandekar