Skip to main content

Docker Deployment

API Docker Deployment​

The configuration for Docker deployment is defined in deploy/docker/.env.

Build the Docker image:​

For a base image with common code and dependencies:

make docker-build-base

For the API image with model dependencies:

make docker-build

Deploy the Docker container:

make docker-deploy-api

Stop the API:

make docker-stop-api

Remove the API container:

make docker-down-api

Predict/Train Docker Deployment​

  1. Build the Docker image and initialize conda environments:

    make docker-build

  2. Deploy Docker containers:

    • For prediction: 
      make docker-deploy-predict
    • For training: 
      make docker-deploy-train

  3. Remove containers:

    • Prediction container:

      make docker-down-predict

    • Training container:

      make docker-down-train

These updated sections provide a more concise and clear overview of the MLflow integration and Docker deployment processes for the EvoML pipeline.