endpoint from a command

endpoint using a raw start command that works with any programming language or framework.

• name: name of the deployment endpoint, this will be the final part of the URL.
• image: the Docker image that will be used as the deployment environment.
• server-command: command that runs an HTTP server
• port: (optional) where should Valohai expect to find the web server, defaults to 8000. Note the server must be bound to 0.0.0.0 (all interfaces), not only localhost.
• description: (optional) more detailed human-readable description of the endpoint
• files: (optional) files that will be loaded into the image, for example, the trained model. The files will be in the same directory as your code, modified by the path property.

- endpoint:
    name: server-endpoint
    image: python:3.6
    server-command: python run_server.py
    port: 1453
    files:
      - name: model
        description: Model output file from TensorFlow
        path: model.pb

Installing additional packages to deployment images

Valohai automatically runs pip install --user -r requirements-deployment.txt in case you have additional dependencies defined in requirements-deployment.txt in your project root.

If you don’t have a requirements-deployment.txt file, Valohai will check if there is a requirements.txt file and then run pip install --user -r requirements.txt in your project root.

Commands from installed packages end up in ~/.local/bin, following the standards for pip --user installation.