hissain/ml_docker

minimal docker reference code to deploy ML models

ML model deployment example

Complete code (including a trained model) to deploy and inference a machine learning model (built on the iris dataset) using Docker and FastAPI.

1. With terminal navigate to the root of this repository

2. Build docker image

docker build -t image_name .

3. Run container

docker run --name container_name -p 8000:8000 image_name

4. Output will contain

INFO: Uvicorn running on http://0.0.0.0:8000

Use this url in chrome to see the model frontend; use http://0.0.0.0:8000/docs for testing the model in the web interface.

5. Query model

1. Via web interface (chrome):

   http://0.0.0.0:8000/docs -> test model

2. Via python client:

   client.py

3. Via curl request:

   curl -X POST "http://0.0.0.0:8000/predict" -H "accept: application/json" -H "Content-Type: application/json" -d '{"features": [5.1, 3.5, 1.4, 0.2]}'
   

Docker/ Docker compose usage

Basic Docker Compose Commands

Build Services

docker-compose build

This command builds the images specified in the docker-compose.yml file.

Start Services

docker-compose up

This command starts the services defined in the docker-compose.yml file. It builds the images if they are not already built.

Start Services in Detached Mode

docker-compose up -d

This command starts the services in the background (detached mode).

Stop Services

docker-compose stop

This command stops the running services without removing the containers.

Restart Services

docker-compose restart

This command restarts the running services.

Remove Services

docker-compose down

This command stops and removes the containers, networks, and volumes defined in the docker-compose.yml file.

Remove Services with Volumes

docker-compose down -v

This command stops and removes the containers, networks, and volumes defined in the docker-compose.yml file, including the volumes associated with the services.

Credits: https://github.com/DanilZherebtsov/ml-docker-flask-api