Step-by-Step Guide: How to Deploy Machine Learning Models Using FastAPI, Docker, and Heroku

In this tutorial, we will walk you through the process of deploying a machine learning model using FastAPI, Docker, and Heroku.

FastAPI is a modern, fast (high-performance) web framework for building APIs with Python 3.6+ based on standard Python type hints. Docker is a platform for developing, shipping, and running applications in containers. Heroku is a cloud platform that lets you build, deploy, and scale applications quickly and easily.

By the end of this tutorial, you will have a fully functioning API that can serve predictions from your machine learning model. Let’s get started!

Step 1: Create a machine learning model
For this tutorial, we will use a simple machine learning model that predicts the sentiment of a text message (positive or negative). You can use any machine learning model of your choice for this tutorial.

Step 2: Create a FastAPI app
First, install FastAPI using pip:

pip install fastapi

Next, create a new Python file (e.g., app.py) and import the necessary libraries:

from fastapi import FastAPI
app = FastAPI()

Next, define a route that will accept text input and return a prediction using your machine learning model:

@app.get("/predict/{text}")
def predict_sentiment(text: str):
    # Make prediction using your machine learning model
    prediction = model.predict(text)
    return {"sentiment": prediction}

Step 3: Build a Docker image
Next, create a Dockerfile in the same directory as your app.py file:

FROM tiangolo/uvicorn-gunicorn-fastapi:python3.8

COPY ./app /app

This Dockerfile specifies a base image for running Python applications with FastAPI and copies the contents of the app directory to the container’s /app directory.

Build the Docker image using the following command:

docker build -t my_fastapi_app .

Step 4: Run the Docker container
Run the Docker container and map the container’s port to the host machine’s port:

docker run -d -p 8000:80 my_fastapi_app

You can now access your FastAPI app at http://localhost:8000.

Step 5: Deploy to Heroku
Create a Procfile in the root directory of your application with the following content:

web: uvicorn app:app --host=0.0.0.0 --port=$PORT

Next, create a Heroku account if you don’t already have one and install the Heroku CLI.

Login to Heroku using the CLI:

heroku login

Create a new Heroku app:

heroku create your_app_name

Push your Docker image to Heroku’s container registry:

heroku container:push web -a your_app_name

Release the new version of your app:

heroku container:release web -a your_app_name

Your FastAPI app with your machine learning model is now deployed on Heroku and accessible at https://your_app_name.herokuapp.com.

That’s it! You have successfully deployed a machine learning model using FastAPI, Docker, and Heroku. Feel free to customize the app further to suit your needs. Happy coding!