Deploy_ML

A web-based Machine Learning API built with Flask that serves real-time predictions from a trained model. The API is containerised with Docker and deployed on Render, making it accessible for integration with frontend applications or other services.

For a step-by-step description of the application, check out this hands-on demonstration: https://medium.com/@hello.gradientthoughts/building-and-hosting-a-machine-learning-model-with-flask-and-docker-1cd4f89cf256

---

Features

• Serve ML predictions: Host a trained model and provide predictions via a RESTful API.
• JSON-based endpoints: Accepts feature inputs in JSON format.
• Error handling: Returns informative error messages for invalid requests.
• CORS enabled: Supports frontend integration from other domains.
• Dockerized: Ensures environment consistency and easy deployment.

---

Project Structure

Deploy_ML/
│
├── app/
│   ├── main.py         # Flask application
│   ├── model.py        # Model loading and prediction
│
├── models/
│   └── model.joblib    # Trained ML model
│
├── src/
│   └── main_model.py   # Script for data preprocessing and model training
|
├── f1_dnf.csv          # Dataset to train the model
├── test.py             # Mini-test script for predictions
├── requirements.txt    # Required libraries 
├── Dockerfile
└── README.md

---

Installation (Local)

1. Clone the repository:

git clone https://github.com/Pranav-Bhatlapenumarthi/Deploy_ML.git
cd Deploy_ML

2. Install dependencies:

pip install -r requirements.txt

3. Run the app:

python app/main.py

Docker

1. Build the Docker image:

docker build -t ml-flask .

2. Run the Docker container:

docker run -p 8000:8000 ml-flask

---