README.md

metadata

title: Chest Cancer Detection AI
emoji: 🩺
colorFrom: blue
colorTo: indigo
sdk: docker
app_file: app.py
pinned: false

🩺 End-to-End Chest Cancer Classification

An MLOps project demonstrating the complete lifecycle of a deep learning model, from data ingestion to CI/CD-powered deployment.

---

🚀 Live Demo

Experience the deployed application live on Hugging Face Spaces!

➡️ Live Demo Link

(Note: The Space may be asleep if it hasn't been used recently. Please allow a moment for it to wake up.)

🖼️ Application Screenshot

Here is the user interface of the deployed web application.

📖 About The Project

This project implements a complete end-to-end MLOps pipeline for a Chest Cancer image classification task. A deep learning model (based on VGG16/ResNet) is trained to distinguish between Normal and Adenocarcinoma chest CT scans.

The primary focus is not just on the model's accuracy, but on building a robust, reproducible, and automated system using modern MLOps tools.

Key Features:

• Experiment Tracking: Uses MLflow to log parameters, metrics, and model artifacts for every run.
• Data & Model Versioning: Uses DVC to version large data files and models, keeping the Git repository lightweight.
• Automated CI/CD: A GitHub Actions workflow automatically tests, builds, and deploys the application on every push to the main branch.
• Web Application: A user-friendly Flask application serves the trained model for real-time predictions.
• Containerization: The entire application is containerized with Docker for consistent and portable deployment.
• Cloud Deployment: Deployed for free on Hugging Face Spaces.

🛠️ Tech Stack

• Backend: Python, Flask
• Model: TensorFlow, Keras
• MLOps Tools: MLflow, DVC, Dagshub (for remote tracking)
• CI/CD & Deployment: Docker, GitHub Actions, Hugging Face Spaces

🌊 MLOps Workflow

The project follows a structured MLOps workflow, which is fully automated by the CI/CD pipeline.

graph TD
    A[Start: Push Code to GitHub] --> B{GitHub Actions CI/CD};
    B --> C[CI: Install Dependencies & Run Tests];
    C -->|Success| D[CD: Deploy to Hugging Face];
    D --> E[🚀 Live Application];

    subgraph "DVC & MLflow Cycle (Local/Remote)"
        F[1. `dvc repro`] --> G[2. Pull Data (DVC)];
        G --> H[3. Train Model];
        H --> I[4. Log Metrics & Model (MLflow)];
        I --> J[5. Push Model (DVC)];
    end

⚙️ Getting Started - Local Setup

To run this project on your local machine, follow these steps.

Prerequisites

• Git
• Python 3.8+
• A DagsHub account (for MLflow tracking)

Installation & Setup

1. Clone the repository:

   git clone https://github.com/AlyyanAhmed21/End-to-End-Chest-Cancer-Classification-using-MLflow-and-DVC.git
   cd End-to-End-Chest-Cancer-Classification-using-MLflow-and-DVC
   

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Set up MLflow/DagsHub credentials: Create a .env file in the root directory and add your credentials. This file is ignored by Git.

   MLFLOW_TRACKING_URI="https://dagshub.com/YourUsername/YourRepoName.mlflow"
   MLFLOW_TRACKING_USERNAME="YourUsername"
   MLFLOW_TRACKING_PASSWORD="YourDagsHubAccessToken"
   

5. Run the DVC pipeline: This command will execute all stages defined in dvc.yaml (data ingestion, model preparation, training, and evaluation).

   dvc repro
   

6. Run the Flask application:

   python app.py
   

   Open your browser and navigate to http://localhost:8080 to use the app.

📁 Project Structure

.
├── .dvc/                # DVC metadata
├── .github/workflows/   # GitHub Actions CI/CD pipeline
├── artifacts/           # (Generated by DVC) Data, models, etc.
├── config/              # Configuration files (config.yaml)
├── src/                 # Source code for the project
│   └── cnnClassifier/
│       ├── components/    # Individual pipeline components
│       ├── config/        # Configuration management code
│       ├── entity/        # Custom entity definitions
│       ├── pipeline/      # DVC pipeline stage definitions
│       └── utils/         # Utility functions
├── templates/           # HTML templates for the Flask app
├── .gitignore
├── app.py               # Main Flask application entrypoint
├── dvc.yaml             # DVC pipeline definition
├── Dockerfile           # Docker configuration for deployment
├── main.py              # Main project orchestrator
├── params.yaml          # Model parameters
└── requirements.txt     # Python dependencies

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Star the repo if you found it useful! ⭐