A time-efficient YOLO pipeline project for object detection models

A full-scale, object-oriented pipeline for data preprocessing, training, predicting, and exporting YOLO models with Firebase integration.

Features

• Google Firebase integration for dataset retrieval
• Automatic dataset organization and splitting
• YOLO model training with configurable parameters
• Model inference and export functionality
• Continuous monitoring of Firebase for new data
• Pipeline automation based on new data threshold

Project Structure

yolo_pipeline/
├── config/             # Configuration settings
├── data/               # Data management (Firebase, dataset)
├── models/             # YOLO model operations
├── utils/              # Utility functions
├── pipeline/           # Pipeline orchestration
├── tests/              # Unit tests
├── main.py             # Entry point
├── requirements.txt    # Project dependencies
└── README.md           # Project documentation

Setup

1. Create a Python virtual environment:

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

2. Install dependencies:

   pip install -r requirements.txt

3. Set up Firebase credentials:

   • open .env file
   • Set the GOOGLE_APPLICATION_CREDENTIALS environment variable to point to your Firebase credentials JSON file

4. Source the environment variables:

   source .env   

5. Run it using various arguments:

   python3 main.py --monitor --debug --run --test-image=test_images/strawberry.png

6. The results will be present in the data_storage folder under predict and you could use the .torchscript with the cpp script and libtorch library for faster inferencing.