NetAlpha

NetAlpha is a Python package designed to train a multi-branch Convolutional Neural Network (CNN) model that integrates ResNet50V2 and DenseNet169 architectures for superior image classification tasks.

Table of Contents

• Introduction
• Features
• Installation
  • Prerequisites
  • Installation via Pip
• Usage
  • Command-Line Interface
    • Basic Usage
    • Options
    • Help
  • Examples
  • Data Preparation
• Programmatic Usage
• Project Structure
• Documentation
• Contributing
  • Reporting Issues
• License
• Contact
• Acknowledgments

Introduction

NetAlpha offers a hybrid architecture combining the strengths of ResNet50V2 and DenseNet169 for high-performance image classification. By integrating these architectures, NetAlpha provides robust feature extraction and better generalization for diverse datasets.

Features

• Hybrid Architecture: Combines ResNet50V2 and DenseNet169 for enhanced feature extraction.
• Data Augmentation: Advanced augmentation techniques to improve model robustness.
• Early Stopping & Learning Rate Scheduling: Built-in callbacks for better training control.
• Command-Line Interface (CLI): Simplified training and configuration through CLI commands.
• Customizable Hyperparameters: Fine-tune learning rate, batch size, and more.
• Training Visualization: Plot training and validation metrics for better insights.

Installation

Prerequisites

• Python: Version 3.6 or higher.
• TensorFlow: Version 2.x or later.
• pip: Latest version recommended.

Installation via Pip

Install NetAlpha directly from PyPI:

pip install netalpha

Usage

Command-Line Interface

Once installed, you can use the netalpha_train command.

Basic Usage

netalpha_train --train_dir PATH_TO_TRAIN_DIR --val_dir PATH_TO_VAL_DIR [OPTIONS]

Options

• --train_dir: (Required) Path to the training data directory.
• --val_dir: (Required) Path to the validation data directory.
• --epochs: Number of epochs (default: 50).
• --batch_size: Batch size (default: 8).
• --initial_lr: Initial learning rate (default: 1e-4).
• --output_dir: Directory to save model outputs (default: ./).
• --plot: Add this flag to generate training plots.

Help

To see a detailed list of options, run:

netalpha_train --help

Examples

Example 1: Basic Training

netalpha_train   --train_dir ./data/train   --val_dir ./data/val   --epochs 30   --batch_size 16

Example 2: Training with Custom Parameters

netalpha_train   --train_dir ./data/train   --val_dir ./data/val   --epochs 40   --batch_size 32   --initial_lr 5e-5   --output_dir ./outputs   --plot

Data Preparation

Organize your dataset as follows:

dataset/
├── train/
│   ├── class1/
│   │   ├── image1.jpg
│   │   ├── image2.jpg
│   │   └── ...
│   └── class2/
│       ├── image1.jpg
│       ├── image2.jpg
│       └── ...
└── val/
    ├── class1/
    │   ├── image1.jpg
    │   ├── image2.jpg
    │   └── ...
    └── class2/
        ├── image1.jpg
        ├── image2.jpg
        └── ...

Programmatic Usage

NetAlpha can also be used in Python scripts:

from netalpha import build_model

# Build the model
model = build_model(input_shape=(224, 224, 3), num_classes=2)

# Compile the model
model.compile(
    optimizer="adam",
    loss="categorical_crossentropy",
    metrics=["accuracy"]
)

# Train or evaluate as needed

Project Structure

netalpha/
├── netalpha/
│   ├── __init__.py
│   ├── model.py
│   └── utils.py
├── scripts/
│   └── train_model.py
├── setup.py
├── README.md
├── LICENSE
└── requirements.txt

Documentation

For detailed documentation, visit the GitHub repository.

Contributing

Contributions are welcome! Follow these steps to contribute:

1. Fork the repository.
2. Clone your fork:

   git clone https://github.com/yourusername/netalpha.git
   

3. Create a new branch:

   git checkout -b feature/new-feature
   

4. Commit your changes:

   git commit -am "Add new feature"
   

5. Push the branch:

   git push origin feature/new-feature
   

6. Open a pull request.

Reporting Issues

If you encounter any issues, please open an issue.

License

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

Contact

• Author: Ihtesham Jahangir
• Email: ihteshamjahangir21@gmail.com
• GitHub: ihtesham-jahangir

Acknowledgments

• TensorFlow and Keras Teams: For their excellent frameworks.
• Researchers: Behind ResNet and DenseNet architectures.
• Open Source Community: For fostering innovation and collaboration.

This README provides detailed information about NetAlpha, making it easy for users to install, use, and contribute to the project.