dog-cat-cnn-classifier 0.1.0

A deep learning package for dog and cat image classification using CNN

Dog-Cat Classifier

A deep learning-based image classification project that distinguishes between dogs and cats using a VGG16-based CNN architecture.

Project Overview

This project implements an end-to-end deep learning solution for binary classification of dog and cat images. It uses a transfer learning approach with VGG16 as the base model, fine-tuned for the specific task of dog-cat classification.

Features

• Transfer learning using VGG16 pre-trained model
• Data augmentation for improved model robustness
• Binary classification with softmax activation
• Comprehensive logging system
• Modular pipeline architecture
• Model evaluation and metrics tracking

Installation

1. Clone the repository:

git clone https://github.com/nishantbadhautiya100/dog-cat-classifier.git
cd dog-cat-classifier

2. Create a virtual environment (recommended):

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

3. Install dependencies:

pip install -r requirements.txt

Usage

1. Prepare your dataset:

   • Place your dog and cat images in the appropriate directories
   • The structure should be:

     training_data/
     ├── dogs/
     └── cats/
     

2. Run the training pipeline:

python main.py

The pipeline will:

• Prepare the base model (VGG16)
• Train the model on your dataset
• Evaluate the model performance
• Save the trained model

Model Architecture

• Base Model: VGG16 (pre-trained on ImageNet)
• Transfer Learning: Fine-tuning approach
• Output Layer: Dense layer with softmax activation
• Loss Function: Categorical Crossentropy
• Optimizer: Adam

Configuration

The project can be configured through:

• params.yaml: Model parameters and training settings
• config/config.yaml: Pipeline configuration

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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

Acknowledgments

• VGG16 model architecture
• TensorFlow and Keras
• ImageNet dataset for pre-training