An Image Classifier that predicts whether an input image contains a cat or a dog. This repository provides multiple approaches to the project, including a Jupyter Notebook approach, a modular approach, and a standalone local deployment method.
This project aims to build a machine learning model that can accurately classify images of cats and dogs. We've implemented the classifier using different methods to provide a comprehensive understanding of the project.
The notebook directory contains a Jupyter Notebook (ModelCreation.ipynb) that walks through the complete process of building, training, and evaluating the image classifier model. This approach is great for learning and experimenting with the code in an interactive environment.
The src directory contains a modular approach to the project. We've broken down the project into different components for better organization and scalability. Key components include:
data_ingestion_transformation.py: Handles data ingestion, transformation, and splitting into train, test, and validation sets.model_trainer.py: Defines the ModelTrainer class responsible for training the classifier model.
For a standalone local deployment, you can run the Gradio interface to predict whether an image contains a cat or a dog. Follow these steps:
deployment.py: Sets up a Gradio interface for interactive predictions using the trained model.
- Install the required dependencies:
pip install -r requirements.txt python src\components\deployment.py
- Result of model with gradio:
Notebook: Follow the step-by-step instructions in the notebook directory to learn how to build and train the classifier. Modular: Utilize the modular components in the src directory to customize and enhance the project. Standalone Local Deployment: Use the Gradio interface to predict whether an image contains a cat or a dog by following the deployment instructions.