EthnoFace: Ethnicity Classification using Deep Learning

This project trains a ResNet-based CNN on the UTKFace Dataset to classify faces into 5 ethnic categories:

• White
• Black
• Asian
• Indian
• Others

🚀 Features

• Trains on UTKFace dataset with class balancing
• Achieves ~82% validation accuracy
• GPU acceleration (CUDA support)
• Supports inference on new images
• Easy-to-use scripts: train.py, predict.py, evaluate.py

📂 Project Structure

• data_loader.py – custom PyTorch Dataset
• model.py – ResNet-based classifier
• train.py – training with weighted loss & validation tracking
• predict.py – inference on single image
• models/ – saved model checkpoints
• requirements.txt – dependencies

⚡ Usage

Install dependencies

pip install -r requirements.txt