dawit-hopes/ethiopian-food-classifier

A deep learning project that classifies traditional Ethiopian dishes from images using FastAI and PyTorch. # Ethiopian Food Classifier 🍲 A deep learning project that classifies traditional Ethiopian dishes from images. Built with FastAI and PyTorch, this project leverages a dataset of Ethiopian foods to train a convolutional neural network (CNN) for image classification. ## Dataset The dataset is sourced from Hugging Face: - Format: Parquet file containing images in byte format and corresponding labels. - Images represent various traditional Ethiopian dishes. ## Features - Handles in-memory images directly from a DataFrame. - Uses `RandomResizedCrop` for better generalization. - Includes standard data augmentations with `aug_transforms`. - Outputs top-1 classification accuracy for images. ## Installation ```bash git clone github.com cd ethiopian-food-classifier pip install -r requirements.txt