georgegyan/ghana-food-cnn-classifier

# Ghanaian Food Image Classification using EfficientNet-B0 ## Project Overview This project uses Deep Learning and Computer Vision to identify Ghanaian foods from images. The model was trained using the Ghanaian Food Dataset obtained from Hugging Face and implemented using PyTorch and EfficientNet-B0. The system can classify 30 different Ghanaian food categories and predict the food shown in an image. --- ## Dataset Source Dataset: huggingface.co Dataset Statistics: - Training Images: 119 - Validation Images: 30 - Total Classes: 30 Food Classes: 1. Abolo Ghana 2. Agbeli Kaklo 3. Apapransa Ghana 4. Banku Ghana 5. Ebunubunu Ghana 6. Fufu Ghana 7. Gari Soakings 8. Groundnut Soup Ghana 9. Hausa Koko 10. Jollof Rice Ghana 11. Kelewele Ghana 12. Kenkey Ghana 13. Konkonte 14. Kontomire Stew Ghana 15. Koose Ghana 16. Kyinkyinga Ghana 17. Okro Stew Ghana 18. Omotuo Ghana 19. Palmnut Soup Ghana 20. Red Red Ghana 21. Roasted Plantain Ghana 22. Tatale Ghana 23. Tubaani Ghana 24. Tuo Zaafi Ghana 25. Waakye Ghana 26. Yam Porridge Ghana 27. Fante Kenkey 28. Gobɛ Ghana 29. Light Soup Ghana 30. Ɛtɔ Ghana --- ## Technologies Used - Python - PyTorch - Torchvision - EfficientNet-B0 - Pandas - Pillow (PIL) - Matplotlib - Hugging Face Dataset --- ## Project Structure ```text dataset/ │ ├── food_dataset/ │ ├── train/ │ └── validation/ │ ├── ghana_food_model.pth │ ├── train_efficientnet.py ├── train_model.py ├── predict.py ├── evaluate_model.py ├── prepare_dataset.py ├── dataset_info.py ├── explore_dataset.py ├── view_dataset.py ├── check_classes.py ├── food_labels.py │ ├── requirements.txt ├── README.md └── .gitignore ``` --- ## Installation ### Clone the Repository ```bash git clone cd dataset ``` ### Create Virtual Environment ```bash python -m venv .venv ``` Activate: Windows: ```bash .venv\Scripts\activate ``` Linux/Mac: ```bash source .venv/bin/activate ``` ### Install Dependencies ```bash pip install -r require …