🍽 Cuisine Classifier

From a list of countries and their popular cuisines , this project can detect them , powered by Resnet50

Cuisines the model can clasify

Cuisine	Popular Dishes
🇯🇵 Japanese	Okonomiyaki, Ramen, Sushi
🇲🇽 Mexican	Chilaquiles, Nachos, Tacos
🇹🇷 Turkish	Baklava, Kebab, Meze food, Shawarma
🇺🇸 American	Burger, Fried Chicken, Hot Dog
🇮🇳 Indian	Fuchka, Biriyani, Khichuri, Samosa
🇮🇹 Italian	Lasagna, Pizza, Spaghetti

Check out the web app Cuisine Classifier

📊 Dataset Preparation

• Data Collection: Images were gathered using DuckDuckGo and organized into a dynamically generated folder structure.
• DataLoader: Leveraged the fastai DataBlock API to create an efficient DataLoader for the model.
• Data Augmentation: Applied GPU-based default data augmentation provided by fastai, optimizing training performance.

➡️ For more details, refer to the notebook: Food_Classifier_Data_cleaning_to_Model_train.ipynb.

---

🏋️‍♂️ Training and Data Cleaning

• Model: Fine-tuned a ResNet50 model on the dataset.
• Performance: Achieved over 90% accuracy after 7 epochs of training.

---

🚀 Model Deployment

• Deployment: The trained model was deployed on HuggingFace Spaces using a Gradio interface.

• Link: Check out the deployed app here.

• Example UI:

  

---

🌐 API Integration with GitHub Pages

• The model’s API is integrated into a GitHub Pages site. Explore the live classifier here.
• The necessary files for GitHub Pages can be found in the repository under the main branch.