AbdisamadO/Somali_Food_Dataset_Mini

Dataset Card for Somali Food Mini

Dataset Summary

Somali Dataset Mini is a balanced collection of traditional Somali food and drink images.
It showcases the culinary diversity of Somalia through six main meal types and 29 food categories.

Each class contains approximately 100 images, making it ideal for:

• Computer vision experiments
• Image classification projects
• Transfer learning and CNN model training

---

📁 Dataset Structure

This dataset contains images of Somali foods, organized into multiple categories.

It is designed as a mini dataset suitable for experimentation, prototyping, and transfer learning tasks. Larger-scale applications may require additional data, but this dataset is a strong starting point for exploration.

The dataset is organized into train and test directories as follows:

---

Meal Categories

• Quraac (Breakfast)
• Qado (Lunch)
• Casho (Dinner)
• Macmacaan (Desserts)
• Shaah (Tea-time foods)
• Cabitaan (Drinks)

---

Acknowledgment This dataset was collected from open web sources to create a balanced Somali food image collection for research and educational purposes.

Usage

• Image classification
• Computer vision
• Deep learning experiments
• Food recognition tasks

Sample of Using Simple CNN Model

@misc{AbdisamadOmar2025, author = {Abdisamad Omar}, title = {Somali Food Dataset Mini},
year = {2025},
publisher = {Hugging Face}, howpublished = https://huggingface.co/datasets/AbdisamadO/Somali_Food_Dataset_Mini }