Fast Food Classification Dataset - V2

Welcome to the Fast Food Classification Dataset - V2 repository, created for use in our upcoming workshop on machine learning and computer vision!

Workshop Details

• Title: Practical Machine Learning Workshop: Fast Food Classification
• Date: 19th-21se March, 2024
• Venue: LNCT Conference Hall
• Organizer: MLSA LNCTU

Overview

The Fast Food Classification Dataset - V2 is specifically curated for our workshop, where participants will learn hands-on techniques for building and training machine learning models to classify images of various fast food items. This dataset provides a diverse collection of 20,000 images, including burgers, sandwiches, donuts, hot dogs, pizzas, baked potatoes, fries, crispy chicken, tacos, and taquitos.

Dataset Structure

• Version 2: This version extends the previous version (V1) by introducing new classes such as Baked Potato, Crispy Chicken, Fries, Taco, and Taquito.

  • Tensorflow Records: Contains training, validation, and testing data in Tensorflow Records format, each resized to 256 by 256 pixels.

    • Train: 15,000 training images (1,500 images per class).
    • Valid: 3,500 validation images (400 images per class).
    • Test: 1,500 validation images (100/200 images per class).

• Version 1: This version contains images of five different types of fast food: Burger, Donut, Hot Dog, Pizza, and Sandwich.

  • Tensorflow Records: Contains training and validation images resized to 256 by 256 pixels.

    • Training Images: 7,500 training images (1,500 images per class).
    • Validation Images: 2,500 validation images (500 images per class).

Usage

This dataset is exclusively prepared for our workshop participants, providing them with the necessary resources to learn and implement machine learning algorithms for image classification tasks. During the workshop, attendees will explore the dataset, understand the fundamentals of deep learning, and build their own fast food classification models.

License

The Fast Food Classification Dataset - V2 is released under the CC0: Public Domain license, allowing for unrestricted use and distribution.

Workshop Agenda

• Introduction to Machine Learning and Computer Vision
• Exploratory Data Analysis of the Fast Food Classification Dataset
• Building Convolutional Neural Networks (CNNs) for Image Classification
• Training and Evaluating Fast Food Classification Models
• Fine-tuning Models and Hyperparameter Optimization
• Deploying Models for Inference

Feedback

We value your feedback and suggestions for improving our workshop and dataset. Please feel free to share your thoughts and ideas with us during the event.

References

For more information and updates about the workshop, visit https://www.kaggle.com/datasets/utkarshsaxenadn/fast-food-classification-dataset for this dataset.