Built and trained an image classification model utilizing the ResNet50 architecture

️ Project Overview

Dataset: 20,000+ labeled images, covering (10) distinct fast food classes

(? Baked Potato, Burger, Crispy Chicken, Donut, Fries, Hot Dog, Pizza, ? Sandwich, Taco, Taquito).

Train/Validation Split:

Used 65% of the data for training and 35% for validation with stratification to maintain class balance .

️ Model Architecture

Base Model: ResNet50 pretrained on ImageNet, with the top removed

Fine-Tuning: Fully trainable ResNet50 backbone.

Results

Training Accuracy 90.26%

Validation Accuracy 86.11%