food vision

Developed an end-to-end image classification system to distinguish

between pizza, steak, and sushi images.

Built and compared EfficientNet-B0 (89% accuracy) and Vision

Transformer (ViT, 97% accuracy) with Hugging Face.

Applied data augmentation (Albumentations) and monitored training

with Tensorboard.

Deployed the ViT model as a Gradio web app for real-time predictions

with confidence scores