Cross-Dataset Image Classification & Generalization using Fine-Tuned ResNet50

This project presents a production-oriented deep learning pipeline for image classification using transfer learning with ResNet50. The goal is to evaluate model generalization across datasets of varying complexity by comparing performance on:

CIFAR-10 (10 classes)

CIFAR-100 (100 classes)

The project demonstrates end-to-end AI engineering, including data processing, model design, fine-tuning, evaluation, error analysis, and deployment.