Robotics
Optimizing CNNs and Vision Transformers with Generative Data Augmentation (DiffuseMix)
A comparative study on improving Image Classification performance on CIFAR-10. This project implements ResNet-32 and Custom Vision Transformers (ViT) from scratch in PyTorch, utilizing Stable Diffusion (DiffuseMix) to generate synthetic training data and boost model accuracy by up to 8%.
December 10, 2025
