Powerful Design of Small Vision Transformer on CIFAR10

Gent Wu

Vision Transformers (ViTs) have demonstrated remarkable success on large-scale datasets, but their performance on smaller datasets often falls short of convolutional neural network…