HomeAI Tutorial

Vit-on-small-data

Public

lightest Vision Transformer (ViT) trained from scratch to achieve 93.37 ± 0.07%” top-1 accuracy on CIFAR-10 within just 50 epochs.

Creat2025-09-11T13:07:43
Update2025-09-18T10:01:38
4
Stars
0
Stars Increase