Porting vision models to Keras 3 for easily accessibility. Contains MobileViT v1, MobileViT v2, fastvit
FastViT base model for use with Autodistill.