Encoder[[cv-encoder]]
The Vision Transformer (ViT) opened the door to computer vision tasks without convolutions.