Slicing Vision Transformer for Flexibile Inference | Read Paper on Bytez