Vision transformers

Level 6: Transformers and multimodal vision

Coming soon

This page is being written. See the full roadmap for the whole syllabus.

What this chapter will cover

• Images as sequences of patches
• patch embedding
• position embeddings
• the class token
• ViT from scratch
• inductive bias and data hunger
• DeiT and distillation
• hierarchical transformers (Swin)
• CNNs vs ViTs