What is Self-Attention on Patches? Why does it exist? Where does it appear in Vision Transformers (ViT)?
Draw or describe a visual model of Self-Attention on Patches. What does it look like conceptually?
How is Self-Attention on Patches different from related concepts in Vision Transformers (ViT)? What are the trade-offs?
Explain Self-Attention on Patches in plain language using an analogy. Could a non-expert follow it?