What is Image Encoder (CLIP / ViT)? Why does it exist? Where does it appear in Multi-Modal Models?
Draw or describe a visual model of Image Encoder (CLIP / ViT). What does it look like conceptually?
How is Image Encoder (CLIP / ViT) different from related concepts in Multi-Modal Models? What are the trade-offs?
Explain Image Encoder (CLIP / ViT) in plain language using an analogy. Could a non-expert follow it?