Step through how the prompt "a red bicycle" is split into tokens, then turned by a text encoder into vectors an image model can use.
Everything you learned in Card 4 is doing the first half of the work here. A text-to-image model is a language model bolted to an image model.
