Sign inSign in
ConceptAI & ML

MMDiT

An architecture introduced in Stable Diffusion 3 where QKV projections are computed separately for different modalities and concatenated before attention.
as Sayak Paul defines itStanford University
Watch it explained

Know first

Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.

Attention mechanism1Diffusion models1Diffusion transformer1Multi-modality1QKV projections1Transformer1

Lecturers

1

    Connections

    3

    The lectures that define it or assume it, and the concepts next to it.

    1 lecturer. Definitions are written by AI from each lecture’s transcript. A time with a play mark starts the lecture at the moment the lecturer says it.