Preview - rendered architecture diagrams; full validated atlas + dataset coming.
Back to models

VideoMAE V2 pretrain base patch16 (masked video autoencoder ViT)

history videomae v2 validation pending

Architecture diagram

Rendered with TorchLens / Graphviz

Open SVG
Unrolled - full
VideoMAE V2 pretrain base patch16 (masked video autoencoder ViT) architecture diagram