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

mask2former:universal-segmentation

vision_classification-backbone mask2former universal validation pending

Architecture diagram

Rendered with TorchLens / Graphviz

Open SVG
Unrolled - full
mask2former:universal-segmentation monster graph thumbnail

Monster graph detected. Inline rendering is lazy-skipped for page weight; OpenSeadragon tiling is a TODO for the Tier-B atlas.

Open full SVG