CanViT
community
AI & ML interests
None defined yet.
Recent Activity
View all activity
Organization Card
CanViT (Canvas Vision Transformer) is a scalable recurrent architecture for the Active-Vision Foundation Model (AVFM) era.
See https://github.com/m2b3/CanViT-PyTorch and CanViT: Toward Active-Vision Foundation Models for more.
Linear segmentation probes trained on frozen CanViT canvas features for ADE20K semantic segmentation (150 classes).
-
canvit/probe-ade20k-40k-s512-c8-in21k
Image Segmentation • 160k • Updated • 5 -
canvit/probe-ade20k-40k-s512-c9-in21k
Image Segmentation • 160k • Updated • 3 -
canvit/probe-ade20k-40k-s512-c10-in21k
Image Segmentation • 160k • Updated • 3 -
canvit/probe-ade20k-40k-s512-c12-in21k
Image Segmentation • 160k • Updated • 3
Linear segmentation probes trained on frozen CanViT canvas features for ADE20K semantic segmentation (150 classes).
-
canvit/probe-ade20k-40k-s512-c8-in21k
Image Segmentation • 160k • Updated • 5 -
canvit/probe-ade20k-40k-s512-c9-in21k
Image Segmentation • 160k • Updated • 3 -
canvit/probe-ade20k-40k-s512-c10-in21k
Image Segmentation • 160k • Updated • 3 -
canvit/probe-ade20k-40k-s512-c12-in21k
Image Segmentation • 160k • Updated • 3
models 46
canvit/canvitb16-abl-vit-s-2026-03-03
29M • Updated • 6
canvit/canvitb16-abl-rw-stride6-2026-03-03
91.8M • Updated • 6
canvit/canvitb16-abl-no-vpe-2026-03-03
98.1M • Updated • 6
canvit/canvitb16-abl-no-bptt-2026-03-06
98.1M • Updated • 6
canvit/canvitb16-abl-no-fiid-2riid-2026-03-06
98.1M • Updated • 7
canvit/canvitb16-abl-no-reads-2026-03-02
93.4M • Updated • 6
canvit/canvitb16-abl-no-fiid-1riid-2026-03-02
98.1M • Updated • 7
canvit/canvitb16-abl-no-dense-2026-03-02
98.1M • Updated • 6
canvit/canvitb16-abl-dcan256-2026-03-02
90.4M • Updated • 6
canvit/canvitb16-abl-qkvo-dcan384-2026-03-02
93.5M • Updated • 6
datasets 0
None public yet