Label Anything: Multi-Class Few-Shot Semantic Segmentation with Visual Prompts
Paper • 2407.02075 • Published • 1
Label Anything
Label Anything performs multi-class few-shot semantic segmentation using visual prompts (bboxes, points, masks)
-
-
pasqualedem/label_anything_sam_1024_coco
0.1B • Updated • 22 -
pasqualedem/label_anything_256_sam_1024_coco
99.8M • Updated • 3 -
pasqualedem/label_anything_coco_fold0_mae_7a5p0t63
Image Segmentation • 95.3M • Updated • 12 -
pasqualedem/label_anything_256_oneway_sam_1024_coco
98.9M • Updated • 5 -
pasqualedem/label_anything_mae_480_coco
96.6M • Updated • 8