Here are model checkpoints for object recognition models powered by SmoothSA and SPOT.
Using DINO2 S/14 for encoding.
Models are trained on dataset COCO and YTVIS (high quality), with random seeds 42, 43 and 44.
Input resolution is 256x256 (224x224).
Slot matching threshold is 1e-3@IoU for COCO and 1e-1@IoU for YTVIS.