STEP (5-fold)

Reported on 6 benchmarks across 2 tasks

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision6 results

Instance SegmentationonRefCOCO+ val
Overall IoU
48.18
best: 79.4 (MLCD-Seg-7B)
Instance SegmentationonRefCOCO+ test B
Overall IoU
40.41
best: 75.6 (MLCD-Seg-7B)
Instance SegmentationonRefCOCO+ testA
Overall IoU
52.33
best: 83.5 (HyperSeg)
Referring Expression SegmentationonRefCOCO+ val
Overall IoU
48.18
best: 79.4 (MLCD-Seg-7B)
Referring Expression SegmentationonRefCOCO+ test B
Overall IoU
40.41
best: 75.6 (MLCD-Seg-7B)
Referring Expression SegmentationonRefCOCO+ testA
Overall IoU
52.33
best: 83.5 (HyperSeg)