Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

ReasonSeg

Benchmarks

Task NameDataset NameSOTA ResultTrend
Reasoning SegmentationReasonSeg (val)
gIoU79.3
382
Reasoning SegmentationReasonSeg (test)
gIoU73.8
287
Reasoning SegmentationReasonSeg long (test)
cIoU0.733
40
Reasoning SegmentationReasonSeg short (test)
cIoU59.4
40
Reasoning SegmentationReasonSeg
gIoU79.3
35
Visual GroundingReasonSeg
cIoU (Overall)61.6
15
Reasoning SegmentationReasonSeg zero-shot (val)
gIoU63.5
11
Reasoning SegmentationReasonSeg zero-shot (test)
gIoU60.3
10
Referring Expression SegmentationReasonSeg (test)
mIoU64.2
9
Referring Expression SegmentationReasonSeg (val)
mIoU66.3
9
Reasoning SegmentationReasonSeg (combined)
Average Score68.6
6
Reasoning segmentationREASONSEG HARD (test)
gIoU58.6
6
Showing 12 of 12 rows