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

COCO

Benchmarks

Task NameDataset NameSOTA ResultTrend
Object DetectionCOCO 2017 (val)
AP64.6
2,643
Object DetectionCOCO (test-dev)
mAP66
1,239
Instance SegmentationCOCO 2017 (val)
APm50.3
1,201
Object DetectionCOCO (val)
mAP68.3
633
Object DetectionCOCO v2017 (test-dev)
mAP65.4
499
Instance SegmentationCOCO (val)
APmk55
475
Human Pose EstimationCOCO (test-dev)
AP81.1
432
2D Human Pose EstimationCOCO 2017 (val)
AP79.3
386
Instance SegmentationCOCO (test-dev)
APM58.4
380
Object DetectionCOCO 2017
AP (Box)62.6
321
Image-to-Text RetrievalMS-COCO 5K (test)
R@185.4
320
Pose EstimationCOCO (val)
AP79.6
319
Semantic SegmentationCOCO 2014 (val)
mIoU61.8
304
Instance SegmentationCOCO
APmask54.8
291
Instance SegmentationCOCO 2017 (test-dev)
AP (Overall)53
253
Object DetectionCOCO
AP50 (Box)73.2
237
Instance SegmentationCOCO 2017
APm53
226
Panoptic SegmentationCOCO (val)
PQ59.5
219
Panoptic SegmentationCOCO 2017 (val)
PQ59.9
185
Object DetectionCOCO (minival)
mAP57.6
184
Human Pose EstimationCOCO 2017 (test-dev)
AP78.3
180
Concept AttributionCOCO
Average Attribution F155
178
Few-Shot Semantic SegmentationCOCO-20i
mIoU75.8
178
Few-shot SegmentationCOCO 20^i (test)
mIoU60
174
Panoptic SegmentationCOCO (test-dev)
PQ56.2
162
Showing 25 of 1172 rows
...