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

MS-COCO

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image CaptioningMS COCO Karpathy (test)
CIDEr149.1
706
Object DetectionMS COCO (test-dev)
mAP@.578.5
677
Object DetectionMS-COCO 2017 (val)
mAP74
264
Text-to-Image RetrievalMS-COCO 5K (test)
R@168.3
244
Object DetectionMS COCO (val)
mAP0.603
222
Image RetrievalMS-COCO 5K (test)
R@167.2
217
Text-to-image generationMS-COCO (val)
FID1.53
202
Object Hallucination EvaluationMS-COCO (POPE Adversarial)
Accuracy87.83
190
Text-to-Image RetrievalMS-COCO
R@165.7
187
Text RetrievalMS-COCO 5K (test)
R@184.8
182
Image RetrievalMS-COCO
R@1100
172
Image-to-Text RetrievalMS-COCO
R@180.5
168
Object Hallucination EvaluationMS-COCO POPE (Popular)
Accuracy90.76
158
Text-to-image GenerationMS-COCO
FID4.52
145
Text-to-image generationMS-COCO 2014 (val)
FID2.47
143
Object DetectionMS COCO novel classes
nAP2,450
132
Text-to-Image GenerationMS-COCO 2017 (val)
FID19.2591
131
Image RetrievalMS-COCO 1K (test)
R@180.1
128
Object DetectionMS COCO novel classes 2017 (val)
AP22.73
123
Object Hallucination EvaluationMS-COCO POPE Random
Accuracy92.36
121
Image-to-Text RetrievalMS-COCO 1K (test)
R@182
121
Image CaptioningMS COCO (test)
CIDEr140.4
120
Instance SegmentationMS COCO
mAP Mask47.9
111
Image RetrievalMS-COCO (test)
MAP84.08
98
Object DetectionMS-COCO 2017 (test)
AP53.9
82
Showing 25 of 449 rows
...