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

MSCOCO

Benchmarks

Task NameDataset NameSOTA ResultTrend
Text-to-Image RetrievalMSCOCO 5K (test)
R@160.9
308
Image-to-text retrievalMSCOCO
R@181.9
129
Text-to-image retrievalMSCOCO
R@164.3
123
Text-to-Image RetrievalMSCOCO 1K (test)
R@16,390
118
Visual Hallucination EvaluationMSCOCO
CHAIR_i18.2
104
Sentence RetrievalMSCOCO 5k (test)
R@180.9
67
Image-to-text retrievalMSCOCO 5K (test)
R@184.8
64
Object Hallucination EvaluationMSCOCO 2014 (val)
CHAIRs54.6
55
Text-to-Image GenerationMSCOCO 30K
FID6.61
54
Text-to-Image RetrievalMSCOCO (val)
R@138.97
51
Image-to-Text RetrievalMSCOCO (val)
R@158.14
51
Object HallucinationMSCOCO 500 images 2014 (val)
Consistency Score (CS)60.6
50
Object Hallucination EvaluationMSCOCO POPE
Random Accuracy91.63
47
Text-to-Image GenerationMSCOCO 2014
FID (30k)9.29
44
Object DetectionMSCOCO (val)
AP61.3
43
Text-to-image RetrievalMSCOCO (5K)
R@153.98
42
Pointing gameMSCOCO 2014 (val)
Mean Accuracy (All)69.9
42
Object Hallucination EvaluationMSCOCO
Accuracy88.97
41
Image RetrievalMSCOCO @5000 (test)
mAP87.27
39
Object Hallucination AssessmentMSCOCO
CHAIR Instance Score30.2
38
Hallucination EvaluationMSCOCO (val)
CHAIR_i23.04
36
Object Hallucination EvaluationMSCOCO
CHAIR Scene Score56.4
35
Text-to-image generationMSCOCO 30k samples 2014 (val)
FID21.96
35
Text RetrievalMSCOCO
ASR@R1100
33
Object DetectionMSCOCO 2017 (val)
APb47.5
33
Showing 25 of 191 rows
...