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

nocaps

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image Captioningnocaps (val)
CIDEr (Overall)126.9
115
Image CaptioningNoCaps
CIDEr127
111
Image CaptioningNoCaps (test)
CIDEr (overall)126.4
61
Image-Text Alignment Evaluationnocaps out-of-domain (val)
CLIPScore84.1
40
Image CaptioningNoCaps
CIDEr (in-domain)111.3
36
Image CaptioningNoCaps 1.0 (val)
Overall Score127
32
Caption Matching and RetrievalNoCaps (val)
Matching Accuracy99.5
26
Image Captioningnocaps standard (test)
CIDEr124.8
26
Scene captioningnocaps RGBP seen scenes (val)
CIDEr102.67
22
Caption Evaluationnocaps
Win Rate71.1
20
Object Hallucination Detectionnocaps FOIL (Out-Domain)
AP89.1
17
Object Hallucination Detectionnocaps-FOIL (Near-Domain)
AP92.6
17
Object Hallucination Detectionnocaps FOIL In-Domain
AP88.8
17
Object Hallucination Detectionnocaps-FOIL (Overall)
AP91.1
17
Text-to-image retrievalNoCaps
Recall@176.2
17
Image-to-text retrievalNoCaps
R@190.9
17
Out-of-domain Image CaptioningNoCaps
CIDEr1.055
16
Novel Object CaptioningNoCaps (val)
CIDEr (In-Domain)85.4
16
Image CaptioningNocaps
CIDEr83.7
15
Image CaptioningNocaps
Primary Score109.37
14
Text-to-text retrievalnocaps
mAP43.7
12
Image CaptioningNoCaps 4,500 (test)
CIDEr122.1
12
Image CaptioningNocaps
Clean CIDEr105.7
10
Image CaptioningNoCaps
BLEU-447.7
9
Image Captioningnocaps XD (val)
CIDEr106.8
8
Showing 25 of 42 rows