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

Flickr8K

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image Captioning EvaluationFlickr8k-CF
Kendall-b Correlation (tau_b)56.6
99
Image Captioning EvaluationFlickr8k Expert
Kendall Tau-c (tau_c)60.8
82
Image Captioning EvaluationFlickr8K Expert (test)
Kendall tau_c56.4
76
Image SearchFlickr8K
R@13,100
74
Image Captioning EvaluationFlickr8K-CF (test)
Kendall tau_b38.8
65
Text RetrievalFlickr8K (test)
R@584.2
31
Image CaptioningFlickr8K (test)
BLEU@438.3
27
Correlation with Human JudgmentFlickr8K-CF
Tau B37.8
26
Image CaptioningFlickr8k-EX
Tau-c0.597
22
Image-to-Text RetrievalFlickr8k
R@158.5
22
Text-to-Image RetrievalFlickr8K-CN
R@170.1
19
Image-to-Text RetrievalFlickr8K CN
R@183.3
19
Image AnnotationFlickr8K
R@143.4
18
Correlation with human judgmentsFlickr8K (Expert)
Kendall's Tau (τc)56.4
17
Correlation with human judgmentFlickr8K Expert 2013 (full)
Kendall's Tau53
14
Multimodal AlignmentFLICKR8K
Delta+ Mean Distance0.503
12
Image SearchFlickr8k (test)
R@141
11
Image CaptioningFlickr8K (val)
Masked Accuracy39.38
10
Text-to-Image RetrievalFlickr8k Rephrased
Recall@595.9
6
Image-to-Text RetrievalFlickr8k-Rephrased
Recall@589.4
6
Image CaptioningFlickr8k
BLEU@438.4
6
Image RetrievalFlickr8k zero-shot
R@144.4
6
Text RetrievalFlickr8k zero-shot
R@158.5
6
Vision-Language Factuality ControlFlickr8k (test)
ECR97.27
5
Conditional Image GenerationFlickr8k
FID31.15
5
Showing 25 of 28 rows