Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Text-to-image synthesis evaluation (Human Correlation - Error Counting) on COCO 2017
Loading...
0.2138
Kendall's Tau
LLMScore
-0.094768
-0.014659
0.06545
0.145559
May 18, 2023
Kendall's Tau
Spearman's Rho
Updated 3d ago
Evaluation Results
Method
Method
Links
Kendall's Tau
Spearman's Rho
LLMScore
2023.05
0.2138
0.3125
CLIP
2023.05
0.1888
0.2677
BLIP-ITC
2023.05
0.1849
0.262
NegCLIP
2023.05
0.1795
0.2581
BLIP-ITM
2023.05
0.1137
0.1635
DescCLIP
type=LLMScore variant
2023.05
0.0338
0.0477
CapMETEOR
type=LLMScore variant
2023.05
0.0004
0.0013
CapCLIP
type=LLMScore variant
2023.05
-0.0068
-0.0085
DescMETEOR
type=LLMScore variant
2023.05
-0.0829
-0.1198
Feedback
Search any
task
Search any
task