Image Captioning on Conceptual Captions 2018 (val)
0.572Win RateCLIP-S + Grammar
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CLIP-S + GrammarCriteria=Attribute, Comparison Baseline=MLE2022.05 | 0.572 | 0.368 | 0.06 | |
| CLIP-S + GrammarCriteria=Attribute, Comparison Baseline=CIDEr2022.05 | 0.558 | 0.372 | 0.07 | |
| CLIP-S + GrammarCriteria=Object, Comparison Baseline=CIDEr2022.05 | 0.552 | 0.328 | 0.12 | |
| CLIP-S + GrammarCriteria=Background, Comparison Baseline=CIDEr2022.05 | 0.539 | 0.254 | 0.206 | |
| CLIP-S + GrammarCriteria=Background, Comparison Baseline=MLE2022.05 | 0.528 | 0.35 | 0.122 | |
| CLIP-S + GrammarCriteria=Object, Comparison Baseline=MLE2022.05 | 0.52 | 0.366 | 0.114 | |
| CLIP-S + GrammarCriteria=Overall, Comparison Baseline=CIDEr2022.05 | 0.51 | 0.308 | 0.182 | |
| CLIP-S + GrammarCriteria=Relation, Comparison Baseline=CIDEr2022.05 | 0.492 | 0.396 | 0.112 | |
| CLIP-S + GrammarCriteria=Overall, Comparison Baseline=MLE2022.05 | 0.49 | 0.418 | 0.092 | |
| CLIP-S + GrammarCriteria=Relation, Comparison Baseline=MLE2022.05 | 0.446 | 0.442 | 0.112 |