Image Captioning On Nocaps Val
Metrics
CIDEr
SPICE
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | CIDEr | SPICE |
|---|---|---|
| language-models-are-general-purpose | 58.7 | 8.6 |
| prismer-a-vision-language-model-with-an | 107.9 | 14.8 |
| unifying-vision-and-language-tasks-via-text | 4.4 | 5.3 |