Image Captioning on COCO Capt. (CIDEr, Log-PPL)
126CIDErViT-g/14
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ViT-g/14mode=Multitask decoding2023.05 | 126 | 27.9 | |
| SoViT-400Mmode=Multitask decoding2023.05 | 125 | 28.1 | |
| ViT-L/16mode=Multitask decoding2023.05 | 120 | 28.7 |
Loading the SOTA2 catalog…