Scene Text Visual Question Answering on ST-VQA 8 (val)
0.6164AccuracyLaTr
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LaTrModel Size=Large, Joint fine-tuning=true2021.12 | 0.6164 | 0.702 | |
| LaTrModel Size=Base, Joint fine-tuning=true2021.12 | 0.5909 | 0.683 | |
| LaTrModel Size=Base, Joint fine-tuning=false2021.12 | 0.5841 | 0.675 | |
| TAP2021.12 | 0.5083 | 0.598 | |
| LOGOS2021.12 | 0.4863 | 0.581 | |
| SA-M4C2021.12 | 0.4223 | 0.512 | |
| LaAP-Net2021.12 | 0.3974 | 0.497 | |
| M4C2021.12 | 0.3805 | 0.472 |