Visual Question Answering on Visual Genome QA (val)
36.64AccuracyOurs_SAT->ST
Evaluation results
| Method | Links | |
|---|---|---|
| Ours_SAT->STPretraining Data=CC2019.12 | 36.64 | |
| Our SAT->STBackbone=BERTB, Training Strategy=All-Task Pretraining + Single-Task Finetuning, Number of Parameters=3B, Number of Models=12 x 250M2019.12 | 36.64 | |
| Our SATBackbone=BERTB, Training Strategy=All-Task (AT), Number of Parameters=270M, Number of Models=1 x 270M2019.12 | 36.36 | |
| Ours_STPretraining Data=CC2019.12 | 34.38 |