Visual Question Answering on Infoseek human (Unseen Question)
33.6AccuracyAVIS
Evaluation results
| Method | Links | |
|---|---|---|
| AVISLearning Protocol=few-shot2023.06 | 33.6 | |
| FiDLearning Protocol=fine-tune, External Knowledge=CLIP2023.06 | 18.9 | |
| PALMLearning Protocol=few-shot, External Knowledge=CLIP2023.06 | 15.6 | |
| PALILearning Protocol=fine-tune, External Knowledge=None2023.06 | 13.3 | |
| OFALearning Protocol=fine-tune, External Knowledge=None2023.06 | 6.2 | |
| PALMModality=Q-only, Learning Protocol=few-shot, External Knowledge=None2023.06 | 4.8 |