Video Question Answering on NEXT-QA open-form generation
34.3WUPSLLaMA-VQA
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaMA-VQA# external visual-text data samples=02023.10 | 34.3 | |
| Flamingo# external visual-text data samples=2.1B, evaluation protocol=32-shot2023.10 | 33.5 | |
| KcGA# external visual-text data samples=02023.10 | 28.2 | |
| Flamingo# external visual-text data samples=2.1B, evaluation protocol=0-shot2023.10 | 26.7 | |
| HGA# external visual-text data samples=02023.10 | 25.2 |