Video Question Answering On Tvqa
评估指标
Accuracy
评测结果
各个模型在此基准测试上的表现结果
模型名称 | Accuracy | Paper Title | Repository |
|---|---|---|---|
| VindLU | 79.0 | VindLU: A Recipe for Effective Video-and-Language Pretraining | |
| FrozenBiLM | 82 | Zero-Shot Video Question Answering via Frozen Bidirectional Language Models | |
| Hero w/ pre-training | 74.24 | HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training | |
| iPerceive (Chadha et al., 2020) | 76.96 | iPerceive: Applying Common-Sense Reasoning to Multi-Modal Dense Video Captioning and Video Question Answering | - |
| LLaMA-VQA | 82.2 | Large Language Models are Temporal and Causal Reasoners for Video Question Answering | |
| STAGE (Lei et al., 2019) | 70.50 | TVQA+: Spatio-Temporal Grounding for Video Question Answering |
0 of 6 row(s) selected.