Referring Expression Comprehension on GRIT refexp
78.61AccuracyUnified-IO
Evaluation Results
| Method | Links | |
|---|---|---|
| Unified-IOModel type=Generalist Models2023.08 | 78.61 | |
| Unified-IOModel type=Generalist VL SOTAs (w/o finetuning)2023.06 | 78.6 | |
| Qwen-VL-7BModel type=Generalist Models2023.08 | 78.22 | |
| Shikra-7BModel type=Generalist VL SOTAs (w/o finetuning)2023.06 | 69.34 | |
| Shikra-7BModel type=Generalist Models2023.08 | 69.34 | |
| Upper BoundToken=2562024.06 | 69.34 | |
| Shikra-13BModel type=Generalist VL SOTAs (w/o finetuning)2023.06 | 69.03 | |
| Shikra-13BModel type=Generalist Models2023.08 | 69.03 | |
| VoCo-LLaMAToken=82024.06 | 68.59 | |
| VoCo-LLaMAToken=12024.06 | 67.69 | |
| OFA-L*Model type=Generalist VL SOTAs (w/o finetuning)2023.06 | 61.7 | |
| OFA-LModel type=Generalist Models2023.08 | 61.7 | |
| Lower BoundToken=12024.06 | 60.5 | |
| GPV-2Model type=Generalist VL SOTAs (w/o finetuning)2023.06 | 51.5 | |
| GPV-2Model type=Generalist Models2023.08 | 51.5 |