Generalized Referring Expression Comprehension on gRefCOCO (testB)
0.4461Pr@F1ReLA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ReLAVisual Encoder=Swin-B, Textual Encoder=BERT2026.01 | 0.4461 | 0.584 | 0.9589 | |
| UNINEXTVisual Encoder=ResNet-50, Textual Encoder=BERT2026.01 | 0.4291 | 0.4822 | 0.9816 | |
| ReLAVisual Encoder=ResNet-50, Textual Encoder=BERT2026.01 | 0.4285 | 0.5781 | 0.9544 | |
| MDETRVisual Encoder=ResNet-101, Textual Encoder=RoBERTa2026.01 | 0.3652 | 0.3102 | 0.9963 | |
| VLTVisual Encoder=DarkNet-53, Textual Encoder=GRU2026.01 | 0.3024 | 0.3253 | 0.9956 | |
| TransVGVisual Encoder=ResNet-101, Textual Encoder=BERT2026.01 | 0.2844 | 0.3278 | 0.9959 | |
| MCNVisual Encoder=DarkNet-53, Textual Encoder=GRU2026.01 | 0.2676 | 0.3027 | 0.998 |