Visual Question Answering On Gqa
Metriken
Accuracy
Ergebnisse
Leistungsergebnisse verschiedener Modelle zu diesem Benchmark
Modellname | Accuracy | Paper Title | Repository |
|---|---|---|---|
| RelViT | 65.54 | RelViT: Concept-guided Vision Transformer for Visual Relational Reasoning | |
| PEVL+ | 77 | PEVL: Position-enhanced Pre-training and Prompt Tuning for Vision-language Models |
0 of 2 row(s) selected.