Visual Commonsense Tests On Vicomte Color
Metriken
Spearman's Rho
Ergebnisse
Leistungsergebnisse verschiedener Modelle zu diesem Benchmark
Modellname | Spearman's Rho | Paper Title | Repository |
|---|---|---|---|
| Z-LaVI (GPT-J-6B) | 49.6 | Z-LaVI: Zero-Shot Language Solver Fueled by Visual Imagination | - |
| BERT-large | 37.6 | Visual Commonsense in Pretrained Unimodal and Multimodal Models |
0 of 2 row(s) selected.