Visual Reasoning on NLVR2 (test-P)
92.6AccuracyBEiT-3
Evaluation results
| Method | Links | |
|---|---|---|
| BEiT-3# Params=1.9B2023.01 | 92.6 | |
| X-FMbase# Params=327M, Training Data=More Data2023.01 | 88.4 | |
| X-FMModel size=base-size2023.01 | 88.4 | |
| X2-VLMbase# Params=255M, Training Data=More Data2023.01 | 87 | |
| CoCa# Params=2.1B2023.01 | 87 | |
| X2-VLMModel size=base-size2023.01 | 87 | |
| X-FMbase# Params=327M2023.01 | 86.5 | |
| X2-VLMbase# Params=255M2023.01 | 86.1 | |
| X-VLMModel size=base-size2023.01 | 84.8 | |
| X-VLM# Params=216M2023.01 | 84.2 | |
| VLMbase# Params=175M2023.01 | 83.3 | |
| METER# Params=341M2023.01 | 83.1 | |
| BLIPbase# Params=240M2023.01 | 83.1 | |
| VL-BEiT# Params=175M2023.01 | 82.7 | |
| SimVLMbase# Params=273M, Training Data=More Data2023.01 | 81.8 | |
| SimVLMModel size=base-size2023.01 | 81.8 | |
| ALBEF# Params=210M2023.01 | 80.5 | |
| DaVinciModel size=base-size2023.01 | 77.9 | |
| MH-MoEExperts Number=82024.04 | 77.8 | |
| X-MoEExperts Number=82024.04 | 76.1 | |
| Dense2024.04 | 74.2 |