Image Classification on ImageNet-R (20 tasks)
89.9AccuracyTRM
Evaluation results
| Method | Links | |
|---|---|---|
| TRMBackbone=ViT-L/14, Pre-training=WebImageText2026.05 | 89.9 | |
| MagMAXBackbone=ViT-L/14, Pre-training=WebImageText2026.05 | 89.1 | |
| BECAMEBackbone=ViT-L/14, Pre-training=WebImageText2026.05 | 87.1 | |
| TIESBackbone=ViT-L/14, Pre-training=WebImageText2026.05 | 86.9 | |
| Model StockBackbone=ViT-L/14, Pre-training=WebImageText2026.05 | 86.7 | |
| TRMBackbone=ViT-B/16, Pre-trained dataset=LAION-400M2026.05 | 82.5 | |
| MagMAXBackbone=ViT-B/16, Pre-trained dataset=LAION-400M2026.05 | 79.2 | |
| TIESBackbone=ViT-B/16, Pre-trained dataset=LAION-400M2026.05 | 78.4 | |
| Model StockBackbone=ViT-B/16, Pre-trained dataset=LAION-400M2026.05 | 76.4 | |
| BECAMEBackbone=ViT-B/16, Pre-trained dataset=LAION-400M2026.05 | 76.1 |