Image Classification on ImageNet 1K (test) (Top-1 Accuracy and GAIN)
87.1Top-1 AccDINOv3
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DINOv3ARCH=ViT-L/16, PRE-TRAINING=LVD-1689M, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 87.1 | 0.5 | |
| SigLIP2ARCH=ViT-L/16, PRE-TRAINING=WebLI, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 87 | 1.8 | |
| SigLIPARCH=ViT-L/16, PRE-TRAINING=WebLI, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 86.1 | 2 | |
| MAEARCH=ViT-L/16, PRE-TRAINING=IN-1K, Evaluation Protocol=FT, # PAR.=304M2025.06 | 85.9 | — | |
| AIMv2ARCH=ViT-L/14, PRE-TRAINING=custom*, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 85.9 | 1.1 | |
| DINOv2ARCH=ViT-L/14, PRE-TRAINING=LVD-142M, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 85.6 | 0.4 | |
| DINOv3ARCH=ViT-B/16, PRE-TRAINING=LVD-1689M, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 84.4 | 0.4 | |
| FrancaARCH=ViT-L/14, PRE-TRAINING=IN-21k, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 84.3 | 0.5 | |
| DINOv2ARCH=ViT-B/14, PRE-TRAINING=LVD-142M, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 84 | 0.8 | |
| MAEARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=FT, # PAR.=87M2025.06 | 83.6 | — | |
| CAPIARCH=ViT-L/14, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 83.6 | 2.1 | |
| CLIPARCH=ViT-L/14, PRE-TRAINING=WIT, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 83.4 | 1.1 | |
| BEiTv2ARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 81.7 | 2.7 | |
| MAEARCH=ViT-S/16, PRE-TRAINING=IN-1K, Evaluation Protocol=FT, # PAR.=22M2025.06 | 80.6 | — | |
| MAEARCH=ViT-L/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=2.1M2025.06 | 79.3 | 3.3 | |
| iBOTARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 79.2 | 0.5 | |
| DINOARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 77.8 | 0.5 | |
| MAEARCH=ViT-L/16, PRE-TRAINING=IN-1K, Evaluation Protocol=LP, # PAR.=1.0M2025.06 | 76 | — | |
| MAEARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 75.6 | 7.9 | |
| BYOLARCH=RN-50, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=6.3M2025.06 | 75.1 | 0.8 | |
| MAEARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=LP, # PAR.=0.8M2025.06 | 67.7 | — | |
| SimMIMARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=1.4M2025.06 | 65.1 | 13.6 | |
| MAEARCH=ViT-S/16, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=0.5M2025.06 | 64.6 | 17.2 | |
| MAEARCH=ViT-L/16, PRE-TRAINING=IN-1K, Evaluation Protocol=k-NN2025.06 | 58.2 | — | |
| DiTARCH=DiT-XL/2, PRE-TRAINING=IN-1K, Evaluation Protocol=EP, # PAR.=2.5M2025.06 | 57 | 24.3 | |
| MAEARCH=ViT-S/16, PRE-TRAINING=IN-1K, Evaluation Protocol=LP, # PAR.=0.4M2025.06 | 47.4 | — | |
| MAEARCH=ViT-B/16, PRE-TRAINING=IN-1K, Evaluation Protocol=k-NN2025.06 | 46.1 | — | |
| MAEARCH=ViT-S/16, PRE-TRAINING=IN-1K, Evaluation Protocol=k-NN2025.06 | 26.7 | — |