Fine-grained Classification on Cars
91.66Accuracyw/ LaViD
Evaluation Results
| Method | Links | |
|---|---|---|
| w/ LaViDStudent=MNV2, Teacher=QN+R502026.06 | 91.66 | |
| LSStudent=MNV2, Teacher=RN-502026.06 | 91.27 | |
| w/ LaViDStudent=RN-18, Teacher=QN+R502026.06 | 91.2 | |
| LSStudent=RN-18, Teacher=RN-502026.06 | 90.98 | |
| MLKDStudent=MNV2, Teacher=RN-502026.06 | 90.3 | |
| MLKDStudent=RN-18, Teacher=RN-502026.06 | 90.23 | |
| w/ LaViDStudent=SNV2, Teacher=QN+R502026.06 | 89.74 | |
| LSStudent=SNV2, Teacher=RN-502026.06 | 89.68 | |
| MLKDStudent=SNV2, Teacher=RN-502026.06 | 89.44 | |
| DKDStudent=RN-18, Teacher=RN-502026.06 | 89.31 | |
| DKDStudent=MNV2, Teacher=RN-502026.06 | 89.3 | |
| LaViDStudent=RN-18, Teacher=Qwen2026.06 | 88.59 | |
| DKDStudent=SNV2, Teacher=RN-502026.06 | 88.4 | |
| MaKDStudent=RN-18, Teacher=InternVL2026.06 | 87.96 | |
| LaViDStudent=MNV2, Teacher=Qwen2026.06 | 87.93 | |
| KDStudent=MNV2, Teacher=RN-502026.06 | 87.33 | |
| KDStudent=RN-18, Teacher=RN-502026.06 | 86.94 | |
| Ind StudentStudent=MNV2, Teacher=None2026.06 | 86.93 | |
| CRDStudent=RN-18, Teacher=LLaVA2026.06 | 86.86 | |
| MaKDStudent=MNV2, Teacher=InternVL2026.06 | 86.83 | |
| FitNetStudent=MNV2, Teacher=LLaVA2026.06 | 86.77 | |
| CLIPPre-training Dataset=WIT-400M, Pre-training Data Type=Real, Backbone=ViT-B/162023.06 | 86.7 | |
| CRDStudent=MNV2, Teacher=LLaVA2026.06 | 86.59 | |
| RKDStudent=MNV2, Teacher=RN-502026.06 | 86.5 | |
| LaViDStudent=SNV2, Teacher=Qwen2026.06 | 86.37 | |
| FitNetStudent=RN-18, Teacher=LLaVA2026.06 | 85.9 | |
| RKDStudent=RN-18, Teacher=RN-502026.06 | 85.78 | |
| Ind StudentStudent=RN-18, Teacher=None2026.06 | 85.77 | |
| Ind StudentStudent=SNV2, Teacher=None2026.06 | 85.67 | |
| MaKDStudent=SNV2, Teacher=InternVL2026.06 | 85.47 | |
| KDStudent=SNV2, Teacher=RN-502026.06 | 85.3 | |
| FitNetStudent=SNV2, Teacher=LLaVA2026.06 | 85.04 | |
| RKDStudent=SNV2, Teacher=RN-502026.06 | 85 | |
| CRDStudent=SNV2, Teacher=LLaVA2026.06 | 84.83 | |
| StableRepPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=105 epochs, Backbone=ViT-B/162023.06 | 83.9 | |
| StableRepPre-training Dataset=CC12M, Pre-training Data Type=Syn, Pre-training Schedule=105 epochs, Backbone=ViT-B/162023.06 | 83.5 | |
| StableRepPre-training Dataset=CC12M, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 80.9 | |
| StableRepPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 78.8 | |
| CLIPPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 77.3 | |
| GIF-SDBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 75.7 | |
| CLIPPre-training Dataset=RedCaps, Pre-training Data Type=Real, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 74.9 | |
| CLIPBackbone=CLIP, Expansion Ratio=1x, Evaluation Protocol=Zero-shot2022.11 | 55.8 | |
| SimCLRPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 53.7 | |
| GIF-DALLEBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 53.1 | |
| SDBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 51.7 | |
| DALL-E2Backbone=ResNet-50, Expansion Ratio=20x2022.11 | 48.3 | |
| GIF-MAEBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 44.5 | |
| RandAugmentBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 43.2 | |
| SimCLRPre-training Dataset=RedCaps, Pre-training Data Type=Real, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 42.8 | |
| GridMaskBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 28.4 | |
| MAEBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 25.9 | |
| CutoutBackbone=ResNet-50, Expansion Ratio=20x2022.11 | 25.8 | |
| OriginalBackbone=ResNet-50, Expansion Ratio=1x2022.11 | 19.8 | |
| Distillation of CLIPBackbone=ResNet-50, Expansion Ratio=1x, Evaluation Protocol=Knowledge Distillation2022.11 | 18.9 |