Fine-grained Classification on SUN397
78.4Top-1 AccuracyCLIP
Evaluation Results
| Method | Links | |
|---|---|---|
| CLIPPre-training Dataset=WIT-400M, Pre-training Data Type=Real, Backbone=ViT-B/162023.06 | 78.4 | |
| StableRepPre-training Dataset=CC12M, Pre-training Data Type=Syn, Pre-training Schedule=105 epochs, Backbone=ViT-B/162023.06 | 74.3 | |
| StableRepPre-training Dataset=CC12M, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 73 | |
| StableRepPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=105 epochs, Backbone=ViT-B/162023.06 | 72.5 | |
| StableRepPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 70.7 | |
| CLIPPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 67.3 | |
| CoCoOpPre-trained on ImageNet=true2024.10 | 66.9 | |
| ADAPTBackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Transductive2025.08 | 66.66 | |
| CLIPPre-training Dataset=RedCaps, Pre-training Data Type=Real, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 66.2 | |
| Fair Context Learning (FCL)mode=Test-Time Adaptation2026.02 | 65.71 | |
| APMPre-trained on ImageNet=false2024.10 | 65.7 | |
| EnsemblePre-trained on ImageNet=false2024.10 | 65.6 | |
| ZEROmode=Test-Time Adaptation2026.02 | 65.59 | |
| TPTPre-trained on ImageNet=false2024.10 | 65.5 | |
| R-TPTmode=Test-Time Adaptation2026.02 | 65.5 | |
| TPTmode=Test-Time Adaptation2026.02 | 65.45 | |
| MTAmode=Test-Time Adaptation2026.02 | 65.28 | |
| RLCFmode=Test-Time Adaptation2026.02 | 65.27 | |
| TTLmode=Test-Time Adaptation2026.02 | 65.07 | |
| ADAPTBackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Online2025.08 | 64.94 | |
| TPSmode=Test-Time Adaptation2026.02 | 64.68 | |
| C-TPTmode=Test-Time Adaptation2026.02 | 64.58 | |
| DMNBackbone=CLIP-RN50, BP-free=false, Evaluation Protocol=Online2025.08 | 64.39 | |
| StatABackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Transductive2025.08 | 64.3 | |
| DPEBackbone=CLIP-RN50, BP-free=false, Evaluation Protocol=Online2025.08 | 64.23 | |
| CoOpPre-trained on ImageNet=true2024.10 | 64.2 | |
| TransCLIPBackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Transductive2025.08 | 64.2 | |
| DotaBackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Online2025.08 | 63.89 | |
| BCABackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Online2025.08 | 63.38 | |
| DiffTPTBackbone=CLIP-RN50, BP-free=false, Evaluation Protocol=Online2025.08 | 62.72 | |
| TPSBackbone=CLIP-RN50, BP-free=false, Evaluation Protocol=Online2025.08 | 62.7 | |
| CLIP-VIT-B/16Pre-trained on ImageNet=false2024.10 | 62.6 | |
| CLIPmode=Zero-shot2026.02 | 62.56 | |
| TDABackbone=CLIP-RN50, BP-free=true, Evaluation Protocol=Online2025.08 | 62.53 | |
| SimCLRPre-training Dataset=RedCaps, Pre-training Data Type=Syn, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 61.8 | |
| TPTBackbone=CLIP-RN50, BP-free=false, Evaluation Protocol=Online2025.08 | 61.46 | |
| SimCLRPre-training Dataset=RedCaps, Pre-training Data Type=Real, Pre-training Schedule=35 epochs, Backbone=ViT-B/162023.06 | 61.2 | |
| C-TPTBackbone=CLIP-RN50, BP-free=false, Evaluation Protocol=Online2025.08 | 61 | |
| CLIPBackbone=CLIP-RN50, BP-free=N/A, Evaluation Protocol=Online2025.08 | 58.8 |