Image Classification on Pets (Acc. and ECE)
94.5AccuracyCLIP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CLIPPre-training dataset=RedCaps, Pre-training data source=Real, Evaluation protocol=Few-shot2023.06 | 94.5 | — | |
| StableRepPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 94.5 | — | |
| OpenAI-CLIP VIT-LN=400M, Zero-shot=true2024.09 | 93.5 | — | |
| DINOv2-ARLN=20M, Zero-shot=true2024.09 | 92.1 | — | |
| CLIPPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 92 | — | |
| DINOv2-MpNetN=20M, Zero-shot=true2024.09 | 91.7 | — | |
| StableRepPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 91.7 | — | |
| LAION-CLIP VIT-LN=400M, Zero-shot=true2024.09 | 91.5 | — | |
| CLIPPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 91.5 | — | |
| CLIPPre-training dataset=CC12M, Pre-training data source=Real, Evaluation protocol=Few-shot2023.06 | 91.1 | — | |
| OmniVLImg-text pairs=14M*, Evaluation protocol=Linear probing2022.09 | 87.1 | — | |
| CLIP-ViT-B/16Img-text pairs=400M, Evaluation protocol=Linear probing2022.09 | 86.7 | — | |
| SimCLRPre-training dataset=RedCaps, Pre-training data source=Real, Evaluation protocol=Few-shot2023.06 | 85.4 | — | |
| FLAVAImg-text pairs=70M, Evaluation protocol=Linear probing2022.09 | 84.8 | — | |
| TPTBackbone=CLIP-RN502025.03 | 84.5 | 3.65 | |
| C-TPTBackbone=CLIP-RN502025.03 | 84.1 | 2.77 | |
| Robust-adapt-Penalty-CTPTBackbone=CLIP-RN502025.03 | 83.84 | 3.3 | |
| Zero ShotBackbone=CLIP-RN502025.03 | 83.8 | 5.91 | |
| Robust-adapt-SaLs-CTPTBackbone=CLIP-RN502025.03 | 83.73 | 3.6 | |
| Robust-adapt-ZS-CTPTBackbone=CLIP-RN502025.03 | 83.61 | 6.11 | |
| O-TPTBackbone=CLIP-RN502025.03 | 83.4 | 3.5 | |
| LiT L16LN=112M, Zero-shot=true2024.09 | 83.3 | — | |
| FLAVAImg-text pairs=14M, Evaluation protocol=Linear probing2022.09 | 82.3 | — | |
| SimCLRPre-training dataset=RedCaps, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 78.3 | — | |
| SimCLRPre-training dataset=CC12M, Pre-training data source=Syn, Evaluation protocol=Few-shot2023.06 | 74.6 | — | |
| SimCLRPre-training dataset=CC12M, Pre-training data source=Real, Evaluation protocol=Few-shot2023.06 | 74.4 | — | |
| SynerNetZero-shot=true2026.01 | 71.1 | — | |
| ZSNLZero-shot=true2026.01 | 68.7 | — | |
| ALBEFImg-text pairs=14M, Evaluation protocol=Linear probing2022.09 | 68.4 | — | |
| BLIPImg-text pairs=14M, Evaluation protocol=Linear probing2022.09 | 65.3 | — | |
| METER-CLIP-B/16Img-text pairs=400M+4M, Evaluation protocol=Linear probing2022.09 | 40.4 | — | |
| TransCLIPZero-shot=true2026.01 | 20.9 | — | |
| OpenCLIPZero-shot=true2026.01 | 18.7 | — |