Fine-grained Image Classification on Stanford Dogs
8.8ScoreMaxEnt-CNN
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MaxEnt-CNNBackbone=ResNet-50, Multi-cropping operations=true2020.06 | 8.8 | 88 | — | — | — | — | — | |
| ResNet-50Backbone=ResNet-50, Multi-cropping operations=true2020.06 | 8.3 | 88.1 | — | — | — | — | — | |
| MAMC-CNNBackbone=ResNet-502020.06 | 7.7 | 84.8 | — | — | — | — | — | |
| SEFBackbone=ResNet-502020.06 | 4.8 | 88.8 | — | — | — | — | — | |
| Cross-XBackbone=ResNet-502020.06 | 2.8 | 88.9 | — | — | — | — | — | |
| API-NetBackbone=ResNet-502020.06 | 2.3 | 88.3 | — | — | — | — | — | |
| Gemini 1.5 proOptimization strategy=false2025.12 | 0.6933 | — | — | — | — | — | — | |
| GPT-4oOptimization strategy=false2025.12 | 0.6283 | — | — | — | — | — | — | |
| Step 1vOptimization strategy=false2025.12 | 0.6233 | — | — | — | — | — | — | |
| Claude 3.5 sonnetOptimization strategy=false2025.12 | 0.62 | — | — | — | — | — | — | |
| Doubao 1.5 vision proOptimization strategy=false2025.12 | 0.5933 | — | — | — | — | — | — | |
| InternVL 2.5-8BOptimization strategy=true2025.12 | 0.5717 | — | — | — | — | — | — | |
| LLaVA 1.5-7BOptimization strategy=true2025.12 | 0.53 | — | — | — | — | — | — | |
| GLM v plusOptimization strategy=false2025.12 | 0.5233 | — | — | — | — | — | — | |
| Qwen-VL-chat-78BOptimization strategy=false2025.12 | 0.4867 | — | — | — | — | — | — | |
| Hunyuan visionOptimization strategy=false2025.12 | 0.405 | — | — | — | — | — | — | |
| InternVL 2.5-8BOptimization strategy=false2025.12 | 0.2492 | — | — | — | — | — | — | |
| LLaVA 1.5-7BOptimization strategy=false2025.12 | 0.195 | — | — | — | — | — | — | |
| Basetraining_mode=original model2025.02 | — | — | 12.2 | — | — | — | — | |
| ConvNeXt-TinyBackbone=ConvNeXt-Tiny2026.02 | — | — | — | — | — | — | 92 | |
| DeiT-SmallBackbone=DeiT-Small2026.02 | — | — | — | — | — | — | 89.8 | |
| INFODISENTBackbone=ResNet-502026.02 | — | — | — | — | — | — | 86.6 | |
| L+GEtraining_mode=train with labels and general explanations2025.02 | — | — | 73.45 | 77.89 | 78.15 | 76.55 | — | |
| LUCID PPNBackbone=ConvNeXt-Tiny2026.02 | — | — | — | — | — | — | 79.4 | |
| NLtraining_mode=only train with labels2025.02 | — | — | 82.73 | 82.34 | 84.03 | 84.27 | — | |
| PIP-NETBackbone=ConvNeXt-Tiny2026.02 | — | — | — | — | — | — | 80.8 | |
| PROTOPFORMERBackbone=DeiT-Small2026.02 | — | — | — | — | — | — | 90 | |
| ProtoQuantBackbone=ResNet-502026.02 | — | — | — | — | — | — | 88 | |
| ProtoQuantBackbone=ConvNeXt-Tiny2026.02 | — | — | — | — | — | — | 91.2 | |
| ProtoQuantBackbone=DeiT-Small2026.02 | — | — | — | — | — | — | 89.8 | |
| ResNet-50Backbone=ResNet-502026.02 | — | — | — | — | — | — | 89 | |
| SelfSynthXtraining_mode=rejection sampling with detailed visual knowledge2025.02 | — | — | 85.29 | 86.75 | 86.86 | 86.91 | — |