Image Classification on CIFAR-100 1.0 (test)
77.78AccuracyReviewKD
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ReviewKDTeacher=ResNet32x4, Student=ShuffleNetV2, Architecture Style=Different2021.04 | 77.78 | — | |
| ReviewKDTeacher=ResNet32x4, Student=ShuffleNetV1, Architecture Style=Different2021.04 | 77.45 | — | |
| ReviewKDTeacher=WRN40-2, Student=ShuffleNetV1, Architecture Style=Different2021.04 | 77.14 | — | |
| ReviewKDTeacher=WRN40-2, Student=WRN16-2, Architecture Style=Same2021.04 | 76.12 | — | |
| OFDTeacher=ResNet32x4, Student=ShuffleNetV12021.04 | 75.98 | — | |
| ReviewKDTeacher=ResNet32x4, Student=ResNet8x4, Architecture Style=Same2021.04 | 75.63 | — | |
| CRDTeacher=ResNet32x4, Student=ShuffleNetV12021.04 | 75.11 | — | |
| ReviewKDTeacher=WRN40-2, Student=WRN40-1, Architecture Style=Same2021.04 | 75.09 | — | |
| ReviewKDTeacher=VGG13, Student=VGG8, Architecture Style=Same2021.04 | 74.84 | — | |
| KDTeacher=ResNet32x4, Student=ShuffleNetV12021.04 | 74.07 | — | |
| ReviewKDTeacher=ResNet110, Student=ResNet32, Architecture Style=Same2021.04 | 73.89 | — | |
| ReviewKDTeacher=ResNet56, Student=ResNet20, Architecture Style=Same2021.04 | 71.89 | — | |
| CRDTeacher=ResNet56, Student=ResNet202021.04 | 71.16 | — | |
| OFDTeacher=ResNet56, Student=ResNet202021.04 | 70.98 | — | |
| KDTeacher=ResNet56, Student=ResNet202021.04 | 70.66 | — | |
| ATTeacher=ResNet56, Student=ResNet202021.04 | 70.55 | — | |
| VIDTeacher=ResNet56, Student=ResNet202021.04 | 70.38 | — | |
| ReviewKDTeacher=VGG13, Student=MobileNetV2, Architecture Style=Different2021.04 | 70.37 | — | |
| PKTTeacher=ResNet56, Student=ResNet202021.04 | 70.34 | — | |
| ReviewKDTeacher=ResNet50, Student=MobileNetV2, Architecture Style=Different2021.04 | 69.89 | — | |
| RKDTeacher=ResNet56, Student=ResNet202021.04 | 69.61 | — | |
| FitNetTeacher=ResNet56, Student=ResNet202021.04 | 69.21 | — | |
| VS2Backbone=ViT-B/32, Ensemble=true, Test-time Augmentation=true2025.06 | 65.8 | — | |
| VS2Backbone=ViT-B/32, Ensemble=true2025.06 | 65.48 | — | |
| C-TPTBackbone=ViT-B/32, Test-time Prompt Tuning=true2025.06 | 64.86 | — | |
| VS2Backbone=ViT-B/322025.06 | 64.52 | — | |
| TPTBackbone=ViT-B/32, Test-time Prompt Tuning=true2025.06 | 64.09 | — | |
| CLIPBackbone=ViT-B/32, Ensemble=true2025.06 | 63.66 | — | |
| Diff-TPTBackbone=ViT-B/32, Test-time Prompt Tuning=true2025.06 | 63.04 | — | |
| CLIPBackbone=ViT-B/32, Protocol=Zero-shot2025.06 | 61.07 | — | |
| DeiT-smallPre-training Method=Baseline [4], Pre-training Acceleration=1.0x, Fine-tuning Method=Baseline [4], Fine-tuning Acceleration=1.0x2024.03 | — | 90.15 | |
| DeiT-smallPre-training Method=Baseline [4], Pre-training Acceleration=1.3x, Fine-tuning Method=ToE r1=0.5, Fine-tuning Acceleration=1.3x2024.03 | — | 90.19 | |
| DeiT-smallPre-training Method=ToE r1=0.5, Pre-training Acceleration=1.3x, Fine-tuning Method=Baseline [4], Fine-tuning Acceleration=1.0x2024.03 | — | 90.37 | |
| DeiT-smallPre-training Method=ToE r1=0.5, Pre-training Acceleration=1.3x, Fine-tuning Method=ToE r1=0.5, Fine-tuning Acceleration=1.3x2024.03 | — | 90.26 | |
| DeiT-tinyPre-training Method=Baseline [4], Pre-training Acceleration=1.0x, Fine-tuning Method=Baseline [4], Fine-tuning Acceleration=1.0x2024.03 | — | 86.78 | |
| DeiT-tinyPre-training Method=Baseline [4], Pre-training Acceleration=1.0x, Fine-tuning Method=ToE r1=0.5, Fine-tuning Acceleration=1.3x2024.03 | — | 86.74 | |
| DeiT-tinyPre-training Method=ToE r1=0.5, Pre-training Acceleration=1.3x, Fine-tuning Method=Baseline [4], Fine-tuning Acceleration=1.0x2024.03 | — | 87.1 | |
| DeiT-tinyPre-training Method=ToE r1=0.5, Pre-training Acceleration=1.3x, Fine-tuning Method=ToE r1=0.5, Fine-tuning Acceleration=1.3x2024.03 | — | 86.91 |