Image Classification on ImageNet (test) (Accuracy and Calibration)
80.96Top-1 AccuracyConvNeXt-Large
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ConvNeXt-LargeModel Role=Teacher2026.06 | 80.96 | 83.9 | 5 | |
| ViT-B/16Model Role=Teacher2026.06 | 80.05 | 82.75 | 6.3 | |
| ResNet152V2Model Role=Teacher2026.06 | 77.23 | 81.42 | 8 | |
| MobileNetV2Distillation Strategy=KD + MixUp, Teacher Backbone=ConvNeXt-Large2026.06 | 71.6 | 73.31 | 4.6 | |
| MobileNetV2Distillation Strategy=KD + MixUp, Teacher Backbone=ViT-B/162026.06 | 71.58 | 73.29 | 4.2 | |
| MobileNetV2Distillation Strategy=KD + MixUp, Teacher Backbone=ResNet152V22026.06 | 71.52 | 73.62 | 3.8 | |
| MobileNetV2 BaselineDistillation Strategy=None2026.06 | 70.06 | 86 | 12.7 | |
| MobileNetV2Distillation Strategy=KD + MixUp, Teacher Backbone=ConvNeXt-Tiny2026.06 | 69.47 | 75.7 | 19.4 | |
| ConvNeXt-TinyModel Role=Teacher2026.06 | 65.9 | 70.28 | 15.2 |