Knowledge Distillation on CIFAR-10
94.07AccuracyNAYER
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| NAYERTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A100, Training iterations (E)=3002023.09 | 94.07 | 6.78 | 4.47 | |
| NAYERTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A100, Training iterations (E)=2002023.09 | 93.84 | 3.85 | 7.93 | |
| NAYERTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A100, Training iterations (E)=1002023.09 | 93.48 | 2.05 | 14.88 | |
| SpaceshipNetTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A1002023.09 | 93.25 | 22.35 | 1.39 | |
| MADTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A1002023.09 | 92.64 | 13.13 | 1.78 | |
| CMITeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A1002023.09 | 92.52 | 24.01 | 1.3 | |
| FMTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A100, Training iterations (E)=3002023.09 | 92.31 | 7.02 | 4.29 | |
| FMTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A100, Training iterations (E)=2002023.09 | 92.05 | 3.98 | 7.46 | |
| DFQTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A1002023.09 | 92.01 | 3.31 | 9.73 | |
| FMTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A100, Training iterations (E)=1002023.09 | 91.63 | 2.18 | 14.17 | |
| DeepInvTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A1002023.09 | 89.72 | 31.23 | 1 | |
| ZSKTTeacher model=WRN40-2, Student model=WRN16-2, GPU=NVIDIA A1002023.09 | 89.66 | 3.44 | 9.08 |