Image Classification on TCIFAR-100 1.0 (test)
0.6125AccuracyTeacher ResNet-32
Evaluation results
| Method | Links | |
|---|---|---|
| Teacher ResNet-32Params=~0.46M2019.11 | 0.6125 | |
| Proposed Method (KR framework) + KDGANParams=~0.17M2019.11 | 0.6033 | |
| Proposed Method (KR framework)Params=~0.17M2019.11 | 0.6003 | |
| KDGANParams=~0.19M2019.11 | 0.5815 | |
| NTSParams=~0.19M2019.11 | 0.5788 | |
| FSPParams=~0.19M2019.11 | 0.5731 | |
| FitNetParams=~0.19M2019.11 | 0.5677 | |
| Student ResNet-14 (Original)Params=~0.19M2019.11 | 0.5437 |