Image Classification on CIFAR-100 original (test)
78.05AccuracyTeacher
Evaluation results
| Method | Links | |
|---|---|---|
| TeacherTeacher Architecture=resnet-34, Student Architecture=resnet-18, Distillation Data Source=Original Data2021.10 | 78.05 | |
| KDTeacher Architecture=resnet-34, Student Architecture=resnet-18, Distillation Data Source=Original Data2021.10 | 77.87 | |
| ReviewKDTeacher=ResNet-32x4, Student=ShuffleNetV1, Transfer Type=Cross-architectural2021.12 | 77.45 | |
| ITRD (Lcorr + Lmi)Teacher=ResNet-32x4, Student=ShuffleNetV1, Transfer Type=Cross-architectural2021.12 | 76.91 | |
| ReviewKDTeacher=WRN-40-2, Student=WRN-16-2, Transfer Type=Same-architecture2021.12 | 76.12 | |
| ITRD (Lcorr + Lmi)Teacher=WRN-40-2, Student=WRN-16-2, Transfer Type=Same-architecture2021.12 | 76.12 | |
| ITRD (Lcorr)Teacher=WRN-40-2, Student=WRN-16-2, Transfer Type=Same-architecture2021.12 | 75.85 | |
| TeacherTeacher Architecture=wrn40-2, Student Architecture=wrn16-1, Distillation Data Source=Original Data2021.10 | 75.83 | |
| TeacherTeacher Architecture=wrn40-2, Student Architecture=wrn40-1, Distillation Data Source=Original Data2021.10 | 75.83 | |
| TeacherTeacher Architecture=wrn40-2, Student Architecture=wrn16-2, Distillation Data Source=Original Data2021.10 | 75.83 | |
| Teacher (Baseline)Teacher=WRN-40-2, Student=WRN-16-2, Transfer Type=Same-architecture2021.12 | 75.61 | |
| TeacherDistillation Data Source=Original Data2021.10 | 75.37 | |
| CRDTeacher=ResNet-32x4, Student=ShuffleNetV1, Transfer Type=Cross-architectural2021.12 | 75.12 | |
| KDTeacher=WRN-40-2, Student=WRN-16-2, Transfer Type=Same-architecture2021.12 | 74.92 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Entropy, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 74.43 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Least Conf, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 74.32 | |
| ITRD (Lcorr + Lmi)Teacher=ResNet-110, Student=ResNet-32, Transfer Type=Same-architecture2021.12 | 74.26 | |
| KDTeacher=ResNet-32x4, Student=ShuffleNetV1, Transfer Type=Cross-architectural2021.12 | 74.07 | |
| ReviewKDTeacher=ResNet-110, Student=ResNet-32, Transfer Type=Same-architecture2021.12 | 73.89 | |
| CRDTeacher=ResNet-110, Student=ResNet-32, Transfer Type=Same-architecture2021.12 | 73.75 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Entropy, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.7 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Least Conf, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.63 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Least Conf, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.6 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Entropy, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.45 | |
| Student (Baseline)Teacher=WRN-40-2, Student=WRN-16-2, Transfer Type=Same-architecture2021.12 | 73.26 | |
| PruneFusePruning Ratio (p)=0.8, Selection Metric=Least Conf, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.17 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Greedy k, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.1 | |
| Baseline ALSelection Metric=Least Conf, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 73.05 | |
| Baseline ALSelection Metric=Greedy k, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 72.91 | |
| Baseline ALSelection Metric=Entropy, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 72.9 | |
| MutexMatchNumber of labels=2500 labels, k=0.6C2022.03 | 72.82 | |
| PruneFusePruning Ratio (p)=0.8, Selection Metric=Entropy, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 72.82 | |
| ReMixMatchNumber of labels=400 labels, Distribution Alignment=true2022.03 | 72.57 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Greedy k, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 72.49 | |
| CoMatchNumber of labels=2500 labels, Distribution Alignment=true2022.03 | 72.45 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Greedy k, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 71.89 | |
| MutexMatchNumber of labels=2500 labels, Distribution Alignment=true2022.03 | 71.8 | |
| Scratch-BModel=VGG-16, Pruning=l1-norm based filter pruning2021.05 | 71.72 | |
| FixMatchNumber of labels=2500 labels2022.03 | 71.63 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Random, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 71.63 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Random, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 71.46 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Least Conf, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 71.45 | |
| CFMIPC=50, Backbone=ResNet-182025.03 | 71.4 | |
| Scratch-EModel=VGG-16, Pruning=l1-norm based filter pruning2021.05 | 71.37 | |
| ITRD (Lcorr + Lmi)Teacher=ResNet-50, Student=MobileNetV2, Transfer Type=Cross-architectural2021.12 | 71.34 | |
| TeacherTeacher Architecture=vgg-11, Student Architecture=resnet-18, Distillation Data Source=Original Data2021.10 | 71.32 | |
| Cyclic Learning Rate Restarting (CLR)Model=ResNet-110, Version=A, Pruning=l1-norm based filter pruning2021.05 | 71.29 | |
| PruneFusePruning Ratio (p)=0.8, Selection Metric=Greedy k, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 71.29 | |
| SLANumber of labels=2500 labels2022.03 | 71.27 | |
| Unpruned BaselineModel=VGG-162021.05 | 71.26 | |
| Cyclic Learning Rate Restarting (CLR)Model=VGG-16, Pruning=l1-norm based filter pruning2021.05 | 71.18 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Entropy, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 71.18 | |
| Scratch-EModel=ResNet-110, Version=A, Pruning=l1-norm based filter pruning2021.05 | 71.11 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Entropy, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 70.9 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Random, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 70.88 | |
| Scratch-BModel=ResNet-110, Version=B, Pruning=l1-norm based filter pruning2021.05 | 70.79 | |
| Baseline ALSelection Metric=Random, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 70.79 | |
| Fine-tuningModel=ResNet-110, Version=A, Pruning=l1-norm based filter pruning2021.05 | 70.71 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Least Conf, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 70.68 | |
| PruneFusePruning Ratio (p)=0.8, Selection Metric=Random, Label Budget (b)=50%, Architecture=ResNet-1642026.03 | 70.62 | |
| Scratch-BModel=ResNet-110, Version=A, Pruning=l1-norm based filter pruning2021.05 | 70.61 | |
| Scratch-EModel=ResNet-110, Version=B, Pruning=l1-norm based filter pruning2021.05 | 70.61 | |
| Unpruned BaselineModel=ResNet-1102021.05 | 70.59 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Least Conf, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 70.59 | |
| Cyclic Learning Rate Restarting (CLR)Model=ResNet-110, Version=B, Pruning=l1-norm based filter pruning2021.05 | 70.52 | |
| ITRD (Lcorr + Lmi)Teacher=VGG13, Student=MobileNetV2, Transfer Type=Cross-architectural2021.12 | 70.39 | |
| PruneFusePruning Ratio (p)=0.8, Selection Metric=Least Conf, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 70.38 | |
| ReviewKDTeacher=VGG13, Student=MobileNetV2, Transfer Type=Cross-architectural2021.12 | 70.37 | |
| Scratch-EModel=ResNet-56, Version=A, Pruning=l1-norm based filter pruning2021.05 | 70.15 | |
| Baseline ALSelection Metric=Least Conf, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 70.07 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Entropy, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 70.07 | |
| Cyclic Learning Rate Restarting (CLR)Model=ResNet-56, Version=B, Pruning=l1-norm based filter pruning2021.05 | 70.06 | |
| Cyclic Learning Rate Restarting (CLR)Model=ResNet-56, Version=A, Pruning=l1-norm based filter pruning2021.05 | 70.04 | |
| WCORDTeacher=VGG13, Student=MobileNetV2, Transfer Type=Cross-architectural2021.12 | 70.02 | |
| Baseline ALSelection Metric=Entropy, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 69.99 | |
| Fine-tuningModel=VGG-16, Pruning=l1-norm based filter pruning2021.05 | 69.97 | |
| PruneFuseSelection Metric=Entropy, Pruning ratio (p)=0.5, Label Budget (b)=50%, Backbone=ResNet-1102026.03 | 69.96 | |
| CRDTeacher=VGG13, Student=MobileNetV2, Transfer Type=Cross-architectural2021.12 | 69.94 | |
| Scratch-BModel=ResNet-56, Version=B, Pruning=l1-norm based filter pruning2021.05 | 69.91 | |
| ReviewKDTeacher=ResNet-50, Student=MobileNetV2, Transfer Type=Cross-architectural2021.12 | 69.89 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Greedy k, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 69.83 | |
| PruneFuseSelection Metric=Least Conf., Pruning ratio (p)=0.5, Label Budget (b)=50%, Backbone=ResNet-1102026.03 | 69.79 | |
| Scratch-BModel=ResNet-56, Version=A, Pruning=l1-norm based filter pruning2021.05 | 69.74 | |
| Scratch-EModel=ResNet-56, Version=B, Pruning=l1-norm based filter pruning2021.05 | 69.64 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Greedy k, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 69.57 | |
| Fine-tuningModel=ResNet-56, Version=A, Pruning=l1-norm based filter pruning2021.05 | 69.55 | |
| Fine-tuningModel=ResNet-110, Version=B, Pruning=l1-norm based filter pruning2021.05 | 69.54 | |
| Fine-tuningModel=ResNet-56, Version=B, Pruning=l1-norm based filter pruning2021.05 | 69.53 | |
| Unpruned BaselineModel=ResNet-562021.05 | 69.51 | |
| PruneFusePruning Ratio (p)=0.8, Selection Metric=Entropy, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 69.4 | |
| Baseline ALSelection Metric=Greedy k, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 69.25 | |
| PruneFuseSelection Metric=Least Conf., Pruning ratio (p)=0.6, Label Budget (b)=50%, Backbone=ResNet-1102026.03 | 69.03 | |
| Baseline ALSelection Metric=Least Conf., Label Budget (b)=50%, Backbone=ResNet-110, Pruning ratio (p)=02026.03 | 68.91 | |
| PruneFuseSelection Metric=Entropy, Pruning ratio (p)=0.6, Label Budget (b)=50%, Backbone=ResNet-1102026.03 | 68.89 | |
| PruneFuseSelection Metric=Greedy k, Pruning ratio (p)=0.5, Label Budget (b)=50%, Backbone=ResNet-1102026.03 | 68.86 | |
| Baseline ALSelection Metric=Entropy, Label Budget (b)=50%, Backbone=ResNet-110, Pruning ratio (p)=02026.03 | 68.79 | |
| PruneFusePruning Ratio (p)=0.5, Selection Metric=Random, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 68.75 | |
| PruneFuseSelection Metric=Entropy, Pruning ratio (p)=0.7, Label Budget (b)=50%, Backbone=ResNet-1102026.03 | 68.64 | |
| PruneFusePruning Ratio (p)=0.6, Selection Metric=Random, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 68.55 | |
| PruneFusePruning Ratio (p)=0.7, Selection Metric=Greedy k, Label Budget (b)=40%, Architecture=ResNet-1642026.03 | 68.55 |