Image Classification on ImageNet 1k (val) (Accuracy & Inference Speed)
76.89Top-1 AccuracyTeacher
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| TeacherTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 76.89 | 93.37 | — | — | |
| EagleEye-2GBackbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=2.12022.10 | 76.4 | 92.89 | 190 | 1.05 | |
| HALPBackbone=ResNet50, Target Latency=90%, Batch Size=1, Hardware=TITAN V, FLOPs (G)=2.92022.10 | 76.4 | 93.1 | 220 | 1.22 | |
| ResNet50Backbone=ResNet50, Pruning=None, Batch Size=1, Hardware=TITAN V, FLOPs (G)=4.12022.10 | 76.2 | 92.87 | 181 | 1 | |
| AutoSlimBackbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=22022.10 | 75.6 | — | 181 | 1 | |
| MetaPruningBackbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=22022.10 | 75.4 | — | 190 | 1.05 | |
| GReg-2Backbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=1.82022.10 | 75.4 | — | 196 | 1.09 | |
| HALPBackbone=ResNet50, Target Latency=80%, Batch Size=1, Hardware=TITAN V, FLOPs (G)=2.32022.10 | 75.3 | 92.35 | 247 | 1.37 | |
| ResNet50Backbone=ResNet50, Scale=0.75x, Batch Size=1, Hardware=TITAN V, FLOPs (G)=2.32022.10 | 74.8 | — | 192 | 1.06 | |
| EagleEye-1GBackbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=12022.10 | 74.2 | 91.77 | 192 | 1.06 | |
| AutoSlimBackbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=12022.10 | 74 | — | 191 | 1.06 | |
| GReg-2Backbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=1.32022.10 | 73.9 | — | 206 | 1.14 | |
| MetaPruningBackbone=ResNet50, Batch Size=1, Hardware=TITAN V, FLOPs (G)=12022.10 | 73.4 | — | 196 | 1.09 | |
| PEFDTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 72.29 | 90.99 | — | — | |
| ResNet50Backbone=ResNet50, Scale=0.50x, Batch Size=1, Hardware=TITAN V, FLOPs (G)=1.12022.10 | 72 | — | 193 | 1.07 | |
| CIDTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 71.99 | 90.64 | — | — | |
| SRRLTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 71.76 | 90.8 | — | — | |
| CRDTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 70.87 | 89.86 | — | — | |
| SPTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 70.75 | 90.01 | — | — | |
| KDTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 70.38 | 90.12 | — | — | |
| StudentTeacher backbone=DenseNet201, Student backbone=ResNet182022.10 | 69.75 | 89.07 | — | — |