Image Classification on CIFAR-10 (2-bit and 4-bit Quantization Evaluation)
93.1Accuracy (4-bit)EmaQ
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EmaQBackbone=DenseNet-402026.06 | 93.1 | 93 | |
| EmaQBackbone=ResNet-562026.06 | 93 | 91.2 | |
| ZAQBackbone=ResNet-562026.06 | 92.9 | 89.2 | |
| Choi et al.Backbone=ResNet-562026.06 | 92.7 | 88.7 | |
| ZAQBackbone=DenseNet-402026.06 | 92.7 | 91.4 | |
| ZeroQBackbone=DenseNet-402026.06 | 92.6 | 91.3 | |
| ZeroQBackbone=ResNet-562026.06 | 92.5 | 88.1 | |
| Choi et al.Backbone=DenseNet-402026.06 | 92.5 | 91.5 | |
| EmaQBackbone=ResNet-202026.06 | 92.2 | 90.8 | |
| ZAQBackbone=ResNet-202026.06 | 92.1 | 88.9 | |
| AdaSGBackbone=ResNet-202026.06 | 92.1 | — | |
| Choi et al.Backbone=ResNet-202026.06 | 91.9 | 88.1 | |
| ZeroQBackbone=ResNet-202026.06 | 91.8 | 87.9 | |
| IntraQBackbone=ResNet-202026.06 | 91.5 | — | |
| ARC+AITBackbone=ResNet-202026.06 | 90.5 | — | |
| GDFQBackbone=ResNet-202026.06 | 90.3 | — | |
| OCSBackbone=ResNet-202026.06 | 89.1 | — | |
| GZNQBackbone=ResNet-202026.06 | 89.1 | — | |
| LLSQBackbone=DenseNet-402026.06 | 87.2 | 81.5 | |
| LSQBackbone=DenseNet-402026.06 | 85.6 | 79.5 | |
| APoTBackbone=DenseNet-402026.06 | 85.6 | 59.4 | |
| LSQBackbone=ResNet-562026.06 | 85.5 | 79.6 | |
| APoTBackbone=ResNet-562026.06 | 84.8 | 68.3 | |
| LSQBackbone=ResNet-202026.06 | 83.4 | 77.7 | |
| LLSQBackbone=ResNet-202026.06 | 81.5 | 76.9 | |
| APoTBackbone=ResNet-202026.06 | 81 | 65.2 |