Object Recognition on Office-31
50.7Accuracy (Amazon)EmaQ
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| EmaQPrecision (bits)=4, Backbone=ResNet-202026.06 | 50.7 | 73.54 | 73.68 | |
| DoReFaPrecision (bits)=4, Backbone=ResNet-202026.06 | 46.52 | 66.07 | 51.46 | |
| EmaQPrecision (bits)=2, Backbone=ResNet-202026.06 | 45.12 | 72.32 | 73.1 | |
| DoReFaPrecision (bits)=2, Backbone=ResNet-202026.06 | 42.16 | 50 | 47.37 | |
| LSQPrecision (bits)=4, Backbone=ResNet-202026.06 | 37.81 | 34.82 | 50.88 | |
| LSQPrecision (bits)=2, Backbone=ResNet-202026.06 | 36.93 | 23.21 | 42.11 | |
| APoTPrecision (bits)=4, Backbone=ResNet-202026.06 | 35.02 | 42.86 | 46.78 | |
| ZAQPrecision (bits)=4, Backbone=ResNet-202026.06 | 31.88 | 21.43 | 26.32 | |
| ZeroQPrecision (bits)=4, Backbone=ResNet-202026.06 | 30.49 | 34.82 | 32.75 | |
| Meta AugPrecision (bits)=4, Backbone=ResNet-202026.06 | 29.97 | 32.14 | 28.65 | |
| Meta AugPrecision (bits)=2, Backbone=ResNet-202026.06 | 27.7 | 25.89 | 26.9 | |
| EMQNetPrecision (bits)=2, Backbone=ResNet-202026.06 | 17.94 | 61.61 | 70.76 | |
| EMQNetPrecision (bits)=4, Backbone=ResNet-202026.06 | 13.07 | 53.57 | 40.94 | |
| ZeroQPrecision (bits)=2, Backbone=ResNet-202026.06 | 10.98 | 11.61 | 11.7 | |
| ZAQPrecision (bits)=2, Backbone=ResNet-202026.06 | 8.19 | 8.04 | 13.45 | |
| QwTPrecision (bits)=4, Backbone=ResNet-202026.06 | 3.66 | 4.46 | 8.77 | |
| APoTPrecision (bits)=2, Backbone=ResNet-202026.06 | 3.48 | 6.25 | 5.26 | |
| QwTPrecision (bits)=2, Backbone=ResNet-202026.06 | 3.48 | 5.36 | 7.6 |