Image Classification on CIFAR10 (test) (Accuracy and Hardware Consistency)
93.8AccuracyFull-precision
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Full-precisionW/A=32/32, Backbone=VGG-SMALL2024.05 | 93.8 | 100 | 100 | |
| BOLDW/A=1/1, Backbone=VGG-SMALL, Batch Normalization=True2024.05 | 92.37 | 4.87 | 3.71 | |
| BOLDW/A=1/1, Backbone=VGG-SMALL, Batch Normalization=False2024.05 | 90.29 | 3.64 | 2.78 | |
| BINARYCONNECTW/A=1/32, Backbone=VGG-SMALL2024.05 | 90.1 | 38.59 | 48.49 | |
| BINARYNETW/A=1/1, Backbone=VGG-SMALL2024.05 | 89.85 | 32.6 | 43.61 | |
| XNOR-NETW/A=1/1, Backbone=VGG-SMALL2024.05 | 89.83 | 34.21 | 45.68 |