Research Article
Small Network for Lightweight Task in Computer Vision: A Pruning Method Based on Feature Representation
| ā | Network size (M) | Pruned (%) | Test accuracy (%) |
| Li et al. [31] | 21.1 | 64.0 | 90.51 | Single sample pruning | 16.4 | 72.0 | 90.89 | Multisample pruning (0.4) | 21.7 | 63.0 | 90.66 | Multisample pruning (0.5) | 15.1 | 74.2 | 89.98 | Multisample pruning (0.6) | 9.8 | 83.3 | 89.70 |
|
|