| Pruning Filters for Efficient ConvNets | ✓ Link | 78.79 | 2.335 | 14.811 | ResNet50-2.3 GFLOPs | 2016-08-31 |
| Pruning Filters for Efficient ConvNets | ✓ Link | 78.07 | 1.635 | 10.511 | ResNet50-1.5 GFLOPs | 2016-08-31 |
| Knapsack Pruning with Inner Distillation | ✓ Link | 78.0 | 2.5 | | ResNet50 2.5 GFLOPS | 2020-02-19 |
| Group Fisher Pruning for Practical Network Compression | ✓ Link | 77.97 | 1.588 | 9.3 | RegX-1.6G | 2021-08-02 |
| Knapsack Pruning with Inner Distillation | ✓ Link | 77.70 | 2 | | ResNet50 2.0 GFLOPS | 2020-02-19 |
| EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning | ✓ Link | 77.1 | | | ResNet50-3G FLOPs | 2020-07-06 |
| EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning | ✓ Link | 76.4 | | | ResNet50-2G FLOPs | 2020-07-06 |
| Pruning Filters for Efficient ConvNets | ✓ Link | 76.376 | 1.075 | 6.954 | ResNet50-1G FLOPs | 2016-08-31 |
| Network Pruning via Transformable Architecture Search | ✓ Link | 76.20 | 2.3 | | TAS-pruned ResNet-50 | 2019-05-23 |
| Network Pruning That Matters: A Case Study on Retraining Variants | ✓ Link | 75.59 | | | ResNet50 | 2021-05-07 |
| EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning | ✓ Link | 74.2 | | | ResNet50-1G FLOPs | 2020-07-06 |
| EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning | ✓ Link | 74.2 | | | ResNet50-1G FLOPs | 2020-07-06 |
| Group Fisher Pruning for Practical Network Compression | ✓ Link | 73.42 | 0.29 | 3.31 | MobileNetV2 | 2021-08-02 |
| AC/DC: Alternating Compressed/DeCompressed Training of Deep Neural Networks | ✓ Link | 73.14 | | | ResNet50 | 2021-06-23 |
| EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning | ✓ Link | 70.7 | | | MobileNetV1-50% FLOPs | 2020-07-06 |
| SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size | ✓ Link | 57.5% | | 1.24 | SqueezeNet (6-bit Deep Compression) | 2016-02-24 |