Research map / Architectures and efficiency / CNN design
Gate Decorator: Global Filter Pruning Method for Accelerating Deep Convolutional Neural Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use iterative Tick pruning rather than pruning all at once (ablation) | Gate Decorator pruning · VGG-16-M, CIFAR-100, 40% FLOPs pruned | fine-tuned accuracy: 71.8 -> 73.0 | Helped |
| Use iterative Tick pruning rather than pruning all at once (ablation) | Gate Decorator pruning · VGG-16-M, CIFAR-100, 60% FLOPs pruned | fine-tuned accuracy: 62.1 -> 71.4 | Helped |
| Use iterative Tick pruning rather than pruning all at once (ablation) | Gate Decorator pruning · VGG-16-M, CIFAR-100, 80% FLOPs pruned | fine-tuned accuracy: 57.7 -> 68.4 | Helped |
| Add the Tock phase, including sparse gate regularization and fine-tuning (ablation) | Tick-Only pruning · VGG-16-M, CIFAR-100, 40% FLOPs pruned | fine-tuned accuracy: 73.0 -> 74.6 | Helped |
| Add the Tock phase, including sparse gate regularization and fine-tuning (ablation) | Tick-Only pruning · VGG-16-M, CIFAR-100, 60% FLOPs pruned | fine-tuned accuracy: 71.4 -> 73.2 | Helped |
| Add the Tock phase, including sparse gate regularization and fine-tuning (ablation) | Tick-Only pruning · VGG-16-M, CIFAR-100, 80% FLOPs pruned | fine-tuned accuracy: 68.4 -> 71.2 | Helped |
| Use Tick-Tock rather than Tick-Only pruning (ablation) | Gate Decorator pruning · VGG-16-M, CIFAR-100 | scratch accuracy: 73.7; 72.9; 69.6 -> 73.7; 73.0; 69.9 | No clear effect |
| Add Taylor-based gated filter importance ranking | Global filter pruning on VGG-16-M · CUB-200 | test accuracy during pruning | Helped |
| Use gate magnitude alone for filter ranking rather than gradient-aware ranking | Slim pruning · VGG-16-M, CUB-200 | test accuracy during pruning | Hurt |
Closest papers by meaning
- Pruning Filters for Efficient ConvNets (ICLR 2017 · similarity 0.92)
- Learning Filter Pruning Criteria for Deep Convolutional Neural Networks Acceleration (CVPR 2020 · similarity 0.91)
- Interspace Pruning: Using Adaptive Filter Representations To Improve Training of Sparse CNNs (CVPR 2022 · similarity 0.91)
- Auto-Balanced Filter Pruning for Efficient Convolutional Neural Networks (AAAI 2018 · similarity 0.90)
- Rethinking the Pruning Criteria for Convolutional Neural Network (NeurIPS 2021 · similarity 0.90)
- REPrune: Channel Pruning via Kernel Representative Selection (AAAI 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.