Research map / Architectures and efficiency / Pruning and efficient tuning
CHIP: CHannel Independence-based Pruning for Compact 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 |
|---|---|---|---|
| channel independence (nuclear-norm change) criterion instead of HRank rank criterion | filter pruning ResNet-56 · CIFAR-10 | top-1 accuracy: 90.72 -> 92.05 | Helped |
| channel independence criterion vs HRank | ResNet-110 · CIFAR-10 | top-1 accuracy: 94.23 -> 94.44 | Helped |
| channel independence criterion vs HRank at high compression | VGG-16 · CIFAR-10 | top-1 accuracy: 91.23 -> 93.18 | Helped |
| channel independence criterion vs HRank | ResNet-50 · ImageNet | top-1 accuracy: 69.10 -> 73.30 | Helped |
| channel independence criterion vs SCOP at moderate compression | ResNet-50 · ImageNet | top-1 accuracy: 75.95 -> 76.30 | Helped |
| channel independence vs SCOP at high compression | ResNet-50 · ImageNet | top-1 accuracy: 75.26 -> 75.26 | No clear effect |
| learn pruning mask from data after CI-based initialization (ablation) | CHIP · CIFAR-10 / ImageNet | top-1 accuracy | No clear effect |
| nuclear-norm change instead of rank change as independence metric (ablation) | CHIP · feature-map independence | top-1 accuracy | Helped |
Closest papers by meaning
- Revisiting Random Channel Pruning for Neural Network Compression (CVPR 2022 · similarity 0.91)
- Rethinking the Pruning Criteria for Convolutional Neural Network (NeurIPS 2021 · similarity 0.90)
- Provable Filter Pruning for Efficient Neural Networks (ICLR 2020 · similarity 0.90)
- Reborn Filters: Pruning Convolutional Neural Networks with Limited Data (AAAI 2020 · similarity 0.90)
- Discrimination-aware Channel Pruning for Deep Neural Networks (NeurIPS 2018 · similarity 0.89)
- Entropy Induced Pruning Framework for Convolutional Neural Networks (AAAI 2024 · similarity 0.88)
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.