Research map / Architectures and efficiency / RNNs and normalization
Efficient Latency-Aware CNN Depth Compression via Two-Stage Dynamic Programming
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 the proposed depth compression instead of DepthShrinker | MobileNetV2-1.0 and MobileNetV2-1.4 · ImageNet-100 | accuracy: 87.58 -> 87.69; 85.30 -> 85.91; 88.01 -> 88.41; 85.29 -> 85.93 | Helped |
| Use the proposed depth compression instead of DepthShrinker | MobileNetV2-1.0 · ImageNet | accuracy: 72.37 -> 72.83; 69.43 -> 70.65 | Helped |
| Use the proposed depth compression instead of DepthShrinker | MobileNetV2-1.0 · ImageNet | latency (ms), RTX 2080 Ti: 14.82 -> 13.67; 10.53 -> 9.88 | Helped |
| Add knowledge distillation from the pretrained weight during finetuning | proposed depth-compressed MobileNetV2 · ImageNet | accuracy: 72.83 -> 73.00 | Helped |
| Jointly optimize the merge-boundary sets A and S instead of merging only according to A (ablation) | proposed depth compression · MobileNetV2-1.0 on ImageNet | inference latency: about 30% faster | Helped |
Closest papers by meaning
- LayerMerge: Neural Network Depth Compression through Layer Pruning and Merging (ICML 2024 · similarity 0.88)
- LEGACY: A Lightweight Dynamic Gradient Compression Strategy for Distributed Deep Learning (ICLR 2026 · similarity 0.88)
- Towards Compact CNNs via Collaborative Compression (CVPR 2021 · similarity 0.87)
- OPQ: Compressing Deep Neural Networks with One-shot Pruning-Quantization (AAAI 2021 · similarity 0.87)
- Focused Quantization for Sparse CNNs (NeurIPS 2019 · similarity 0.87)
- Layer Compression of Deep Networks with Straight Flows (AAAI 2024 · similarity 0.86)
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.