Research map / Architectures and efficiency / RNNs and normalization
MEC: Memory-efficient Convolution for Deep Neural Network
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 MEC compact lowering instead of conventional im2col convolution | Convolution on ResNet-101 · ResNet-101 (Mobile) | memory: 203.6 MB -> 64.6 MB; 3.2x | Helped |
| Use MEC compact lowering instead of conventional im2col convolution | Convolution on ResNet-101 · ResNet-101 (Mobile) | runtime: 1701.6 msec -> 1391.6 msec; 1.2 | Helped |
| Use MEC instead of conventional im2col convolution | MEC.cpu · Mobile benchmarks | runtime: 20% | Helped |
| Use MEC instead of conventional im2col convolution | MEC.cpu · Server-CPU benchmarks | runtime: 8.8x | Helped |
| Use MEC instead of conventional im2col convolution | MEC.cpu · Mobile and Server-CPU benchmarks | memory overhead: 3.4x; 3.2x | Helped |
| Use MEC instead of conventional im2col convolution | MEC.gpu · Server-GPU benchmarks | lowering runtime: 85% | Helped |
| Use MEC instead of Winograd convolution | MEC.cpu · CV6-CV12 | memory overhead: 5.9x | Helped |
| Use MEC instead of Winograd convolution | MEC.cpu · CV7, CV8, and CV9 on Server-CPU | runtime | Mixed |
Closest papers by meaning
- High Performance Zero-Memory Overhead Direct Convolutions (ICML 2018 · similarity 0.87)
- MemSR: Training Memory-efficient Lightweight Model for Image Super-Resolution (ICML 2022 · similarity 0.86)
- Deep Tensor Convolution on Multicores (ICML 2017 · similarity 0.84)
- ClcNet: Improving the Efficiency of Convolutional Neural Network Using Channel Local Convolutions (CVPR 2018 · similarity 0.84)
- MEET: Towards Memory-Efficient Temporal Sparse Deep Neural Networks (CVPR 2025 · similarity 0.82)
- Multiscale Low-Frequency Memory Network for Improved Feature Extraction in Convolutional Neural Networks (AAAI 2024 · similarity 0.82)
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.