VarGNet: Variable Group Convolutional Neural Network for Efficient Embedded Computing
arXiv:1907.05653
Abstract
In this paper, we propose a novel network design mechanism for efficient embedded computing. Inspired by the limited computing patterns, we propose to fix the number of channels in a group convolution, instead of the existing practice that fixing the total group numbers. Our solution based network, named Variable Group Convolutional Network (VarGNet), can be optimized easier on hardware side, due to the more unified computing schemes among the layers. Extensive experiments on various vision tasks, including classification, detection, pixel-wise parsing and face recognition, have demonstrated the practical value of our VarGNet.
arXiv commentTechnical report
References in corpus (12)
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
- SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
- Neural Architecture Search with Reinforcement Learning
- Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
- DARTS: Differentiable Architecture Search
- ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation
- TVM: An Automated End-to-End Optimizing Compiler for Deep Learning
- Panoptic Feature Pyramid Networks
- IGCV3: Interleaved Low-Rank Group Convolutions for Efficient Deep Neural Networks
- ChamNet: Towards Efficient Network Design through Platform-Aware Model Adaptation
- Compiling Deep Learning Models for Custom Hardware Accelerators
- DNNVM : End-to-End Compiler Leveraging Heterogeneous Optimizations on FPGA-based CNN Accelerators