Deep Fried Convnets
Abstract
The fully connected layers of a deep convolutional neural network typically contain over 90% of the network parameters, and consume the majority of the memory required to store the network parameters. Reducing the number of parameters while preserving essentially the same predictive performance is critically important for operating deep neural networks in memory constrained environments such as GPUs or embedded devices. In this paper we show how kernel methods, in particular a single Fastfood layer, can be used to replace all fully connected layers in a deep convolutional neural network. This novel Fastfood layer is also end-to-end trainable in conjunction with convolutional layers, allowing us to combine them into a new architecture, named deep fried convolutional networks, which substantially reduces the memory footprint of convolutional networks trained on MNIST and ImageNet with no drop in predictive performance.
Cite
@article{arxiv.1412.7149,
title = {Deep Fried Convnets},
author = {Zichao Yang and Marcin Moczulski and Misha Denil and Nando de Freitas and Alex Smola and Le Song and Ziyu Wang},
journal= {arXiv preprint arXiv:1412.7149},
year = {2015}
}
Comments
svd experiments included