Deep Learning with Limited Numerical Precision

Autor:	Gupta, Suyog, Agrawal, Ankur, Gopalakrishnan, Kailash, Narayanan, Pritish
Rok vydání:	2015
Předmět:	Computer Science - Learning Computer Science - Neural and Evolutionary Computing Statistics - Machine Learning
Druh dokumentu:	Working Paper
Popis:	Training of large-scale deep neural networks is often constrained by the available computational resources. We study the effect of limited precision data representation and computation on neural network training. Within the context of low-precision fixed-point computations, we observe the rounding scheme to play a crucial role in determining the network's behavior during training. Our results show that deep networks can be trained using only 16-bit wide fixed-point number representation when using stochastic rounding, and incur little to no degradation in the classification accuracy. We also demonstrate an energy-efficient hardware accelerator that implements low-precision fixed-point arithmetic with stochastic rounding. Comment: 10 pages, 6 figures, 1 table
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/1502.02551 Zobrazit plný text záznamu View this record from Arxiv