Paper Info
Title
Joint Training of Low-Precision Neural Network with Quantization Interval Parameters.
Abstract
Optimization for low-precision neural network is an important technique for deep convolutional neural network models to be deployed to mobile devices. In order to realize convolutional layers with the simple bit-wise operations, both activation and weight parameters need to be quantized with a low bit-precision. In this paper, we propose a novel optimization method for low-precision neural network which trains both activation quantization parameters and the quantized model weights. We parameterize the quantization intervals of the weights and the activations and train the parameters with the full-precision weights by directly minimizing the training loss rather than minimizing the quantization error. Thanks to the joint optimization of quantization parameters and model weights, we obtain the highly accurate low-precision network given a target bitwidth. We demonstrated the effectiveness of our method on two benchmarks: CIFAR-10 and ImageNet.
Year
Venue
Field
2018
arXiv: Computer Vision and Pattern Recognition
Convolutional neural network,Computer science,Algorithm,Mobile device,Quantization (physics),Train,Quantization (signal processing),Artificial neural network
DocType
Volume
Citations 
Journal
abs/1808.05779
3
PageRank 
References 
Authors
0.39
24
7
Name
Order
Citations
PageRank
Sangil Jung1161.57
Changyong Son2161.24
Seohyung Lee3160.90
JinWoo Son4160.90
Youngjun Kwak5242.37
Jae-Joon Han67412.34
Changkyu Choi715914.99