Dual Precision Deep Neural Network

Abstract

On-line Precision scalability of the deep neural networks(DNNs) is a critical feature to support accuracy and complexity trade-off during the DNN inference. In this paper, we propose dual-precision DNN that includes two different precision modes in a single model, thereby supporting an on-line precision switch without re-training. The proposed two-phase training process optimizes both low- and high-precision modes.

5 pages, 4 figures, 2 tables

References in corpus (6)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Deep Learning with Limited Numerical Precision
• Ternary Weight Networks
• Training and Inference with Integers in Deep Neural Networks
• Trained Quantization Thresholds for Accurate and Efficient Fixed-Point Inference of Deep Neural Networks
• Post-training Quantization with Multiple Points: Mixed Precision without Mixed Precision