Data processing method using corrected neural network quantization operation and data processing apparatus thereby

Abstract

A data processing method and a data processing apparatus using an enhanced neural network quantization operation are provided, comprising the steps of: quantizing the weights of a neural network to obtain quantized weights; obtaining a quantization error which is the difference between the weights and the quantized weights; obtaining input data for the neural network; performing a convolution operation between the quantized weights and the input data to obtain a first convolution operation result; performing a convolution operation between the quantization error and the input data to obtain a second convolution operation result, and obtaining a scaled second convolution operation result by scaling the second convolution operation result using a bit shift operation; and obtaining output data using the first convolution operation result and the scaled second convolution operation result.

Images