Data processing method of neural network model, medium and electronic equipment

Abstract

The invention relates to the field of artificial intelligence, and provides a data processing method of a neural network model, a medium and electronic equipment. The method comprises the following steps: detecting a first type of operation on an operand in the process of operating a neural network model, wherein the first type of operation comprises the operation of solving a power exponent of the operand and then solving a reciprocal of the operand; from the multiplication factor result set, selecting operation result values of a plurality of multiplication factors corresponding to the operation of solving the reciprocal after solving the power exponent of the operand; and multiplying the obtained operation result values of the multiplication factors to obtain an operation result of the first type of operation on the operands. According to the method provided by the invention, the electronic equipment can obtain the operation result of the first type of operation through search and multiplication, so that the calculation speed of the electronic equipment on the first type of operation is improved, and the speed of the electronic equipment for operating the neural network model is further improved.

Description

technical field [0001] The present application relates to the field of artificial intelligence, in particular to a data processing method, medium and electronic equipment of a neural network model. Background technique [0002] With the rapid development of artificial intelligence (AI), neural network models are more and more widely used in the field of artificial intelligence. In order to improve the operating speed of the neural network, various operations in the neural network model, such as normalization and convolution, are usually implemented through fixed-point operations. However, the computing unit used to run the neural network model, such as the neural network processor (Neural-Network Processing Unit, NPU), cannot directly implement the fixed-point operation of the nonlinear operation through the hardware, such as exponentiating the operand to the exponent and then finding The reciprocal operation (such as the reciprocal of the square root), when the NPU calcula...

AI Technical Summary

Problems solved by technology

However, the calculation process of Newton’s method needs to be iteratively solved, which is slow and takes up a lot of hardware resources of the NPU, which reduces the speed of the electronic device to run the neural network model; while the Taylor series expansion method has low accuracy and cannot ensure the accuracy of the neural network model. Model running accuracy

Images