Singular value decomposition operation implementation method and device and related equipment

Abstract

The invention discloses a singular value decomposition operation implementation method and device and related equipment, and the method comprises the steps: constructing a singular value decomposition operator which is used for carrying data in target equipment and carrying out singular value decomposition operation; deploying the singular value decomposition operator in the target equipment, and the target equipment is an Ascend AI processor; and obtaining to-be-processed data, and performing singular value decomposition operation on the to-be-processed data based on the deployed singular value decomposition operator. Compared with the prior art, in the scheme of the invention, the singular value decomposition operator capable of carrying the data in the Ascend AI processor and carrying out singular value decomposition operation is constructed, and the singular value decomposition operator is deployed in the Ascend AI processor, so that the calculation capability of the Ascend processor is fully utilized; and SVD operation is performed on the to-be-processed data directly based on the Ascend AI processor.

Description

technical field [0001] The present invention relates to the technical field of singular value decomposition, in particular to a singular value decomposition operation realization method, device and related equipment. Background technique [0002] The development of AI technology often relies on the processing of massive data, which places very high requirements on computing power. The Ascend AI processor is a high-computing processor launched for the characteristics of such computationally intensive tasks. The Ascend AI processor has achieved good results in AI task processing, but its underlying operator support is still lacking. [0003] Singular value decomposition (SVD, Singularity Value Decompose) operation is a relatively common matrix calculation process in mathematics, which can be used to accelerate the calculation of matrix inversion, etc., in signal processing, image compression, tensor network, second-order optimization, etc. play an important role. In the pri...

AI Technical Summary

Problems solved by technology

[0005] The main purpose of the present invention is to provide a singular value decomposition operation implementation method, device and related equipment, aiming at solving the problem that the singular value decomposition operator in the prior art cannot be run on the Ascend AI processor, and the singular value decomposition is required. The calculated data cannot be directly based on the Ascend AI processor for singular value decomposition calculations, which is not conducive to making full use of the computing power of the Ascend AI processor.

Images