Embedded artificial intelligence computing framework and application method

Abstract

The invention provides an embedded artificial intelligence computing framework and an application method, and belongs to the technical field of computers and artificial intelligence, and the computing framework comprises: a first interface layer which is an external interface, is used for providing a front-end interface or a user development interface, and is used for providing a unified call interface for an application service; an intelligent layer which is connected with the first interface layer and is used for providing an interface and a core module which are used for constructing an intelligent model and training and reasoning the constructed intelligent model; a second interface layer which is an internal interface, is connected with the intelligent layer and is used for providing an interface for calling the hardware layer to perform data calculation for the calculation task; and a hardware layer which is connected with the second interface layer and is used for calculating the data related to the intelligent layer through hardware provided by the system. According to the embedded artificial intelligence computing framework provided by the invention, the problems that an open source framework in the prior art is serious in dependence on a third party, the version is difficult to control, codes cannot be autonomously controlled and the like are solved.

Description

technical field [0001] The invention relates to the technical field of computers and artificial intelligence, in particular to an embedded artificial intelligence computing framework and an application method. Background technique [0002] The wide application of artificial intelligence algorithms in various fields has brought about the rapid development of various intelligent computing framework technologies. Both academia and industry have launched corresponding intelligent computing frameworks, and more and more hardware architectures and software platforms are supported. . At present, the computing frameworks of machine learning intelligent algorithms mainly include: TensorFlow, PyTorch, Caffe, MXNet, CNTK, Theano, etc. [0003] The current open source framework faces serious third-party dependence, difficult version control, and uncontrollable code, which hinders subsequent performance optimization and adaptive improvement. At the same time, most intelligent computing ...

AI Technical Summary

Problems solved by technology

[0004] The present invention provides an embedded artificial intelligence computing framework and application method, which are used to solve the problems faced by open source frameworks in the prior art, such as serious third-party dependence, difficult version control, uncontrollable code, etc., and based on the embedded artificial intelligence The intelligent computing framework realizes the construction of the intelligent model and trains and infers the intelligent model to obtain the inference result of the intelligent model

Images