Visual modeling method and device for neural network model

Abstract

The present invention provides a visual modeling method and device for a neural network model, by constructing a mapping relationship between each operation method in the deep learning framework and each object / instance, and constructing a corresponding visual control for each object / instance, and inputting and outputting The relationship is mapped to a visual directed arc. When an operation on at least one visual control is received, the node set is determined according to the object / instance corresponding to each visual control and the mapping relationship; when at least one visual directed arc is received During the operation, according to each of the visualized directed arcs and the input-output relationship of the operation, the set of dependencies is determined; when a modeling request is received, the pre-built source code file is initialized; based on the deep learning framework and the initialized The source code file uses the node set and dependency set to generate the corresponding neural network model. The solution provided by the invention realizes the visual construction of the neural network model.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a visual modeling method and device of a neural network model. Background technique [0002] The self-learning function of the neural network model makes it have extensive and attractive prospects in the fields of system identification, pattern recognition, and intelligent control. However, the current way to build neural network models is to write program codes for business processes based on deep learning frameworks and programming languages ​​supported by deep learning frameworks, and convert them into corresponding neural network models according to the written program codes and deep learning frameworks. . At present, the construction process of this neural network model can only be understood by users after the neural network model is converted into a visual graph such as a decision tree. For users who do not understand the program code and deep learning framework, they ...

AI Technical Summary

Problems solved by technology

For users who do not understand the program code and deep learning framework, they can only check whether the neural network model has errors based on visual graphics such as decision trees. If there are errors, the program code needs to be re-inspected, etc.

Images