Model visualization method and device and storage medium

Abstract

The embodiment of the invention discloses a model visualization method and device, and a storage medium, and the method comprises the steps: obtaining a connection relationship between neurons of a target network model, mapping the neurons into nodes, mapping the connection relationship between the neurons into connection lines between the corresponding nodes to construct a node relationship graphcorresponding to the target network model, wherein the connecting lines can represent the weights of the corresponding connecting relationship, generating a model image corresponding to the target network model based on the node relationship graph, and outputting the model image.

Description

technical field [0001] The present application relates to deep learning technology, and in particular to a model visualization method, device and storage medium. Background technique [0002] At present, in the field of artificial intelligence, deep learning network models trained through deep learning are increasingly playing an important role. Especially in the fields of natural language processing, image recognition, and recommendation systems, deep learning network models have helped various fields to make significant progress. However, deep learning network models are often difficult to interpret. [0003] In related technologies, represented by the visualization tool tensorboard, it shows the model structure of the deep learning network model and the value of the loss function. It can start from the network layer and show the layer structure of the deep learning network model, but it cannot show the deep learning in detail. The case for network models. Contents of ...

AI Technical Summary

Problems solved by technology

However, deep learning network models are often difficult to interpret

[0003] In related technologies, represented by the visualization tool tensorboard, it shows the model structure of the deep learning network model and the value of the loss function. It can start from the network layer and show the layer structure of the deep learning network model, but it cannot show the deep learning in detail. The case for network models

Images