Self-adaptive loading method of embedded product AI model

A self-adaptive loading method of an embedded product AI model comprises the steps that a microprocessor of an embedded product is subjected to partition structure, different AI models are loaded in different areas, a periodic timed task 1 is set, the microprocessor of the embedded product is made t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: QU HAOYANG, WANG JUN, ZHANG JIAXIN, SHI SAIFEI, LIU HU, YE MAOQING
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:A self-adaptive loading method of an embedded product AI model comprises the steps that a microprocessor of an embedded product is subjected to partition structure, different AI models are loaded in different areas, a periodic timed task 1 is set, the microprocessor of the embedded product is made to conduct self-inspection on equipment parameters and a working environment, and the self-adaptive loading of the embedded product AI model is achieved through the change of the parameters. Different AI models are selected to adapt to the current working environment and equipment parameters; the method comprises the following steps of: establishing a periodic timed task 1, setting a periodic timed task 2, periodically uploading equipment information and working environment parameters to a server side, retraining a set of embedded AI model capable of adapting to environment change at the server side by analyzing the parameters, and updating new AI model parameters into an embedded product in a network communication