Model training and data processing method and device, equipment and storage medium

• Main Authors: SHEN LIANG, GONG WEIBAO, WU ZHIHUA, YU DIANHAI, WU TIAN
• Format: Patent
• Language: Chinese; English
• Published: 02.08.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a model training and data processing method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical field of deep learning and the like. The model training method comprises the steps of obtaining a current step number; wherein the current step number is determined based on current convergence degree information of the to-be-trained model, and the current step number and the current convergence degree information are in a negative correlation relationship; and based on the current step number, executing an updating operation on the model parameters on each computing resource. The model precision and the training efficiency can be balanced. 本公开提供了一种模型训练及数据处理方法、装置、设备和存储介质，涉及人工智能技术领域，具体涉及深度学习等技术领域。模型训练方法包括：获取当前步数；其中，所述当前步数基于所述待训练模型的当前收敛程度信息确定，且，所述当前步数与所述当前收敛程度信息成负相关关系；基于所述当前步数，对所述各个计算资源上的所述模型参数执行更新操作。本公开可以均衡模型精度与训练效率。