Model training method and device and computer readable storage medium

Abstract

The invention relates to the technical field of machine learning, particularly provides a model training method and device and a computer readable storage medium, and aims to solve the technical problem of how to improve the model training effect. According to the method provided by the embodiment of the invention, the preset data processing model can be trained by utilizing the initial training set to obtain the first data processing model; a first model loss difference is acquired between the first data processing model and each second data processing model on the test set, the second data processing model is obtained by training according to different sub-training sets under the initial training set, and the difference between the different sub-training sets is one or more different deleted samples; and finally, an abnormal sample is acquired according to the difference value to optimize the initial training set, and the first data processing model is trained by using the optimizedinitial training set. Based on the steps, the abnormal samples can be quickly and accurately screened out from the training set, and the model training effect is greatly improved.

Description

technical field [0001] The present invention relates to the technical field of machine learning, in particular to a model training method, device and computer-readable storage medium. Background technique [0002] Supervised learning in the field of machine learning technology mainly uses training samples and sample labels to train the model, and in order to improve the training effect of the model, it is necessary to use large-scale training samples such as millions of training samples and prepare for each training in advance. The samples are marked with accurate sample labels to ensure that the trained model has high model performance. For example: use millions of training samples and the corresponding category labels of each training sample to train the data classification model, so that the trained data classification model has higher classification performance. Because the order of magnitude of the training samples is too large, when labeling the training samples, it i...

AI Technical Summary

Problems solved by technology

Because the order of magnitude of the training samples is too large, when labeling the training samples, it is impossible to guarantee that each training sample is accurately labeled. If these noise samples with wrong labels are used for model training, the training effect of the model will be reduced.

Images