Methods and device for model fusion

Abstract

The invention discloses a method and device for model fusion, and relates to the field of artificial intelligence deep learning and image processing. The method specifically comprises the following steps: obtaining at least two models; obtaining a preset sample set; randomly deleting the sample set to obtain at least two different candidate sample sets; for each of the at least two models, performing model training by using one candidate sample set in at least two different candidate sample sets to obtain a trained model; and fusing the models trained through different candidate sample sets toobtain a fused model. According to the embodiment, the model training speed can be increased, and the image recognition precision is improved.

Description

technical field [0001] The embodiments of the present disclosure relate to the field of computer technology, and specifically to the fields of deep learning of artificial intelligence and image processing. Background technique [0002] Image recognition technology is a technology that extracts the features of images by means of machine learning and distinguishes different images through the extracted features. Image recognition technology is widely used in various visual tasks, such as plant classification, dish recognition, landmark recognition, etc. In the field of image recognition, how to improve the accuracy of existing models has always been one of the most worthwhile exploration points in academia and industry. [0003] Existing techniques use model fusion to improve model accuracy. The way of model fusion is that multiple models are trained on a unified data set, and the data itself will determine the quality and tendency of the model, so multiple models trained on...

AI Technical Summary

Problems solved by technology

The way of model fusion is that multiple models are trained on a unified data set, and the data itself will determine the quality and tendency of the model, so multiple models trained on a unified data set will cause multiple models to remain consistent The tendency of the fusion, the result after fusion will also have a certain tendency, so this kind of fusion can only improve the result to a certain extent.

Images