Image classification model training method and device, image classification method and device and calculation equipment

Abstract

The invention relates to the technical field of artificial intelligence, and provides an image classification model training method, an image classification method, an image classification device andcalculation equipment, which are used for improving the processing capability of a trained model for a difficult sample. The method comprises the steps of training an image prediction model based on afirst image sample set; obtaining a difficulty estimation value of each second image sample in the second image sample set based on the image prediction model, wherein each second image sample has atarget classification label; training a first image classification model based on the second image sample set until the training loss of the first image classification model meets the target loss, andobtaining a second image classification model; wherein the training loss of the first image classification model is obtained by weighting the classification loss of each second image sample in the second image sample set, and the weighting weight corresponding to the classification loss of each second image sample is determined according to the difficulty estimation value of each second image sample.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular to an image classification model training method, image classification method, device and computing equipment. Background technique [0002] The image classification model generally refers to the model used for classification. At present, most image classification models are trained in advance through supervised learning, that is, the classification results corresponding to the training samples are predicted by the image classification model, and the model parameters are adjusted to make the output of the image classification model It is closer to the real classification result. [0003] However, during the training process of the image classification model, if the training sample set contains unbalanced samples, it is easy to overfit the model, which makes the trained model less effective on some samples. Contents of the invention [0004] Embodimen...

AI Technical Summary

Problems solved by technology

[0003] However, during the training process of the image classification model, if the training sample set contains unbalanced samples, it is easy to overfit the model, which makes the trained model less effective on some samples.

Images