Classification model training method and device suitable for text auditing

The invention provides a training method and device of a classification model suitable for text auditing, and relates to the technical field of artificial intelligence, in particular to the technical field of deep learning and computers. The method comprises the steps of performing different types of prediction on text samples in an ith round of text sample set based on a pre-training language model to obtain enhanced text samples of different prediction types; based on the first classification model after the jth round of training, obtaining a label and a confidence coefficient of the enhanced text sample; screening the enhanced text samples according to the prediction types of the enhanced text samples and the labels and confidence of the enhanced text samples, updating the i-th round of text sample set according to the screened enhanced text samples, obtaining the i + 1-th round of text sample set, training the first classification model, and obtaining a second classification model after the (j + 1)-th rou