| abstract |
The present invention provides a method for training a description text generation model, a method and a device for generating a description text, wherein the method for training a description text generation model includes: obtaining training data, and the training data includes: entity words, first words of entity words Descriptive text and the second descriptive text, wherein the second descriptive text is a simplified expression of the first descriptive text; the first descriptive text of the entity word and the entity word is used as the input of the seq2seq model, and the first descriptive text of the entity word is 2. The description text is used as the output of the seq2seq model, and the seq2seq model is trained to obtain a description text generation model. The descriptive text generation model obtained in the present invention can generate a condensed descriptive text for entity words through in-depth understanding. |