Text model training method, text error correction method, electronic equipment and storage medium

Abstract

The embodiment of the invention relates to the field of deep learning, and discloses a text model training method, a text error correction method, electronic equipment and a storage medium. Accordingto the method, text conversion is carried out on a voice conversation to obtain an error text, error correction is carried out on the error text to obtain a correct text, and word segmentation is carried out on the error text and the correct text according to a hot word list; and then, encoding the text sample after word segmentation to obtain a text training sample, and inputting the text training sample into a deep neural network for training to obtain a target text model. Compared with text error correction based on a rule model or a statistical model in the prior art, the word segmentationaccuracy is higher, and the error correction accuracy of the target text model obtained through model training is improved. Besides, error correction is carried out based on the established target text model, correct sentences can be obtained only by inputting wrong sentences in the error correction process, and the workload and cost in the error correction process are reduced.

Description

technical field [0001] The embodiments of the present invention relate to the field of deep learning, and in particular to a text model training method, a text error correction method, electronic equipment, and a storage medium. Background technique [0002] At present, with the development of respective social networking sites, voice conversations are becoming increasingly popular in social software. In order to enable the receiving party to receive the content sent by the other party in a timely manner when it is inconvenient to receive the voice, the function of converting voice conversations to text has also begun to appear. However, due to the relatively colloquial nature of voice dialogue, it is easy to make wrong sentences during the conversion process, which will eventually cause the receiver to be unclear about the other party's intentions. Therefore, in order to realize the function of text error correction in the prior art, it is mainly to use a rule-based model ...

AI Technical Summary

Problems solved by technology

[0003] The inventors have found that there are at least the following problems in the prior art: in the process of text error correction based on rule-based models or statistical models, the accuracy of text error correction is low, and the effect of text error correction cannot meet the needs of users at this stage. functional requirements

Images