Speech recognition model training method and device, equipment and medium

Abstract

The invention relates to the field of artificial intelligence, and provides a speech recognition model training method and device, equipment and a medium, and the method comprises the steps: obtaining a speech sample set containing speech samples; inputting the speech sample into an initial recognition model; obtaining a to-be-processed audio clip through audio enhancement processing; performing teacher acoustic feature extraction through a teacher network in the initial recognition model to obtain a first feature vector, and performing student acoustic feature extraction through a student network in the initial recognition model to obtain a second feature vector; performing alignment comparison processing in combination with a dynamic queue in a teacher network to obtain a loss value; and when the loss value does not reach a preset convergence condition, carrying out iterative updating until convergence, and obtaining a trained speech recognition model. According to the invention, common speech recognition through the teacher network and the student network is realized, and the training efficiency is improved. The method is suitable for the field of artificial intelligence, and can further promote the construction of a smart city.

Description

technical field [0001] The invention relates to the field of speech recognition of artificial intelligence, in particular to a speech recognition model training method, device, computer equipment and storage medium. Background technique [0002] Speech translation is the process of converting one natural language (source language) into another natural language (target language). Unlike traditional machine translation, the input of voice translation is directly voice, and the output is text. With international communication With the increase of the number of people, different languages ​​are used to communicate more and more frequently. In order to overcome language communication barriers, online voice translation based on clients has been widely used. [0003] Online speech translation generally involves two steps. The first is to perform speech recognition, that is, to convert the speech signal of the first language input by the user into text; the second is to translate th...

AI Technical Summary

Problems solved by technology

[0003] Online speech translation generally involves two steps. The first is to perform speech recognition, that is, to convert the speech signal of the first language input by the user into text; the second is to translate the text online through a machine translation device to obtain the first translation result. The text of the second language, and finally provide the user with the text or voice information of the second language. However, the voice recognition in the existing scheme is usually obtained by using a large number of voice samples marked by artificial inefficiency, and the trained voice recognition model The complex structure and large amount of calculation lead to low efficiency of the output text, and finally there is a long translation lag time, resulting in poor real-time online voice translation and low user experience satisfaction

Images