Voice processing model training method and device thereof

Abstract

The invention provides a voice processing model training method and a device thereof, and the method comprises the steps: obtaining the training data of a target voice recognition scene; wherein the training data comprises a first type of voice data and a second type of voice data, and the matching degree of the first type of voice data and the voice features of the target voice recognition sceneis greater than the matching degree of the second type of voice data and the voice features; generating a joint model structure; inputting the training data into a joint model structure, training thetraining data through an input layer and a shared hidden layer, training a first type of voice data output by the shared hidden layer through a first hidden layer and a first output layer, and training a second type of voice data output by the shared hidden layer in combination with a second hidden layer and a second output layer; and generating a voice processing model according to the model parameters correspondingly trained by the input layer, the shared hidden layer, the first hidden layer and the first output layer. Therefore, the training data is split and trained at the hidden layer, sothat the training accuracy of the voice processing model is improved.

Description

technical field [0001] The present application relates to the technical field of data processing, in particular to a method and device for training a speech processing model. Background technique [0002] At present, model fine-tuning training usually adopts the strategy of using a single output layer for all data, that is, input all data into the input layer, shared hidden layer and output layer of the training model for training and output all data labels. However, there are some The data does not match the speech features of the target scene to a low degree, which leads to deviations in the model training results. Contents of the invention [0003] This application aims to solve one of the technical problems in the related art at least to a certain extent. [0004] For this reason, this application proposes a speech processing model training method and device, which is used to solve the technical problem that some data in all data in the prior art do not match the spee...

AI Technical Summary

Problems solved by technology

[0004] For this reason, this application proposes a speech processing model training method and device, which is used to solve the technical problem that some data in all data in the prior art do not match the speech features of the target scene to a low degree, resulting in deviations in the model training results. By splitting the training data at the hidden layer, the accuracy of the speech processing model training is improved.

Images