How to create and apply a custom model in the console-Intelligent Speech Interaction(ISI)-阿里云帮助中心

Prerequisites

You have activated the Intelligent Speech Interaction service. For more information, see Activate Service.

Training corpus

Limitations

This feature is available free of charge to all users of the free and commercial editions of Intelligent Speech Interaction.
The training corpus must contain domain-specific text. The closer the corpus is to the audio that you want to recognize, the better the optimization results.
Save the corpus as a plain text file encoded in UTF-8 (without BOM). The file size cannot exceed 10 MB.
Each user can create up to 10 models.
Place each sentence or keyword that you want to emphasize on a separate line. Each line can contain up to 500 characters.
Replace digits with the corresponding Chinese words based on their pronunciation. For example, replace "58.9元" with "五十八点九元".
The file must contain at least one sentence with more than four words.
Use only commas (，), periods (。), question marks (？), and exclamation points (！), and end each sentence with a punctuation mark. Remove punctuation marks such as book title marks (《》) and double quotation marks (“”).

Optimization tips

If a keyword is recognized inaccurately, repeat the sentence that contains the keyword, or the keyword itself on a separate line, multiple times, such as 10 times. If the result is still unsatisfactory, increase the number of repetitions.

Note

Before you adjust the training corpus, make sure that the inaccurate keyword recognition is not caused by unclear pronunciation or poor audio quality.
Test and refine the model carefully to prevent misrecognition of other words that have similar pronunciations.

Procedure

Important

Custom model names must be unique.

Log on to the Intelligent Speech Interaction console.
In the left-side navigation pane, choose Self-Learning Platform > Custom Linguistic Models. If you use this feature for the first time, follow the on-screen instructions to activate it.
On the Models tab, click Create Model.
1. Select Basic Model: Enter a Model Name, Select Basic Model, and then click Next.
2. Upload Corpus: Select Upload Dataset or Select Existing Dataset, and then click OK.
Click Training Model next to the created model. After training is complete, the model status automatically changes to Deployed, and the model can be used in a project.
In the left-side navigation pane, click All Projects, and then click Project Settings next to the project.
In the Custom Model section, select the created model for 语言模型.

Note
The base model selected during language model customization must match the speech recognition model configured for the current project. Otherwise, the custom model does not appear in the drop-down list.