ai speech csr model create
Usage
Options
| Option | Description |
|---|---|
--key KEY |
Specifies the key for the custom speech service connection. |
--region REGION |
Specifies the region of the custom speech service connection. |
--project URL |
URL of the project to which the custom model should be added. |
--name NAME |
Name of the custom speech recognition model. |
--text @FILE |
Path to the text file containing training data. |
--base URL |
URL of the base model to use for creating the custom model. |
--dataset URL |
URL of the dataset to be used for training. |
--datasets URL1;URL2 |
Semicolon-separated URLs of multiple datasets to use for training. |
--language LANGUAGE |
Language code for the custom speech recognition model. |
--description DESCRIPTION |
Description of the custom model being created. |
--output json FILENAME |
Save the output in JSON format to the specified file. |
--output url @@FILE |
Save the output URL to the specified file. |
--output id @@FILE |
Save the output ID to the specified file. |
--wait [TIMEOUT] |
Wait for the operation to complete with an optional timeout. |
--input path PATH |
Path to the input data for creating the model. |
--output path PATH |
Specifies the output path for the resultant model. |
--foreach in @FILENAME |
Apply the command to each item in the specified file. |
--save FILENAME |
Save the command output to the specified file. |
--zip ZIPFILE |
Create a ZIP file of the output data. |
Examples
Create a custom speech recognition model with a specified project and name
ai speech csr model create --project https://myprojecturl --name "MyCustomModel" --key myapikey --region westus --text @/path/to/textfile --base https://basemodelurl --dataset https://dataseturl --language en-US --description "Custom model for specific domain"
Create a custom speech recognition model with multiple datasets
ai speech csr model create --project https://myprojecturl --name "MyCustomModelMultiDataset" --key myapikey --region westus --text @/path/to/textfile --base https://basemodelurl --datasets https://dataseturl1;https://dataseturl2 --language en-US --description "Custom model using multiple datasets"