On the Keyword Spotting Configuration page, define the commands your model will recognize. To add a command, complete the following steps:
- In the Configure Commands section, enter a word in the Add New Command field.
- Optionally, select Play Sample to hear how the command will be pronounced. If the pronunciation is not as expected, adjust the capitalization or spacing.
- Select Add to add the command to your list.
- Repeat for each command you want the model to recognize.
When you have added all your commands, select Start to begin training. Training starts automatically and typically takes 2–3 hours to complete, depending on the number of commands. You can safely leave the page — the results will be available after training is complete.