Batch

Available in Classic and VPC

You can batch process multiple speech recognition tasks by creating a batch. Batch operates automatically after a certain period after creation, recognizes the speech of media files in the specified path, and creates a recognition result file and an execution result file.

Note

• The fees for batch execution are charged per file to be recognized.
• An execution result file (CSV) is created for every 10 tasks.
  • If there are fewer than 10 tasks, an execution result file will be created if recognition is not executed within a certain period.
  • The execution result file contains the following information:
    • Work sequence
    • Recognition targets and name of the result file
    • Time required for recognition
    • Task start/finish time
    • Execution result

The Batch menu allows you to create and manage batches.

Batch screen

The basic description of the Batch menu for using CLOVA Speech is as follows:

Area	Description
① Menu name	Name of the menu currently being viewed
② Basic features	Batch creation, CLOVA Speech detailed information check, batch screen refresh
③ Post-creation features	Delete batch
④ Search area	Initialize batch search and search results
⑤ Batch list	List of batches being searched

Check batch list

The following describes how to check the information on the created batch.

1. Click the environment you are using in the Region menu and Platform menu of the NAVER Cloud Platform console.
2. Click Services > AI Services > CLOVA Speech menus in that order.
3. Click the Batch menu.
4. When the batch list appears, check the batch information
   • Batch ID: batch ID that is automatically assigned
   • Batch name: batch name entered when creating a batch
   • Domain: domain name/domain code of the domain where the batch was created
   • File format: typo of the recognition result file generated by batch
   • Recognition target storage path: object storage path that stores recognition target files
   • Result file storage path: an object storage path that stores recognition result files and execution result files
   • Status: batch availability
     • Normal: available
     • Error: unavailable
       • You can mouseover to check the cause of the error.
   • [Setting] button: click it to change the settings of the batch (see Change batch settings)

Create batch

The following describes how to create a batch.

Note

• First, you need to create a basic domain to connect the batch to. For more information on domain creation, see Create domain.
  • You cannot create a batch in free domains.
• You can create 1 batch per domain, up to a maximum of 10 batches.

1. Click the environment you are using in the Region menu and Platform menu of the NAVER Cloud Platform console.
2. Click Services > AI Services > CLOVA Speech menus in that order.
3. Click the Batch menu.
4. Click the [Create batch] button.
5. Enter batch information.
   • Batch name: enter batch name
   • Domain ID: click the domain to connect the batch to in the domain list that appears after clicking the input field
     • You can search for a domain by entering the domain name or code.
     • Free domains and domains to which batches are already connected are not displayed in the list.
   • File format: select the format of the recognition result file that the batch will generate
   • Language settings: select the language to use for the recognition task
   • Use the speaker separation function: select whether or not to use speaker separation
   • Keyword boosting: select whether or not to use the keyword boosting function (For more information on keyword boosting, see See keyword boosting)
6. Enter the object storage path to save the recognition target and result file.
   • Recognition target storage path: after clicking the input field, select the object storage path to save the file to be recognized in the pop-up window.
   • Result file storage path: after clicking the input field, select the object storage path to save the recognition result file and execution result file in the pop-up window.
7. If necessary, click of the advanced settings, and set up Webhook.
   • URL: enter the URL to which batch execution results are sent.

Note

Whenever a task is completed, a webhook including the recognition result is sent to the set URL.

8. Click the [Create] button.
   • It may take some time for the batch to operate normally.
   • You can check the recognition result file and execution result file in the result file storage path.
   • Files that have been recognized are moved to a folder with the same name as the automatically created batch in the recognition target storage path.

Change batch settings

The following describes how to change the settings of a created batch.

1. Click the environment you are using in the Region menu and Platform menu of the NAVER Cloud Platform console.
2. Click Services > AI Services > CLOVA Speech menus in that order.
3. Click the Batch menu.
4. In the batch list, click the [Settings] button of the batch whose settings you want to change.
5. Change batch settings.
   • For details of information items, see Create batch.
6. Click the [Edit] button.

Delete batch

The following describes how to delete a created batch.

1. Click the environment you are using in the Region menu and Platform menu of the NAVER Cloud Platform console.
2. Click Services > AI Services > CLOVA Speech menus in that order.
3. Click the Batch menu.
4. Select the batch to delete from the batch list, and click the [Delete batch] button.
5. When the Alert pop-up window appears, click the [Delete batch] button.