Skip to main content

Deploy a Large Vocabulary Streaming Transcription service

A streaming transcription service deploys a API that accept an audio stream and return a real-time transcription. "Large vocabulary" implies that the transcription won't be limited to preset sentences and intend to cover an entire language..

Prerequisistes​

  • You have a deployed stack with LinTO Platform Service Manager running.
  • You have access to the LinTO Platform Service Manager Swagger at [LINTO_STACK_DOMAIN]/stt-manager/api-doc/.
  • You have created an acoustic model for your language. If not see Create an acoustic model

Creating the service​

In the following steps we will create a large vocabulary streaming transcription service.

  1. Create a large vocabulary language model
  2. Create a streaming transcription service
  3. Submit transcription requests

1- Create a large vocabulary language model​

In this step you will create a large vocabulary language model.

A language model takes the phonemes extracted from the audio signal by the acoustic model and map them with known sentences.

If you already created a large vocabuly language model you do not need to create an other. You can skip this step.

Connect to the STT Service Manager Swagger.

  • lang: The intended language.
  • file or link: Set the path or the link of a large language model.
  • acousticModel : The acoustic model you created for the intended language.
  • type: Select lvcsr for large vocabulary.

Tap Execute and the STT-Manager should return a code 200.

Now you have a large vocabulary language model set in your STT Service Manager.

2- Create an online transcription service​

In this step we will create a large vocabulary streaming transcription service .

On the STT Service Manager Swagger.

  • serviceId: Service Name.
  • replica: Set the number of instance you wish to create.
  • tag : Set to online for a streaming transcription service.
  • languageModel : Set the modelId of the LV language model created on the previous step.

You have now a live large vocabulary online transcription service deployed on your stack.

3- Connect and submit request​

For now, use the streaming service through LinTO applications, documentation pending.