Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Update] Google Speech-to-Text API updates transcription models #392

Open
Tsu-HaoLiu opened this issue Sep 28, 2023 · 1 comment
Open

[Update] Google Speech-to-Text API updates transcription models #392

Tsu-HaoLiu opened this issue Sep 28, 2023 · 1 comment

Comments

@Tsu-HaoLiu
Copy link

Looks like google is making an update to the speech to text api.


Dear Speech-to-Text user,

We’re writing to let you know about the changes coming to Google Cloud Speech-to-Text API. We’ll migrate our classical speech models to our conformer-based models, aiming to improve speech recognition accuracy and performance across a range of use-cases.

You are receiving this notification because we have detected that one or more of your projects has Speech-to-Text API enabled.

What do you need to know?

Since this is a significant model architecture change, we expect it to improve alphanumeric character recognition, enhanced biasing effectiveness, and overall transcription robustness.

As part of this migration, we are updating the all of our models that are exposed through Speech-to-Text V1 API in the corresponding languages and locales:

Model Identifier in V1 Model Identifier in V2 BCP-47 codes
latest_long long de-DE, en-AU, en-GB, en-IN, en-US, es-ES, es-US, fr-CA, fr-FR, it-IT, ja-JP, nl-NL, pt-BR
latest_short short
command_and_search short
phone_call telephony
video long
default long

What do you need to do?

This change is an internal speech model migration and as such no action is needed from your side in order to use these models. Customers who have migrated to Speech-to-Text V2 API are already using the latest version of the models, under the identifiers presented in the table above. Customers who have not migrated to the Speech-to-Text V2 API and are still on V1 will be automatically migrated starting October 17, 2023, using the existing V1 identifiers. This will not break any backwards compatibility.

If you want to be migrated automatically, no action is required on your part, as this will happen in accordance to the timeline above.

If you want to opt-out temporarily and migrate in your own time, you can do so by November 10, 2023. Using our Google Cloud Speech console, navigate to the “Preview features” section in the navigation bar on the left and enable the dedicated toggle to opt-out.

We’re here to help

If you have any questions or require assistance, please contact Google Speech-toText support. If you’d like to learn more about Speech-to-Text and our latest V2 API, please check Speech-to-Text v2 API resources.

Thanks for choosing Google Speech-to-Text.
— The Google Speech-to-Text Team

@dessant
Copy link
Owner

dessant commented May 20, 2024

Thanks for the heads up regarding the internal model changes! We could also support the new API version, but it looks like Speech-to-Text V2 does not a have a free tier. The lack of a free monthly quota makes it useless for Buster while Speech-to-Text V1 is still available, because V1 works very well and users typically only use up some of the free quota.

https://cloud.google.com/speech-to-text/pricing

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants