What is the minimal resource required for Speech on device?

Google Cloud Build with AI AI APIs
,
1

I have read through this article intruducing Speech on device solution. I have some questions about this solution.

  1. What is the minimal resource (CPU/RAM/Disk space etc.) required for Speech on device ?

  2. If the device is unable to connect to the internet, will the speech to text result degrade ?

  3. What is the datatype of the model (float32/64)?

I can’t find any answer in Google Cloud document, if anyone knows the detail please tell me, thank you!

1 Like
2

I believe this solution is based on Google’s Cloud STT so it has minimal effect on its system requirements. The documentation also stated that its offline function is comparable to its server made possible with Google’s new modelling technique.

With these said I would recommend to contact sales for further inquiry about the solution, this will be helpful and possible to align this solution to your use case: https://cloud.google.com/contact

2 Likes