Is Coral/TPU supported for acceleration of model/compute?

[typoworx-de]

I'm having coral tpu device (USB) and I'm asking myself if this device could help speeding up compute for the models? Is this supported somehow?

Is this announcement related to this project?

Introducing Coral: Our platform for development with local AI
https://developers.googleblog.com/2019/03/introducing-coral-our-platform-for.html

[imightbelosthere]

Hi,
I'm looking for the same type of information... did you ever found a way to get LocalAI working with the Coral TPU?

[Coolwrt]

HI, thank you WELL!

[MarvinSchenkel]

Also would love some insight in this. I am currently running Local AI on a NUC that does not have a way to add GPU acceleration

[zentaos]

Is there any progress on this? It would be awesome to boost AI without GPU

[BlackRockCity]

I'm interested in this too! Specifically, running Local AI in a Docker container and passing a USB Coral TPU (or perhaps M2) to the Docker container.