In summer 2018, Google announced two Edge TPU devices for machine learning. These are now available under the Coral brand. I bought the Coral USB Accelerator. It is a portable USB accessory that brings machine learning inferencing to existing systems and it is compatible with Raspberry Pi and other Linux systems.
I wanted to test the performance of the Edge-TPU on my Raspberry Pi3, and as I have every application running on Docker on my Raspberry Pi, I've created a Docker image with everything inside that you need to start using the Coral USB accelerator.
Also I included two example. An live object identification script, and a Jupyter Notebook for a teachable machine.
You can find more detailed information on the following links: