Google Coral USB Edge TPU ML Accelerator coprocessor for Raspberry Pi and Other Embedded Single Board Computers

Product Information

Once you start adding accessories like the sensor board and camera, your IoT projects will have no limits. This cookie is installed by Google Universal Analytics to restrain request rate and thus limit the collection of data on high traffic sites. This on-device ML processing reduces latency, increases data privacy, and removes the need for a constant internet connection. Google does provide some documentation on that but it’s much more advanced, far too much for me to include in this blog post.

Note: The first inference on Edge TPU is slow because it includes loading the model into Edge TPU memory. This is essential to build AI inferencing solutions in the field, with many distributed devices in a challenging setting (temporary power and network constraints). You can calculate the maximum performance of your Coral based on the inference speed reported by Frigate. A few weeks ago, Google released “Coral”, a super fast, “no internet required” development board and USB accelerator that enables deep learning practitioners to deploy their models “on the edge” and “closer to the data”.

Stream images from a camera and run classification or detection models with the TensorFlow Lite API. If the item comes direct from a manufacturer, it may be delivered in non-retail packaging, such as a plain or unprinted box or plastic bag. The information does not usually directly identify you, but it can give you a more personalized web experience. At last year’s Google Next conference in San Francisco Google announced two new upcoming hardware products both built around Google’s Edge TPU, their purpose-built ASIC designed to run machine learning inferencing at the edge. After quantization, you need to convert your model from Tensorflow to Tensorflow Lite and compile it using the Edge TPU compiler.

Contact us for press inquiries and free product samples for accredited journalists, bloggers, and YouTubers.

The Coral USB Accelerator comes in at 65x30x8mm, making it slightly smaller than its competitor, the Intel Movidius Neural Compute Stick. You could easily modify the script to ignore detections with < 50% probability (we’ll work on custom object detection with the Google coral next month). Best of all, you can manage the Python packages inside your your virtual environment inside with pip (Python’s package manager). If you only need a single camera input, perhaps the OAK cameras with integrated MyraidX could fill the bill.