Google continues to expand its range of AI products and services with a trio of new hardware devices aimed at the development community.
It appears that the devices have not yet been officially announced and were first detected by Hackster . They are being introduced under a brand new Google Coral (which is still in "beta"), and include a development board that sells for $ 149.99, a USB accelerator that costs $ 74.99 and a 5-megapixel camera. available for $ 24.99.
Both the development board and the accelerator are powered by Google's Edge TPU chips, which are ASIC processors that are no larger than the nail and are designed to run AI models without breaking the sweat. The camera, meanwhile, is a complement to the development board.
None of these products is good for training machine learning algorithms. That is still a heavy task that requires GPU or processing capacity in the cloud. Instead, they facilitate the "inference" part of the AI, which comes after you have trained your model and are using it in nature. For that purpose, both the accelerator and the development board are compatible with TensorFlow Lite, a lightweight version of the company's AI framework that is designed for mobile devices and IoT.
Of course, you can run AI models on any old CPU, but they will be slow and in a bad mood. Specially designed chips, such as Google TPU processors, do the job much more efficiently without the need for a dedicated Internet connection or a bulky power source.
The USB accelerator can increase inference on any Linux machine, while the series of pins and ports on the development board make it perfect for hardware prototyping and other experimental applications. ( Hackaday notes that its configuration is almost identical to that of the Raspberry Pi.)
Google is not the first company to offer this type of product to developers, since Intel launched its own USB Accelerator version a few years ago, the Neural Compute Stick by Movidius. Products like this help companies reach the developer community, driving the adoption of their other AI services.
For more information on these devices, you can find fact sheets and complete resources, including tutorials and test models, on the new Google Coral homepage.