TPU — Tensor Processing Unit is an ASIC AI accelerator used for machine learning tasks specifically developed using Tensorflow. Edge TPU is designed and released in early 2019 and used in ML Inference acceleration use cases at the Edge. One such Edge TPU device is Coral USB Accelerator.

We can connect the Edge TPU Device( i.e. Coral USB Accelerator) to a computer and perform inference actions using the same. Coral USB Accelerator — Edge TPU is capable of performing 4 trillion operations per second using 2W.
Coral USB Edge TPU suppo
rts only TensorFlow Lite models that are fully 8-bit quantized and then compiled specifically for the Edge TPU. TF Lite is a lightweight version of TensorFlow designed for mobile and embedded devices. It achieves low-latency inference. We cannot train a model directly with TensorFlow Lite, instead, we must convert the model from a TensorFlow file (such as a .pb file) to a TensorFlow Lite file (a .tflite file), using the TensorFlow Lite converter.

#machine-learning #ai #coral #tpu #asics