Since 2015, Google has been using its own internally-developed Tensor Processing Unit (TPU) chips to run machine-learning applications inside its data centers at speeds 15 to 30 times faster than traditional CPUs and GPUs.
Collaborate Disseminate
Since 2015, Google has been using its own internally-developed Tensor Processing Unit (TPU) chips to run machine-learning applications inside its data centers at speeds 15 to 30 times faster than traditional CPUs and GPUs.