It implements matrix operations and reduced-precision arithmetic efficiently, with software that maps model graphs across one or more devices. Workload shape and compiler support determine how much of its capacity is used.
A tensor processing unit is a specialized accelerator designed for high-throughput tensor computation in machine learning.
It implements matrix operations and reduced-precision arithmetic efficiently, with software that maps model graphs across one or more devices. Workload shape and compiler support determine how much of its capacity is used.