Glossary term
Glossary term
Infrastructure and Serving
An application-specific integrated circuit (ASIC) that optimizes the performance of machine learning workloads. These ASICs are deployed as multiple TPU chips on a TPU device.
Created for this library
An ML platform team uses TPUs for training its largest language models so a single training job completes within a day.
A computer vision team uses TPUs for both training and serving its production detector to lower cost per request.
A research lab uses TPUs for large-batch contrastive pretraining where matrix multiplications dominate the workload.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License