Glossary term
Glossary term
Infrastructure and Serving
A specialized hardware accelerator designed to speed up machine learning workloads on Google Cloud.
Created for this library
An ML platform team provisions Cloud TPU pods for training its largest language model so a single training job completes overnight.
A computer vision team uses Cloud TPU v5e for both training and serving its production object detector to reduce inference cost per request.
A research lab uses Cloud TPU 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