easy5 pts · first answer countsgpu-computingdistributed-training

What data parallelism does

In data-parallel training on 4 GPUs, how is the work split?

A.Each GPU holds the full model and gets different dataB.Each GPU holds one quarter of the layersC.One GPU trains while three wait as backupsD.Each GPU trains a different model

Discussion 0

Join free to post in the discussion.

What data parallelism does

Discussion 0

Related questions