Practice free →
HomeGCP ML Engineercloudcomputingmletraining › Which Vertex AI capability should an ML Engineer…

Which Vertex AI capability should an ML Engineer use to train a model on data PARTITIONED across multiple workers — for datasets too large to fit on a single accelerator?

ACloud Memorystore
BSingle-worker training only
CCloud DNS
DVertex AI distributed training (data parallelism across workers, optionally with model parallelism)
Answer & Solution
Correct answer: D. Vertex AI distributed training (data parallelism across workers, optionally with model parallelism)
Distributed training enables data parallelism + model parallelism (per PMLE §3.2). The other options can't scale.
Solve this in the app — GCP ML Engineer practice & 24k+ MCQs →
Related questions