PMLE Scaling Prototypes into ML Models Practice Question

An ML engineer has a prototype that trains a TensorFlow model on a single CPU machine using Vertex AI custom training. The job now needs to train on a larger dataset and must use multiple GPUs on one machine. The training script already uses tf.distribute.MirroredStrategy. What change is required to scale the job?

⚠ Common exam trap

The trap here is assuming that multi-GPU training always requires a multi-worker strategy, when a single machine with multiple accelerators works with MirroredStrategy.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Set the worker pool machine type to a GPU machine and specify the number of accelerators and accelerator type.

Scaling a single-machine TensorFlow job that already uses MirroredStrategy to multiple GPUs is primarily a resource configuration task. You set the machine type to a GPU-capable machine and specify the accelerator type and count in the worker pool spec. The strategy code detects the local GPUs automatically. Multi-worker strategies, TPU strategies, and disk or image changes are not required for this scenario.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Add a second worker pool with one replica and set the distribution strategy to MultiWorkerMirroredStrategy.
Why it's wrong here
MultiWorkerMirroredStrategy is used for multi-worker, multi-node training, not for scaling within a single machine. The scenario requires multiple GPUs on one machine, and the script already uses MirroredStrategy, which handles local GPUs. Adding a second worker pool would introduce unnecessary multi-node configuration and would not match the stated requirement.
✗
Enable TPU training by specifying a TPU machine type and changing the script to use TPUStrategy.
Why it's wrong here
TPUs are a different accelerator family and require the script to use TPUStrategy rather than MirroredStrategy. The requirement is to use multiple GPUs on one machine, so switching to TPUs changes the hardware target and the code. This would be a larger change than needed and does not match the stated GPU requirement.
✓
Set the worker pool machine type to a GPU machine and specify the number of accelerators and accelerator type.
Why this is correct
Vertex AI custom training uses the worker pool specification to select the machine type, accelerator type, and accelerator count. Because the script already uses MirroredStrategy, it will automatically detect the local GPUs and replicate training across them. Providing a GPU machine type with the desired accelerator count is the only configuration change needed to scale this single-machine job.
✗
Increase the boot disk size and set the training container to use the GPU-enabled base image.
Why it's wrong here
Disk size and base image affect storage and runtime environment, but they do not attach GPUs to the worker pool. Without specifying an accelerator type and count, the job will not have GPUs available, and MirroredStrategy will fall back to CPU devices. This choice does not address the core requirement of provisioning GPUs for the training job.

Go deeper

PMLE Scaling Prototypes into ML Models Practice Question

Why each option matters

Option-by-option breakdown

Related to this question