A dock-camera slip classifier sits on ml.g4dn.xlarge because “vision uses GPU.” Nobody has benchmarked families or shapes for cost and latency. What should they run?
Select an answer to reveal the explanation.
Short Explanation
The slip classifier sits on a GPU because vision uses GPU, and nobody has benchmarked families. Run SageMaker Inference Recommender. Compute Optimizer does not replace that job for a hosted endpoint, and guessing a larger g5 skips the look.
Full Explanation
SageMaker Inference Recommender benchmarks instance families and shapes for a hosted endpoint’s cost and latency. Compute Optimizer rightsizes EC2 and similar compute; it is not the SageMaker inference recommender. Lex is not that job, and guessing a larger GPU skips the benchmark.