A company deploys a TensorFlow model on Vertex AI Prediction with a single node. During peak hours, inference latency increases. What should they do first to reduce latency?
Select one:
The trap here is that candidates often confuse improving throughput (batching or bigger machines) wi...