Practice Test 1 | Google Cloud Certified Professional Data Engineer | Dumps | Mock Test
You are deploying a Tensorflow model built by the data science team to the cloud. Based on the requirements provided by data scientists, the model should be able to return the output as soon as possible to minimize the latency of serving predictions. Input will be passed as JSON.
Which of the following approaches are best for this scenario?
A. Use Google Kubernetes Engine to deploy the model. Use online prediction to pass input data to the model hosted in cloud.
B. Use Google Kubernetes Engine to deploy the model. Use batch prediction to pass input data to the model hosted in cloud.
C. Use Vertex AI to deploy the model. Use batch prediction to pass input data to the model hosted in cloud.
D. Use Vertex AI to deploy the model. Use online prediction to pass input data to the model hosted in cloud.
Answer: D
Vertex AI provides two ways to get predictions from trained models: online prediction (sometimes called HTTP prediction), and batch prediction. In both cases, you pass input data to a cloud-hosted machine-learning model and get inferences for each data instance.
Online prediction passes input as a JSON string and returns the output as soon as possible.
Options A and B are incorrect: GKE is not a recommended option to deploy the model.
Option C is incorrect: Batch prediction doesn’t support returning the output as soon as possible. Input is passed indirectly as one or more URIs of files in Google Storage.
Source(s):
Online vs. Batch Prediction: https://cloud.google.com/ml-engine/docs/tensorflow/online-vs-batch- prediction
Comments are closed, but trackbacks and pingbacks are open.