Developing ML solutionsFull notesSummaryRingkasanStoriesPracticeModel Deployment8 exam-style questions on this lesson.Question 1 of 8A team wants to avoid managing infrastructure and get automatic scaling and HTTPS endpoints for its model. Which deployment approach fits?AManaged API with SageMaker AIBRunning the model on a laptopCSelf-hosted API on its own serversDBatch jobs on premisesCheck answerQuestion 2 of 8A company needs maximum control and customization over its model hosting and accepts the extra operational work. Which approach fits?AManaged APIBSelf-hosted APICSageMaker Serverless InferenceDAmazon BedrockCheck answerQuestion 3 of 8A mobile app needs predictions in milliseconds for every user action, with a persistent endpoint. Which SageMaker inference option fits?AServerless inference with long idle periodsBAsynchronous inferenceCReal-time inferenceDBatch transformCheck answerQuestion 4 of 8A company needs predictions for a 50 GB dataset once a week and doesn't want a persistent endpoint. Which option fits?AReal-time inferenceBServerless inferenceCAsynchronous inferenceDBatch transformCheck answerQuestion 5 of 8Each request sends a 700 MB video file, and processing takes about 40 minutes. Users can wait, and requests should be queued. Which option fits?AAsynchronous inferenceBReal-time inferenceCBatch transformDServerless inferenceCheck answerQuestion 6 of 8An internal tool gets a few requests a day with long idle periods between them. The team can accept a short delay on the first request after idle time and doesn't want to manage infrastructure. Which option fits?AReal-time inferenceBServerless inferenceCAsynchronous inferenceDSelf-hosted APICheck answerQuestion 7 of 8. Select two.Which TWO statements about SageMaker inference options are correct? (Select TWO.)ABatch transform gives the lowest latency per requestBBatch transform doesn't need a persistent endpointCAsynchronous inference queues requestsDServerless inference has no cold startsEReal-time inference is best for 1 GB payloads that take an hourCheck answer0 of 2 selectedQuestion 8 of 8What does model deployment mean?ACollecting and storing the training dataBChoosing hyperparameters before trainingCPutting the model into production to make predictionsDLabeling the dataset with the correct answersCheck answerMachine Learning Models Performance Evaluation12 exam-style questions on this lesson.Fundamental Concepts of MLOps10 exam-style questions on this lesson.