MLA-C01 ML Model Development Practice Test 3 — 15 Questions

Question 1

A machine learning engineer is using SageMaker Debugger to detect if a neural network has dead ReLU units during training. Which built-in rule should they enable?

Accepted Answer

DeadRelu. The 'DeadRelu' rule in Debugger monitors the fraction of ReLU activations that are zero and alerts if too many neurons are dead.

Answer

Overfit

Answer

ExplodingGradients

Answer

LossNotDecreasing

Question 2

A data scientist is using SageMaker Autopilot for a regression problem. They want to see which data preprocessing steps Autopilot applied. Which TWO sources can they use to find this information?

Accepted Answer

Candidate definition notebook. The two sources that show the data preprocessing steps applied by SageMaker Autopilot are the candidate definition notebook and the data exploration report. The candidate definition notebook contains the generated Python code for the entire pipeline, including all preprocessing transformations. The data exploration report provides a summary of the data and includes information about the transformations that were automatically applied. The model leaderboard only lists the performance metrics of the trials, not the preprocessing steps. The Autopilot job description in AWS CloudTrail logs the API calls made, but does not capture the preprocessing steps. The explainability report shows feature importance values, not the preprocessing steps.

Answer

Model leaderboard

Answer

Autopilot job description in AWS CloudTrail

Answer

Explainability report

Question 3

A company is using SageMaker to train a large model using data parallelism with the SageMaker distributed data parallelism library. They notice that the training throughput is not scaling linearly with the number of GPUs. Which THREE factors could be causing this?

Accepted Answer

I/O bottleneck from reading data from Amazon S3. Communication overhead from gradient synchronization, I/O bottlenecks from reading data, and an inefficient loss scaling strategy can all limit scaling. Model size alone is not a scaling issue if it fits on GPUs. Instance type differences affect speed but not scaling linearity directly.

Answer

Using different instance types across the cluster

Answer

Model size too large for the GPUs

Question 4

A data scientist wants to quickly build a binary classification model without writing any code. Which SageMaker feature is MOST suitable?

Accepted Answer

SageMaker Autopilot

Answer

SageMaker Debugger

Answer

SageMaker Model Monitor

Answer

SageMaker Ground Truth

Question 5

Which SageMaker built-in algorithm is designed for time series forecasting?

Accepted Answer

DeepAR

Answer

Linear Learner

Answer

Factorisation Machines

Answer

BlazingText

Question 6

A machine learning engineer wants to reduce training costs by using excess EC2 capacity. Which instance purchasing option should they choose for SageMaker training jobs?

Accepted Answer

Spot Instances

Answer

Reserved Instances

Answer

On-Demand Instances

Answer

Dedicated Instances

Question 7

A team is training a large language model and needs to split the model layers across multiple GPUs due to memory constraints. Which distributed training strategy should they use?

Accepted Answer

Model parallelism

Answer

Data parallelism

Answer

Hyperparameter tuning

Answer

Autopilot

Question 8

A company uses SageMaker Experiments to track training runs. They want to compare different hyperparameter configurations and identify the best run. Which SageMaker Experiments component should they use to organize related runs?

Accepted Answer

Experiment + Trial

Answer

Trial

Answer

Experiment

Answer

Trial Component

Question 9

A data scientist is using SageMaker Automatic Model Tuning to optimize hyperparameters for an XGBoost model. They want to maximize AUC. Which search strategy is MOST appropriate for efficient exploration?

Accepted Answer

Bayesian optimization

Answer

Random search

Answer

Grid search

Answer

Hyperband

Question 10

A team is fine-tuning a Hugging Face transformer model on SageMaker. They need to use a custom training script with the Hugging Face Estimator. Which SageMaker feature does this represent?

Accepted Answer

Script mode

Answer

Built-in algorithm

Answer

SageMaker Autopilot

Answer

SageMaker Debugger

Question 11

A fraud detection model is being trained on imbalanced data. The team wants to ensure the model's precision is optimized. Which objective metric should be used in automatic model tuning?

Accepted Answer

Precision

Answer

F1

Answer

AUC

Answer

Recall

Question 12

A machine learning engineer runs a training job and notices the loss is NaN after a few steps. Which SageMaker Debugger rule can help identify this issue?

Accepted Answer

Exploding gradients

Answer

Overfit

Answer

Dead ReLU

Answer

Class imbalance

Question 13

A team is fine-tuning a foundation model using LoRA for a text summarization task. They want to reduce memory footprint during training. Which technique should they combine with LoRA?

Accepted Answer

QLoRA

Answer

Data parallelism

Answer

Gradient checkpointing

Answer

Mixed precision

Question 14

A company needs to detect bias in a pre-trained model before deployment. They want to compute metrics like disparate impact and equal opportunity difference. Which AWS service should they use?

Accepted Answer

SageMaker Clarify

Answer

Amazon Rekognition

Answer

SageMaker Model Monitor

Answer

SageMaker Debugger

Question 15

A team is training a large model on SageMaker using the SageMaker distributed training library with model parallelism. They need to choose the most cost-effective instance type. Which instance family offers the best balance of performance and cost for large model training?

Accepted Answer

ml.trn1

Answer

ml.g4dn

Answer

ml.p3

Answer

ml.c5