NVIDIA-Certified-Professional Accelerated Data Science - NCP-ADS

You are processing a large dataset with UNIX timestamps (seconds since Jan 1, 1970) ranging from Jan 1, 2000, to the present.
What is the most memory-efficient data type for the timestamp column in a GPU-accelerated cloud environment?

• A. df['timestamp'] = df['timestamp'].astype('datetime64[ms]')
• B. df['timestamp'] = df['timestamp'].astype('float32')
• C. df['timestamp'] = df['timestamp'].astype('int64')
• D. df['timestamp'] = df['timestamp'].astype('int32')

You are comparing the performance of GPU-accelerated deep learning models on two cloud platforms: AWS EC2 and Google Cloud Platform (GCP). You want to design a benchmark that evaluates GPU resource utilization, processing time, and cost-efficiency for training models with large datasets.
Which actions should you take to implement an effective benchmark? (Select two)

• A. Only consider the cost of GPU resources for each cloud provider without factoring in data storage and transfer costs.
• B. Limit the benchmarking tests to a single model architecture for each cloud platform.
• C. Measure both GPU utilization and network latency between the cloud platform and your data storage location.
• D. Run each model on the same GPU type and instance size across both cloud platforms.

Which of the following best describes a key advantage of using cloud-based GPU instances for machine learning model training?

• A. Cloud-based GPU instances offer lower latency and better network performance compared to on- premise deployments, regardless of geographical location.
• B. Cloud GPU instances cannot support containerized workloads, limiting their applicability for MLOps and CI/CD pipelines.
• C. Cloud GPUs provide dynamically scalable resources, allowing users to increase or decrease compute power based on demand without upfront hardware investment.
• D. Cloud GPUs are always more cost-effective than on-premise GPUs, as they do not incur long-term usage costs.

A data engineer is preparing a dataset for training a deep learning model. The dataset contains numerical features with missing values, outliers, and inconsistent units.
Which of the following strategies is the most appropriate for ensuring a standardized and clean dataset?

• A. Remove all rows with missing values and outliers to ensure only clean data is used.
• B. Standardize the dataset using the mean and standard deviation, but keep missing values and outliers unchanged to avoid data manipulation.
• C. Use the median to fill missing values, convert all numerical values into categorical bins, and apply Min-Max scaling.
• D. Replace missing values with the mean, apply z-score normalization, and clip extreme outliers based on a threshold (e.g., 3 standard deviations).

You are preprocessing a dataset using NVIDIA RAPIDS cuDF and need to handle missing values in the column temperature by replacing them with the column's median value.
Which of the following approaches correctly achieves this in an optimized manner?

• A. df['temperature'].dropna(inplace=True)
• B. 1. df['temperature'] = df['temperature'].map(2. lambda x: df['temperature'].median() if x is None else x
  3.)
• C. df['temperature'].fillna(df['temperature'].mean(), inplace=True)
• D. df['temperature'].fillna(df['temperature'].median(), inplace=True)