NVIDIA Certified Associate – AI Infrastructure and Operations — Question 1

A company is implementing a new network architecture and needs to consider the requirements and considerations for training and inference. Which of the following statements is true about training and inference architecture?

Answer options

• A. Training architecture and inference architecture have the same requirements and considerations.
• B. Training architecture is only concerned with hardware requirements, while inference architecture is only concerned with software requirements.
• C. Training architecture is focused on optimizing performance while inference architecture is focused on reducing latency.
• D. Training architecture and inference architecture cannot be the same.

Correct answer: C

Explanation

Option C is correct because training architecture prioritizes optimizing performance to handle the extensive computations involved in training models, while inference architecture is designed to minimize latency for quick predictions. Options A and B are incorrect as they misrepresent the distinct requirements of training and inference. Option D is also incorrect as, while they serve different purposes, training and inference architectures can share components.