H13-311_V3.5 · Question #123
What are the implementation modes ofTensorflow? (Multiple Choice)
The correct answer is A. Stand-alone mode B. D1stnbuted mode. TensorFlow supports two primary implementation modes: Stand-alone mode (A), where the entire computation runs on a single machine using one or more local CPUs/GPUs, and Distributed mode (B), where computation is spread across multiple machines or devices in a cluster for…
Question
What are the implementation modes ofTensorflow? (Multiple Choice)
Options
- AStand-alone mode
- BD1stnbuted mode
- CReverse mode
- DForward mode
How the community answered
(27 responses)- A85% (23)
- C4% (1)
- D11% (3)
Explanation
TensorFlow supports two primary implementation modes: Stand-alone mode (A), where the entire computation runs on a single machine using one or more local CPUs/GPUs, and Distributed mode (B), where computation is spread across multiple machines or devices in a cluster for large-scale training.
Why C and D are wrong: "Reverse mode" and "Forward mode" refer to automatic differentiation techniques (reverse-mode autodiff is actually what TensorFlow uses internally for backpropagation), not deployment/implementation modes. These are concepts from calculus/optimization, not TensorFlow runtime configurations.
Memory tip: Think of the modes in terms of where the computation runs - Stand-alone = Single machine, Distributed = Different machines. The two autodiff options (C & D) are a trap for students who've studied how gradients are computed internally.
Topics
Community Discussion
No community discussion yet for this question.