Question 18
Domain 3: NVIDIA Tools, Performance, and DeploymentWhich model format is supported by TensorRT for optimization?
Correct answer: C
Explanation
TensorRT supports multiple model formats for optimization, including ONNX, because it is designed to import and optimize models from common deep learning frameworks. ONNX is widely used as an interchange format, so TensorRT can take an ONNX model and build an optimized inference engine from it.
Why each option is right or wrong
A. Only PyTorch models
B. Only TensorFlow models
C. Multiple formats including ONNX
TensorRT’s documented import path includes ONNX models via the ONNX parser, which is part of the TensorRT API and allows conversion into an optimized inference engine. NVIDIA’s TensorRT documentation also notes support for models from multiple frameworks through ONNX as an interchange format, rather than limiting optimization to a single native model type.
D. Only custom NVIDIA formats