ONNX and Model Interoperability Quiz

Answers and explanations appear as you take the quiz.

1. What core problem does ONNX primarily solve?
   • Moving models between different frameworks and runtimes via a common format
   • Labeling training data automatically
   • Generating new neural network architectures
   • Encrypting model weights for security
2. How does an ONNX file represent a model?
   • As raw training data
   • As a computation graph of standardized operators
   • As a single floating-point number
   • As source code in Python
3. What does the acronym ONNX stand for?
   • Optimized Network Node Execution
   • Open Neural Network Extension
   • Open Neural Network Exchange
   • Operator Network Normalization Exchange
4. In ONNX Runtime, what is an 'execution provider'?
   • A cloud billing account
   • The person who trained the model
   • A type of loss function
   • A backend that runs supported operators on specific hardware like CPU, CUDA, or TensorRT
5. What does the 'opset' (operator set) version in an ONNX model define?
   • Which version of the standardized operators the graph uses
   • The model's accuracy on a test set
   • The number of GPUs required
   • The training learning rate
6. Which graph-level operations can ONNX Runtime apply?
   • Automatic retraining and label correction
   • Constant folding and node fusion, among other optimizations
   • Changing a model’s architecture to a new framework
   • Generating new training examples from inference requests
7. What can happen when an ONNX Runtime execution provider supports only part of a model graph?
   • The provider must execute every node or fail by definition
   • The model is automatically retrained
   • The runtime can assign supported nodes or subgraphs to that provider and others to a different provider
   • All unsupported operators are deleted
8. Why should an ONNX deployment check the model’s opset against the target runtime?
   • The opset specifies the model’s training dataset size
   • An opset number guarantees identical performance
   • The opset controls the number of GPU devices
   • The graph uses versioned operator definitions that the runtime must support