[Feature Request] using SetSeed on C++ inference

Describe the feature request

Will a future update be able to fix the random number seed during inference as well?

Describe scenario use case

Hi, I am making a C++ program for inference with a neural network model.
The neural network model includes a module that generates random numbers (RandomNormalLike).
Therefore, the output values are slightly different even for the same input.

I want to implement a function that turns this randomness on and off.
Example: We would like to store and use the seed of an excellent result.
Looking at the API reference, it seems that the SetSeed() function can be used to fix the random number seed during training (onnxruntime_training_c_api.h), but it is not available during inference (onnxruntime_cxx_api.h).
https://onnxruntime.ai/docs/api/c/namespace_ort.html

Compare the existing SetSeed() API in onnxruntime_training_c_api.h with the inference interface in onnxruntime_cxx_api.h, using the linked API reference as context. Determine the intended way to control RandomNormalLike randomness during C++ inference, including how enabling and disabling fixed seeds should behave. Done means the inference API supports the requested seed control with documented behavior.