Hi, I would like to know about how to generate inferences in real time with nvidia modulus or if there is any document or example where inference is made in real time.
Right now we largely leave how to deploy the trained model up to the user. Its a PyTorch checkpoint so there are many ways to perform inference both inside and outside of the Modulus framework. What exactly is considered real time largely depends on the problem, the model and other factors. For a deployment example of real time inference, check out the Modulus Omniverse connector.