Hi,
At the moment, After I produced the synthetic images and trained a YOLO model. I then rendered an animation contaning the target and ran the model using other inference engine outside Omniverse.
Is it possible to directly inference the YOLO model (in pytorch format) directly in Issac Sim or USD Composer? e.g. in a seperate viewport where when camera moves, it feeds the camera view to the model & draws bounding box in the viewport. That would speed up the model test significantly.
thank you.