I am thinking reinforcement on top of Blender would be straight forward with unlimited synthetic data potential. I’ve come across people incorporating SD into r

I am thinking reinforcement on top of Blender would be straight forward with unlimited synthetic data potential. I’ve come across people incorporating SD into rendering workflow so tools are all there.