Inference Speed
rolpotamias opened this issue · comments
Rolandos Potamias commented
Could you please prove the script that you used to evaluate the inference speed of the model.
I haven't managed to achieve anything more than 25FPS on A100, V100, or RTX2080. Have you modified any part of the model to achieve 70 FPS?
patienceFromZhou commented
did you apply reparameterization before inference?
stanginch commented