Now that I've regressed the SMPL parameters of the sequence in my model, can I just feed them into SmoothNet for smoothing?