luost26/score-denoise

PCNet result

rsy6318 opened this issue · 1 comments

Hi, I have run your pretrained model to test the PCNet data, but I cannot get the same result as paper. Did I use wrong hyperparameters?

The scale and point sampling method affect the absolute value of results. Please contact Haolan for the original PCNet dataset we used.

His email can be found in this paper: https://arxiv.org/abs/2111.02045