RandLA-Net icon indicating copy to clipboard operation
RandLA-Net copied to clipboard

Published 20 hours ago • verified publisher icon QingyongHu

NaN values

Open ayoubOUSSM2021 opened this issue 3 years ago • 2 comments

Please can you help me solve this problem, I am trying to apply the model on my data which contains 6 classes, but either I change the number of classes to 6 or I keep 8 classes because I prepared my data as Semantic3D I still receive NaN values.

Captureerror

ayoubOUSSM2021 avatar Oct 12 '21 12:10 ayoubOUSSM2021

can you please tell your running environment? OS, cuda, cudnn and tensorflow version?

youssef962 avatar Oct 15 '21 10:10 youssef962

I met the same problem and I find the classes of validation data's labels are not equal to the train，maybe missing a few。when I modified it , this problem was solved.

liujingangel avatar Jun 09 '22 03:06 liujingangel