You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello luis-gonzales,
first of all thank you very much for your code! This is really some good inspiration.
I was wondering if it is possible to implement variable sized point clouds? I saw the points you used have all a fixed length of 1024. Is there any way to make it work on the original data set (point cloud sizes varies from just a few hundred to 150k pts)? I guess batching will be troublesome, maybe it could work with padding or as ragged tensor?
The text was updated successfully, but these errors were encountered:
@cmb87 I think this could be done, just adjust the hard-coded value and if you point clouds are variable in size, just find the upper-bound (150K points) and pad all of the point clouds with less than 150K points via point duplication (e.g., over sampling).
Hello luis-gonzales,
first of all thank you very much for your code! This is really some good inspiration.
I was wondering if it is possible to implement variable sized point clouds? I saw the points you used have all a fixed length of 1024. Is there any way to make it work on the original data set (point cloud sizes varies from just a few hundred to 150k pts)? I guess batching will be troublesome, maybe it could work with padding or as ragged tensor?
The text was updated successfully, but these errors were encountered: