GitLab

The early break may miss some points in voxel_num < max_voxels.
Use continue instead of break will be OK.

1. use dispatch instead of template
2. set use_hash default to False

1. add cuda hash support for cuda indice generation.
2. use hash table instead of dense table in CPU code.
3. add CPU-only build support.

2. fix SparseInverseConv problem
3. fix unittest problem