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
I don't want to fix this problem by just deleting compute_chunk_start(), because it causes strange packed tensors and ray_indices and t_starts (t_ends) have mismatched shapes.
I think the best way is delete redundant 0s in intervals and samples in grid.cu, but I don't know an efficient way.
The text was updated successfully, but these errors were encountered:
#197 implements over-allocation mode, but it seems that
samples.packed_info
mismatches tosamples.ray_indices
on https://github.com/KAIR-BAIR/nerfacc/blob/10315043bb6abd5a132deee39c2807afb684e13b/nerfacc/grid.py#L190samples.ray_indices
contains so many 0s, because its chunks are aligned bytraverse_step_limit
here. This is reasonable to parallelizetraverse_grids_kernel
, but re-calculatedchunk_starts
at https://github.com/KAIR-BAIR/nerfacc/blob/10315043bb6abd5a132deee39c2807afb684e13b/nerfacc/cuda/csrc/grid.cu#L402-L404 ignores this redundancy.I don't want to fix this problem by just deleting
compute_chunk_start()
, because it causes strange packed tensors andray_indices
andt_starts (t_ends)
have mismatched shapes.I think the best way is delete redundant 0s in
intervals
andsamples
ingrid.cu
, but I don't know an efficient way.The text was updated successfully, but these errors were encountered: