mirror of
https://github.com/facebookresearch/pytorch3d.git
synced 2025-12-20 22:30:35 +08:00
fix default settings for point rasterization and update benchmark
Summary: Fixes the default setting of `max_points_per_bin` in `rasterize_points.py`. For large batches with large size pointclouds this was a causing the rasterizer to be very slow. Expanded the pointcloud rendering benchmarks to include larger size pointclouds and fixed cuda synchronization issue in benchmark. Reviewed By: gkioxari Differential Revision: D22301185 fbshipit-source-id: 5077c1ba2c43d73efc1c659f0ec75959ceddf893
This commit is contained in:
committed by
Facebook GitHub Bot
parent
b636f2950d
commit
88f579389f
@@ -98,7 +98,7 @@ def rasterize_points(
|
||||
)
|
||||
|
||||
if max_points_per_bin is None:
|
||||
max_points_per_bin = int(max(10000, points_packed.shape[0] / 5))
|
||||
max_points_per_bin = int(max(10000, pointclouds._P / 5))
|
||||
|
||||
# Function.apply cannot take keyword args, so we handle defaults in this
|
||||
# wrapper and call apply with positional args only
|
||||
|
||||
Reference in New Issue
Block a user