Getting worse results when not using `--cuda_ray`?

[yongsiang-fb]

Hi,

I found that when testing nerf_synthetic_lego dataset, if I remove --cuda_ray, the result seems to become worse. What's the difference between the pytorch ray marching & cuda ray marching? Shouldn't they only differ in speed? Thanks!

[kiui]

@yongsiang-fb Hi, they are quite different in implementation (e.g., the occupancy grid). You may increase num_steps and upsample_steps for non cuda ray mode for better quality.