The near and far parameter of render_image during evaluation is not consistent with training stage. Currently this is not a real problem since there is a default near and far value in ray_marching for unbound scenes.
Other modifications:
reuse the MIPNERF360_UNBOUNDED_SCENES and NERF_SYNTHETIC_SCENES.
save the model at the end of the training. Since it still takes some time for MLP to converge, it would be annoying for first-time user who forgot to generate results.
The
near
andfar
parameter ofrender_image
during evaluation is not consistent with training stage. Currently this is not a real problem since there is a default near and far value inray_marching
for unbound scenes.Other modifications:
MIPNERF360_UNBOUNDED_SCENES
andNERF_SYNTHETIC_SCENES
.