tink/nerfstudio
4
0
Fork 0
mirror of https://github.com/nerfstudio-project/nerfstudio.git synced 2026-11-11 18:44:21 +08:00
Code

Merge pull request #65 from ethanweber/tancik/nerf_alignment

Browse Source 
Better align with nerf
This commit is contained in:
Matthew Tancik 2022-05-10 10:27:14 -07:00 committed by GitHub
commit 6340f975db
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 7 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
configs/graph/default_graph.yaml
View File

@ -12,7 +12,7 @@ resume_train:
# NOTE(ethan): currently setting to nothing b/c we're still designing the config format...
network:
_target_: mattport.nerf.graph.vanilla_nerf.NeRFGraph
near_plane: 1.0
near_plane: 2.0
far_plane: 6.0
num_coarse_samples: 64
num_importance_samples: 128

5
mattport/nerf/graph/vanilla_nerf.py
View File

@ -48,7 +48,8 @@ class NeRFField(nn.Module):
self.mlp_rgb = MLP(
in_dim=self.mlp_base.get_out_dim() + self.encoding_dir.get_out_dim(),
out_dim=self.layer_width // 2,
num_layers=1,
num_layers=2,
layer_width=self.layer_width // 2,
activation=nn.ReLU(),
)
self.field_output_rgb = RGBFieldHead(in_dim=self.mlp_rgb.get_out_dim())
@ -84,7 +85,7 @@ class NeRFGraph(Graph):
self,
intrinsics=None,
camera_to_world=None,
near_plane=1.0,
near_plane=2.0,
far_plane=6.0,
num_coarse_samples=64,
num_importance_samples=128,

2
mattport/nerf/sampler.py
View File

@ -125,7 +125,7 @@ class PDFSampler(nn.Module):
u = torch.rand(size=(*cdf.shape[:-1], num_samples), device=cdf.device)
else:
u = torch.linspace(0.0, 1.0, steps=num_samples, device=cdf.device)
u = torch.expand(size=(*cdf.shape[:-1], num_samples))
u = u.expand(size=(*cdf.shape[:-1], num_samples))
u = u.contiguous()
indicies = torch.searchsorted(cdf, u, right=True)

2
mattport/nerf/trainer.py
View File

@ -240,8 +240,10 @@ class Trainer:
@profiler.time_function
def test_image(self, image_idx, step):
"""Test a specific image."""
self.graph.eval()
intrinsics = self.val_image_intrinsics[image_idx]
camera_to_world = self.val_image_camera_to_world[image_idx]
outputs = self.graph.get_outputs_for_camera(intrinsics, camera_to_world)
image = self.val_image_dataset[image_idx]["image"].to(self.device)
self.graph.log_test_image_outputs(image_idx, step, image, outputs)
self.graph.train()