Fix pyright errors

nerfstudio/models/gaussian_splatting.py

@@ -402,7 +402,7 @@ def refinement_after(self, optimizers: Optimizers, step):
 
                     # After a guassian is split into two new gaussians, the original one should be also pruned.
                     splits_mask = torch.cat(
-                        (splits, torch.zeros(nsamps * splits.sum() + dups.sum(), device="cuda", dtype=bool))
+                        (splits, torch.zeros(nsamps * splits.sum() + dups.sum(), device="cuda", dtype=torch.bool))
                     )
                     deleted_mask = self.cull_gaussians(splits_mask)
                     param_groups = self.get_gaussian_param_groups()
@@ -649,7 +649,7 @@ def get_outputs(self, camera: Cameras) -> Dict[str, Union[torch.Tensor, List]]:
             H,
             W,
             tile_bounds,
-        )
+        )  # type: ignore
         if (self.radii).sum() == 0:
             return {"rgb": background.repeat(int(camera.height.item()), int(camera.width.item()), 1)}
 
@@ -661,9 +661,8 @@ def get_outputs(self, camera: Cameras) -> Dict[str, Union[torch.Tensor, List]]:
             viewdirs = means_crop.detach() - camera.camera_to_worlds.detach()[..., :3, 3]  # (N, 3)
             viewdirs = viewdirs / viewdirs.norm(dim=-1, keepdim=True)
             n = min(self.step // self.config.sh_degree_interval, self.config.sh_degree)
-            n_coefs_use = num_sh_bases(n)
-            rgbs = SphericalHarmonics.apply(n, viewdirs, colors_crop[:, :n_coefs_use, :])
-            rgbs = torch.clamp(rgbs + 0.5, min=0.0)
+            rgbs = SphericalHarmonics.apply(n, viewdirs, colors_crop)
+            rgbs = torch.clamp(rgbs + 0.5, min=0.0)  # type: ignore
         else:
             rgbs = torch.sigmoid(colors_crop[:, 0, :])
 
@@ -685,8 +684,8 @@ def get_outputs(self, camera: Cameras) -> Dict[str, Union[torch.Tensor, List]]:
             H,
             W,
             background,
-        )
-        rgb = torch.clamp(rgb, max=1.0)
+        )  # type: ignore
+        rgb = torch.clamp(rgb, max=1.0)  # type: ignore
         depth_im = None
         if not self.training:
             depth_im = RasterizeGaussians.apply(  # type: ignore
@@ -700,9 +699,11 @@ def get_outputs(self, camera: Cameras) -> Dict[str, Union[torch.Tensor, List]]:
                 H,
                 W,
                 torch.ones(3, device=self.device) * 10,
-            )[..., 0:1]
+            )[
+                ..., 0:1
+            ]  # type: ignore
 
-        return {"rgb": rgb, "depth": depth_im}
+        return {"rgb": rgb, "depth": depth_im}  # type: ignore
 
     def get_metrics_dict(self, outputs, batch) -> Dict[str, torch.Tensor]:
         """Compute and returns metrics.

nerfstudio/scripts/exporter.py

@@ -39,6 +39,7 @@
 from nerfstudio.exporter import texture_utils, tsdf_utils
 from nerfstudio.exporter.exporter_utils import collect_camera_poses, generate_point_cloud, get_mesh_from_filename
 from nerfstudio.exporter.marching_cubes import generate_mesh_with_multires_marching_cubes
+from nerfstudio.fields.sdf_field import SDFField  # noqa
 from nerfstudio.models.gaussian_splatting import GaussianSplattingModel
 from nerfstudio.pipelines.base_pipeline import Pipeline, VanillaPipeline
 from nerfstudio.utils.eval_utils import eval_setup

nerfstudio/scripts/gsplat_render.py

@@ -39,8 +39,8 @@ class CameraInfo(NamedTuple):
     uid: int
     viewmat: torch.Tensor
     projmat: torch.Tensor
-    FovY: np.array
-    FovX: np.array
+    FovY: float
+    FovX: float
     image_name: str
     width: int
     height: int
@@ -149,15 +149,15 @@ def render(self, cam_info: CameraInfo, background: torch.Tensor):
             cam_info.height,
             cam_info.width,
             tile_bounds,
-        )
+        )  # type: ignore
         torch.cuda.synchronize()
 
         if self.n_sh_level > 0:
             c2w = torch.inverse(viewmat)
             viewdirs = self.xyz - c2w[None, :3, 3]  # (N, 3)
             viewdirs = viewdirs / viewdirs.norm(dim=-1, keepdim=True)
             rgbs = SphericalHarmonics.apply(self.n_sh_level, viewdirs, self.sh_coefs)
-            rgbs = torch.clamp(rgbs + 0.5, min=0.0)
+            rgbs = torch.clamp(rgbs + 0.5, min=0.0)  # type: ignore
         else:
             rgbs = self.colors
         rgb = RasterizeGaussians.apply(
@@ -171,10 +171,10 @@ def render(self, cam_info: CameraInfo, background: torch.Tensor):
             cam_info.height,
             cam_info.width,
             background,
-        )
+        )  # type: ignore
         torch.cuda.synchronize()
 
-        return torch.clamp(rgb, max=1.0)
+        return torch.clamp(rgb, max=1.0)  # type: ignore
 
 
 def projection_matrix(znear, zfar, fovx, fovy, device="cpu"):
@@ -203,7 +203,7 @@ def focal2fov(focal, pixels):
     return 2 * math.atan(pixels / (2 * focal))
 
 
-def load_trajectory(filename: Path) -> List[CameraInfo]:
+def load_trajectory(filename: Path):
     """Loads the cameras from a trajectory.json file like the ones used for rendering in nerfstudio
 
     Parameters
@@ -351,7 +351,7 @@ def main(
         assert len(bg_color) == 3
         bg = bg_color
 
-    render_set(render_path, point_cloud, cam_infos, bg)
+    render_set(render_path=render_path, point_cloud=point_cloud, cam_infos=cam_infos, bg_color=bg)
 
 
 def entrypoint():