replace view with reshape, check for nans

Summary: Replace view with reshape, add check for nans before mesh sampling

Reviewed By: nikhilaravi

Differential Revision: D20548456

fbshipit-source-id: c4e1b88e033ecb8f0f3a8f3a33a04ce13a5b5043

pytorch3d/ops/sample_points_from_meshes.py

@@ -41,6 +41,8 @@ def sample_points_from_meshes(
         raise ValueError("Meshes are empty.")
 
     verts = meshes.verts_packed()
+    if not torch.isfinite(verts).all():
+        raise ValueError("Meshes contain nan or inf.")
     faces = meshes.faces_packed()
     mesh_to_face = meshes.mesh_to_faces_packed_first_idx()
     num_meshes = len(meshes)

pytorch3d/renderer/mesh/texturing.py

@@ -53,7 +53,7 @@ def interpolate_texture_map(fragments, meshes) -> torch.Tensor:
     N, H_in, W_in, C = texture_maps.shape  # 3 for RGB
 
     # pixel_uvs: (N, H, W, K, 2) -> (N, K, H, W, 2) -> (NK, H, W, 2)
-    pixel_uvs = pixel_uvs.permute(0, 3, 1, 2, 4).view(N * K, H_out, W_out, 2)
+    pixel_uvs = pixel_uvs.permute(0, 3, 1, 2, 4).reshape(N * K, H_out, W_out, 2)
 
     # textures.map:
     #   (N, H, W, C) -> (N, C, H, W) -> (1, N, C, H, W)
@@ -81,7 +81,7 @@ def interpolate_texture_map(fragments, meshes) -> torch.Tensor:
     if texture_maps.device != pixel_uvs.device:
         texture_maps = texture_maps.to(pixel_uvs.device)
     texels = F.grid_sample(texture_maps, pixel_uvs, align_corners=False)
-    texels = texels.view(N, K, C, H_out, W_out).permute(0, 3, 4, 1, 2)
+    texels = texels.reshape(N, K, C, H_out, W_out).permute(0, 3, 4, 1, 2)
     return texels
 
 

pytorch3d/structures/utils.py

@@ -191,7 +191,7 @@ def padded_to_packed(
             "Only one of split_size or pad_value should be provided."
         )
 
-    x_packed = x.view(-1, D)  # flatten padded
+    x_packed = x.reshape(-1, D)  # flatten padded
 
     if pad_value is None and split_size is None:
         return x_packed

tests/test_sample_points_from_meshes.py

@@ -291,6 +291,26 @@ def test_multinomial_weights(self):
             if sampled_weights.min() <= 0:
                 return False
         return True
+
+    def test_verts_nan(self):
+        num_verts = 30
+        num_faces = 50
+        for device in ["cpu", "cuda:0"]:
+            for invalid in ["nan", "inf"]:
+                verts = torch.rand(
+                    (num_verts, 3), dtype=torch.float32, device=device
+                )
+                # randomly assign an invalid type 
+                verts[torch.randperm(num_verts)[:10]] = float(invalid)
+                faces = torch.randint(
+                    num_verts, size=(num_faces, 3), dtype=torch.int64, device=device
+                )
+                meshes = Meshes(verts=[verts], faces=[faces])
+
+                with self.assertRaisesRegex(ValueError, "Meshes contain nan or inf."):
+                    sample_points_from_meshes(
+                        meshes, num_samples=100, return_normals=True
+                    )
 
     @staticmethod
     def sample_points_with_init(