Fix CQS signal modernize-use-using in fbcode/vision/fair

Reviewed By: bottler

Differential Revision: D94879733

fbshipit-source-id: fc35eaaa723a2a035b3b204732add7ba8b225c57

pytorch3d/common/workaround/symeig3x3.py

@@ -82,10 +82,12 @@ class _SymEig3x3(nn.Module):
         q = inputs_trace / 3.0
 
         # Calculate squared sum of elements outside the main diagonal / 2
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        p1 = (
-        p1 = ((inputs**2).sum(dim=(-1, -2)) - (inputs_diag**2).sum(-1)) / 2
+            torch.square(inputs).sum(dim=(-1, -2)) - torch.square(inputs_diag).sum(-1)
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        ) / 2
-        p2 = ((inputs_diag - q[..., None]) ** 2).sum(dim=-1) + 2.0 * p1.clamp(self._eps)
+        p2 = torch.square(inputs_diag - q[..., None]).sum(dim=-1) + 2.0 * p1.clamp(
+            self._eps
+        )
 
         p = torch.sqrt(p2 / 6.0)
         B = (inputs - q[..., None, None] * self._identity) / p[..., None, None]
@@ -104,7 +106,9 @@ class _SymEig3x3(nn.Module):
         # Soft dispatch between the degenerate case (diagonal A) and general.
         # diag_soft_cond -> 1.0 when p1 < 6 * eps and diag_soft_cond -> 0.0 otherwise.
         # We use 6 * eps to take into account the error accumulated during the p1 summation
-        diag_soft_cond = torch.exp(-((p1 / (6 * self._eps)) ** 2)).detach()[..., None]
+        diag_soft_cond = torch.exp(-torch.square(p1 / (6 * self._eps))).detach()[
+            ..., None
+        ]
 
         # Eigenvalues are the ordered elements of main diagonal in the degenerate case
         diag_eigenvals, _ = torch.sort(inputs_diag, dim=-1)
@@ -199,8 +203,7 @@ class _SymEig3x3(nn.Module):
             cross_products[..., :1, :]
         )
 
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        norms_sq = torch.square(cross_products).sum(dim=-1)
-        norms_sq = (cross_products**2).sum(dim=-1)
         max_norms_index = norms_sq.argmax(dim=-1)
 
         # Pick only the cross-product with highest squared norm for each input

pytorch3d/csrc/pulsar/include/renderer.destruct.device.h

@@ -18,68 +18,89 @@ namespace Renderer {
 
 template <bool DEV>
 HOST void destruct(Renderer* self) {
-  if (self->result_d != NULL)
+  if (self->result_d != NULL) {
     FREE(self->result_d);
+  }
   self->result_d = NULL;
-  if (self->min_depth_d != NULL)
+  if (self->min_depth_d != NULL) {
     FREE(self->min_depth_d);
+  }
   self->min_depth_d = NULL;
-  if (self->min_depth_sorted_d != NULL)
+  if (self->min_depth_sorted_d != NULL) {
     FREE(self->min_depth_sorted_d);
+  }
   self->min_depth_sorted_d = NULL;
-  if (self->ii_d != NULL)
+  if (self->ii_d != NULL) {
     FREE(self->ii_d);
+  }
   self->ii_d = NULL;
-  if (self->ii_sorted_d != NULL)
+  if (self->ii_sorted_d != NULL) {
     FREE(self->ii_sorted_d);
+  }
   self->ii_sorted_d = NULL;
-  if (self->ids_d != NULL)
+  if (self->ids_d != NULL) {
     FREE(self->ids_d);
+  }
   self->ids_d = NULL;
-  if (self->ids_sorted_d != NULL)
+  if (self->ids_sorted_d != NULL) {
     FREE(self->ids_sorted_d);
+  }
   self->ids_sorted_d = NULL;
-  if (self->workspace_d != NULL)
+  if (self->workspace_d != NULL) {
     FREE(self->workspace_d);
+  }
   self->workspace_d = NULL;
-  if (self->di_d != NULL)
+  if (self->di_d != NULL) {
     FREE(self->di_d);
+  }
   self->di_d = NULL;
-  if (self->di_sorted_d != NULL)
+  if (self->di_sorted_d != NULL) {
     FREE(self->di_sorted_d);
+  }
   self->di_sorted_d = NULL;
-  if (self->region_flags_d != NULL)
+  if (self->region_flags_d != NULL) {
     FREE(self->region_flags_d);
+  }
   self->region_flags_d = NULL;
-  if (self->num_selected_d != NULL)
+  if (self->num_selected_d != NULL) {
     FREE(self->num_selected_d);
+  }
   self->num_selected_d = NULL;
-  if (self->forw_info_d != NULL)
+  if (self->forw_info_d != NULL) {
     FREE(self->forw_info_d);
+  }
   self->forw_info_d = NULL;
-  if (self->min_max_pixels_d != NULL)
+  if (self->min_max_pixels_d != NULL) {
     FREE(self->min_max_pixels_d);
+  }
   self->min_max_pixels_d = NULL;
-  if (self->grad_pos_d != NULL)
+  if (self->grad_pos_d != NULL) {
     FREE(self->grad_pos_d);
+  }
   self->grad_pos_d = NULL;
-  if (self->grad_col_d != NULL)
+  if (self->grad_col_d != NULL) {
     FREE(self->grad_col_d);
+  }
   self->grad_col_d = NULL;
-  if (self->grad_rad_d != NULL)
+  if (self->grad_rad_d != NULL) {
     FREE(self->grad_rad_d);
+  }
   self->grad_rad_d = NULL;
-  if (self->grad_cam_d != NULL)
+  if (self->grad_cam_d != NULL) {
     FREE(self->grad_cam_d);
+  }
   self->grad_cam_d = NULL;
-  if (self->grad_cam_buf_d != NULL)
+  if (self->grad_cam_buf_d != NULL) {
     FREE(self->grad_cam_buf_d);
+  }
   self->grad_cam_buf_d = NULL;
-  if (self->grad_opy_d != NULL)
+  if (self->grad_opy_d != NULL) {
     FREE(self->grad_opy_d);
+  }
   self->grad_opy_d = NULL;
-  if (self->n_grad_contributions_d != NULL)
+  if (self->n_grad_contributions_d != NULL) {
     FREE(self->n_grad_contributions_d);
+  }
   self->n_grad_contributions_d = NULL;
 }
 

pytorch3d/csrc/pulsar/include/renderer.norm_sphere_gradients.device.h

@@ -64,8 +64,9 @@ GLOBAL void norm_sphere_gradients(Renderer renderer, const int num_balls) {
   // The sphere only contributes to the camera gradients if it is
   // large enough in screen space.
   if (renderer.ids_sorted_d[idx] > 0 && ii.max.x >= ii.min.x + 3 &&
-      ii.max.y >= ii.min.y + 3)
+      ii.max.y >= ii.min.y + 3) {
     renderer.ids_sorted_d[idx] = 1;
+  }
   END_PARALLEL_NORET();
 };
 

pytorch3d/csrc/pulsar/include/renderer.render.device.h

@@ -139,8 +139,9 @@ GLOBAL void render(
       coord_y < cam_norm.film_border_top + cam_norm.film_height) {
     // Initialize the result.
     if (mode == 0u) {
-      for (uint c_id = 0; c_id < cam_norm.n_channels; ++c_id)
+      for (uint c_id = 0; c_id < cam_norm.n_channels; ++c_id) {
         result[c_id] = bg_col[c_id];
+      }
     } else {
       result[0] = 0.f;
     }
@@ -190,20 +191,22 @@ GLOBAL void render(
             "render|found intersection with sphere %u.\n",
             sphere_id_l[write_idx]);
       }
-      if (ii.min.x == MAX_USHORT)
+      if (ii.min.x == MAX_USHORT) {
         // This is an invalid sphere (out of image). These spheres have
         // maximum depth. Since we ordered the spheres by earliest possible
         // intersection depth we re certain that there will no other sphere
         // that is relevant after this one.
         loading_done = true;
+      }
     }
     // Reset n_pixels_done.
     n_pixels_done = 0;
     thread_block.sync(); // Make sure n_loaded is updated.
     if (n_loaded > RENDER_BUFFER_LOAD_THRESH) {
       // The load buffer is full enough. Draw.
-      if (thread_block.thread_rank() == 0)
+      if (thread_block.thread_rank() == 0) {
         n_balls_loaded += n_loaded;
+      }
       max_closest_possible_intersection = 0.f;
       // This excludes threads outside of the image boundary. Also, it reduces
       // block artifacts.
@@ -290,8 +293,9 @@ GLOBAL void render(
       uint warp_done = thread_warp.ballot(done);
       int warp_done_bit_cnt = POPC(warp_done);
 #endif //__CUDACC__ && __HIP_PLATFORM_AMD__
-      if (thread_warp.thread_rank() == 0)
+      if (thread_warp.thread_rank() == 0) {
         ATOMICADD_B(&n_pixels_done, warp_done_bit_cnt);
+      }
       // This sync is necessary to keep n_loaded until all threads are done with
       // painting.
       thread_block.sync();
@@ -299,8 +303,9 @@ GLOBAL void render(
     }
     thread_block.sync();
   }
-  if (thread_block.thread_rank() == 0)
+  if (thread_block.thread_rank() == 0) {
     n_balls_loaded += n_loaded;
+  }
   PULSAR_LOG_DEV_PIX(
       PULSAR_LOG_RENDER_PIX,
       "render|loaded %d balls in total.\n",
@@ -386,8 +391,9 @@ GLOBAL void render(
             static_cast<float>(tracker.get_n_hits());
   } else {
     float sm_d_normfac = FRCP(FMAX(sm_d, FEPS));
-    for (uint c_id = 0; c_id < cam_norm.n_channels; ++c_id)
+    for (uint c_id = 0; c_id < cam_norm.n_channels; ++c_id) {
       result[c_id] *= sm_d_normfac;
+    }
     int write_loc = (coord_y - cam_norm.film_border_top) * cam_norm.film_width *
             (3 + 2 * n_track) +
         (coord_x - cam_norm.film_border_left) * (3 + 2 * n_track);

pytorch3d/csrc/pulsar/pytorch/renderer.cpp

@@ -860,8 +860,9 @@ std::tuple<torch::Tensor, torch::Tensor> Renderer::forward(
             ? (cudaStream_t) nullptr
 #endif
             : (cudaStream_t) nullptr);
-    if (mode == 1)
+    if (mode == 1) {
       results[batch_i] = results[batch_i].slice(2, 0, 1, 1);
+    }
     forw_infos[batch_i] = from_blob(
         this->renderer_vec[batch_i].forw_info_d,
         {this->renderer_vec[0].cam.film_height,

pytorch3d/csrc/pulsar/pytorch/renderer.h

@@ -128,8 +128,9 @@ struct Renderer {
     stream << "pulsar::Renderer[";
     // Device info.
     stream << self.device_type;
-    if (self.device_index != -1)
+    if (self.device_index != -1) {
       stream << ", ID " << self.device_index;
+    }
     stream << "]";
     return stream;
   }

pytorch3d/csrc/rasterize_meshes/rasterize_meshes_cpu.cpp

@@ -106,6 +106,8 @@ auto ComputeFaceAreas(const torch::Tensor& face_verts) {
   return face_areas;
 }
 
+namespace {
+
 // Helper function to use with std::find_if to find the index of any
 // values in the top k struct which match a given idx.
 struct IsNeighbor {
@@ -118,7 +120,6 @@ struct IsNeighbor {
   int neighbor_idx;
 };
 
-namespace {
 void RasterizeMeshesNaiveCpu_worker(
     const int start_yi,
     const int end_yi,

pytorch3d/csrc/utils/vec2.h

@@ -19,7 +19,7 @@ template <
         std::is_same<T, double>::value || std::is_same<T, float>::value>>
 struct vec2 {
   T x, y;
-  typedef T scalar_t;
+  using scalar_t = T;
   vec2(T x, T y) : x(x), y(y) {}
 };
 

pytorch3d/csrc/utils/vec3.h

@@ -18,7 +18,7 @@ template <
         std::is_same<T, double>::value || std::is_same<T, float>::value>>
 struct vec3 {
   T x, y, z;
-  typedef T scalar_t;
+  using scalar_t = T;
   vec3(T x, T y, T z) : x(x), y(y), z(z) {}
 };
 

pytorch3d/implicitron/dataset/sql_dataset.py

@@ -483,9 +483,10 @@ class SqlIndexDataset(DatasetBase, ReplaceableBase):
             *self._get_pick_filters(),
             *self._get_exclude_filters(),
         ]
-        if self.pick_sequences_sql_clause:
+        if pick_sequences_sql_clause := self.pick_sequences_sql_clause:
             print("Applying the custom SQL clause.")
-            where_conditions.append(sa.text(self.pick_sequences_sql_clause))
+            # pyre-ignore[6]: TextClause is compatible with where conditions
+            where_conditions.append(sa.text(pick_sequences_sql_clause))
 
         def add_where(stmt):
             return stmt.where(*where_conditions) if where_conditions else stmt
@@ -505,6 +506,7 @@ class SqlIndexDataset(DatasetBase, ReplaceableBase):
 
             subquery = add_where(subquery).subquery()
             stmt = sa.select(subquery.c.sequence_name).where(
+                # pyre-ignore[6]: SQLAlchemy column comparison returns ColumnElement, not bool
                 subquery.c.row_number <= self.limit_sequences_per_category_to
             )
 
@@ -633,9 +635,10 @@ class SqlIndexDataset(DatasetBase, ReplaceableBase):
                     )
                 )
 
-        if self.pick_frames_sql_clause:
+        if pick_frames_sql_clause := self.pick_frames_sql_clause:
             logger.info("Applying the custom SQL clause.")
-            pick_frames_criteria.append(sa.text(self.pick_frames_sql_clause))
+            # pyre-ignore[6]: TextClause is compatible with where conditions
+            pick_frames_criteria.append(sa.text(pick_frames_sql_clause))
 
         if pick_frames_criteria:
             index = self._pick_frames_by_criteria(index, pick_frames_criteria)
@@ -698,9 +701,10 @@ class SqlIndexDataset(DatasetBase, ReplaceableBase):
                 )
             )
 
-        if self.pick_frames_sql_clause:
+        if pick_frames_sql_clause := self.pick_frames_sql_clause:
             logger.info("  applying custom SQL clause")
-            where_conditions.append(sa.text(self.pick_frames_sql_clause))
+            # pyre-ignore[6]: TextClause is compatible with where conditions
+            where_conditions.append(sa.text(pick_frames_sql_clause))
 
         if where_conditions:
             stmt = stmt.where(*where_conditions)

pytorch3d/loss/mesh_laplacian_smoothing.py

@@ -114,9 +114,7 @@ def mesh_laplacian_smoothing(meshes, method: str = "uniform"):
             if method == "cot":
                 norm_w = torch.sparse.sum(L, dim=1).to_dense().view(-1, 1)
                 idx = norm_w > 0
-                # pyre-fixme[58]: `/` is not supported for operand types `float` and
+                norm_w[idx] = torch.reciprocal(norm_w[idx])
-                #  `Tensor`.
-                norm_w[idx] = 1.0 / norm_w[idx]
             else:
                 L_sum = torch.sparse.sum(L, dim=1).to_dense().view(-1, 1)
                 norm_w = 0.25 * inv_areas

pytorch3d/loss/point_mesh_distance.py

@@ -6,6 +6,7 @@
 
 # pyre-unsafe
 
+import torch
 from pytorch3d import _C
 from pytorch3d.structures import Meshes, Pointclouds
 from torch.autograd import Function
@@ -302,8 +303,7 @@ def point_mesh_edge_distance(meshes: Meshes, pcls: Pointclouds):
     point_to_cloud_idx = pcls.packed_to_cloud_idx()  # (sum(P_i), )
     num_points_per_cloud = pcls.num_points_per_cloud()  # (N,)
     weights_p = num_points_per_cloud.gather(0, point_to_cloud_idx)
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    weights_p = torch.reciprocal(weights_p.float())
-    weights_p = 1.0 / weights_p.float()
     point_to_edge = point_to_edge * weights_p
     point_dist = point_to_edge.sum() / N
 
@@ -377,8 +377,7 @@ def point_mesh_face_distance(
     point_to_cloud_idx = pcls.packed_to_cloud_idx()  # (sum(P_i),)
     num_points_per_cloud = pcls.num_points_per_cloud()  # (N,)
     weights_p = num_points_per_cloud.gather(0, point_to_cloud_idx)
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    weights_p = torch.reciprocal(weights_p.float())
-    weights_p = 1.0 / weights_p.float()
     point_to_face = point_to_face * weights_p
     point_dist = point_to_face.sum() / N
 

pytorch3d/ops/cameras_alignment.py

@@ -223,8 +223,7 @@ def _align_camera_extrinsics(
         # of centered A and centered B
         Ac = A - Amu
         Bc = B - Bmu
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        align_t_s = (Ac * Bc).mean() / torch.square(Ac).mean().clamp(eps)
-        align_t_s = (Ac * Bc).mean() / (Ac**2).mean().clamp(eps)
     else:
         # set the scale to identity
         align_t_s = 1.0

pytorch3d/ops/laplacian_matrices.py

@@ -55,11 +55,9 @@ def laplacian(verts: torch.Tensor, edges: torch.Tensor) -> torch.Tensor:
     # We construct the Laplacian matrix by adding the non diagonal values
     # i.e. L[i, j] = 1 ./ deg(i) if (i, j) is an edge
     deg0 = deg[e0]
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    deg0 = torch.where(deg0 > 0.0, torch.reciprocal(deg0), deg0)
-    deg0 = torch.where(deg0 > 0.0, 1.0 / deg0, deg0)
     deg1 = deg[e1]
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    deg1 = torch.where(deg1 > 0.0, torch.reciprocal(deg1), deg1)
-    deg1 = torch.where(deg1 > 0.0, 1.0 / deg1, deg1)
     val = torch.cat([deg0, deg1])
     L = torch.sparse_coo_tensor(idx, val, (V, V), dtype=torch.float32)
 
@@ -137,8 +135,7 @@ def cot_laplacian(
     val = torch.stack([area] * 3, dim=1).view(-1)
     inv_areas.scatter_add_(0, idx, val)
     idx = inv_areas > 0
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    inv_areas[idx] = torch.reciprocal(inv_areas[idx])
-    inv_areas[idx] = 1.0 / inv_areas[idx]
     inv_areas = inv_areas.view(-1, 1)
 
     return L, inv_areas

pytorch3d/ops/points_alignment.py

@@ -182,8 +182,7 @@ def iterative_closest_point(
         t_history.append(SimilarityTransform(R, T, s))
 
         # compute the root mean squared error
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        Xt_sq_diff = torch.square(Xt - Xt_nn_points).sum(2)
-        Xt_sq_diff = ((Xt - Xt_nn_points) ** 2).sum(2)
         rmse = oputil.wmean(Xt_sq_diff[:, :, None], mask_X).sqrt()[:, 0, 0]
 
         # compute the relative rmse

pytorch3d/ops/sample_farthest_points.py

@@ -179,9 +179,7 @@ def sample_farthest_points_naive(
             # and all the other points. If a point has already been selected
             # it's distance will be 0.0 so it will not be selected again as the max.
             dist = points[n, selected_idx, :] - points[n, : lengths[n], :]
-            # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and
+            dist_to_last_selected = torch.square(dist).sum(-1)  # (P - i)
-            #  `int`.
-            dist_to_last_selected = (dist**2).sum(-1)  # (P - i)
 
             # If closer than currently saved distance to one of the selected
             # points, then updated closest_dists

pytorch3d/renderer/cameras.py

@@ -629,10 +629,8 @@ class FoVPerspectiveCameras(CamerasBase):
         # so the so the z sign is 1.0.
         z_sign = 1.0
 
-        # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+        K[:, 0, 0] = torch.div(2.0 * znear, max_x - min_x)
-        K[:, 0, 0] = 2.0 * znear / (max_x - min_x)
+        K[:, 1, 1] = torch.div(2.0 * znear, max_y - min_y)
-        # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
-        K[:, 1, 1] = 2.0 * znear / (max_y - min_y)
         K[:, 0, 2] = (max_x + min_x) / (max_x - min_x)
         K[:, 1, 2] = (max_y + min_y) / (max_y - min_y)
         K[:, 3, 2] = z_sign * ones
@@ -1178,9 +1176,7 @@ class PerspectiveCameras(CamerasBase):
         xy_inv_depth = torch.cat(
             # pyre-fixme[6]: For 1st argument expected `Union[List[Tensor],
             #  tuple[Tensor, ...]]` but got `Tuple[Tensor, float]`.
-            # pyre-fixme[58]: `/` is not supported for operand types `float` and
+            (xy_depth[..., :2], torch.reciprocal(xy_depth[..., 2:3])),
-            #  `Tensor`.
-            (xy_depth[..., :2], 1.0 / xy_depth[..., 2:3]),
             dim=-1,  # type: ignore
         )
         return unprojection_transform.transform_points(xy_inv_depth)

pytorch3d/renderer/mesh/clip.py

@@ -434,13 +434,7 @@ def clip_faces(
     # These will then be filled in for each case.
     ###########################################
     F_clipped = (
-        F
+        F + int(faces_delta_cum[-1].item()) + int(faces_delta[-1].item())
-        # pyre-fixme[58]: `+` is not supported for operand types `int` and
-        #  `Union[bool, float, int]`.
-        + faces_delta_cum[-1].item()
-        # pyre-fixme[58]: `+` is not supported for operand types `int` and
-        #  `Union[bool, float, int]`.
-        + faces_delta[-1].item()
     )  # Total number of faces in the new Meshes
     face_verts_clipped = torch.zeros(
         (F_clipped, 3, 3), dtype=face_verts_unclipped.dtype, device=device

pytorch3d/renderer/mesh/textures.py

@@ -71,9 +71,7 @@ def _list_to_padded_wrapper(
         # pyre-fixme[6]: For 2nd param expected `int` but got `Union[bool, float, int]`.
         x_reshaped.append(y.reshape(-1, D))
     x_padded = list_to_padded(x_reshaped, pad_size=pad_size, pad_value=pad_value)
-    # pyre-fixme[58]: `+` is not supported for operand types `Tuple[int, int]` and
+    return x_padded.reshape((N, -1) + tuple(reshape_dims))
-    #  `Size`.
-    return x_padded.reshape((N, -1) + reshape_dims)
 
 
 def _padded_to_list_wrapper(
@@ -104,9 +102,7 @@ def _padded_to_list_wrapper(
     # pyre-fixme[6]: For 3rd param expected `int` but got `Union[bool, float, int]`.
     x_reshaped = x.reshape(N, M, D)
     x_list = padded_to_list(x_reshaped, split_size=split_size)
-    # pyre-fixme[58]: `+` is not supported for operand types `Tuple[typing.Any]` and
+    x_list = [xl.reshape((xl.shape[0],) + tuple(reshape_dims)) for xl in x_list]
-    #  `Size`.
-    x_list = [xl.reshape((xl.shape[0],) + reshape_dims) for xl in x_list]
     return x_list
 
 

pytorch3d/renderer/splatter_blend.py

@@ -132,15 +132,13 @@ def _get_splat_kernel_normalization(
 
     epsilon = 0.05
     normalization_constant = torch.exp(
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        -torch.square(offsets).sum(dim=1) / (2 * sigma**2)
-        -(offsets**2).sum(dim=1) / (2 * sigma**2)
     ).sum()
 
     # We add an epsilon to the normalization constant to ensure the gradient will travel
     # through non-boundary pixels' normalization factor, see Sec. 3.3.1 in "Differentia-
     # ble Surface Rendering via Non-Differentiable Sampling", Cole et al.
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    return torch.div(1 + epsilon, normalization_constant)
-    return (1 + epsilon) / normalization_constant
 
 
 def _compute_occlusion_layers(
@@ -264,8 +262,9 @@ def _compute_splatting_colors_and_weights(
         torch.floor(pixel_coords_screen[..., :2]) - pixel_coords_screen[..., :2] + 0.5
     ).view((N, H, W, K, 1, 2))
 
-    # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+    dist2_p_q = torch.sum(
-    dist2_p_q = torch.sum((q_to_px_center + offsets) ** 2, dim=5)  # (N, H, W, K, 9)
+        torch.square(q_to_px_center + offsets), dim=5
+    )  # (N, H, W, K, 9)
     splat_weights = torch.exp(-dist2_p_q / (2 * sigma**2))
     alpha = colors[..., 3:4]
     splat_weights = (alpha * splat_kernel_normalization * splat_weights).unsqueeze(
@@ -417,12 +416,12 @@ def _normalize_and_compose_all_layers(
     device = splatted_colors_per_occlusion_layer.device
 
     # Normalize each of bg/surface/fg splat layers separately.
-    normalization_scales = 1.0 / (
+    normalization_scales = torch.div(
-        # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+        1.0,
         torch.maximum(
             splatted_weights_per_occlusion_layer,
             torch.tensor([1.0], device=device),
-        )
+        ),
     )  # (N, H, W, 1, 3)
 
     normalized_splatted_colors = (

pytorch3d/renderer/utils.py

@@ -269,9 +269,7 @@ class TensorProperties(nn.Module):
                         # to have the same shape as the input tensor.
                         new_dims = len(tensor_dims) - len(idx_dims)
                         new_shape = idx_dims + (1,) * new_dims
-                        # pyre-fixme[58]: `+` is not supported for operand types
+                        expand_dims = (-1,) + tuple(tensor_dims[1:])
-                        # `Tuple[int]` and `torch._C.Size`
-                        expand_dims = (-1,) + tensor_dims[1:]
                         _batch_idx = _batch_idx.view(*new_shape)
                         _batch_idx = _batch_idx.expand(*expand_dims)
 

pytorch3d/transforms/rotation_conversions.py

@@ -52,8 +52,7 @@ def quaternion_to_matrix(quaternions: torch.Tensor) -> torch.Tensor:
         Rotation matrices as tensor of shape (..., 3, 3).
     """
     r, i, j, k = torch.unbind(quaternions, -1)
-    # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+    two_s = torch.div(2.0, (quaternions * quaternions).sum(-1))
-    two_s = 2.0 / (quaternions * quaternions).sum(-1)
 
     o = torch.stack(
         (
@@ -137,18 +136,18 @@ def matrix_to_quaternion(matrix: torch.Tensor) -> torch.Tensor:
     # we produce the desired quaternion multiplied by each of r, i, j, k
     quat_by_rijk = torch.stack(
         [
-            # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and
+            torch.stack(
-            #  `int`.
+                [torch.square(q_abs[..., 0]), m21 - m12, m02 - m20, m10 - m01], dim=-1
-            torch.stack([q_abs[..., 0] ** 2, m21 - m12, m02 - m20, m10 - m01], dim=-1),
+            ),
-            # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and
+            torch.stack(
-            #  `int`.
+                [m21 - m12, torch.square(q_abs[..., 1]), m10 + m01, m02 + m20], dim=-1
-            torch.stack([m21 - m12, q_abs[..., 1] ** 2, m10 + m01, m02 + m20], dim=-1),
+            ),
-            # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and
+            torch.stack(
-            #  `int`.
+                [m02 - m20, m10 + m01, torch.square(q_abs[..., 2]), m12 + m21], dim=-1
-            torch.stack([m02 - m20, m10 + m01, q_abs[..., 2] ** 2, m12 + m21], dim=-1),
+            ),
-            # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and
+            torch.stack(
-            #  `int`.
+                [m10 - m01, m20 + m02, m21 + m12, torch.square(q_abs[..., 3])], dim=-1
-            torch.stack([m10 - m01, m20 + m02, m21 + m12, q_abs[..., 3] ** 2], dim=-1),
+            ),
         ],
         dim=-2,
     )

pytorch3d/transforms/se3.py

@@ -195,15 +195,15 @@ def _se3_V_matrix(
     V = (
         torch.eye(3, dtype=log_rotation.dtype, device=log_rotation.device)[None]
         + log_rotation_hat
-        # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+        * ((1 - torch.cos(rotation_angles)) / torch.square(rotation_angles))[
-        * ((1 - torch.cos(rotation_angles)) / (rotation_angles**2))[:, None, None]
+            :, None, None
+        ]
         + (
             log_rotation_hat_square
-            # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and
+            * (
-            #  `int`.
+                (rotation_angles - torch.sin(rotation_angles))
-            * ((rotation_angles - torch.sin(rotation_angles)) / (rotation_angles**3))[
+                / torch.pow(rotation_angles, 3)
-                :, None, None
+            )[:, None, None]
-            ]
         )
     )
 
@@ -215,8 +215,7 @@ def _get_se3_V_input(log_rotation: torch.Tensor, eps: float = 1e-4):
     A helper function that computes the input variables to the `_se3_V_matrix`
     function.
     """
-    # pyre-fixme[58]: `**` is not supported for operand types `Tensor` and `int`.
+    nrms = torch.square(log_rotation).sum(-1)
-    nrms = (log_rotation**2).sum(-1)
     rotation_angles = torch.clamp(nrms, eps).sqrt()
     log_rotation_hat = hat(log_rotation)
     log_rotation_hat_square = torch.bmm(log_rotation_hat, log_rotation_hat)

pytorch3d/transforms/transform3d.py

@@ -623,9 +623,7 @@ class Scale(Transform3d):
         Return the inverse of self._matrix.
         """
         xyz = torch.stack([self._matrix[:, i, i] for i in range(4)], dim=1)
-        # pyre-fixme[58]: `/` is not supported for operand types `float` and `Tensor`.
+        ixyz = torch.reciprocal(xyz)
-        ixyz = 1.0 / xyz
-        # pyre-fixme[6]: For 1st param expected `Tensor` but got `float`.
         imat = torch.diag_embed(ixyz, dim1=1, dim2=2)
         return imat
 

tests/test_transforms.py

@@ -679,8 +679,8 @@ class TestTranslate(TestCaseMixin, unittest.TestCase):
         im = t.inverse()._matrix
         im_2 = t._matrix.inverse()
         im_comp = t.get_matrix().inverse()
-        self.assertClose(im, im_comp)
+        self.assertClose(im, im_comp, atol=1e-4)
-        self.assertClose(im, im_2)
+        self.assertClose(im, im_2, atol=1e-4)
 
     def test_get_item(self, batch_size=5):
         device = torch.device("cuda:0")