making sorting for K >1 optional in KNN points function

Summary: Added `sorted` argument to the `knn_points` function. This came up during the benchmarking against Faiss - sorting added extra memory usage. Match the memory usage of Faiss by making sorting optional.

Reviewed By: bottler, gkioxari

Differential Revision: D22329070

fbshipit-source-id: 0828ff9b48eefce99ce1f60089389f6885d03139

pytorch3d/ops/knn.py

@@ -18,7 +18,9 @@ class _knn_points(Function):
     """
 
     @staticmethod
-    def forward(ctx, p1, p2, lengths1, lengths2, K, version):
+    def forward(
+        ctx, p1, p2, lengths1, lengths2, K, version, return_sorted: bool = True
+    ):
         """
         K-Nearest neighbors on point clouds.
 
@@ -36,6 +38,8 @@ def forward(ctx, p1, p2, lengths1, lengths2, K, version):
             K: Integer giving the number of nearest neighbors to return.
             version: Which KNN implementation to use in the backend. If version=-1,
                 the correct implementation is selected based on the shapes of the inputs.
+            return_sorted: (bool) whether to return the nearest neighbors sorted in
+                ascending order of distance.
 
         Returns:
             p1_dists: Tensor of shape (N, P1, K) giving the squared distances to
@@ -52,7 +56,7 @@ def forward(ctx, p1, p2, lengths1, lengths2, K, version):
         idx, dists = _C.knn_points_idx(p1, p2, lengths1, lengths2, K, version)
 
         # sort KNN in ascending order if K > 1
-        if K > 1:
+        if K > 1 and return_sorted:
             if lengths2.min() < K:
                 P1 = p1.shape[1]
                 mask = lengths2[:, None] <= torch.arange(K, device=dists.device)[None]
@@ -84,7 +88,7 @@ def backward(ctx, grad_dists, grad_idx):
         grad_p1, grad_p2 = _C.knn_points_backward(
             p1, p2, lengths1, lengths2, idx, grad_dists
         )
-        return grad_p1, grad_p2, None, None, None, None
+        return grad_p1, grad_p2, None, None, None, None, None
 
 
 def knn_points(
@@ -95,6 +99,7 @@ def knn_points(
     K: int = 1,
     version: int = -1,
     return_nn: bool = False,
+    return_sorted: bool = True,
 ):
     """
     K-Nearest neighbors on point clouds.
@@ -113,7 +118,9 @@ def knn_points(
         K: Integer giving the number of nearest neighbors to return.
         version: Which KNN implementation to use in the backend. If version=-1,
             the correct implementation is selected based on the shapes of the inputs.
-        return_nn: If set to True returns the K nearest neighors in p2 for each point in p1.
+        return_nn: If set to True returns the K nearest neighbors in p2 for each point in p1.
+        return_sorted: (bool) whether to return the nearest neighbors sorted in
+            ascending order of distance.
 
     Returns:
         dists: Tensor of shape (N, P1, K) giving the squared distances to
@@ -158,7 +165,9 @@ def knn_points(
         lengths2 = torch.full((p1.shape[0],), P2, dtype=torch.int64, device=p1.device)
 
     # pyre-fixme[16]: `_knn_points` has no attribute `apply`.
-    p1_dists, p1_idx = _knn_points.apply(p1, p2, lengths1, lengths2, K, version)
+    p1_dists, p1_idx = _knn_points.apply(
+        p1, p2, lengths1, lengths2, K, version, return_sorted
+    )
 
     p2_nn = None
     if return_nn:

tests/test_knn.py

@@ -49,7 +49,7 @@ def _knn_points_naive(p1, p2, lengths1, lengths2, K: int) -> torch.Tensor:
 
         return _KNN(dists=dists, idx=idx, knn=None)
 
-    def _knn_vs_python_square_helper(self, device):
+    def _knn_vs_python_square_helper(self, device, return_sorted):
         Ns = [1, 4]
         Ds = [3, 5, 8]
         P1s = [8, 24]
@@ -70,7 +70,24 @@ def _knn_vs_python_square_helper(self, device):
 
                 # forward
                 out1 = self._knn_points_naive(x, y, lengths1=None, lengths2=None, K=K)
-                out2 = knn_points(x_cuda, y_cuda, K=K, version=version)
+                out2 = knn_points(
+                    x_cuda, y_cuda, K=K, version=version, return_sorted=return_sorted
+                )
+                if K > 1 and not return_sorted:
+                    # check out2 is not sorted
+                    self.assertFalse(torch.allclose(out1[0], out2[0]))
+                    self.assertFalse(torch.allclose(out1[1], out2[1]))
+                    # now sort out2
+                    dists, idx, _ = out2
+                    if P2 < K:
+                        dists[..., P2:] = float("inf")
+                        dists, sort_idx = dists.sort(dim=2)
+                        dists[..., P2:] = 0
+                    else:
+                        dists, sort_idx = dists.sort(dim=2)
+                    idx = idx.gather(2, sort_idx)
+                    out2 = _KNN(dists, idx, None)
+
                 self.assertClose(out1[0], out2[0])
                 self.assertTrue(torch.all(out1[1] == out2[1]))
 
@@ -86,11 +103,13 @@ def _knn_vs_python_square_helper(self, device):
 
     def test_knn_vs_python_square_cpu(self):
         device = torch.device("cpu")
-        self._knn_vs_python_square_helper(device)
+        self._knn_vs_python_square_helper(device, return_sorted=True)
 
     def test_knn_vs_python_square_cuda(self):
         device = get_random_cuda_device()
-        self._knn_vs_python_square_helper(device)
+        # Check both cases where the output is sorted and unsorted
+        self._knn_vs_python_square_helper(device, return_sorted=True)
+        self._knn_vs_python_square_helper(device, return_sorted=False)
 
     def _knn_vs_python_ragged_helper(self, device):
         Ns = [1, 4]