diff --git a/pytorch3d/transforms/transform3d.py b/pytorch3d/transforms/transform3d.py
index 03ae253a..8cbb3e9d 100644
--- a/pytorch3d/transforms/transform3d.py
+++ b/pytorch3d/transforms/transform3d.py
@@ -421,8 +421,9 @@ class Transform3d:
         if self.device != device:
             other.device = device
             other._matrix = self._matrix.to(device=device, dtype=dtype)
-            for t in other._transforms:
-                t.to(device, copy=copy, dtype=dtype)
+            other._transforms = [
+                t.to(device, copy=copy, dtype=dtype) for t in other._transforms
+            ]
         return other
 
     def cpu(self):
diff --git a/tests/test_transforms.py b/tests/test_transforms.py
index 62404d7a..6c2642d8 100644
--- a/tests/test_transforms.py
+++ b/tests/test_transforms.py
@@ -20,13 +20,17 @@ class TestTransform(TestCaseMixin, unittest.TestCase):
     def test_to(self):
         tr = Translate(torch.FloatTensor([[1.0, 2.0, 3.0]]))
         R = torch.FloatTensor([[0.0, 1.0, 0.0], [0.0, 0.0, 1.0], [1.0, 0.0, 0.0]])
+        cpu_points = torch.rand(9, 3)
+        cuda_points = cpu_points.cuda()
         R = Rotate(R)
         t = Transform3d().compose(R, tr)
         for _ in range(3):
-            t.cpu()
-            t.cuda()
-            t.cuda()
-            t.cpu()
+            t = t.cpu()
+            t.transform_points(cpu_points)
+            t = t.cuda()
+            t.transform_points(cuda_points)
+            t = t.cuda()
+            t = t.cpu()
 
     def test_clone(self):
         """