transform test fix

Summary: Indexing with a big matrix now fails with a ValueError, possibly because of pytorch improvements. Remove the testcase for it.

Reviewed By: davidsonic

Differential Revision: D42609741

fbshipit-source-id: 0a5a6632ed199cb942bfc4cc4ed347b72e491125

Author: bottler

tests/test_transforms.py

@@ -464,9 +464,6 @@ def test_get_item(self, batch_size=5):
         for invalid_index in (
             torch.tensor([1, 0, 1], dtype=torch.float32, device=device),  # float tensor
             1.2,  # float index
-            torch.tensor(
-                [[1, 0, 1], [1, 0, 1]], dtype=torch.int32, device=device
-            ),  # multidimensional tensor
         ):
             with self.assertRaises(IndexError):
                 t3d_selected = t3d[invalid_index]