diff --git a/tests/torchtune/models/clip/test_pos_embedding_interpolation.py b/tests/torchtune/models/clip/test_pos_embedding_interpolation.py
index 559403d9f5..c46a4a28a9 100644
--- a/tests/torchtune/models/clip/test_pos_embedding_interpolation.py
+++ b/tests/torchtune/models/clip/test_pos_embedding_interpolation.py
@@ -201,6 +201,7 @@ def test_tile_resize_position_embedding(self, params):
             embedding, tgt_max_num_tiles
         )
 
+        assert resized_pos_embed.is_contiguous()
         assert_expected(resized_pos_embed, expected_output, atol=1e-3, rtol=1e-4)
 
     @pytest.mark.parametrize("params", local_pos_emb_test_cases)
@@ -215,6 +216,7 @@ def test_resize_local_position_embedding(self, params):
             )
         )
 
+        assert resized_pos_embed.is_contiguous()
         assert_expected(resized_pos_embed, expected_output, atol=1e-3, rtol=1e-4)
 
     @pytest.mark.parametrize("params", global_pos_emb_test_cases)
@@ -230,6 +232,7 @@ def test_resize_global_position_embedding(self, params):
             )
         )
 
+        assert resized_pos_embed.is_contiguous()
         assert_expected(resized_pos_embed, expected_output, atol=1e-3, rtol=1e-4)
 
     @pytest.mark.parametrize(
diff --git a/torchtune/models/clip/_position_embeddings.py b/torchtune/models/clip/_position_embeddings.py
index cd1ea5947c..09a98862e1 100644
--- a/torchtune/models/clip/_position_embeddings.py
+++ b/torchtune/models/clip/_position_embeddings.py
@@ -319,7 +319,7 @@ def _resize_local_position_embedding(
         # add cls token back in
         local_pos_embed = torch.cat([cls_token, local_pos_embed], dim=0)
 
-        return local_pos_embed
+        return local_pos_embed.contiguous()
 
     # TODO: Switch to public method after 2.5 is stable
     @staticmethod
@@ -436,7 +436,7 @@ def _resize_global_position_embedding(
         # add cls token back in
         global_pos_embed = torch.cat([cls_embed, pos_embed], dim=2)
 
-        return global_pos_embed
+        return global_pos_embed.contiguous()
 
     def forward(self, x: torch.Tensor, aspect_ratio: torch.Tensor) -> torch.Tensor:
         """
@@ -633,7 +633,7 @@ def _resize_position_embedding(
         )
         # permute to the original shape
         embedding = embedding.permute(2, 3, 0, 1)
-        return embedding
+        return embedding.contiguous()
 
     def forward(self, x: torch.Tensor, aspect_ratio: torch.Tensor) -> torch.Tensor:
         """