[fbsync] proper comparison of input and outpus size before resizing (#7519)

Reviewed By: vmoens

Differential Revision: D45523937

fbshipit-source-id: e01d76b0eded5dc71206a523eab95ab06674efea

Co-authored-by: Liron <[email protected]>
Co-authored-by: Philip Meier <[email protected]>

Author: 3 people

test/test_transforms.py

@@ -463,6 +463,16 @@ def test_resize_size_equals_small_edge_size(height, width):
     assert max(result.size) == max_size
 
 
+def test_resize_equal_input_output_sizes():
+    # Regression test for https://github.com/pytorch/vision/issues/7518
+    height, width = 28, 27
+    img = Image.new("RGB", size=(width, height))
+
+    t = transforms.Resize((height, width), antialias=True)
+    result = t(img)
+    assert result is img
+
+
 class TestPad:
     @pytest.mark.parametrize("fill", [85, 85.0])
     def test_pad(self, fill):

torchvision/transforms/functional.py

@@ -478,7 +478,7 @@ def resize(
         size = [size]
     output_size = _compute_resized_output_size((image_height, image_width), size, max_size)
 
-    if (image_height, image_width) == output_size:
+    if [image_height, image_width] == output_size:
         return img
 
     antialias = _check_antialias(img, antialias, interpolation)