Relax tolerances for FFmpeg v4.1 and earlier (#1045)

Author: scotts

test/test_transform_ops.py

@@ -28,6 +28,8 @@
     assert_frames_equal,
     assert_tensor_close_on_at_least,
     AV1_VIDEO,
+    get_ffmpeg_major_version,
+    get_ffmpeg_minor_version,
     H265_VIDEO,
     NASA_VIDEO,
     needs_cuda,
@@ -254,7 +256,13 @@ def test_resize_ffmpeg(self):
 
             assert frame_resize.shape == expected_shape
             assert frame_ref.shape == expected_shape
-            assert_frames_equal(frame_resize, frame_ref)
+
+            if get_ffmpeg_major_version() <= 4 and get_ffmpeg_minor_version() <= 1:
+                # FFmpeg version 4.1 and before appear to have a different
+                # resize implementation.
+                torch.testing.assert_close(frame_resize, frame_ref, rtol=0, atol=2)
+            else:
+                assert_frames_equal(frame_resize, frame_ref)
 
     def test_resize_transform_fails(self):
         decoder = create_from_file(str(NASA_VIDEO.path))