fix atol in resize_image kernel test (#8133)

pytorch · Nov 23, 2023 · 1fda0e8 · 1fda0e8
1 parent ec7f80b
commit 1fda0e8
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/test/test_transforms_v2.py b/test/test_transforms_v2.py
@@ -599,7 +599,7 @@ def test_kernel_image(self, size, interpolation, use_max_size, antialias, dtype,
 
         # In contrast to CPU, there is no native `InterpolationMode.BICUBIC` implementation for uint8 images on CUDA.
         # Internally, it uses the float path. Thus, we need to test with an enormous tolerance here to account for that.
-        atol = 30 if transforms.InterpolationMode.BICUBIC and dtype is torch.uint8 else 1
+        atol = 30 if (interpolation is transforms.InterpolationMode.BICUBIC and dtype is torch.uint8) else 1
         check_cuda_vs_cpu_tolerances = dict(rtol=0, atol=atol / 255 if dtype.is_floating_point else atol)
 
         check_kernel(