8274-mitigate-gpu-load-check (#8275)

Fixes #8274 .

### Description

I tried to use A100 with same container to test, but could not reproduce
the issue. Therefore, I think here we can do a bit change on the test,
and if there are still same issues, I will try to check more.

### Types of changes
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [ ] New tests added to cover the changes.
- [ ] Integration tests passed locally by running `./runtests.sh -f -u
--net --coverage`.
- [ ] Quick tests passed locally by running `./runtests.sh --quick
--unittests --disttests`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated, tested `make html` command in the `docs/`
folder.

Signed-off-by: Yiheng Wang <[email protected]>

Author: yiheng-wang-nv

tests/test_load_image.py

@@ -233,7 +233,7 @@ def test_nibabel_reader_gpu(self, input_param, filenames, expected_shape):
             input_param_cpu = input_param.copy()
             input_param_cpu["to_gpu"] = False
             result_cpu = LoadImage(image_only=True, **input_param_cpu)(filenames)
-            self.assertTrue(torch.equal(result_cpu, result.cpu()))
+            self.assertTrue(torch.allclose(result_cpu, result.cpu(), atol=1e-6))
 
     @parameterized.expand([TEST_CASE_6, TEST_CASE_7, TEST_CASE_8, TEST_CASE_8_1, TEST_CASE_9])
     def test_itk_reader(self, input_param, filenames, expected_shape):