build(deps): bump monai from 1.3.2 to 1.4.0 in /requirements (#2795)

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Nicki Skafte <[email protected]>

requirements/segmentation_test.txt

@@ -2,4 +2,4 @@
 #  in case you want to preserve/enforce restrictions on the latest compatible version, add "strict" as an in-line comment
 
 scipy >1.0.0, <1.15.0
-monai ==1.3.2
+monai ==1.4.0

tests/unittests/_helpers/testers.py

@@ -37,7 +37,7 @@ def _assert_allclose(tm_result: Any, ref_result: Any, atol: float = 1e-8, key: O
             ref_result.detach().cpu().numpy() if isinstance(ref_result, Tensor) else ref_result,
             atol=atol,
             equal_nan=True,
-        )
+        ), f"tm_result: {tm_result}, ref_result: {ref_result}"
     # multi output compare
     elif isinstance(tm_result, Sequence):
         for pl_res, ref_res in zip(tm_result, ref_result):
@@ -50,7 +50,7 @@ def _assert_allclose(tm_result: Any, ref_result: Any, atol: float = 1e-8, key: O
             ref_result.detach().cpu().numpy() if isinstance(ref_result, Tensor) else ref_result,
             atol=atol,
             equal_nan=True,
-        )
+        ), f"tm_result: {tm_result}, ref_result: {ref_result}"
     else:
         raise ValueError("Unknown format for comparison")
 

tests/unittests/segmentation/test_generalized_dice_score.py

@@ -51,10 +51,10 @@ def _reference_generalized_dice(
     if input_format == "index":
         preds = torch.nn.functional.one_hot(preds, num_classes=NUM_CLASSES).movedim(-1, 1)
         target = torch.nn.functional.one_hot(target, num_classes=NUM_CLASSES).movedim(-1, 1)
-    val = compute_generalized_dice(preds, target, include_background=include_background)
+    val = compute_generalized_dice(preds, target, include_background=include_background, sum_over_classes=True)
     if reduce:
         val = val.mean()
-    return val
+    return val.squeeze()
 
 
 @pytest.mark.parametrize(