Skip to content

3672 Add more messages for AUC warning #3676

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Nic-Ma merged 10 commits into from
Jan 19, 2022
+49 −4
Merged

3672 Add more messages for AUC warning #3676

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
42a45e0
Merge pull request #19 from Project-MONAI/master
Nic-Ma Feb 1, 2021
cd16a13
Merge pull request #32 from Project-MONAI/master
Nic-Ma Feb 24, 2021
6f87afd
Merge pull request #180 from Project-MONAI/dev
Nic-Ma Jul 22, 2021
f398298
Merge pull request #214 from Project-MONAI/dev
Nic-Ma Sep 8, 2021
a0fc4da
Merge pull request #353 from Project-MONAI/dev
Nic-Ma Jan 18, 2022
4896db3
[DLMED] change to warning
Nic-Ma Jan 18, 2022
2255080
[DLMED] add more error message
Nic-Ma Jan 18, 2022
ab202e7
Merge branch 'dev' into 3672-change-auc-warning
Nic-Ma Jan 18, 2022
26df393
Merge branch 'dev' into 3672-change-auc-warning
Nic-Ma Jan 18, 2022
0b36580
[DLMED] simplify the messages
Nic-Ma Jan 19, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions monai/metrics/rocauc.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,9 +66,14 @@ def aggregate(self): # type: ignore
def _calculate(y_pred: torch.Tensor, y: torch.Tensor) -> float:
if not (y.ndimension() == y_pred.ndimension() == 1 and len(y) == len(y_pred)):
raise AssertionError("y and y_pred must be 1 dimension data with same length.")
if not y.unique().equal(torch.tensor([0, 1], dtype=y.dtype, device=y.device)):
warnings.warn("y values must be 0 or 1, can not be all 0 or all 1, skip AUC computation and return `Nan`.")
y_unique = y.unique()
if len(y_unique) == 1:
warnings.warn(f"y values can not be all {y_unique.item()}, skip AUC computation and return `Nan`.")
return float("nan")
if not y_unique.equal(torch.tensor([0, 1], dtype=y.dtype, device=y.device)):
warnings.warn(f"y values must be 0 or 1, but in {y_unique.tolist()}, skip AUC computation and return `Nan`.")
return float("nan")

n = len(y)
indices = y_pred.argsort()
y = y[indices].cpu().numpy()
Expand Down
44 changes: 42 additions & 2 deletions tests/test_compute_roc_auc.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,10 +77,39 @@
float("nan"),
]

TEST_CASE_9 = [
torch.tensor([[0.1, 0.9], [0.3, 1.4], [0.2, 0.1], [0.1, 0.5]]),
torch.tensor([[1], [1], [1], [1]]),
True,
2,
"macro",
float("nan"),
]

TEST_CASE_10 = [
torch.tensor([[0.1, 0.9], [0.3, 1.4], [0.2, 0.1], [0.1, 0.5]]),
torch.tensor([[0, 0], [1, 1], [2, 2], [3, 3]]),
True,
None,
"macro",
float("nan"),
]


class TestComputeROCAUC(unittest.TestCase):
@parameterized.expand(
[TEST_CASE_1, TEST_CASE_2, TEST_CASE_3, TEST_CASE_4, TEST_CASE_5, TEST_CASE_6, TEST_CASE_7, TEST_CASE_8]
[
TEST_CASE_1,
TEST_CASE_2,
TEST_CASE_3,
TEST_CASE_4,
TEST_CASE_5,
TEST_CASE_6,
TEST_CASE_7,
TEST_CASE_8,
TEST_CASE_9,
TEST_CASE_10,
]
)
def test_value(self, y_pred, y, softmax, to_onehot, average, expected_value):
y_pred_trans = Compose([ToTensor(), Activations(softmax=softmax)])
Expand All @@ -91,7 +120,18 @@ def test_value(self, y_pred, y, softmax, to_onehot, average, expected_value):
np.testing.assert_allclose(expected_value, result, rtol=1e-5)

@parameterized.expand(
[TEST_CASE_1, TEST_CASE_2, TEST_CASE_3, TEST_CASE_4, TEST_CASE_5, TEST_CASE_6, TEST_CASE_7, TEST_CASE_8]
[
TEST_CASE_1,
TEST_CASE_2,
TEST_CASE_3,
TEST_CASE_4,
TEST_CASE_5,
TEST_CASE_6,
TEST_CASE_7,
TEST_CASE_8,
TEST_CASE_9,
TEST_CASE_10,
]
)
def test_class_value(self, y_pred, y, softmax, to_onehot, average, expected_value):
y_pred_trans = Compose([ToTensor(), Activations(softmax=softmax)])
Expand Down