Skip to content

Fix metric tests #223

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

Open
sananakther6642 wants to merge 4 commits into
base: main
Choose a base branch
from
Open

Fix metric tests #223

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
7288216
Fix flaky multiclass accuracy test when class has no samples
Dec 14, 2025
7700955
Skip VGGish FAD tests when torchaudio prototype is unavailable
Dec 14, 2025
44a8af6
Fix BinaryAUROC tests dtype mismatch
Dec 14, 2025
c38003d
Fix flaky functional multiclass accuracy test when class has no samples
Dec 14, 2025
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
4 changes: 4 additions & 0 deletions tests/metrics/audio/test_fad.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,10 @@
import numpy as np
import torch
from torcheval.metrics import FrechetAudioDistance
import pytest

pytest.importorskip("torchaudio.prototype.pipelines")



# pyre-fixme[24]: Generic type `np.ndarray` expects 2 type parameters.
Expand Down
13 changes: 9 additions & 4 deletions tests/metrics/classification/test_accuracy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -118,10 +118,15 @@ def test_accuracy_class_average(self) -> None:
target_flattened = target.flatten()
accuracy_per_class = np.empty(num_classes)
for i in range(num_classes):
accuracy_per_class[i] = accuracy_score(
target_flattened[target_flattened == i].numpy(),
input_flattened[target_flattened == i].numpy(),
)
mask = target_flattened == i
if mask.sum().item() == 0:
accuracy_per_class[i] = np.nan
else:
accuracy_per_class[i] = accuracy_score(
target_flattened[mask].numpy(),
input_flattened[mask].numpy(),
)


self.run_class_implementation_tests(
metric=MulticlassAccuracy(num_classes=num_classes, average="macro"),
Expand Down
18 changes: 11 additions & 7 deletions tests/metrics/classification/test_auroc.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,15 +37,16 @@ def _test_auroc_class_with_input(
weight_tensors = weight.reshape(-1, 1) if weight is not None else None

if compute_result is None:
compute_result = (
torch.tensor(
roc_auc_score(
target_tensors, input_tensors, sample_weight=weight_tensors
)
)
score = (
roc_auc_score(target_tensors, input_tensors, sample_weight=weight_tensors)
if weight_tensors is not None
else torch.tensor(roc_auc_score(target_tensors, input_tensors))
else roc_auc_score(target_tensors, input_tensors)
)
compute_result = torch.tensor(score, dtype=torch.float64)




if weight is not None:
self.run_class_implementation_tests(
metric=BinaryAUROC(num_tasks=num_tasks, use_fbgemm=use_fbgemm),
Expand Down Expand Up @@ -147,7 +148,10 @@ def test_auroc_class_update_input_shape_different(self) -> None:
torch.cat(update_target, dim=0),
torch.cat(update_input, dim=0),
sample_weight=torch.cat(update_weight, dim=0),

),
dtype=torch.float64,

)

self.run_class_implementation_tests(
Expand Down
13 changes: 9 additions & 4 deletions tests/metrics/functional/classification/test_accuracy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -117,10 +117,15 @@ def test_accuracy_class_average(self) -> None:
target_flattened = target.flatten()
accuracy_per_class = np.empty(num_classes)
for i in range(num_classes):
accuracy_per_class[i] = accuracy_score(
target_flattened[target_flattened == i].numpy(),
input_flattened[target_flattened == i].numpy(),
)
mask = target_flattened == i
if mask.sum().item() == 0:
accuracy_per_class[i] = np.nan
else:
accuracy_per_class[i] = accuracy_score(
target_flattened[mask].numpy(),
input_flattened[mask].numpy(),
)


torch.testing.assert_close(
multiclass_accuracy(
Expand Down