Merge branch 'master' into bugfix/map_tm_to_coco

CHANGELOG.md

@@ -18,6 +18,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Added `input_format` argument to segmentation metrics ([#2572](https://github.com/Lightning-AI/torchmetrics/pull/2572))
 
 
+- Added better error messages for intersection detection metrics for wrong user input ([#2577](https://github.com/Lightning-AI/torchmetrics/pull/2577))
+
+
 ### Changed
 
 -

requirements/_tests.txt

@@ -16,5 +16,5 @@ fire <=0.6.0
 
 cloudpickle >1.3, <=3.0.0
 scikit-learn >=1.1.1, <1.3.0; python_version < "3.9"
-scikit-learn >=1.4.0, <1.5.0; python_version >= "3.9"
+scikit-learn >=1.4.0, <1.6.0; python_version >= "3.9"
 cachier ==3.0.0

src/torchmetrics/functional/detection/ciou.py

@@ -24,6 +24,11 @@
 def _ciou_update(
     preds: torch.Tensor, target: torch.Tensor, iou_threshold: Optional[float], replacement_val: float = 0
 ) -> torch.Tensor:
+    if preds.ndim != 2 or preds.shape[-1] != 4:
+        raise ValueError(f"Expected preds to be of shape (N, 4) but got {preds.shape}")
+    if target.ndim != 2 or target.shape[-1] != 4:
+        raise ValueError(f"Expected target to be of shape (N, 4) but got {target.shape}")
+
     from torchvision.ops import complete_box_iou
 
     iou = complete_box_iou(preds, target)

src/torchmetrics/functional/detection/diou.py

@@ -24,6 +24,11 @@
 def _diou_update(
     preds: torch.Tensor, target: torch.Tensor, iou_threshold: Optional[float], replacement_val: float = 0
 ) -> torch.Tensor:
+    if preds.ndim != 2 or preds.shape[-1] != 4:
+        raise ValueError(f"Expected preds to be of shape (N, 4) but got {preds.shape}")
+    if target.ndim != 2 or target.shape[-1] != 4:
+        raise ValueError(f"Expected target to be of shape (N, 4) but got {target.shape}")
+
     from torchvision.ops import distance_box_iou
 
     iou = distance_box_iou(preds, target)

src/torchmetrics/functional/detection/giou.py

@@ -24,6 +24,11 @@
 def _giou_update(
     preds: torch.Tensor, target: torch.Tensor, iou_threshold: Optional[float], replacement_val: float = 0
 ) -> torch.Tensor:
+    if preds.ndim != 2 or preds.shape[-1] != 4:
+        raise ValueError(f"Expected preds to be of shape (N, 4) but got {preds.shape}")
+    if target.ndim != 2 or target.shape[-1] != 4:
+        raise ValueError(f"Expected target to be of shape (N, 4) but got {target.shape}")
+
     from torchvision.ops import generalized_box_iou
 
     iou = generalized_box_iou(preds, target)

src/torchmetrics/functional/detection/iou.py

@@ -24,6 +24,12 @@
 def _iou_update(
     preds: torch.Tensor, target: torch.Tensor, iou_threshold: Optional[float], replacement_val: float = 0
 ) -> torch.Tensor:
+    """Compute the IoU matrix between two sets of boxes."""
+    if preds.ndim != 2 or preds.shape[-1] != 4:
+        raise ValueError(f"Expected preds to be of shape (N, 4) but got {preds.shape}")
+    if target.ndim != 2 or target.shape[-1] != 4:
+        raise ValueError(f"Expected target to be of shape (N, 4) but got {target.shape}")
+
     from torchvision.ops import box_iou
 
     iou = box_iou(preds, target)

src/torchmetrics/metric.py

@@ -107,7 +107,8 @@ def __init__(
         # see (https://github.com/pytorch/pytorch/blob/3e6bb5233f9ca2c5aa55d9cda22a7ee85439aa6e/
         # torch/nn/modules/module.py#L227)
         torch._C._log_api_usage_once(f"torchmetrics.metric.{self.__class__.__name__}")
-
+        # magic patch for `RuntimeError: DataLoader worker (pid(s) 104) exited unexpectedly`
+        self._TORCH_GREATER_EQUAL_2_1 = bool(_TORCH_GREATER_EQUAL_2_1)
         self._device = torch.device("cpu")
         self._dtype = torch.get_default_dtype()
 
@@ -441,7 +442,7 @@ def _sync_dist(self, dist_sync_fn: Callable = gather_all_tensors, process_group:
 
             # cornor case in distributed settings where a rank have not received any data, create empty to concatenate
             if (
-                _TORCH_GREATER_EQUAL_2_1
+                self._TORCH_GREATER_EQUAL_2_1
                 and reduction_fn == dim_zero_cat
                 and isinstance(input_dict[attr], list)
                 and len(input_dict[attr]) == 0

tests/unittests/detection/test_intersection.py

@@ -314,6 +314,20 @@ def test_error_on_wrong_input(self, class_metric, functional_metric, reference_m
                 [{"boxes": Tensor(), "labels": []}],
             )
 
+    def test_functional_error_on_wrong_input_shape(self, class_metric, functional_metric, reference_metric):
+        """Test functional input validation."""
+        with pytest.raises(ValueError, match="Expected preds to be of shape.*"):
+            functional_metric(torch.randn(25), torch.randn(25, 4))
+
+        with pytest.raises(ValueError, match="Expected target to be of shape.*"):
+            functional_metric(torch.randn(25, 4), torch.randn(25))
+
+        with pytest.raises(ValueError, match="Expected preds to be of shape.*"):
+            functional_metric(torch.randn(25, 25), torch.randn(25, 4))
+
+        with pytest.raises(ValueError, match="Expected target to be of shape.*"):
+            functional_metric(torch.randn(25, 4), torch.randn(25, 25))
+
 
 def test_corner_case():
     """See issue: https://github.com/Lightning-AI/torchmetrics/issues/1921."""