add install note

src/torchmetrics/clustering/cluster_accuracy.py

@@ -50,6 +50,11 @@ class ClusterAccuracy(Metric):
 
     - ``acc_score`` (:class:`~torch.Tensor`): A tensor with the Cluster Accuracy score
 
+    .. note:: This metric requires you to have `torch_linear_assignment` package installed for solving the
+        linear sum assignment problem. You can install it directly from source with the following command:
+        ``pip install git+https://github.com/ivan-chai/torch-linear-assignment`` or install it with
+        ``pip install torchmetrics[clustering]``.
+
     Args:
         num_classes: number of classes
         kwargs: Additional keyword arguments, see :ref:`Metric kwargs` for more info.

src/torchmetrics/functional/clustering/cluster_accuracy.py

@@ -33,6 +33,11 @@ def _cluster_accuracy_compute(confmat: Tensor) -> Tensor:
 def cluster_accuracy(preds: Tensor, target: Tensor, num_classes: int) -> Tensor:
     """Computes the clustering accuracy between the predicted and target clusters.
 
+    .. note:: This metric requires you to have `torch_linear_assignment` package installed for solving the
+        linear sum assignment problem. You can install it directly from source with the following command:
+        ``pip install git+https://github.com/ivan-chai/torch-linear-assignment`` or install it with
+        ``pip install torchmetrics[clustering]``.
+
     Args:
         preds: predicted cluster labels
         target: ground truth cluster labels