Ensure LinaerKernel stores ard_num_dims property. (#2635)

[Fixes #2633]
cornellius-gp · Feb 7, 2025 · b017b9c · b017b9c
1 parent da70269
commit b017b9c
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 1 deletion.
diff --git a/gpytorch/kernels/linear_kernel.py b/gpytorch/kernels/linear_kernel.py
@@ -52,7 +52,7 @@ def __init__(
         variance_constraint: Optional[Interval] = None,
         **kwargs,
     ):
-        super().__init__(**kwargs)
+        super().__init__(ard_num_dims=ard_num_dims, **kwargs)
         if variance_constraint is None:
             variance_constraint = Positive()
         self.register_parameter(

diff --git a/test/kernels/test_linear_kernel.py b/test/kernels/test_linear_kernel.py
@@ -98,6 +98,7 @@ class TestLinearKernelARD(TestLinearKernel):
     def test_kernel_ard(self) -> None:
         self.kernel_kwargs = {"ard_num_dims": 2}
         kernel = self.create_kernel_no_ard()
+        self.assertEqual(kernel.ard_num_dims, 2)
         self.assertEqual(kernel.variance.shape, torch.Size([1, 2]))
 
     def test_computes_linear_function_rectangular(self):