Bugfixes for Normal distribution in torch backend

Author: marvinpfoertner

src/probnum/backend/linalg/_torch.py

@@ -26,15 +26,15 @@ def solve_triangular(
             upper=not lower,
             transpose=transpose,
             unitriangular=unit_diagonal,
-        )[:, 0]
+        ).solution[:, 0]
 
     return torch.triangular_solve(
         b,
         A,
         upper=not lower,
         transpose=transpose,
         unitriangular=unit_diagonal,
-    )
+    ).solution
 
 
 def solve_cholesky(

src/probnum/randvars/_normal.py

@@ -204,9 +204,9 @@ def cov_cholesky(self) -> ArrayType:
 
         return self._cov_cholesky
 
-    @property
+    @functools.cached_property
     def _cov_matrix_cholesky(self) -> ArrayType:
-        return self._cov_op_cholesky.todense()
+        return backend.asarray(self._cov_op_cholesky.todense())
 
     @property
     def _cov_op_cholesky(self) -> ArrayType:
@@ -453,7 +453,9 @@ def _sample(self, seed: SeedLike, sample_shape: ShapeType = ()) -> ArrayType:
             dtype=self.dtype,
         )
 
-        samples = self._cov_op_cholesky(backend.to_numpy(samples), axis=-1)
+        samples = backend.asarray(
+            self._cov_op_cholesky(backend.to_numpy(samples), axis=-1)
+        )
         samples += self.dense_mean
 
         return samples.reshape(sample_shape + self.shape)
@@ -489,7 +491,7 @@ def _logpdf(self, x: ArrayType) -> ArrayType:
         # TODO (#569): Replace `solve_triangular` with:
         # self._cov_op_cholesky.inv() @ x_centered[..., None]
         x_whitened = backend.linalg.solve_triangular(
-            self._cov_matrix_cholesky,
+            backend.asarray(self._cov_matrix_cholesky),
             x_centered[..., None],
             lower=True,
         )[..., 0]