Allow monotonic rejection GP to initialize its inducing point allocat…

…or from config. (facebookresearch#485)

Summary:
Pull Request resolved: facebookresearch#485

Monotonic rejection GP used to completely ignore any inducing point config, now it follows other models.

Reviewed By: crasanders

Differential Revision: D67225873

fbshipit-source-id: 843d4a129e81ac8b2f3fb4b3b63e6d24bd51e3a5

aepsych/models/monotonic_rejection_gp.py

@@ -96,12 +96,9 @@ def __init__(
             dim=self.dim
         )
 
-        # TODO: This allocator *must* be SobolAllocator and not the set one. This
-        # suggests that this model doesn't actually properly use data for inducing
-        # points properly.
-        inducing_points = SobolAllocator(
-            bounds=torch.stack([lb, ub]), dim=self.dim
-        ).allocate_inducing_points(num_inducing=self.inducing_size)
+        inducing_points = self.inducing_point_method.allocate_inducing_points(
+            num_inducing=self.inducing_size
+        )
 
         inducing_points_aug = self._augment_with_deriv_index(inducing_points, 0)
         variational_distribution = CholeskyVariationalDistribution(