Support initialising Linear when A is a Parameter (#2096)

GPflow · Dec 19, 2023 · dca094b · dca094b
1 parent 5d4c222
commit dca094b
Show file tree

Hide file tree

Showing 4 changed files with 34 additions and 1 deletion.
diff --git a/CONTRIBUTORS.md b/CONTRIBUTORS.md
@@ -89,5  89,6 @@ Because GitHub's [graph of contributors](http://github.com/GPflow/GPflow/graphs/
 [@khurram-ghani](https://github.com/khurram-ghani)
 [@partev](https://github.com/partev)
 [@uri-granta](https://github.com/uri-granta)
 [@jschmiegel](https://github.com/jschmiegel)
 
 Add yourself when you first contribute to GPflow's code, tests, or documentation!
diff --git a/RELEASE.md b/RELEASE.md
@@ -56,6  56,7 @@ This release contains contributions from:
 ## Bug Fixes and Other Changes
 
 * Don't round small values in kernel summary printout
 * allowing to set a prior for A in the Linear() mean function
 
 ## Thanks to our Contributors
 

diff --git a/gpflow/functions.py b/gpflow/functions.py
@@ -110,7  110,15 @@ def __init__(self, A: TensorType = None, b: TensorType = None) -> None:
         MeanFunction.__init__(self)
         A = np.ones((1, 1), dtype=default_float()) if A is None else A
         b = np.zeros(1, dtype=default_float()) if b is None else b
-        self.A = Parameter(np.atleast_2d(A))
         if isinstance(A, Parameter):
             if len(A._shape) >= 2:
                 self.A = A
             else:
                 raise ValueError(
                     "Error 'gpflow.funcitons.Linear()' mean function. A has not the correct shape (at least 2d)."
                 )
         else:
             self.A = Parameter(np.atleast_2d(A))
         self.b = Parameter(b)
 
     @inherit_check_shapes

diff --git a/tests/gpflow/test_functions.py b/tests/gpflow/test_functions.py
@@ -16,6  16,7 @@
 
 import numpy as np
 import pytest
 import tensorflow_probability as tfp
 from check_shapes import check_shapes
 from numpy.testing import assert_allclose
 
@@ -371,3  372,25 @@ def test_models_with_mean_functions_changes(model_class: Type[Any]) -> None:
     assert np.all(var_zero.numpy() == var_non_zero.numpy())
     # predictive mean changes after modifying mean function
     assert not np.all(mu_zero.numpy() == mu_non_zero.numpy())
 
 
 class TestIssue2091EnsureParameterTypeForLinearMeanFunction:
     """
     See github issue #2091. These are tests to ensure that the prior is kept if A is already given as Parameter.
     """
 
     # test_parameter_with_correct_shape
     def test_parameter_with_correct_shape(self) -> None:
         "Check that Linear copies the prior from a correctly shaped A Parameter"
         A = gpflow.Parameter(
             np.ones((1, 1)), dtype=np.float64, prior=tfp.distributions.Normal(0.0, 1.0)
         )
         linear_function = Linear(A, 1)
         assert linear_function.A.prior is not None
 
     # test_parameter_with_incorrect_shape
     def test_parameter_with_incorrect_shape(self) -> None:
         "Check that Linear throws an error when the A Parameter is not correctly shaped"
         A = gpflow.Parameter(np.zeros(20), dtype=np.float64)
         with pytest.raises(ValueError):
             Linear(A, 1)