def test_gpflow_predictor_get_observation_noise_raises_for_likelihood_with_variance(
gpflow_interface_factory: ModelFactoryType,
) -> None:
data = mock_data()
model, _ = gpflow_interface_factory(*data)
model.model.likelihood = gpflow.likelihoods.Gaussian() # has variance attribute
model.get_observation_noise()
model.model.likelihood = gpflow.likelihoods.Bernoulli() # does not have variance attribute
with pytest.raises(NotImplementedError):
model.get_observation_noise()
def test_gaussian_process_regression_default_optimize(
gpr_interface_factory: ModelFactoryType, ) -> None:
data = mock_data()
model, _ = gpr_interface_factory(*data)
internal_model = model.model
if isinstance(internal_model, SVGP):
args = {"data": data}
else:
args = {}
loss = internal_model.training_loss(**args)
model.optimize(Dataset(*data))
assert internal_model.training_loss(**args) < loss
def test_gaussian_process_regression_optimize(
optimizer: gpflow.optimizers.Scipy | tf.optimizers.Optimizer,
gpflow_interface_factory: ModelFactoryType,
compile: bool,
) -> None:
data = mock_data()
optimizer_wrapper = create_optimizer(optimizer, dict(compile=compile))
model, _ = gpflow_interface_factory(*data, optimizer=optimizer_wrapper)
internal_model = model.model
if isinstance(internal_model, SVGP):
args = {"data": data}
else:
args = {}
loss = internal_model.training_loss(**args)
model.optimize(Dataset(*data))
assert internal_model.training_loss(**args) < loss
def test_sparse_variational_model_attribute() -> None:
model = svgp_model(*mock_data())
sv = SparseVariational(model)
assert sv.model is model
def test_sgpr_raises_for_covariance_between_points() -> None:
data = mock_data()
model = GaussianProcessRegression(sgpr_model(*data))
with pytest.raises(NotImplementedError):
model.covariance_between_points(data[0], data[0])