mirror of
https://github.com/SheffieldML/GPy.git
synced 2026-05-24 14:15:14 +02:00
format on save
This commit is contained in:
Martin Bubel
parent
03fcf7311d
commit
a02f4039fa
1 changed files with 501 additions and 211 deletions
|
|
@ -15,7 +15,8 @@ verbose = 0
|
|||
|
||||
try:
|
||||
from ..kern.src import coregionalize_cython
|
||||
cython_coregionalize_working = config.getboolean('cython', 'working')
|
||||
|
||||
cython_coregionalize_working = config.getboolean("cython", "working")
|
||||
except ImportError:
|
||||
cython_coregionalize_working = False
|
||||
|
||||
|
|
@ -26,8 +27,9 @@ class Kern_check_model(GPy.core.Model):
|
|||
gradients of a given kernel are implemented correctly. It enables
|
||||
checkgrad() to be called independently on a kernel.
|
||||
"""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None, X2=None):
|
||||
super(Kern_check_model, self).__init__('kernel_test_model')
|
||||
super(Kern_check_model, self).__init__("kernel_test_model")
|
||||
if kernel == None:
|
||||
kernel = GPy.kern.RBF(1)
|
||||
kernel.randomize(loc=1, scale=0.1)
|
||||
|
|
@ -55,13 +57,17 @@ class Kern_check_model(GPy.core.Model):
|
|||
def log_likelihood(self):
|
||||
return np.sum(self.dL_dK * self.kernel.K(self.X, self.X2))
|
||||
|
||||
|
||||
class Kern_check_dK_dtheta(Kern_check_model):
|
||||
"""
|
||||
This class allows gradient checks for the gradient of a kernel with
|
||||
respect to parameters.
|
||||
"""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None, X2=None):
|
||||
super(Kern_check_dK_dtheta, self).__init__(kernel=kernel,dL_dK=dL_dK, X=X, X2=X2)
|
||||
super(Kern_check_dK_dtheta, self).__init__(
|
||||
kernel=kernel, dL_dK=dL_dK, X=X, X2=X2
|
||||
)
|
||||
self.link_parameter(self.kernel)
|
||||
|
||||
def parameters_changed(self):
|
||||
|
|
@ -73,8 +79,11 @@ class Kern_check_dKdiag_dtheta(Kern_check_model):
|
|||
This class allows gradient checks of the gradient of the diagonal of a
|
||||
kernel with respect to the parameters.
|
||||
"""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None):
|
||||
super(Kern_check_dKdiag_dtheta, self).__init__(kernel=kernel,dL_dK=dL_dK, X=X, X2=None)
|
||||
super(Kern_check_dKdiag_dtheta, self).__init__(
|
||||
kernel=kernel, dL_dK=dL_dK, X=X, X2=None
|
||||
)
|
||||
self.link_parameter(self.kernel)
|
||||
|
||||
def log_likelihood(self):
|
||||
|
|
@ -83,20 +92,26 @@ class Kern_check_dKdiag_dtheta(Kern_check_model):
|
|||
def parameters_changed(self):
|
||||
self.kernel.update_gradients_diag(np.diag(self.dL_dK), self.X)
|
||||
|
||||
|
||||
class Kern_check_dK_dX(Kern_check_model):
|
||||
"""This class allows gradient checks for the gradient of a kernel with respect to X."""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None, X2=None):
|
||||
super(Kern_check_dK_dX, self).__init__(kernel=kernel, dL_dK=dL_dK, X=X, X2=X2)
|
||||
self.X = Param('X',X)
|
||||
self.X = Param("X", X)
|
||||
self.link_parameter(self.X)
|
||||
|
||||
def parameters_changed(self):
|
||||
self.X.gradient[:] = self.kernel.gradients_X(self.dL_dK, self.X, self.X2)
|
||||
|
||||
|
||||
class Kern_check_dKdiag_dX(Kern_check_dK_dX):
|
||||
"""This class allows gradient checks for the gradient of a kernel diagonal with respect to X."""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None, X2=None):
|
||||
super(Kern_check_dKdiag_dX, self).__init__(kernel=kernel,dL_dK=dL_dK, X=X, X2=None)
|
||||
super(Kern_check_dKdiag_dX, self).__init__(
|
||||
kernel=kernel, dL_dK=dL_dK, X=X, X2=None
|
||||
)
|
||||
|
||||
def log_likelihood(self):
|
||||
return (np.diag(self.dL_dK) * self.kernel.Kdiag(self.X)).sum()
|
||||
|
|
@ -104,11 +119,15 @@ class Kern_check_dKdiag_dX(Kern_check_dK_dX):
|
|||
def parameters_changed(self):
|
||||
self.X.gradient[:] = self.kernel.gradients_X_diag(self.dL_dK.diagonal(), self.X)
|
||||
|
||||
|
||||
class Kern_check_d2K_dXdX(Kern_check_model):
|
||||
"""This class allows gradient checks for the secondderivative of a kernel with respect to X."""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None, X2=None):
|
||||
super(Kern_check_d2K_dXdX, self).__init__(kernel=kernel,dL_dK=dL_dK, X=X, X2=X2)
|
||||
self.X = Param('X',X.copy())
|
||||
super(Kern_check_d2K_dXdX, self).__init__(
|
||||
kernel=kernel, dL_dK=dL_dK, X=X, X2=X2
|
||||
)
|
||||
self.X = Param("X", X.copy())
|
||||
self.link_parameter(self.X)
|
||||
self.Xc = X.copy()
|
||||
|
||||
|
|
@ -123,28 +142,37 @@ class Kern_check_d2K_dXdX(Kern_check_model):
|
|||
if self.X2 is None:
|
||||
grads = -self.kernel.gradients_XX(self.dL_dK, self.X).sum(1).sum(1)
|
||||
else:
|
||||
grads = -self.kernel.gradients_XX(self.dL_dK.T, self.X2, self.X).sum(0).sum(1)
|
||||
grads = (
|
||||
-self.kernel.gradients_XX(self.dL_dK.T, self.X2, self.X).sum(0).sum(1)
|
||||
)
|
||||
self.X.gradient[:] = grads
|
||||
|
||||
|
||||
class Kern_check_d2Kdiag_dXdX(Kern_check_model):
|
||||
"""This class allows gradient checks for the second derivative of a kernel with respect to X."""
|
||||
|
||||
def __init__(self, kernel=None, dL_dK=None, X=None):
|
||||
super(Kern_check_d2Kdiag_dXdX, self).__init__(kernel=kernel, dL_dK=dL_dK, X=X)
|
||||
self.X = Param('X',X)
|
||||
self.X = Param("X", X)
|
||||
self.link_parameter(self.X)
|
||||
self.Xc = X.copy()
|
||||
|
||||
def log_likelihood(self):
|
||||
l = 0.
|
||||
l = 0.0
|
||||
for i in range(self.X.shape[0]):
|
||||
l += self.kernel.gradients_X(self.dL_dK[[i],[i]], self.X[[i]], self.Xc[[i]]).sum()
|
||||
l += self.kernel.gradients_X(
|
||||
self.dL_dK[[i], [i]], self.X[[i]], self.Xc[[i]]
|
||||
).sum()
|
||||
return l
|
||||
|
||||
def parameters_changed(self):
|
||||
grads = -self.kernel.gradients_XX_diag(self.dL_dK.diagonal(), self.X)
|
||||
self.X.gradient[:] = grads.sum(-1)
|
||||
|
||||
def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verbose=False, fixed_X_dims=None):
|
||||
|
||||
def check_kernel_gradient_functions(
|
||||
kern, X=None, X2=None, output_ind=None, verbose=False, fixed_X_dims=None
|
||||
):
|
||||
"""
|
||||
This function runs on kernels to check the correctness of their
|
||||
implementation. It checks that the covariance function is positive definite
|
||||
|
|
@ -174,9 +202,15 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Positive definite check failed for " + kern.name + " covariance function."))
|
||||
print(
|
||||
(
|
||||
"Positive definite check failed for "
|
||||
+ kern.name
|
||||
+ " covariance function."
|
||||
)
|
||||
)
|
||||
pass_checks = False
|
||||
assert(result)
|
||||
assert result
|
||||
return False
|
||||
|
||||
if verbose:
|
||||
|
|
@ -185,10 +219,16 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of K(X, X) wrt theta failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of K(X, X) wrt theta failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
Kern_check_dK_dtheta(kern, X=X, X2=None).checkgrad(verbose=True)
|
||||
pass_checks = False
|
||||
assert(result)
|
||||
assert result
|
||||
return False
|
||||
|
||||
if verbose:
|
||||
|
|
@ -198,14 +238,25 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
except NotImplementedError:
|
||||
result = True
|
||||
if verbose:
|
||||
print(("update_gradients_full, with differing X and X2, not implemented for " + kern.name))
|
||||
print(
|
||||
(
|
||||
"update_gradients_full, with differing X and X2, not implemented for "
|
||||
+ kern.name
|
||||
)
|
||||
)
|
||||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of K(X, X) wrt theta failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of K(X, X) wrt theta failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
Kern_check_dK_dtheta(kern, X=X, X2=X2).checkgrad(verbose=True)
|
||||
pass_checks = False
|
||||
assert(result)
|
||||
assert result
|
||||
return False
|
||||
|
||||
if verbose:
|
||||
|
|
@ -219,10 +270,16 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of Kdiag(X) wrt theta failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of Kdiag(X) wrt theta failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
Kern_check_dKdiag_dtheta(kern, X=X).checkgrad(verbose=True)
|
||||
pass_checks = False
|
||||
assert(result)
|
||||
assert result
|
||||
return False
|
||||
|
||||
if verbose:
|
||||
|
|
@ -239,9 +296,15 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of K(X, X) wrt X failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of K(X, X) wrt X failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
testmodel.checkgrad(verbose=True)
|
||||
assert(result)
|
||||
assert result
|
||||
pass_checks = False
|
||||
return False
|
||||
|
||||
|
|
@ -259,9 +322,15 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of K(X, X2) wrt X failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of K(X, X2) wrt X failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
testmodel.checkgrad(verbose=True)
|
||||
assert(result)
|
||||
assert result
|
||||
pass_checks = False
|
||||
return False
|
||||
|
||||
|
|
@ -279,10 +348,16 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of Kdiag(X) wrt X failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of Kdiag(X) wrt X failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
Kern_check_dKdiag_dX(kern, X=X).checkgrad(verbose=True)
|
||||
pass_checks = False
|
||||
assert(result)
|
||||
assert result
|
||||
return False
|
||||
|
||||
if verbose:
|
||||
|
|
@ -299,9 +374,15 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of dK(X, X2) wrt X failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of dK(X, X2) wrt X failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
testmodel.checkgrad(verbose=True)
|
||||
assert(result)
|
||||
assert result
|
||||
pass_checks = False
|
||||
return False
|
||||
|
||||
|
|
@ -319,9 +400,15 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of dK(X, X) wrt X with full cov in dimensions failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of dK(X, X) wrt X with full cov in dimensions failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
testmodel.checkgrad(verbose=True)
|
||||
assert(result)
|
||||
assert result
|
||||
pass_checks = False
|
||||
return False
|
||||
|
||||
|
|
@ -339,71 +426,114 @@ def check_kernel_gradient_functions(kern, X=None, X2=None, output_ind=None, verb
|
|||
if result and verbose:
|
||||
print("Check passed.")
|
||||
if not result:
|
||||
print(("Gradient of dKdiag(X, X) wrt X with cov in dimensions failed for " + kern.name + " covariance function. Gradient values as follows:"))
|
||||
print(
|
||||
(
|
||||
"Gradient of dKdiag(X, X) wrt X with cov in dimensions failed for "
|
||||
+ kern.name
|
||||
+ " covariance function. Gradient values as follows:"
|
||||
)
|
||||
)
|
||||
testmodel.checkgrad(verbose=True)
|
||||
assert(result)
|
||||
assert result
|
||||
pass_checks = False
|
||||
return False
|
||||
|
||||
return pass_checks
|
||||
|
||||
|
||||
|
||||
class KernelGradientTestsContinuous(unittest.TestCase):
|
||||
def setUp(self):
|
||||
self.N, self.D = 10, 5
|
||||
self.X = np.random.randn(self.N, self.D + 1)
|
||||
self.X2 = np.random.randn(self.N + 10, self.D + 1)
|
||||
|
||||
continuous_kerns = ['RBF', 'Linear']
|
||||
continuous_kerns = ["RBF", "Linear"]
|
||||
self.kernclasses = [getattr(GPy.kern, s) for s in continuous_kerns]
|
||||
|
||||
def test_MLP(self):
|
||||
k = GPy.kern.MLP(self.D, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Matern32(self):
|
||||
k = GPy.kern.Matern32(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Prod(self):
|
||||
k = GPy.kern.Matern32(2, active_dims=[2,3]) * GPy.kern.RBF(2, active_dims=[0,4]) + GPy.kern.Linear(self.D)
|
||||
k = GPy.kern.Matern32(2, active_dims=[2, 3]) * GPy.kern.RBF(
|
||||
2, active_dims=[0, 4]
|
||||
) + GPy.kern.Linear(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Prod1(self):
|
||||
k = GPy.kern.RBF(self.D) * GPy.kern.Linear(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Prod2(self):
|
||||
k = GPy.kern.RBF(2, active_dims=[0, 4]) * GPy.kern.Linear(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Prod3(self):
|
||||
k = GPy.kern.RBF(self.D) * GPy.kern.Linear(self.D) * GPy.kern.Bias(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Prod4(self):
|
||||
k = GPy.kern.RBF(2, active_dims=[0,4]) * GPy.kern.Linear(self.D) * GPy.kern.Matern32(2, active_dims=[0,1])
|
||||
k = (
|
||||
GPy.kern.RBF(2, active_dims=[0, 4])
|
||||
* GPy.kern.Linear(self.D)
|
||||
* GPy.kern.Matern32(2, active_dims=[0, 1])
|
||||
)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Add(self):
|
||||
k = GPy.kern.Matern32(2, active_dims=[2,3]) + GPy.kern.RBF(2, active_dims=[0,4]) + GPy.kern.Linear(self.D)
|
||||
k += GPy.kern.Matern32(2, active_dims=[2,3]) + GPy.kern.RBF(2, active_dims=[0,4]) + GPy.kern.Linear(self.D)
|
||||
k = (
|
||||
GPy.kern.Matern32(2, active_dims=[2, 3])
|
||||
+ GPy.kern.RBF(2, active_dims=[0, 4])
|
||||
+ GPy.kern.Linear(self.D)
|
||||
)
|
||||
k += (
|
||||
GPy.kern.Matern32(2, active_dims=[2, 3])
|
||||
+ GPy.kern.RBF(2, active_dims=[0, 4])
|
||||
+ GPy.kern.Linear(self.D)
|
||||
)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Add_dims(self):
|
||||
k = GPy.kern.Matern32(2, active_dims=[2,self.D]) + GPy.kern.RBF(2, active_dims=[0,4]) + GPy.kern.Linear(self.D)
|
||||
k = (
|
||||
GPy.kern.Matern32(2, active_dims=[2, self.D])
|
||||
+ GPy.kern.RBF(2, active_dims=[0, 4])
|
||||
+ GPy.kern.Linear(self.D)
|
||||
)
|
||||
k.randomize()
|
||||
self.assertRaises(IndexError, k.K, self.X[:, : self.D])
|
||||
k = GPy.kern.Matern32(2, active_dims=[2,self.D-1]) + GPy.kern.RBF(2, active_dims=[0,4]) + GPy.kern.Linear(self.D)
|
||||
k = (
|
||||
GPy.kern.Matern32(2, active_dims=[2, self.D - 1])
|
||||
+ GPy.kern.RBF(2, active_dims=[0, 4])
|
||||
+ GPy.kern.Linear(self.D)
|
||||
)
|
||||
k.randomize()
|
||||
# assert it runs:
|
||||
try:
|
||||
|
|
@ -414,101 +544,135 @@ class KernelGradientTestsContinuous(unittest.TestCase):
|
|||
def test_Matern52(self):
|
||||
k = GPy.kern.Matern52(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_RBF(self):
|
||||
k = GPy.kern.RBF(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_OU(self):
|
||||
k = GPy.kern.OU(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Cosine(self):
|
||||
# Don't test Cosine directly as it fails positive definite test.
|
||||
k = GPy.kern.RBF(self.D - 1, ARD=False) * GPy.kern.Cosine(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_ExpQuadCosine(self):
|
||||
k = GPy.kern.ExpQuadCosine(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Sinc(self):
|
||||
k = GPy.kern.Sinc(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_RatQuad(self):
|
||||
k = GPy.kern.RatQuad(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_ExpQuad(self):
|
||||
k = GPy.kern.ExpQuad(self.D - 1, ARD=True)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_integral(self):
|
||||
k = GPy.kern.Integral(1)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_multidimensional_integral_limits(self):
|
||||
k = GPy.kern.Multidimensional_Integral_Limits(2)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_integral_limits(self):
|
||||
k = GPy.kern.Integral_Limits(2)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Linear(self):
|
||||
k = GPy.kern.Linear(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_LinearFull(self):
|
||||
k = GPy.kern.LinearFull(self.D, self.D - 1)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Fixed(self):
|
||||
cov = np.dot(self.X, self.X.T)
|
||||
X = np.arange(self.N).reshape(self.N, 1)
|
||||
k = GPy.kern.Fixed(1, cov)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=X, X2=None, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=X, X2=None, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_Poly(self):
|
||||
k = GPy.kern.Poly(self.D, order=5)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_WhiteHeteroscedastic(self):
|
||||
k = GPy.kern.WhiteHeteroscedastic(self.D, self.X.shape[0])
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_standard_periodic(self):
|
||||
k = GPy.kern.StdPeriodic(self.D)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_symmetric_even(self):
|
||||
k_base = GPy.kern.Linear(1) + GPy.kern.RBF(1)
|
||||
transform = -np.array([[1.0]])
|
||||
k = GPy.kern.Symmetric(k_base, transform, 'even')
|
||||
k = GPy.kern.Symmetric(k_base, transform, "even")
|
||||
self.assertTrue(check_kernel_gradient_functions(k))
|
||||
|
||||
def test_symmetric_odd(self):
|
||||
k_base = GPy.kern.Linear(1) + GPy.kern.RBF(1)
|
||||
transform = -np.array([[1.0]])
|
||||
k = GPy.kern.Symmetric(k_base, transform, 'odd')
|
||||
k = GPy.kern.Symmetric(k_base, transform, "odd")
|
||||
self.assertTrue(check_kernel_gradient_functions(k))
|
||||
|
||||
def test_MultioutputKern(self):
|
||||
|
|
@ -520,7 +684,11 @@ class KernelGradientTestsContinuous(unittest.TestCase):
|
|||
k = GPy.kern.MultioutputKern([k1, k2])
|
||||
Xt, _, _ = GPy.util.multioutput.build_XY([self.X, self.X])
|
||||
X2t, _, _ = GPy.util.multioutput.build_XY([self.X2, self.X2])
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=Xt, X2=X2t, verbose=verbose, fixed_X_dims=-1))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
k, X=Xt, X2=X2t, verbose=verbose, fixed_X_dims=-1
|
||||
)
|
||||
)
|
||||
|
||||
def test_Precomputed(self):
|
||||
Xall = np.concatenate([self.X, self.X2])
|
||||
|
|
@ -529,42 +697,73 @@ class KernelGradientTestsContinuous(unittest.TestCase):
|
|||
X2 = np.arange(self.N, 2 * self.N + 10).reshape(self.N + 10, 1)
|
||||
k = GPy.kern.Precomputed(1, cov)
|
||||
k.randomize()
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=X, X2=X2, verbose=verbose, fixed_X_dims=[0]))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
k, X=X, X2=X2, verbose=verbose, fixed_X_dims=[0]
|
||||
)
|
||||
)
|
||||
|
||||
def test_basis_func_linear_slope(self):
|
||||
start_stop = np.random.uniform(self.X.min(0), self.X.max(0), (4, self.X.shape[1])).T
|
||||
start_stop = np.random.uniform(
|
||||
self.X.min(0), self.X.max(0), (4, self.X.shape[1])
|
||||
).T
|
||||
start_stop.sort(axis=1)
|
||||
ks = []
|
||||
for i in range(start_stop.shape[0]):
|
||||
start, stop = np.split(start_stop[i], 2)
|
||||
ks.append(GPy.kern.LinearSlopeBasisFuncKernel(1, start, stop, ARD=i%2==0, active_dims=[i]))
|
||||
ks.append(
|
||||
GPy.kern.LinearSlopeBasisFuncKernel(
|
||||
1, start, stop, ARD=i % 2 == 0, active_dims=[i]
|
||||
)
|
||||
)
|
||||
k = GPy.kern.Add(ks)
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_basis_func_changepoint(self):
|
||||
points = np.random.uniform(self.X.min(0), self.X.max(0), (self.X.shape[1]))
|
||||
ks = []
|
||||
for i in range(points.shape[0]):
|
||||
ks.append(GPy.kern.ChangePointBasisFuncKernel(1, points[i], ARD=i%2==0, active_dims=[i]))
|
||||
ks.append(
|
||||
GPy.kern.ChangePointBasisFuncKernel(
|
||||
1, points[i], ARD=i % 2 == 0, active_dims=[i]
|
||||
)
|
||||
)
|
||||
k = GPy.kern.Add(ks)
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_basis_func_poly(self):
|
||||
ks = []
|
||||
for i in range(self.X.shape[1]):
|
||||
ks.append(GPy.kern.PolynomialBasisFuncKernel(1, 5, ARD=i%2==0, active_dims=[i]))
|
||||
ks.append(
|
||||
GPy.kern.PolynomialBasisFuncKernel(
|
||||
1, 5, ARD=i % 2 == 0, active_dims=[i]
|
||||
)
|
||||
)
|
||||
k = GPy.kern.Add(ks)
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
def test_basis_func_domain(self):
|
||||
start_stop = np.random.uniform(self.X.min(0), self.X.max(0), (4, self.X.shape[1])).T
|
||||
start_stop = np.random.uniform(
|
||||
self.X.min(0), self.X.max(0), (4, self.X.shape[1])
|
||||
).T
|
||||
start_stop.sort(axis=1)
|
||||
ks = []
|
||||
for i in range(start_stop.shape[0]):
|
||||
start, stop = np.split(start_stop[i], 2)
|
||||
ks.append(GPy.kern.DomainKernel(1, start, stop, ARD=i%2==0, active_dims=[i]))
|
||||
ks.append(
|
||||
GPy.kern.DomainKernel(1, start, stop, ARD=i % 2 == 0, active_dims=[i])
|
||||
)
|
||||
k = GPy.kern.Add(ks)
|
||||
self.assertTrue(check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(k, X=self.X, X2=self.X2, verbose=verbose)
|
||||
)
|
||||
|
||||
|
||||
class KernelTestsMiscellaneous(unittest.TestCase):
|
||||
def setUp(self):
|
||||
|
|
@ -581,9 +780,24 @@ class KernelTestsMiscellaneous(unittest.TestCase):
|
|||
# self.sumkern.randomize()
|
||||
|
||||
def test_which_parts(self):
|
||||
self.assertTrue(np.allclose(self.sumkern.K(self.X, which_parts=[self.linear, self.matern]), self.linear.K(self.X)+self.matern.K(self.X)))
|
||||
self.assertTrue(np.allclose(self.sumkern.K(self.X, which_parts=[self.linear, self.rbf]), self.linear.K(self.X)+self.rbf.K(self.X)))
|
||||
self.assertTrue(np.allclose(self.sumkern.K(self.X, which_parts=self.sumkern.parts[0]), self.rbf.K(self.X)))
|
||||
self.assertTrue(
|
||||
np.allclose(
|
||||
self.sumkern.K(self.X, which_parts=[self.linear, self.matern]),
|
||||
self.linear.K(self.X) + self.matern.K(self.X),
|
||||
)
|
||||
)
|
||||
self.assertTrue(
|
||||
np.allclose(
|
||||
self.sumkern.K(self.X, which_parts=[self.linear, self.rbf]),
|
||||
self.linear.K(self.X) + self.rbf.K(self.X),
|
||||
)
|
||||
)
|
||||
self.assertTrue(
|
||||
np.allclose(
|
||||
self.sumkern.K(self.X, which_parts=self.sumkern.parts[0]),
|
||||
self.rbf.K(self.X),
|
||||
)
|
||||
)
|
||||
|
||||
def test_active_dims(self):
|
||||
np.testing.assert_array_equal(self.sumkern.active_dims, [0, 1, 2, 3, 7, 9])
|
||||
|
|
@ -604,6 +818,7 @@ class KernelTestsMiscellaneous(unittest.TestCase):
|
|||
np.testing.assert_array_equal(tmp.active_dims, [0, 1, 2, 3, 7, 9])
|
||||
np.testing.assert_array_equal(tmp._all_dims_active, range(10))
|
||||
|
||||
|
||||
class KernelTestsNonContinuous(unittest.TestCase):
|
||||
def setUp(self):
|
||||
N0 = 3
|
||||
|
|
@ -622,46 +837,84 @@ class KernelTestsNonContinuous(unittest.TestCase):
|
|||
self.X2[(N0 * 2) :, -1] = 1
|
||||
|
||||
def test_IndependentOutputs(self):
|
||||
k = [GPy.kern.RBF(1, active_dims=[1], name='rbf1'), GPy.kern.RBF(self.D, active_dims=range(self.D), name='rbf012'), GPy.kern.RBF(2, active_dims=[0,2], name='rbf02')]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name='ind_split')
|
||||
k = [
|
||||
GPy.kern.RBF(1, active_dims=[1], name="rbf1"),
|
||||
GPy.kern.RBF(self.D, active_dims=range(self.D), name="rbf012"),
|
||||
GPy.kern.RBF(2, active_dims=[0, 2], name="rbf02"),
|
||||
]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name="ind_split")
|
||||
np.testing.assert_array_equal(kern.active_dims, [-1, 0, 1, 2])
|
||||
np.testing.assert_array_equal(kern._all_dims_active, [0, 1, 2, -1])
|
||||
|
||||
def testIndependendGradients(self):
|
||||
k = GPy.kern.RBF(self.D, active_dims=range(self.D))
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, 'ind_single')
|
||||
self.assertTrue(check_kernel_gradient_functions(kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1))
|
||||
k = [GPy.kern.RBF(1, active_dims=[1], name='rbf1'), GPy.kern.RBF(self.D, active_dims=range(self.D), name='rbf012'), GPy.kern.RBF(2, active_dims=[0,2], name='rbf02')]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name='ind_split')
|
||||
self.assertTrue(check_kernel_gradient_functions(kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1))
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, "ind_single")
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1
|
||||
)
|
||||
)
|
||||
k = [
|
||||
GPy.kern.RBF(1, active_dims=[1], name="rbf1"),
|
||||
GPy.kern.RBF(self.D, active_dims=range(self.D), name="rbf012"),
|
||||
GPy.kern.RBF(2, active_dims=[0, 2], name="rbf02"),
|
||||
]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name="ind_split")
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1
|
||||
)
|
||||
)
|
||||
|
||||
def test_Hierarchical(self):
|
||||
k = [GPy.kern.RBF(2, active_dims=[0,2], name='rbf1'), GPy.kern.RBF(2, active_dims=[0,2], name='rbf2')]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name='ind_split')
|
||||
k = [
|
||||
GPy.kern.RBF(2, active_dims=[0, 2], name="rbf1"),
|
||||
GPy.kern.RBF(2, active_dims=[0, 2], name="rbf2"),
|
||||
]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name="ind_split")
|
||||
np.testing.assert_array_equal(kern.active_dims, [-1, 0, 2])
|
||||
np.testing.assert_array_equal(kern._all_dims_active, [0, 1, 2, -1])
|
||||
|
||||
def test_Hierarchical_gradients(self):
|
||||
k = [GPy.kern.RBF(2, active_dims=[0,2], name='rbf1'), GPy.kern.RBF(2, active_dims=[0,2], name='rbf2')]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name='ind_split')
|
||||
self.assertTrue(check_kernel_gradient_functions(kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1))
|
||||
|
||||
k = [
|
||||
GPy.kern.RBF(2, active_dims=[0, 2], name="rbf1"),
|
||||
GPy.kern.RBF(2, active_dims=[0, 2], name="rbf2"),
|
||||
]
|
||||
kern = GPy.kern.IndependentOutputs(k, -1, name="ind_split")
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1
|
||||
)
|
||||
)
|
||||
|
||||
def test_ODE_UY(self):
|
||||
kern = GPy.kern.ODE_UY(2, active_dims=[0, self.D])
|
||||
X = self.X[self.X[:, -1] != 2]
|
||||
X2 = self.X2[self.X2[:, -1] != 2]
|
||||
self.assertTrue(check_kernel_gradient_functions(kern, X=X, X2=X2, verbose=verbose, fixed_X_dims=-1))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
kern, X=X, X2=X2, verbose=verbose, fixed_X_dims=-1
|
||||
)
|
||||
)
|
||||
|
||||
def test_Coregionalize(self):
|
||||
kern = GPy.kern.Coregionalize(1, output_dim=3, active_dims=[-1])
|
||||
self.assertTrue(check_kernel_gradient_functions(kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1))
|
||||
self.assertTrue(
|
||||
check_kernel_gradient_functions(
|
||||
kern, X=self.X, X2=self.X2, verbose=verbose, fixed_X_dims=-1
|
||||
)
|
||||
)
|
||||
|
||||
@unittest.skipIf(not cython_coregionalize_working,"Cython coregionalize module has not been built on this machine")
|
||||
|
||||
@unittest.skipIf(
|
||||
not cython_coregionalize_working,
|
||||
"Cython coregionalize module has not been built on this machine",
|
||||
)
|
||||
class Coregionalize_cython_test(unittest.TestCase):
|
||||
"""
|
||||
Make sure that the coregionalize kernel work with and without cython enabled
|
||||
"""
|
||||
|
||||
def setUp(self):
|
||||
self.k = GPy.kern.Coregionalize(1, output_dim=12)
|
||||
self.N1, self.N2 = 100, 200
|
||||
|
|
@ -691,12 +944,12 @@ class Coregionalize_cython_test(unittest.TestCase):
|
|||
def test_nonsym(self):
|
||||
dL_dK = np.random.randn(self.N1, self.N2)
|
||||
K_cython = self.k._K_cython(self.X, self.X2)
|
||||
self.k.gradient = 0.
|
||||
self.k.gradient = 0.0
|
||||
self.k.update_gradients_full(dL_dK, self.X, self.X2)
|
||||
grads_cython = self.k.gradient.copy()
|
||||
|
||||
K_numpy = self.k._K_numpy(self.X, self.X2)
|
||||
self.k.gradient = 0.
|
||||
self.k.gradient = 0.0
|
||||
# Same hack as in test_sym (Line 639)
|
||||
_gradient_reduce_cython = self.k._gradient_reduce_cython
|
||||
self.k._gradient_reduce_cython = self.k._gradient_reduce_numpy
|
||||
|
|
@ -709,27 +962,26 @@ class Coregionalize_cython_test(unittest.TestCase):
|
|||
self.assertTrue(np.allclose(grads_numpy, grads_cython))
|
||||
|
||||
|
||||
|
||||
class KernelTestsProductWithZeroValues(unittest.TestCase):
|
||||
|
||||
def setUp(self):
|
||||
self.X = np.array([[0, 1], [1, 0]])
|
||||
self.k = GPy.kern.Linear(2) * GPy.kern.Bias(2)
|
||||
|
||||
def test_zero_valued_kernel_full(self):
|
||||
self.k.update_gradients_full(1, self.X)
|
||||
self.assertFalse(np.isnan(self.k['linear.variances'].gradient),
|
||||
"Gradient resulted in NaN")
|
||||
self.assertFalse(
|
||||
np.isnan(self.k["linear.variances"].gradient), "Gradient resulted in NaN"
|
||||
)
|
||||
|
||||
def test_zero_valued_kernel_gradients_X(self):
|
||||
target = self.k.gradients_X(1, self.X)
|
||||
self.assertFalse(np.any(np.isnan(target)),
|
||||
"Gradient resulted in NaN")
|
||||
self.assertFalse(np.any(np.isnan(target)), "Gradient resulted in NaN")
|
||||
|
||||
|
||||
class Kernel_Psi_statistics_GradientTests(unittest.TestCase):
|
||||
|
||||
def setUp(self):
|
||||
from GPy.core.parameterization.variational import NormalPosterior
|
||||
|
||||
N, M, Q = 100, 20, 3
|
||||
|
||||
X = np.random.randn(N, Q)
|
||||
|
|
@ -746,9 +998,16 @@ class Kernel_Psi_statistics_GradientTests(unittest.TestCase):
|
|||
|
||||
def test_kernels(self):
|
||||
from GPy.kern import RBF, Linear, MLP, Bias, White
|
||||
|
||||
Q = self.Z.shape[1]
|
||||
kernels = [RBF(Q,ARD=True), Linear(Q,ARD=True),MLP(Q,ARD=True), RBF(Q,ARD=True)+Linear(Q,ARD=True)+Bias(Q)+White(Q)
|
||||
,RBF(Q,ARD=True)+Bias(Q)+White(Q), Linear(Q,ARD=True)+Bias(Q)+White(Q)]
|
||||
kernels = [
|
||||
RBF(Q, ARD=True),
|
||||
Linear(Q, ARD=True),
|
||||
MLP(Q, ARD=True),
|
||||
RBF(Q, ARD=True) + Linear(Q, ARD=True) + Bias(Q) + White(Q),
|
||||
RBF(Q, ARD=True) + Bias(Q) + White(Q),
|
||||
Linear(Q, ARD=True) + Bias(Q) + White(Q),
|
||||
]
|
||||
|
||||
for k in kernels:
|
||||
k.randomize()
|
||||
|
|
@ -760,50 +1019,69 @@ class Kernel_Psi_statistics_GradientTests(unittest.TestCase):
|
|||
self._test_qX(k, psi2n=True)
|
||||
|
||||
def _test_kernel_param(self, kernel, psi2n=False):
|
||||
|
||||
def f(p):
|
||||
kernel.param_array[:] = p
|
||||
psi0 = kernel.psi0(self.Z, self.qX)
|
||||
psi1 = kernel.psi1(self.Z, self.qX)
|
||||
if not psi2n:
|
||||
psi2 = kernel.psi2(self.Z, self.qX)
|
||||
return (self.w1*psi0).sum() + (self.w2*psi1).sum() + (self.w3*psi2).sum()
|
||||
return (
|
||||
(self.w1 * psi0).sum()
|
||||
+ (self.w2 * psi1).sum()
|
||||
+ (self.w3 * psi2).sum()
|
||||
)
|
||||
else:
|
||||
psi2 = kernel.psi2n(self.Z, self.qX)
|
||||
return (self.w1*psi0).sum() + (self.w2*psi1).sum() + (self.w3n*psi2).sum()
|
||||
return (
|
||||
(self.w1 * psi0).sum()
|
||||
+ (self.w2 * psi1).sum()
|
||||
+ (self.w3n * psi2).sum()
|
||||
)
|
||||
|
||||
def df(p):
|
||||
kernel.param_array[:] = p
|
||||
kernel.update_gradients_expectations(self.w1, self.w2, self.w3 if not psi2n else self.w3n, self.Z, self.qX)
|
||||
kernel.update_gradients_expectations(
|
||||
self.w1, self.w2, self.w3 if not psi2n else self.w3n, self.Z, self.qX
|
||||
)
|
||||
return kernel.gradient.copy()
|
||||
|
||||
from GPy.models import GradientChecker
|
||||
|
||||
m = GradientChecker(f, df, kernel.param_array.copy())
|
||||
m.checkgrad(verbose=1)
|
||||
self.assertTrue(m.checkgrad())
|
||||
|
||||
def _test_Z(self, kernel, psi2n=False):
|
||||
|
||||
def f(p):
|
||||
psi0 = kernel.psi0(p, self.qX)
|
||||
psi1 = kernel.psi1(p, self.qX)
|
||||
psi2 = kernel.psi2(p, self.qX)
|
||||
if not psi2n:
|
||||
psi2 = kernel.psi2(p, self.qX)
|
||||
return (self.w1*psi0).sum() + (self.w2*psi1).sum() + (self.w3*psi2).sum()
|
||||
return (
|
||||
(self.w1 * psi0).sum()
|
||||
+ (self.w2 * psi1).sum()
|
||||
+ (self.w3 * psi2).sum()
|
||||
)
|
||||
else:
|
||||
psi2 = kernel.psi2n(p, self.qX)
|
||||
return (self.w1*psi0).sum() + (self.w2*psi1).sum() + (self.w3n*psi2).sum()
|
||||
return (
|
||||
(self.w1 * psi0).sum()
|
||||
+ (self.w2 * psi1).sum()
|
||||
+ (self.w3n * psi2).sum()
|
||||
)
|
||||
|
||||
def df(p):
|
||||
return kernel.gradients_Z_expectations(self.w1, self.w2, self.w3 if not psi2n else self.w3n, p, self.qX)
|
||||
return kernel.gradients_Z_expectations(
|
||||
self.w1, self.w2, self.w3 if not psi2n else self.w3n, p, self.qX
|
||||
)
|
||||
|
||||
from GPy.models import GradientChecker
|
||||
|
||||
m = GradientChecker(f, df, self.Z.copy())
|
||||
self.assertTrue(m.checkgrad())
|
||||
|
||||
def _test_qX(self, kernel, psi2n=False):
|
||||
|
||||
def f(p):
|
||||
self.qX.param_array[:] = p
|
||||
self.qX._trigger_params_changed()
|
||||
|
|
@ -811,22 +1089,34 @@ class Kernel_Psi_statistics_GradientTests(unittest.TestCase):
|
|||
psi1 = kernel.psi1(self.Z, self.qX)
|
||||
if not psi2n:
|
||||
psi2 = kernel.psi2(self.Z, self.qX)
|
||||
return (self.w1*psi0).sum() + (self.w2*psi1).sum() + (self.w3*psi2).sum()
|
||||
return (
|
||||
(self.w1 * psi0).sum()
|
||||
+ (self.w2 * psi1).sum()
|
||||
+ (self.w3 * psi2).sum()
|
||||
)
|
||||
else:
|
||||
psi2 = kernel.psi2n(self.Z, self.qX)
|
||||
return (self.w1*psi0).sum() + (self.w2*psi1).sum() + (self.w3n*psi2).sum()
|
||||
return (
|
||||
(self.w1 * psi0).sum()
|
||||
+ (self.w2 * psi1).sum()
|
||||
+ (self.w3n * psi2).sum()
|
||||
)
|
||||
|
||||
def df(p):
|
||||
self.qX.param_array[:] = p
|
||||
self.qX._trigger_params_changed()
|
||||
grad = kernel.gradients_qX_expectations(self.w1, self.w2, self.w3 if not psi2n else self.w3n, self.Z, self.qX)
|
||||
grad = kernel.gradients_qX_expectations(
|
||||
self.w1, self.w2, self.w3 if not psi2n else self.w3n, self.Z, self.qX
|
||||
)
|
||||
self.qX.set_gradients(grad)
|
||||
return self.qX.gradient.copy()
|
||||
|
||||
from GPy.models import GradientChecker
|
||||
|
||||
m = GradientChecker(f, df, self.qX.param_array.copy())
|
||||
self.assertTrue(m.checkgrad())
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
print("Running unit tests, please be (very) patient...")
|
||||
unittest.main()
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue