apunkt/GPy
1
0
Fork 0
mirror of https://github.com/SheffieldML/GPy.git synced 2026-06-05 14:55:15 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

objective function seperate from calls for optimizer

Browse source
This commit is contained in:
Max Zwiessele 2014-03-24 13:32:56 +00:00
parent 6b8e418597
commit 3db095338d
1 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

10
GPy/core/model.py
View file

@ -176,7 +176,7 @@ class Model(Parameterized):
(including the MAP prior), so we return it here. If your model is not
probabilistic, just return your gradient here!
"""
return self._log_likelihood_gradients() + self._log_prior_gradients()
return -(self._log_likelihood_gradients() + self._log_prior_gradients())
def _grads(self, x):
"""
@ -191,13 +191,13 @@ class Model(Parameterized):
"""
try:
self._set_params_transformed(x)
obj_grads = -self._transform_gradients(self.objective_function_gradients())
obj_grads = self._transform_gradients(self.objective_function_gradients())
self._fail_count = 0
except (LinAlgError, ZeroDivisionError, ValueError):
if self._fail_count >= self._allowed_failures:
raise
self._fail_count += 1
obj_grads = np.clip(-self._transform_gradients(self.objective_function_gradients()), -1e100, 1e100)
obj_grads = np.clip(self._transform_gradients(self.objective_function_gradients()), -1e100, 1e100)
return obj_grads
def _objective(self, x):
@ -226,14 +226,14 @@ class Model(Parameterized):
def _objective_grads(self, x):
try:
self._set_params_transformed(x)
obj_f, obj_grads = self.objective_function(), self.objective_function_gradients()
obj_f, obj_grads = self.objective_function(), self._transform_gradients(self.objective_function_gradients())
self._fail_count = 0
except (LinAlgError, ZeroDivisionError, ValueError):
if self._fail_count >= self._allowed_failures:
raise
self._fail_count += 1
obj_f = np.inf
obj_grads = np.clip(-self._transform_gradients(self.objective_function_gradients()), -1e100, 1e100)
obj_grads = np.clip(self._transform_gradients(self.objective_function_gradients()), -1e100, 1e100)
return obj_f, obj_grads
def optimize(self, optimizer=None, start=None, **kwargs):