apunkt/GPy
1
0
Fork 0
mirror of https://github.com/SheffieldML/GPy.git synced 2026-04-26 21:36:23 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'devel' of github.com:SheffieldML/GPy into devel

Browse source
This commit is contained in:
Ricardo 2014-08-06 14:45:20 +01:00
commit 140fdc8211
11 changed files with 63 additions and 56 deletions
Download patch file Download diff file Expand all files Collapse all files

8
GPy/core/model.py
View file

@ -61,7 +61,7 @@ class Model(Parameterized):
on the current machine.
"""
initial_parameters = self.optimizer_array
initial_parameters = self.optimizer_array.copy()
if parallel:
try:
@ -97,9 +97,9 @@ class Model(Parameterized):
if len(self.optimization_runs):
i = np.argmin([o.f_opt for o in self.optimization_runs])
self._set_params_transformed(self.optimization_runs[i].x_opt)
self.optimizer_array = self.optimization_runs[i].x_opt
else:
self._set_params_transformed(initial_parameters)
self.optimizer_array = initial_parameters
def ensure_default_constraints(self, warning=True):
"""
@ -349,7 +349,7 @@ class Model(Parameterized):
numerical_gradient = (f1 - f2) / (2 * step)
if np.all(gradient[xind] == 0): ratio = (f1 - f2) == gradient[xind]
else: ratio = (f1 - f2) / (2 * step * gradient[xind])
difference = np.abs((f1 - f2) / 2 / step - gradient[xind])
difference = np.abs(numerical_gradient - gradient[xind])
if (np.abs(1. - ratio) < tolerance) or np.abs(difference) < tolerance:
formatted_name = "\033[92m {0} \033[0m".format(names[nind])

28
GPy/core/parameterization/parameter_core.py
View file

@ -699,36 +699,10 @@ class OptimizationHandlable(Indexable):
def _get_params_transformed(self):
raise DeprecationWarning, "_get|set_params{_optimizer_copy_transformed} is deprecated, use self.optimizer array insetad!"
# # transformed parameters (apply un-transformation rules)
# p = self.param_array.copy()
# [np.put(p, ind, c.finv(p[ind])) for c, ind in self.constraints.iteritems() if c != __fixed__]
# if self.has_parent() and self.constraints[__fixed__].size != 0:
# fixes = np.ones(self.size).astype(bool)
# fixes[self.constraints[__fixed__]] = FIXED
# return p[fixes]
# elif self._has_fixes():
# return p[self._fixes_]
# return p
#
def _set_params_transformed(self, p):
raise DeprecationWarning, "_get|set_params{_optimizer_copy_transformed} is deprecated, use self.optimizer array insetad!"
# """
# Set parameters p, but make sure they get transformed before setting.
# This means, the optimizer sees p, whereas the model sees transformed(p),
# such that, the parameters the model sees are in the right domain.
# """
# if not(p is self.param_array):
# if self.has_parent() and self.constraints[__fixed__].size != 0:
# fixes = np.ones(self.size).astype(bool)
# fixes[self.constraints[__fixed__]] = FIXED
# self.param_array.flat[fixes] = p
# elif self._has_fixes(): self.param_array.flat[self._fixes_] = p
# else: self.param_array.flat = p
# [np.put(self.param_array, ind, c.f(self.param_array.flat[ind]))
# for c, ind in self.constraints.iteritems() if c != __fixed__]
# self._trigger_params_changed()
def _trigger_params_changed(self, trigger_parent=True):
"""
First tell all children to update,
@ -736,7 +710,7 @@ class OptimizationHandlable(Indexable):
If trigger_parent is True, we will tell the parent, otherwise not.
"""
[p._trigger_params_changed(trigger_parent=False) for p in self.parameters]
[p._trigger_params_changed(trigger_parent=False) for p in self.parameters if not p.is_fixed]
self.notify_observers(None, None if trigger_parent else -np.inf)
def _size_transformed(self):

8
GPy/core/sparse_gp.py
View file

@ -60,10 +60,14 @@ class SparseGP(GP):
self.likelihood.update_gradients(self.grad_dict['dL_dthetaL'])
if isinstance(self.X, VariationalPosterior):
#gradients wrt kernel
dL_dKmm = self.grad_dict.pop('dL_dKmm')
dL_dKmm = self.grad_dict['dL_dKmm']
self.kern.update_gradients_full(dL_dKmm, self.Z, None)
target = self.kern.gradient.copy()
self.kern.update_gradients_expectations(variational_posterior=self.X, Z=self.Z, dL_dpsi0=self.grad_dict['dL_dpsi0'], dL_dpsi1=self.grad_dict['dL_dpsi1'], dL_dpsi2=self.grad_dict['dL_dpsi2'])
self.kern.update_gradients_expectations(variational_posterior=self.X,
Z=self.Z,
dL_dpsi0=self.grad_dict['dL_dpsi0'],
dL_dpsi1=self.grad_dict['dL_dpsi1'],
dL_dpsi2=self.grad_dict['dL_dpsi2'])
self.kern.gradient += target
#gradients wrt Z

10
GPy/inference/latent_function_inference/var_dtc.py
View file

@ -194,7 +194,7 @@ class VarDTC(LatentFunctionInference):
return post, log_marginal, grad_dict
class VarDTCMissingData(LatentFunctionInference):
const_jitter = 1e-6
const_jitter = 1e-10
def __init__(self, limit=1, inan=None):
from ...util.caching import Cacher
self._Y = Cacher(self._subarray_computations, limit)
@ -289,13 +289,6 @@ class VarDTCMissingData(LatentFunctionInference):
Lm = jitchol(Kmm)
if uncertain_inputs: LmInv = dtrtri(Lm)
#VVT_factor_all = np.empty(Y.shape)
#full_VVT_factor = VVT_factor_all.shape[1] == Y.shape[1]
#if not full_VVT_factor:
# psi1V = np.dot(Y.T*beta_all, psi1_all).T
#logger.info('computing dimension-wise likelihood and derivatives')
#size = len(Ys)
size = Y.shape[1]
next_ten = 0
for i, [y, v, trYYT] in enumerate(itertools.izip(Ys, self._inan.T, traces)):
@ -348,7 +341,6 @@ class VarDTCMissingData(LatentFunctionInference):
VVT_factor, Cpsi1Vf, DBi_plus_BiPBi,
psi1, het_noise, uncertain_inputs)
#import ipdb;ipdb.set_trace()
dL_dpsi0_all[v] += dL_dpsi0
dL_dpsi1_all[v, :] += dL_dpsi1
if uncertain_inputs:

4
GPy/kern/_src/add.py
View file

@ -10,7 +10,7 @@ class Add(CombinationKernel):
"""
Add given list of kernels together.
propagates gradients through.
This kernel will take over the active dims of it's subkernels passed in.
"""
def __init__(self, subkerns, name='add'):
@ -40,7 +40,7 @@ class Add(CombinationKernel):
return reduce(np.add, (p.Kdiag(X) for p in which_parts))
def update_gradients_full(self, dL_dK, X, X2=None):
[p.update_gradients_full(dL_dK, X, X2) for p in self.parts]
[p.update_gradients_full(dL_dK, X, X2) for p in self.parts if not p.is_fixed]
def update_gradients_diag(self, dL_dK, X):
[p.update_gradients_diag(dL_dK, X) for p in self.parts]

3
GPy/plotting/matplot_dep/dim_reduction_plots.py
View file

@ -30,7 +30,7 @@ def most_significant_input_dimensions(model, which_indices):
def plot_latent(model, labels=None, which_indices=None,
resolution=50, ax=None, marker='o', s=40,
fignum=None, plot_inducing=False, legend=True,
plot_limits=None,
plot_limits=None,
aspect='auto', updates=False, predict_kwargs={}, imshow_kwargs={}):
"""
:param labels: a np.array of size model.num_data containing labels for the points (can be number, strings, etc)
@ -84,6 +84,7 @@ def plot_latent(model, labels=None, which_indices=None,
cmap=pb.cm.binary, **imshow_kwargs)
# make sure labels are in order of input:
labels = np.asarray(labels)
ulabels = []
for lab in labels:
if not lab in ulabels:

4
GPy/util/config.py
View file

@ -8,9 +8,9 @@ config = ConfigParser.ConfigParser()
# This is the default configuration file that always needs to be present.
default_file = os.path.abspath(os.path.join(os.path.dirname( __file__ ), '..', 'defaults.cfg'))
# These files are optional
# These files are optional
# This specifies configurations that are typically specific to the machine (it is found alongside the GPy installation).
local_file = os.path.abspath(os.path.join(os.path.dirname( __file__ ), '..', 'machine.cfg'))
local_file = os.path.abspath(os.path.join(os.path.dirname( __file__ ), '..', 'installation.cfg'))
# This specifies configurations specific to the user (it is found in the user home directory)
home = os.getenv('HOME') or os.getenv('USERPROFILE')

16
GPy/util/datasets.py
View file

@ -659,6 +659,22 @@ def ripley_synth(data_set='ripley_prnn_data'):
ytest = test[:, 2:3]
return data_details_return({'X': X, 'Y': y, 'Xtest': Xtest, 'Ytest': ytest, 'info': 'Synthetic data generated by Ripley for a two class classification problem.'}, data_set)
def global_average_temperature(data_set='global_temperature', num_train=1000, refresh_data=False):
path = os.path.join(data_path, data_set)
if data_available(data_set) and not refresh_data:
print 'Using cached version of the data set, to use latest version set refresh_data to True'
else:
download_data(data_set)
data = np.loadtxt(os.path.join(data_path, data_set, 'GLBTS.long.data'))
print 'Most recent data observation from month ', data[-1, 1], ' in year ', data[-1, 0]
allX = data[data[:, 3]!=-99.99, 2:3]
allY = data[data[:, 3]!=-99.99, 3:4]
X = allX[:num_train, 0:1]
Xtest = allX[num_train:, 0:1]
Y = allY[:num_train, 0:1]
Ytest = allY[num_train:, 0:1]
return data_details_return({'X': X, 'Y': Y, 'Xtest': Xtest, 'Ytest': Ytest, 'info': "Mauna Loa data with " + str(num_train) + " values used as training points."}, data_set)
def mauna_loa(data_set='mauna_loa', num_train=545, refresh_data=False):
path = os.path.join(data_path, data_set)
if data_available(data_set) and not refresh_data:

28
GPy/util/linalg.py
View file

@ -15,6 +15,7 @@ import scipy
import warnings
import os
from config import *
import logging
_scipyversion = np.float64((scipy.__version__).split('.')[:2])
_fix_dpotri_scipy_bug = True
@ -93,14 +94,20 @@ def jitchol(A, maxtries=5):
raise linalg.LinAlgError, "not pd: non-positive diagonal elements"
jitter = diagA.mean() * 1e-6
while maxtries > 0 and np.isfinite(jitter):
print 'Warning: adding jitter of {:.10e}'.format(jitter)
try:
return linalg.cholesky(A + np.eye(A.shape[0]).T * jitter, lower=True)
L = linalg.cholesky(A + np.eye(A.shape[0]) * jitter, lower=True)
except:
jitter *= 10
finally:
maxtries -= 1
raise linalg.LinAlgError, "not positive definite, even with jitter."
import traceback
try: raise
except:
logging.warning('\n'.join(['Added jitter of {:.10e}'.format(jitter),
' in '+traceback.format_list(traceback.extract_stack(limit=2)[-2:-1])[0][2:]]))
import ipdb;ipdb.set_trace()
return L
@ -110,7 +117,7 @@ def jitchol(A, maxtries=5):
# """
# Wrapper for lapack dtrtri function
# Inverse of L
#
#
# :param L: Triangular Matrix L
# :param lower: is matrix lower (true) or upper (false)
# :returns: Li, info
@ -122,10 +129,17 @@ def dtrtrs(A, B, lower=1, trans=0, unitdiag=0):
"""
Wrapper for lapack dtrtrs function
DTRTRS solves a triangular system of the form
A * X = B or A**T * X = B,
where A is a triangular matrix of order N, and B is an N-by-NRHS
matrix. A check is made to verify that A is nonsingular.
:param A: Matrix A(triangular)
:param B: Matrix B
:param lower: is matrix lower (true) or upper (false)
:returns:
:returns: Solution to A * X = B or A**T * X = B
"""
A = np.asfortranarray(A)
@ -146,11 +160,11 @@ def dpotrs(A, B, lower=1):
def dpotri(A, lower=1):
"""
Wrapper for lapack dpotri function
DPOTRI - compute the inverse of a real symmetric positive
definite matrix A using the Cholesky factorization A =
U**T*U or A = L*L**T computed by DPOTRF
:param A: Matrix A
:param lower: is matrix lower (true) or upper (false)
:returns: A inverse
@ -159,7 +173,7 @@ def dpotri(A, lower=1):
if _fix_dpotri_scipy_bug:
assert lower==1, "scipy linalg behaviour is very weird. please use lower, fortran ordered arrays"
lower = 0
A = force_F_ordered(A)
R, info = lapack.dpotri(A, lower=lower) #needs to be zero here, seems to be a scipy bug

4
MANIFEST.in
View file

@ -2,3 +2,7 @@ include *.txt
recursive-include doc *.txt
include *.md
recursive-include doc *.md
include *.cfg
recursive-include doc *.cfg
include *.json
recursive-include doc *.json

6
setup.py
View file

@ -1,12 +1,13 @@
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import os
#import os
from setuptools import setup
# Version number
version = '0.4.6'
from pkg_resources import Requirement, resource_string
def read(fname):
return open(os.path.join(os.path.dirname(__file__), fname)).read()
@ -20,7 +21,7 @@ setup(name = 'GPy',
url = "http://sheffieldml.github.com/GPy/",
packages = ["GPy.models", "GPy.inference.optimization", "GPy.inference", "GPy.inference.latent_function_inference", "GPy.likelihoods", "GPy.mappings", "GPy.examples", "GPy.core.parameterization", "GPy.core", "GPy.testing", "GPy", "GPy.util", "GPy.kern", "GPy.kern._src.psi_comp", "GPy.kern._src", "GPy.plotting.matplot_dep.latent_space_visualizations.controllers", "GPy.plotting.matplot_dep.latent_space_visualizations", "GPy.plotting.matplot_dep", "GPy.plotting"],
package_dir={'GPy': 'GPy'},
package_data = {'GPy': ['GPy/examples']},
package_data = {'GPy': ['defaults.cfg', 'installation.cfg', 'util/data_resources.json', 'util/football_teams.json']},
py_modules = ['GPy.__init__'],
long_description=read('README.md'),
install_requires=['numpy>=1.6', 'scipy>=0.9','matplotlib>=1.1', 'nose'],
@ -29,4 +30,5 @@ setup(name = 'GPy',
},
classifiers=[
"License :: OSI Approved :: BSD License"],
zip_safe = False
)