GPflow vs. GPy

gpflow_vs_gpy.py

import GPy
import GPflow
import numpy as np

rng = np.random.RandomState(0)
X = rng.randn(100,2)
Y = rng.randn(100,1)
Xtest = rng.randn(10,2)

# GPy
m_gpy = GPy.models.GPRegression(X, Y)

# GPflow
m_gpflow = GPflow.gpr.GPR(X, Y, GPflow.kernels.RBF(1))

print m_gpy
print m_gpflow
print m_gpy.predict(Xtest)
print m_gpflow.predict_y(Xtest)

# Expected output
"""
Name : GP regression
Objective : 144.20800383
Number of Parameters : 3
Number of Optimization Parameters : 3
Updates : True
Parameters:
  GP_regression.           |  value  |  constraints  |  priors
  rbf.variance             |    1.0  |      +ve      |        
  rbf.lengthscale          |    1.0  |      +ve      |        
  Gaussian_noise.variance  |    1.0  |      +ve      |        
model.likelihood.variance transform:+ve prior:None
[ 1.]

model.kern.lengthscales transform:+ve prior:None
[ 1.]
model.kern.variance transform:+ve prior:None
[ 1.]
(array([[-0.15199914],
       [-0.26766372],
       [-0.38295276],
       [ 0.01115626],
       [-0.31434469],
       [-0.46805395],
       [ 0.03996755],
       [-0.14441124],
       [-0.30056849],
       [-0.27920849]]), array([[ 1.17194185],
       [ 1.07740217],
       [ 1.0767739 ],
       [ 1.2536379 ],
       [ 1.06700106],
       [ 1.090615  ],
       [ 1.12538894],
       [ 1.06045394],
       [ 1.06845334],
       [ 1.06448736]]))
[array([[ 0.03072267],
       [-0.32096233],
       [-0.19599798],
       [ 0.03379334],
       [-0.18174901],
       [-0.31749606],
       [ 0.00391937],
       [-0.23998849],
       [-0.33504122],
       [-0.33190934]]), array([[ 1.04870813],
       [ 1.02989084],
       [ 1.03347398],
       [ 1.0515597 ],
       [ 1.03393089],
       [ 1.02983322],
       [ 1.03524757],
       [ 1.02843413],
       [ 1.03012201],
       [ 1.03007054]])]
"""