cpp_gp_loglikelihood = cppGaussianProcessLogLikelihoodMCMC( historical_data=init_data, derivatives=derivatives, prior=prior, chain_length=1000, burnin_steps=2000, n_hypers=2 ** 4, noisy=noisy, ) cpp_gp_loglikelihood.train() py_sgd_params_ps = pyGradientDescentParameters( max_num_steps=1000, max_num_restarts=3, num_steps_averaged=15, gamma=0.7, pre_mult=1.0, max_relative_change=0.02, tolerance=1.0e-10, ) cpp_sgd_params_ps = cppGradientDescentParameters( num_multistarts=1, max_num_steps=6, max_num_restarts=1, num_steps_averaged=3, gamma=0.0, pre_mult=1.0, max_relative_change=0.1, tolerance=1.0e-10, )
# noisy = False means the underlying function being optimized is noise-free cpp_gp_loglikelihood = cppGaussianProcessLogLikelihoodMCMC( historical_data=init_data, derivatives=derivatives, prior=prior, chain_length=1000, burnin_steps=2000, n_hypers=2**4, noisy=True) cpp_gp_loglikelihood.train() py_sgd_params_ps = pyGradientDescentParameters(max_num_steps=100, max_num_restarts=3, num_steps_averaged=15, gamma=0.7, pre_mult=0.01, max_relative_change=0.02, tolerance=1.0e-8) py_sgd_params_acquisition = pyGradientDescentParameters( max_num_steps=50, max_num_restarts=1, num_steps_averaged=0, gamma=0.7, pre_mult=1.0, max_relative_change=0.1, tolerance=1.0e-8) cpp_sgd_params_ps = cppGradientDescentParameters(num_multistarts=1, max_num_steps=6,