diff --git a/src/bayesvalidrox/surrogate_models/meta_model_engine.py b/src/bayesvalidrox/surrogate_models/meta_model_engine.py
index 3985da8a3d4ff124428c9f8710b4d029ca03716c..2df2dee5390ae4e6dc7eb88343c2469dbd88aad6 100644
--- a/src/bayesvalidrox/surrogate_models/meta_model_engine.py
+++ b/src/bayesvalidrox/surrogate_models/meta_model_engine.py
@@ -19,7 +19,6 @@ import os
import gc
import seaborn as sns
from joblib import Parallel, delayed
-import resource
from .exploration import Exploration
@@ -230,7 +229,6 @@ class MetaModelEngine():
print(f'\n>>>> Iteration number {itr_no} <<<<')
# Save the metamodel prediction before updating
- m_1 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
prevMetaModel_dict[itr_no] = deepcopy(self.MetaModel)
if itr_no > 1:
pc_model = prevMetaModel_dict[itr_no-1]
@@ -243,7 +241,6 @@ class MetaModelEngine():
Xnew, updatedPrior = self.opt_SeqDesign(TotalSigma2,
n_canddidate,
util_f)
- m_2 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
S = np.min(distance.cdist(Xinit, Xnew, 'euclidean'))
self.MetaModel.seqMinDist.append(S)
print(f"\nmin Dist from OldExpDesign: {S:2f}")
@@ -298,7 +295,6 @@ class MetaModelEngine():
# Train the surrogate model for new ExpDesign
self.MetaModel.train_norm_design(parallel=False)
- m_3 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
# -------- Evaluate the retrained surrogate model -------
# Extract Modified LOO from Output
@@ -340,7 +336,6 @@ class MetaModelEngine():
if len(self.MetaModel.valid_model_runs) != 0:
SeqValidError = np.vstack(
(SeqValidError, ValidError))
- m_4 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
# -------- Caclulation of BME as accuracy metric -------
# Check if data is provided
if len(obs_data) != 0:
@@ -361,7 +356,6 @@ class MetaModelEngine():
self.__posteriorPlot(Posterior, parNames,
f'SeqPosterior_{postcnt}')
postcnt += 1
- m_5 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
# Check the convergence of the Mean&Std
if mc_ref and pce:
@@ -386,17 +380,9 @@ class MetaModelEngine():
for LOO in ModifiedLOO):
break
- print(f"Memory itr {itr_no}: I: {m_2-m_1:.2f} MB")
- print(f"Memory itr {itr_no}: II: {m_3-m_2:.2f} MB")
- print(f"Memory itr {itr_no}: III: {m_4-m_3:.2f} MB")
- print(f"Memory itr {itr_no}: IV: {m_5-m_4:.2f} MB")
- m_6 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
- print(f"Memory itr {itr_no}: total: {m_6:.2f} MB")
-
# Clean up
if len(obs_data) != 0:
del out
- gc.collect()
print()
print('-'*50)
print()
@@ -610,7 +596,6 @@ class MetaModelEngine():
del likelihoods
del Y_MC
del std_MC
- gc.collect(generation=2)
return -1 * U_J_d # -1 is for minimization instead of maximization
@@ -909,7 +894,8 @@ class MetaModelEngine():
logBME = np.log(np.nanmean(likelihoods))
# Posterior-based expectation of likelihoods
- postLikelihoods = likelihoods[accepted] / np.nansum(likelihoods[accepted])
+ postLikelihoods = likelihoods[accepted]
+ postLikelihoods /= np.nansum(likelihoods[accepted])
postExpLikelihoods = np.mean(np.log(postLikelihoods))
# Posterior-based expectation of prior densities
@@ -938,7 +924,6 @@ class MetaModelEngine():
del likelihoods
del Y_MC
del std_MC
- gc.collect(generation=2)
return -1 * U_J_d # -1 is for minimization instead of maximization
@@ -1907,10 +1892,7 @@ class MetaModelEngine():
)
# Monte Carlo simulation for the candidate design
- m_1 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
Y_MC, std_MC = MetaModel.eval_metamodel(samples=X_MC)
- m_2 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
- print(f"\nMemory eval_metamodel in BME: {m_2-m_1:.2f} MB")
# Likelihood computation (Comparison of data and
# simulation results via PCE with candidate design)
@@ -2039,10 +2021,7 @@ class MetaModelEngine():
valid_model_runs = MetaModel.valid_model_runs
# Run the PCE model with the generated samples
- m_1 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
valid_PCE_runs, _ = MetaModel.eval_metamodel(samples=valid_samples)
- m_2 = resource.getrusage(resource.RUSAGE_SELF).ru_maxrss/1024
- print(f"\nMemory eval_metamodel: {m_2-m_1:.2f} MB")
rms_error = {}
valid_error = {}
diff --git a/src/bayesvalidrox/surrogate_models/surrogate_models.py b/src/bayesvalidrox/surrogate_models/surrogate_models.py
index 009ba7105c0f3113148b244e7263db8d5538fb73..c51ea8c6679aaeec685da069154e8f460c7c4450 100644
--- a/src/bayesvalidrox/surrogate_models/surrogate_models.py
+++ b/src/bayesvalidrox/surrogate_models/surrogate_models.py
@@ -14,7 +14,6 @@ import sklearn.linear_model as lm
from sklearn.gaussian_process import GaussianProcessRegressor
import sklearn.gaussian_process.kernels as kernels
import os
-import sys
from joblib import Parallel, delayed
import copy
@@ -25,7 +24,6 @@ from .reg_fast_ard import RegressionFastARD
from .reg_fast_laplace import RegressionFastLaplace
from .orthogonal_matching_pursuit import OrthogonalMatchingPursuit
from .bayes_linear import VBLinearRegression, EBLinearRegression
-from .sequential_design import SeqDesign
warnings.filterwarnings("ignore")
# Load the mplstyle
plt.style.use(os.path.join(os.path.split(__file__)[0],
@@ -65,7 +63,7 @@ class MetaModel():
is `'fast'`. It means that in each iteration except the first one, only
the coefficent are recalculated with the ordinary least square method.
n_bootstrap_itrs : int
- Number of iterations for the bootstrap sampling. The default is `100`.
+ Number of iterations for the bootstrap sampling. The default is `1`.
pce_deg : int or list of int
Polynomial degree(s). If a list is given, an adaptive algorithm is used
to find the best degree with the lowest Leave-One-Out cross-validation
@@ -102,7 +100,7 @@ class MetaModel():
def __init__(self, input_obj, model_obj, meta_model_type='PCE',
pce_reg_method='OLS', bootstrap_method='fast',
- n_bootstrap_itrs=100, pce_deg=1, pce_q_norm=1.0,
+ n_bootstrap_itrs=1, pce_deg=1, pce_q_norm=1.0,
dim_red_method='no', verbose=False):
self.input_obj = input_obj
@@ -137,8 +135,9 @@ class MetaModel():
self.pce_deg = np.array(self.pce_deg)
if self.ExpDesign.method == 'sequential':
- self.train_norm_design(Model)
- metamodel = SeqDesign().train_seq_design(self)
+ raise Exception(
+ "Please use MetaModelEngine class for the sequential design!"
+ )
elif self.ExpDesign.method == 'normal':
self.ExpDesignFlag = 'normal'