From 6048baadae87abd5721b2487a4b52cdc3e008b51 Mon Sep 17 00:00:00 2001
From: Farid Mohammadi <farid.mohammadi@iws.uni-stuttgart.de>
Date: Mon, 12 Sep 2022 09:21:49 +0200
Subject: [PATCH] [surrogate] remove float16 to preserve the correct
computation.
---
.../surrogate_models/eval_rec_rule.py | 2 +-
.../surrogate_models/surrogate_models.py | 19 ++++++++++---------
2 files changed, 11 insertions(+), 10 deletions(-)
diff --git a/src/bayesvalidrox/surrogate_models/eval_rec_rule.py b/src/bayesvalidrox/surrogate_models/eval_rec_rule.py
index 730a3818d..b583c7eb2 100644
--- a/src/bayesvalidrox/surrogate_models/eval_rec_rule.py
+++ b/src/bayesvalidrox/surrogate_models/eval_rec_rule.py
@@ -183,7 +183,7 @@ def eval_univ_basis(x, max_deg, poly_types, apoly_coeffs=None):
"""
# Initilize the output array
n_samples, n_params = x.shape
- univ_vals = np.zeros((n_samples, n_params, max_deg+1), dtype=np.float16)
+ univ_vals = np.zeros((n_samples, n_params, max_deg+1))
for i in range(n_params):
diff --git a/src/bayesvalidrox/surrogate_models/surrogate_models.py b/src/bayesvalidrox/surrogate_models/surrogate_models.py
index e82f7b3f9..341ecf566 100644
--- a/src/bayesvalidrox/surrogate_models/surrogate_models.py
+++ b/src/bayesvalidrox/surrogate_models/surrogate_models.py
@@ -197,7 +197,7 @@ class MetaModel():
# Define an array containing the degrees
n_samples, ndim = CollocationPoints.shape
self.deg_array = self.__select_degree(ndim, n_samples)
- print("deg_array:", self.deg_array)
+
# Generate all basis indices
self.allBasisIndices = self.auto_vivification()
for deg in self.deg_array:
@@ -305,7 +305,7 @@ class MetaModel():
else:
self.univ_p_val = univ_p_val[b_indices]
if parallel and (not fast_bootstrap or b_i == 0):
- out = Parallel(n_jobs=-1, backend='loky')(
+ out = Parallel(n_jobs=-1, backend='multiprocessing')(
delayed(self.adaptive_regression)(X_train_b,
target[:, idx],
idx)
@@ -601,7 +601,7 @@ class MetaModel():
)
# Preallocate the Psi matrix for performance
- psi = np.ones((n_samples, n_terms), dtype=np.float16)
+ psi = np.ones((n_samples, n_terms))
# Assemble the Psi matrix
for m in range(basis_indices.shape[1]):
aa = np.where(basis_indices[:, m] > 0)[0]
@@ -748,7 +748,7 @@ class MetaModel():
"""
- NrSamples, n_params = ED_X.shape
+ n_samples, n_params = ED_X.shape
# Initialization
qAllCoeffs, AllCoeffs = {}, {}
qAllIndices_Sparse, AllIndices_Sparse = {}, {}
@@ -967,8 +967,8 @@ class MetaModel():
# Build the projection matrix
PsiTPsi = np.dot(psi_sparse.T, psi_sparse)
- if np.linalg.cond(PsiTPsi) > 1e-12 and \
- np.linalg.cond(PsiTPsi) < 1/sys.float_info.epsilon:
+ if np.linalg.cond(PsiTPsi) > 1e-12: #and \
+ # np.linalg.cond(PsiTPsi) < 1/sys.float_info.epsilon:
# faster
M = sp.linalg.solve(PsiTPsi,
sp.sparse.eye(PsiTPsi.shape[0]).toarray())
@@ -1199,6 +1199,7 @@ class MetaModel():
samples,
method='user'
)
+ # print(samples)
# Compute univariate bases for the given samples
if self.meta_model_type.lower() != 'gpe':
@@ -1223,8 +1224,8 @@ class MetaModel():
std_pred = {}
for output, values in model_dict.items():
- mean = np.empty((len(samples), len(values)), dtype=np.float16)
- std = np.empty((len(samples), len(values)), dtype=np.float16)
+ mean = np.empty((len(samples), len(values)))
+ std = np.empty((len(samples), len(values)))
idx = 0
for in_key, InIdxValues in values.items():
@@ -1235,7 +1236,7 @@ class MetaModel():
y_mean, y_std = gp.predict(X_T, return_std=True)
else:
- # Perdiction with PCE or pcekriging
+ # Perdiction with PCE
# Assemble Psi matrix
basis = self.basis_dict[f'b_{b_i+1}'][output][in_key]
psi = self.create_psi(basis, univ_p_val)
--
GitLab