def test_fit_3():
"""
This function tests the thrid way of fitting the descriptor: the data is passed directly to the fit function.
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load(test_dir + "/data/local_slatm_ch4cn_light.npz")
descriptor = data["arr_0"]
classes = data["arr_1"]
energies = data["arr_2"]
estimator = ARMP()
estimator.fit(x=descriptor, y=energies, classes=classes)
def test_predict_3():
test_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load(test_dir + "/data/local_slatm_ch4cn_light.npz")
descriptor = data["arr_0"]
classes = data["arr_1"]
energies = data["arr_2"]
estimator = ARMP()
estimator.fit(x=descriptor, y=energies, classes=classes)
energies_pred = estimator.predict(x=descriptor, classes=classes)
assert energies.shape == energies_pred.shape
def test_set_properties():
"""
This test checks that the set_properties function sets the correct properties.
:return:
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
energies = np.loadtxt(test_dir + '/CN_isobutane/prop_kjmol_training.txt',
usecols=[1])
estimator = ARMP(representation='slatm')
assert estimator.properties == None
estimator.set_properties(energies)
assert np.all(estimator.properties == energies)
def test_fit_1():
"""
This function tests the first way of fitting the descriptor: the data is passed by first creating compounds and then
the descriptors are created from the compounds.
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
filenames = glob.glob(test_dir + "/CN_isobutane/*.xyz")
energies = np.loadtxt(test_dir + '/CN_isobutane/prop_kjmol_training.txt',
usecols=[1])
filenames.sort()
estimator = ARMP(representation="acsf")
estimator.generate_compounds(filenames[:50])
estimator.set_properties(energies[:50])
estimator.generate_representation()
idx = np.arange(0, 50)
estimator.fit(idx)
def test_fit_2():
"""
This function tests the second way of fitting the descriptor: the data is passed by storing the compounds in the
class.
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load(test_dir + "/data/local_slatm_ch4cn_light.npz")
descriptor = data["arr_0"]
classes = data["arr_1"]
energies = data["arr_2"]
estimator = ARMP()
estimator.set_representations(representations=descriptor)
estimator.set_classes(classes=classes)
estimator.set_properties(energies)
idx = np.arange(0, 100)
estimator.fit(idx)
def test_set_descriptor():
"""
This test checks that the set_descriptor function works as expected.
:return:
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
data_incorrect = np.load(test_dir + "/data/CN_isopent_light_UCM.npz")
data_correct = np.load(test_dir + "/data/local_slatm_ch4cn_light.npz")
descriptor_correct = data_correct["arr_0"]
descriptor_incorrect = data_incorrect["arr_0"]
estimator = ARMP()
assert estimator.representation == None
estimator.set_representations(representations=descriptor_correct)
assert np.all(estimator.representation == descriptor_correct)
# Pass a descriptor with the wrong shape
try:
estimator.set_representations(representations=descriptor_incorrect)
raise Exception
except InputError:
pass
def test_set_representation():
"""
This function tests the function _set_representation.
"""
try:
ARMP(representation='slatm',
representation_params={'slatm_sigma12': 0.05})
raise Exception
except InputError:
pass
try:
ARMP(representation='coulomb_matrix')
raise Exception
except InputError:
pass
try:
ARMP(representation='slatm',
representation_params={'slatm_alchemy': 0.05})
raise Exception
except InputError:
pass
parameters = {
'slatm_sigma1': 0.07,
'slatm_sigma2': 0.04,
'slatm_dgrid1': 0.02,
'slatm_dgrid2': 0.06,
'slatm_rcut': 5.0,
'slatm_rpower': 7,
'slatm_alchemy': True
}
estimator = ARMP(representation='slatm', representation_params=parameters)
assert estimator.representation_name == 'slatm'
assert estimator.slatm_parameters == parameters
zs = np.array(data["zs"][-n_samples:], dtype=np.int32)
# Creating the estimator
acsf_params = {
"nRs2": 5,
"nRs3": 5,
"nTs": 5,
"rcut": 5,
"acut": 5,
"zeta": 220.127,
"eta": 30.8065
}
estimator = ARMP(iterations=200,
representation_name='acsf',
representation_params=acsf_params,
tensorboard=True,
store_frequency=10,
l1_reg=0.0001,
l2_reg=0.005,
learning_rate=0.0005)
estimator.set_properties(ene)
estimator.generate_representation(xyz, zs)
saved_dir = "saved_model"
estimator.load_nn(saved_dir)
idx = list(range(n_samples))
estimator.fit(idx)
def test_score_3():
"""
This function tests that all the scoring functions work.
"""
test_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load(test_dir + "/data/local_slatm_ch4cn_light.npz")
descriptor = data["arr_0"]
classes = data["arr_1"]
energies = data["arr_2"]
estimator_1 = ARMP(scoring_function='mae')
estimator_1.fit(x=descriptor, y=energies, classes=classes)
estimator_1.score(x=descriptor, y=energies, classes=classes)
estimator_2 = ARMP(scoring_function='r2')
estimator_2.fit(x=descriptor, y=energies, classes=classes)
estimator_2.score(x=descriptor, y=energies, classes=classes)
estimator_3 = ARMP(scoring_function='rmse')
estimator_3.fit(x=descriptor, y=energies, classes=classes)
estimator_3.score(x=descriptor, y=energies, classes=classes)
import shutil
from qml.aglaia.aglaia import ARMP
import tensorflow as tf
xyz = np.array([[[0, 1, 0], [0, 1, 1], [1, 0, 1]],
[[1, 2, 2], [3, 1, 2], [1, 3, 4]],
[[4, 1, 2], [0.5, 5, 6], [-1, 2, 3]]])
zs = np.array([[1, 2, 3],
[1, 2, 3],
[1, 2, 3]])
ene_true = np.array([0.5, 0.9, 1.0])
estimator = ARMP(iterations=10, l1_reg=0.0001, l2_reg=0.005, learning_rate=0.0005, representation='acsf',
representation_params={"radial_rs": np.arange(0, 10, 5), "angular_rs": np.arange(0, 10, 5),
"theta_s": np.arange(0, 3.14, 3)},
tensorboard=True, store_frequency=10
)
estimator.set_properties(ene_true)
estimator.generate_representation(xyz, zs)
idx = list(range(xyz.shape[0]))
estimator.fit(idx)
estimator.save_nn(save_dir="temp")
pred1 = estimator.predict(idx)
estimator.loaded_model = True
## ------------- ** Setting up the estimator ** ---------------
acsf_params = {
"nRs2": 5,
"nRs3": 5,
"nTs": 5,
"rcut": 5,
"acut": 5,
"zeta": 220.127,
"eta": 30.8065
}
estimator = ARMP(iterations=5000,
representation_name='acsf',
representation_params=acsf_params,
tensorboard=True,
learning_rate=0.075,
l1_reg=0.0,
l2_reg=0.0)
estimator.generate_compounds(filenames)
estimator.set_properties(energies)
estimator.generate_representation(method="fortran")
print("The shape of the representation is: %s" %
(str(estimator.representation.shape)))
## ------------- ** Fitting to the data ** ---------------
idx = np.arange(0, 100)
idx_train, idx_test = modsel.train_test_split(idx,
from qml.aglaia.aglaia import ARMP import numpy as np import os ## ------------- ** Loading the data ** --------------- current_dir = os.path.dirname(os.path.realpath(__file__)) data = np.load(current_dir + '/../test/data/local_slatm_ch4cn_light.npz') descriptor = data["arr_0"] zs = data["arr_1"] energies = data["arr_2"] ## ------------- ** Setting up the estimator ** --------------- estimator = ARMP(iterations=100, l2_reg=0.0) estimator.set_representations(representations=descriptor) estimator.set_classes(zs) estimator.set_properties(energies) ## ------------- ** Fitting to the data ** --------------- idx = np.arange(0, 100) estimator.fit(idx) ## ------------- ** Predicting and scoring ** --------------- score = estimator.score(idx)
print("%i files were loaded." % (n_samples))
acsf_params = {
"nRs2": 5,
"nRs3": 5,
"nTs": 5,
"rcut": 5,
"acut": 5,
"zeta": 220.127,
"eta": 30.8065
}
estimator = ARMP(iterations=6000,
representation_name='acsf',
representation_params=acsf_params,
l1_reg=0.0,
l2_reg=0.0,
scoring_function="rmse",
tensorboard=False,
store_frequency=10,
learning_rate=0.075)
estimator.set_properties(energies[:100])
estimator.generate_compounds(filenames[:100])
estimator.generate_representation(method="tf")
print(estimator.representation.shape)
idx = list(range(100))
idx_train, idx_test = modsel.train_test_split(idx,
test_size=0,
random_state=42,
# Creating the estimator
acsf_params = {
"nRs2": 14,
"nRs3": 14,
"nTs": 14,
"rcut": 3.29,
"acut": 3.29,
"zeta": 100.06564927139748,
"eta": 39.81824764370754
}
estimator = ARMP(iterations=2633,
batch_size=22,
l1_reg=1.46e-05,
l2_reg=0.0001,
learning_rate=0.0013,
representation_name='acsf',
representation_params=acsf_params,
tensorboard=True,
store_frequency=25,
hidden_layer_sizes=(185, ))
# Loading the model previously trained
estimator.load_nn("../trained_nn/vr-nn")
estimator.set_properties(ene_surface)
# Generating the representation
start = time.time()
estimator.generate_representation(xyz_surface, zs_surface, method="fortran")
end = time.time()
print("The time taken to generate the representations is %s s" %
(str(end - start)))
],
'radial_cutoff':
5,
'angular_cutoff':
5,
'zeta':
17.8630,
'eta':
2.5148
}
# Generate estimator
estimator = ARMP(iterations=1,
l1_reg=0.0001,
l2_reg=0.005,
learning_rate=0.0005,
representation_name='acsf',
representation_params=acsf_params,
tensorboard=False,
store_frequency=10)
estimator.load_nn()
data_squal = h5py.File(
"/Volumes/Transcend/data_sets/CN_squalane/dft/squalane_cn_dft.hdf5", "r")
xyz_squal = np.array(data_squal.get("xyz")[:10])
zs_squal = np.array(data_squal.get("zs")[:10], dtype=np.int32)
ene_squal = np.array(data_squal.get("ene")[:10]) * 2625.50
pred1 = estimator.predict_from_xyz(xyz_squal, zs_squal)
print(pred1)
from qml.aglaia.aglaia import ARMP
import glob
import numpy as np
import os
## ------------- ** Loading the data ** ---------------
current_dir = os.path.dirname(os.path.realpath(__file__))
filenames = glob.glob(current_dir + '/../test/CN_isobutane/*.xyz')
energies = np.loadtxt(current_dir + '/../test/CN_isobutane/prop_kjmol_training.txt', usecols=[1])
filenames.sort()
## ------------- ** Setting up the estimator ** ---------------
estimator = ARMP(iterations=10, representation='acsf', representation_params={"radial_rs": np.arange(0, 10, 1), "angular_rs": np.arange(0.5, 10.5, 1),
"theta_s": np.arange(0, 5, 1)}, tensorboard=False)
estimator.generate_compounds(filenames)
estimator.set_properties(energies)
estimator.generate_representation()
## ------------- ** Fitting to the data ** ---------------
idx = np.arange(0,100)
estimator.fit(idx)
## ------------- ** Predicting and scoring ** ---------------
## ------------- ** Loading the data ** ---------------
current_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load(
"/Volumes/Transcend/repositories/my_qml_fork/qml/test/data/local_slatm_ch4cn_light.npz"
)
descriptor = data["arr_0"]
zs = data["arr_1"]
energies = data["arr_2"]
## ------------- ** Setting up the estimator ** ---------------
estimator = ARMP(iterations=3000,
learning_rate=0.075,
l1_reg=0.0,
l2_reg=0.0,
tensorboard=True,
store_frequency=50)
estimator.set_representations(representations=descriptor)
estimator.set_classes(zs)
estimator.set_properties(energies)
## ------------- ** Fitting to the data ** ---------------
idx = np.arange(0, 100)
estimator.fit(idx)
## ------------- ** Predicting and scoring ** ---------------
from sklearn import model_selection as modsel
import tensorflow as tf
# Getting the dataset
data = h5py.File("/Volumes/Transcend/data_sets/CN_isopentane/pruned_dft_with_forces/pruned_isopentane_cn_dft.hdf5", "r")
n_samples = 500
xyz = np.array(data.get("xyz")[-n_samples:])
ene = np.array(data.get("ene")[-n_samples:])*2625.50
ene = ene - data.get("ene")[0]*2625.50
zs = np.array(data["zs"][-n_samples:], dtype=np.int32)
# Creating the estimator
acsf_param = {"nRs2": 5, "nRs3": 5, "nTs": 5, "rcut": 5, "acut": 5, "zeta": 220.127, "eta": 30.8065}
estimator = ARMP(iterations=1000, batch_size=512, l1_reg=0.0, l2_reg=0.0, learning_rate=0.001, representation_name='acsf',
representation_params=acsf_param, tensorboard=False, store_frequency=50)
estimator.set_properties(ene)
estimator.generate_representation(xyz, zs, method='fortran')
print(estimator.g.shape)
# Doing cross validation
idx = list(range(n_samples))
idx_train, idx_test = modsel.train_test_split(idx, test_size=0.15, random_state=42, shuffle=False)
print("Starting the fitting...")
estimator.fit(idx_train)
# estimator.save_nn("saved_model")
pred1 = estimator.predict(idx_train)
pred2 = estimator.predict_from_xyz(xyz[idx_train], zs[idx_train])
n_samples = 300
xyz = np.array(data.get("xyz")[-n_samples:])
ene = np.array(data.get("ene")[-n_samples:]) * 2625.50
ene = ene - data.get("ene")[0] * 2625.50
zs = np.array(data["zs"][-n_samples:], dtype=np.int32)
# Creating the estimator
estimator = ARMP(iterations=100,
l1_reg=0.0001,
l2_reg=0.005,
learning_rate=0.0005,
representation='acsf',
representation_params={
"radial_rs": np.arange(0, 10, 3),
"angular_rs": np.arange(0, 10, 3),
"theta_s": np.arange(0, 3.14, 3)
},
tensorboard=True,
store_frequency=10,
tensorboard_subdir="tb")
estimator.set_properties(ene)
estimator.generate_representation(xyz, zs)
saved_dir = "saved_model"
estimator.load_nn(saved_dir)
idx = list(range(n_samples))
# List of properties
Y = np.array([mol.properties for mol in training])
Ys = np.array([mol.properties for mol in test])
## ------------- ** Setting up the estimator ** ---------------
print(Z)
print(Z.shape)
estimator = ARMP(
iterations=10,
l1_reg=0.0,
l2_reg=0.0,
hidden_layer_sizes=(40, 20, 10),
tensorboard=True,
store_frequency=10,
# batch_size=400,
batch_size=n_train,
learning_rate=0.1,
# scoring_function="mae",
)
estimator.set_representations(representations=X)
estimator.set_classes(Z)
estimator.set_properties(Y)
# idx = np.arange(0,100)
# estimator.fit(idx)
# score = estimator.score(idx)
ene_isopent = ene_isopent - ref_ene
zs_isopent = np.array(data["zs"], dtype=np.int32)
# Shuffling the indices of the data and then selecting the first 9625 data points
idx = list(range(len(ene_isopent)))
shuffle(idx)
idx = idx[:7621]
# Appending the true energies to a list
predictions = [ene_isopent[idx]]
# Creating the estimator
acsf_params = {"nRs2":10, "nRs3":10, "nTs":10, "rcut":3.18, "acut":3.18, "zeta":52.779232035094125, "eta":1.4954812022150898}
estimator = ARMP(iterations=5283, batch_size=37, l1_reg=8.931599068573057e-06, l2_reg=3.535679697949907e-05,
learning_rate=0.0008170485394812195, representation_name='acsf', representation_params=acsf_params,
tensorboard=True, store_frequency=25, hidden_layer_sizes=(15,88))
# Putting the data into the model
estimator.set_properties(ene_isopent)
estimator.generate_representation(xyz_isopent, zs_isopent, method="fortran")
estimator.load_nn("md-nn")
# Predicting the energies
predictions.append(estimator.predict(idx))
# Scoring the results
score = estimator.score(idx)
print(score)
# Saving the predictions to a npz file
from qml.aglaia.aglaia import ARMP
import glob
import numpy as np
from sklearn import model_selection as modsel
test_dir = "/Volumes/Transcend/repositories/my_qml_fork/qml/test/"
filenames = glob.glob(test_dir + "/qm7/*.xyz")
energies = np.loadtxt(test_dir + '/data/hof_qm7.txt', usecols=[1])
filenames.sort()
n_samples = 500
estimator = ARMP(representation_name="acsf", iterations=100)
estimator.generate_compounds(filenames[:n_samples])
estimator.set_properties(energies[:n_samples])
estimator.generate_representation(method="fortran")
idx = np.arange(0, n_samples)
idx_train, idx_test = modsel.train_test_split(idx,
random_state=42,
shuffle=True,
test_size=0.1)
estimator.fit(idx_train)
estimator.score(idx_train)
# Creating the estimator
acsf_params = {
"nRs2": 14,
"nRs3": 14,
"nTs": 14,
"rcut": 3.29,
"acut": 3.29,
"zeta": 100.06564927139748,
"eta": 39.81824764370754
}
estimator = ARMP(iterations=2633,
batch_size=22,
l1_reg=1.46e-05,
l2_reg=0.0001,
learning_rate=0.0013,
representation_name='acsf',
representation_params=acsf_params,
tensorboard=True,
store_frequency=25,
hidden_layer_sizes=(185, ))
estimator.set_properties(ene_isopent)
estimator.generate_representation(xyz_isopent, zs_isopent, method="fortran")
# Training the model on 3 folds of n data points
for n in n_samples:
cv_idx = idx_train[:n]
splitter = modsel.KFold(n_splits=3, random_state=42, shuffle=True)
indices = splitter.split(cv_idx)
acsf_params = {
"nRs2": 15,
"nRs3": 15,
"nTs": 15,
"rcut": 5,
"acut": 5,
"zeta": 220.127,
"eta": 30.8065
}
# Generate estimator
estimator = ARMP(iterations=10,
l1_reg=0.0001,
l2_reg=0.005,
learning_rate=0.0005,
representation_name='acsf',
representation_params=acsf_params,
tensorboard=True,
store_frequency=2,
hidden_layer_sizes=(50, 30, 10),
batch_size=200)
estimator.set_properties(ene_isopent)
estimator.generate_representation(pad_xyz, pad_zs, method='fortran')
print("Generated the representations")
print(estimator.representation.shape)
idx = list(range(n_samples))
idx_train, idx_test = modsel.train_test_split(idx,
random_state=42,
shuffle=True)
print("The l1 regularisation values:")
print(l1_reg)
print("The l2 regularisation values:")
print(l2_reg)
acsf_params = {
"radial_rs": np.arange(0, 10, 0.5),
"angular_rs": np.arange(0, 10, 0.5),
"theta_s": np.arange(0, 3.14, 0.25)
}
estimator = ARMP(iterations=2000,
batch_size=256,
l1_reg=0.0001,
l2_reg=0.005,
learning_rate=0.00015,
representation='acsf',
representation_params=acsf_params,
tensorboard=True,
store_frequency=50)
estimator.set_properties(ene)
estimator.generate_representation(xyz, zs)
idx = list(range(n_samples))
idx_train, idx_test = modsel.train_test_split(idx,
test_size=0.15,
random_state=42,
shuffle=True)
all_scores = []
import os
import numpy as np
## ------------- ** Loading the data ** ---------------
current_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load(current_dir + '/../test/data/local_slatm_ch4cn_light.npz')
representation = data["arr_0"]
zs = data["arr_1"]
energies = data["arr_2"]
## ------------- ** Setting up the estimator ** ---------------
estimator = ARMP(iterations=150,
l2_reg=0.0,
learning_rate=0.005,
hidden_layer_sizes=(40, 20, 10))
## ------------- ** Fitting to the data ** ---------------
estimator.fit(x=representation, y=energies, classes=zs)
## ------------- ** Predicting and scoring ** ---------------
score = estimator.score(x=representation, y=energies, classes=zs)
print("The mean absolute error is %s kJ/mol." % (str(-score)))
energies_predict = estimator.predict(x=representation, classes=zs)
from qml.aglaia.aglaia import ARMP
import os
import numpy as np
## ------------- ** Loading the data ** ---------------
current_dir = os.path.dirname(os.path.realpath(__file__))
data = np.load("/Volumes/Transcend/repositories/my_qml_fork/qml/test/data/local_slatm_ch4cn_light.npz")
representation = data["arr_0"]
zs = data["arr_1"]
energies = data["arr_2"]
## ------------- ** Setting up the estimator ** ---------------
estimator = ARMP(iterations=3000, learning_rate=0.075, l1_reg=0.0, l2_reg=0.0, tensorboard=True, store_frequency=50)
## ------------- ** Fitting to the data ** ---------------
estimator.fit(x=representation, y=energies, classes=zs)
## ------------- ** Predicting and scoring ** ---------------
score = estimator.score(x=representation, y=energies, classes=zs)
print("The mean absolute error is %s kJ/mol." % (str(-score)))
energies_predict = estimator.predict(x=representation, classes=zs)
