import os
os.environ['CUDA_VISIBLE_DEVICES'] = '' #disables GPU detection, as multithreaded BLAS on CPU is faster in most cases; remove this line to enable the use of GPUs
import sys; sys.path.append('..') #makes this script runnable from the /examples subdir without adding adding /Outer to the pythonpath
import InnerModel.train as train_inner
target_col = 'expt'
input_data_csv = '../data/SAMPL.csv'
data_name = 'FreeSolv'
HyperParams = {'mparam3': 33, 'mparam2': 20, 'mparam1': 24, 'batch_size': 5, 'lr': 0.0011192729, 'MAX_epochs': 145, 'weight_decay_factor': 7.0894428e-05}
Output_dir = '{}_model_output/'.format(data_name)
training_scores_dict, validation_scores_dict = train_inner.main(output_dir=Output_dir,
model_name='my_{}_model'.format(data_name), logp_col='', add_logp=False, training_file=input_data_csv,
validation_file=None, smile_col='smiles', target_col=target_col,
crossval_total_num_splits=10, experiment_name='{}'.format(data_name),
regression=1, binary_classification=0, batch_size = HyperParams['batch_size'], clip_gradient=0,
model_params = [HyperParams['mparam1'], HyperParams['mparam2'], HyperParams['mparam3']], contract_rings=0, learning_rate = HyperParams['lr'],
max_epochs=HyperParams['MAX_epochs'], enable_plotting=0,weight_decay_factor = HyperParams['weight_decay_factor'])
text = '<Training set scores>:\n{}\n\n<Validation set scores>:\n{}'.format('\n'.join(map(str,training_scores_dict)), '\n'.join(map(str,validation_scores_dict)))
train_inner.utils.save_text('{}{}_crossvalidation.txt'.format(Output_dir, data_name), text)
'lr': 0.00052664289,
'contract_rings': 0
}
training_scores_dict, validation_scores_dict = train_inner.main(
output_dir=Output_dir,
model_name='my_tox21_model_1',
logp_col='',
add_logp=False,
training_file=input_data_csv,
validation_file=None,
smile_col='smiles',
target_col=target_col,
crossval_total_num_splits=10,
initial_crossvalidation_index=0,
experiment_name='tox21',
regression=0,
binary_classification=1,
batch_size=HyperParams['batch_size'],
clip_gradient=0,
model_params=[
HyperParams['mparam1'], HyperParams['mparam2'], HyperParams['mparam3']
],
contract_rings=0,
learning_rate=HyperParams['lr'],
max_epochs=100,
enable_plotting=0)
text = '<Training set scores>:\n{}\n\n<Validation set scores>:\n{}'.format(
'\n'.join(map(str, training_scores_dict)),
'\n'.join(map(str, validation_scores_dict)))
import os
os.environ['CUDA_VISIBLE_DEVICES'] = '' #disables GPU detection, as multithreaded BLAS on CPU is faster in most cases; remove this line to enable the use of GPUs
import sys; sys.path.append('..') #makes this script runnable from the /examples subdir without adding adding /Outer to the pythonpath
import InnerModel.train as train_inner
target_col = 'p_np'
input_data_csv = '../data/BBBP.csv'
data_name = 'BBBP'
HyperParams = {'mparam3': 65, 'mparam2': 58, 'mparam1': 6, 'lr': 0.002, 'batch_size': 10, 'train_n_epochs':100}
Output_dir='{}_model_output/'.format(data_name)
training_scores_dict, validation_scores_dict = train_inner.main(output_dir=Output_dir,
model_name='my_{}_model'.format(data_name), training_file=input_data_csv,
validation_file=None, smile_col='smiles', target_col=target_col,
crossval_total_num_splits=10, experiment_name=data_name,
regression=False, binary_classification=True,
batch_size = HyperParams['batch_size'], clip_gradient=False,
model_params = [HyperParams['mparam1'], HyperParams['mparam2'], HyperParams['mparam3']],
contract_rings=False, learning_rate = HyperParams['lr'],
max_epochs=HyperParams['train_n_epochs'], enable_plotting=False)
text = '<Training set scores>:\n{}\n\n<Validation set scores>:\n{}'.format('\n'.join(map(str,training_scores_dict)), '\n'.join(map(str,validation_scores_dict)))
train_inner.utils.save_text('{}{}_crossvalidation.txt'.format(Output_dir, data_name), text)