def get_hyperparam_config(dataset):
c = initialise_lstm_arguments()
c['mode'] = 'train'
c['exp_name'] = 'StandardLSTM'
if dataset == 'MIMIC':
c['no_diag'] = True
c['dataset'] = dataset
c = best_global(c)
# hyper-parameter grid
param_grid = {
'n_layers': [1, 2, 3, 4],
'learning_rate':
list(np.logspace(np.log10(0.0001), np.log10(0.01), base=10, num=100)),
'batch_size':
list(
int(x)
for x in np.logspace(np.log2(4), np.log2(512), base=2, num=8)),
'lstm_dropout_rate':
[0, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5],
'hidden_size':
list(
int(x)
for x in np.logspace(np.log2(16), np.log2(256), base=2, num=5)),
}
c['n_layers'] = random.choice(param_grid['n_layers'])
c['learning_rate'] = round(random.choice(param_grid['learning_rate']), 5)
c['batch_size'] = random.choice(param_grid['batch_size'])
c['lstm_dropout_rate'] = random.choice(param_grid['lstm_dropout_rate'])
c['hidden_size'] = random.choice(param_grid['hidden_size'])
return c
def get_hyperparam_config(dataset):
c = initialise_lstm_arguments()
c['mode'] = 'train'
c['exp_name'] = 'ChannelwiseLSTM'
if dataset == 'MIMIC':
c['no_diag'] = True
c['dataset'] = dataset
c = best_lstm(
c) # get best hyperparams from standard LSTM for this dataset
c['channelwise'] = True
hidden_size_choice = list(
int(x) for x in np.logspace(np.log2(4), np.log2(16), base=2, num=6))
c['hidden_size'] = random.choice(hidden_size_choice)
return c
from eICU_preprocessing.split_train_test import create_folder
from models.run_lstm import BaselineLSTM
from models.initialise_arguments import initialise_lstm_arguments
from models.final_experiment_scripts.best_hyperparameters import best_lstm
if __name__ == '__main__':
c = initialise_lstm_arguments()
c['exp_name'] = 'StandardLSTM'
c['dataset'] = 'MIMIC'
c['task'] = 'mortality'
c = best_lstm(c)
log_folder_path = create_folder('models/experiments/final/MIMIC/mortality',
c.exp_name)
baseline_lstm = BaselineLSTM(
config=c,
n_epochs=c.n_epochs,
name=c.exp_name,
base_dir=log_folder_path,
explogger_kwargs={'folder_format': '%Y-%m-%d_%H%M%S{run_number}'})
baseline_lstm.run()
