def _save_times(times, parms):
times = [{**{'output': parms['output']}, **times}]
log_file = os.path.join('output_files', 'execution_times.csv')
if not os.path.exists(log_file):
open(log_file, 'w').close()
if os.path.getsize(log_file) > 0:
sup.create_csv_file(times, log_file, mode='a')
else:
sup.create_csv_file_header(times, log_file)
def _save_times(times, settings, temp_output):
if times:
times = [{**{'output': settings['output']}, **times}]
log_file = os.path.join(temp_output, 'execution_times.csv')
if not os.path.exists(log_file):
open(log_file, 'w').close()
if os.path.getsize(log_file) > 0:
sup.create_csv_file(times, log_file, mode='a')
else:
sup.create_csv_file_header(times, log_file)
def on_train_end(self, logs={}):
log_file = self.log_path
data = [{
'output_folder': self.output_folder,
'train_epochs': len(self.logs),
'avg_time': np.mean(self.logs),
'min_time': np.min(self.logs),
'max_time': np.max(self.logs)
}]
if os.path.exists(log_file):
sup.create_csv_file(data, log_file, mode='a')
else:
sup.create_csv_file_header(data, log_file)
def _define_response(self, settings, status, sim_values, **kwargs) -> None:
response = dict()
measurements = list()
data = {
'alg_manag': settings['alg_manag'],
'gate_management': settings['gate_management'],
'output': settings['output']
}
# Miner parms
if settings['mining_alg'] in ['sm1', 'sm3']:
data['epsilon'] = settings['epsilon']
data['eta'] = settings['eta']
elif settings['mining_alg'] == 'sm2':
data['concurrency'] = settings['concurrency']
else:
raise ValueError(settings['mining_alg'])
similarity = 0
# response['params'] = settings
response['output'] = settings['output']
if status == STATUS_OK:
similarity = np.mean([x['sim_val'] for x in sim_values])
loss = (1 - similarity)
response['loss'] = loss
response['status'] = status if loss > 0 else STATUS_FAIL
for sim_val in sim_values:
measurements.append({
**{
'similarity': sim_val['sim_val'],
'sim_metric': sim_val['metric'],
'status': response['status']
},
**data
})
else:
response['status'] = status
measurements.append({
**{
'similarity': 0,
'sim_metric': 'dl',
'status': response['status']
},
**data
})
if os.path.getsize(self.file_name) > 0:
sup.create_csv_file(measurements, self.file_name, mode='a')
else:
sup.create_csv_file_header(measurements, self.file_name)
return response
def _define_response(self, parms, loss, **kwargs) -> None:
measurements = list()
measurements.append({
'loss': loss,
'sim_metric': 'val_loss',
'n_size': parms['n_size'],
'l_size': parms['l_size'],
'lstm_act': parms['lstm_act'],
'dense_act': parms['dense_act'],
'optim': parms['optim'],
'output': parms['output'][len(os.getcwd()) + 1:]
})
if os.path.getsize(self.res_files) > 0:
sup.create_csv_file(measurements, self.res_files, mode='a')
else:
sup.create_csv_file_header(measurements, self.res_files)
def _define_response(self, parms, status, loss, **kwargs) -> None:
print(loss)
response = dict()
measurements = list()
data = {
'n_size': parms['n_size'],
'l_size': parms['l_size'],
'lstm_act': parms['lstm_act'],
'dense_act': parms['dense_act'],
'optim': parms['optim'],
'scale_args': parms['scale_args'],
'output': parms['output']
}
response['output'] = parms['output']
if status == STATUS_OK:
response['loss'] = loss
response['status'] = status if loss > 0 else STATUS_FAIL
measurements.append({
**{
'loss': loss,
'sim_metric': 'val_loss',
'status': response['status']
},
**data
})
else:
response['status'] = status
measurements.append({
**{
'loss': 1,
'sim_metric': 'val_loss',
'status': response['status']
},
**data
})
if os.path.getsize(self.file_name) > 0:
sup.create_csv_file(measurements, self.file_name, mode='a')
else:
sup.create_csv_file_header(measurements, self.file_name)
return response