def objective(params):
#optimize_model = build_lstm_v1.lstm_model_102(params, train_data.shape[2], 24, 24)
#optimize_model = build_lstm_v1.lstm_model_106(params, train_data.shape[2], 24)
optimize_model = build_lstm_v1.lstm_model_110b(params, train_data.shape[2],
24)
#for epochs in range(5):
for ep in range(20):
#optimize_history = optimize_model.fit(X_seq, Y_seq, batch_size=1, nb_epoch=3, validation_split=(X_seq, Y_seq), shuffle=False)
optimize_history = optimize_model.fit(train_data,
H_t,
batch_size=1,
nb_epoch=1,
validation_data=(val_data,
H_val),
shuffle=False)
#optimize_history = optimize_model.fit(train_data, H_t, batch_size=1, nb_epoch=1, validation_split=0.3, shuffle=False)
optimize_model.reset_states()
loss_v = optimize_history.history['val_loss']
print loss_v
loss_out = loss_v[-1]
return {'loss': loss_out, 'status': STATUS_OK}
def objective(params):
optimize_model = build_lstm_v1.lstm_model_110b(params, train_data.shape[2], 24)
loss_out = NNFunctions.model_optimizer_101(optimize_model, train_data, H_t, val_data, H_val, 10)
return {'loss': loss_out, 'status': STATUS_OK}
optimize_model = build_lstm_v1.lstm_model_110b(params, train_data.shape[2], 24)
loss_out = NNFunctions.model_optimizer_101(optimize_model, train_data, H_t, val_data, H_val, 10)
return {'loss': loss_out, 'status': STATUS_OK}
trials = Trials()
best = fmin(objective, space, algo=tpe.suggest, trials=trials, max_evals=5)
#Building Stateful Model
lstm_hidden = hyperopt.space_eval(space, best)
print lstm_hidden
tsteps = 24
out_dim = 24
#lstm_model = build_lstm_v1.lstm_model_102(lstm_hidden, train_data.shape[2], out_dim, tsteps)
lstm_model = build_lstm_v1.lstm_model_110b(lstm_hidden, train_data.shape[2], tsteps)
save_model = lstm_model
##callbacks for Early Stopping
callbacks = [EarlyStopping(monitor='val_loss', patience=3)]
#parameters for simulation
attempt_max = 3
epoch_max = 20
min_epoch = 10
#Criterion for early stopping
tau = 5
e_mat = numpy.zeros((epoch_max, attempt_max))
e_temp = numpy.zeros((tau, ))