import lstm
import parameters
import plot
%load_ext autoreload
%autoreload 2
# Initialization of seeds
set_random_seed(2)
seed(2)
(params, _, yraw, y, yhat, num_errors) =\
lstm.predict('params_3y_1L256_09i.yaml')
plot.prediction(y, yhat,
yraw, num_errors, params,
inv_scale=False, inv_diff=False, inv_log=False)
#
# -- Averaging predictions.
#
# y_avg = ((yhat1 + yhat2 + yhat3) / 3.0)
# rmse, num_errors_avg = compute.error(y1, y_avg)
# plot.prediction(y1, y_avg, num_errors_avg, params1)
#
# -- Single value prediction.
#
# prediction = lstm.single_predict(model1, X_test[31], Y_test[31], params)
# print(prediction)
raw = data.read(params)
print('Original dataset num samples:', raw.shape)
adjusted = parameters.adjust(raw, params)
X_train, Y_train, X_test, Y_test = data.prepare(adjusted, params)
# Build the model and train it.
# model = lstm.load('../../data/networks/20180115_0832.h5')
model = lstm.build(params)
train_loss = lstm.fit(model, X_train, Y_train, params)
plot.history(train_loss)
# Plot the test values for Y, and Y_hat, without scaling (inverted)
Y_hat = model.predict(X_test, batch_size=params['lstm_batch_size'])
rmse, num_errors = compute.error(Y_test, Y_hat)
plot.prediction(params['y_scaler'].inverse_transform(Y_test),
params['y_scaler'].inverse_transform(Y_hat),
num_errors)
# Save the model
saved_model_name = lstm.save(model)
# How to make a prediction
# We need a vector of size (timesteps x num_features)
# That vector is replicated 'batch_size' times, and feed into the network
# The result is an array o 'batch_size' predictions, and we're only interested
# in the first one, which is the one we give back
p = repeat(X_test[0], 8).reshape((8, 6, 8))
p.shape
model.predict(p)[0]
Y_hat[0]
# s e t u p
#
params = parameters.read()
adjusted = parameters.adjust(read(params), params)
X, Y, Xtest, ytest = prepare(normalize(adjusted, params), params)
#
# t r a i n i n g
#
model = setup(params)
parameters.summary(params)
model.summary()
lstm.stateless_fit(model, X, Y, Xtest, ytest, params)
EarlyStopping(
monitor='val_loss', min_delta=0.0,
patience=0, verbose=0, mode='auto')
save(model, params, prefix='5y', additional_epocs=0)
#
# r e b u i l d & p r e d i c t
#
pred = lstm.build(params, batch_size=1)
pred.set_weights(model.get_weights())
(yhat, rmse, num_errors) = lstm.range_predict(pred, Xtest, ytest, params)
#
# p l o t
#
# plot.history(train_loss)
plot.prediction(ytest, yhat, rmse, num_errors, params)