saver = tf.train.Saver()
# Train the network
with tf.Session() as sess:
init = tf.global_variables_initializer()
sess.run(init)
#saver.restore(sess, 'models/%s/%s.ckpt' % (eid, 533))
record_loss(sess, train_dataset.X)
k = 1
new_iters = int(n_iters / k)
for it in range(new_iters):
# Shuffle data once for each epoch
if it % int(new_iters / n_epochs) == 0:
train_dataset.shuffle()
# Train
for i in range(k):
next_x, _ = train_dataset.next_batch()
draw_prior_samples = draw_multivariate_gaussian(
latent_dim, batch_size)
sess.run(train_discriminator,
feed_dict={
x: next_x,
prior_samples: draw_prior_samples
})
sess.run(train_generator, feed_dict={x: next_x})
sess.run(train_reconstruct, feed_dict={x: next_x})
# Show loss
if it % show_steps == 0:
print('Iterations %5d: ' % (it + 1), end='')
record_loss(sess, train_dataset.X)
extract_encoded_data(sess)
extract_image(sess)
if __name__ == '__main__':
num_cur_best = 0
saver = tf.train.Saver()
# Launch the graph
with tf.Session() as sess:
init = tf.global_variables_initializer()
sess.run(init)
# Restore model
#saver.restore(sess, 'models/try_2.ckpt')
#print("Model restored.")
#print('Before train:\t', end="")
record_error(sess)
for it in range(n_iters):
# Train next batch
next_x, next_y = train_dataset.next_batch()
sess.run(train_step, feed_dict={x: next_x, y: next_y})
# Record loss
if it % show_steps == 0:
print('Iterations %4d:\t' %(it+1) , end="")
loss_this_iter = record_error(sess)
# Save the model
if log['best_loss'] - loss_this_iter > 0.0000000001:
num_cur_best += 1
print('Find and save %d current best loss model. %.10f' %(num_cur_best, loss_this_iter))
log['best_loss'] = loss_this_iter
if not os.path.exists(save_directory):
os.makedirs(save_directory)
save_path = saver.save(sess, '%s/model.ckpt' % (save_directory))
# Shuffle data once for each epoch
def calculate_error_with_real_price(sess, predict_op, X, Y, real_last_one,
real_last_two, statistcs):
# Predict prices
dataset = Dataset(X, Y, batch_size)
n_batch = int(X.shape[0] / batch_size)
for i in range(n_batch):
next_x, next_y = dataset.next_batch()
if i == 0:
predict_log_return = sess.run(predict,
feed_dict={
x: next_x,
y: next_y
}).tolist()
else:
predict_log_return = np.append(predict_log_return,
sess.run(predict,
feed_dict={
x: next_x,
y: next_y
}).tolist(),
axis=0)
predict_last_one = np.exp(predict_log_return + np.log(real_last_two))
# Export prices to statistcs_file
statistcs['real_price'] = real_last_one.tolist()
statistcs['predict_price'] = predict_last_one.tolist()
# Show last 5 statistics
print('last price is: ', real_last_two[-5:])
print('predict price is: ', predict_last_one[-5:])
print('real price is: ', real_last_one[-5:])
# Calculate MSE
MSE = np.mean((real_last_one - predict_last_one)**2)
statistcs['MSE'] = MSE
print('MSE = %.10f' % MSE)
# Caluculate Variance
mean_real = np.mean(real_last_one)
var_real = np.var(real_last_one)
print('Real mean=%.4f, var=%.5f' % (mean_real, var_real))
mean_predict = np.mean(predict_last_one)
var_predict = np.var(predict_last_one)
print('Predict mean=%.4f, var=%.5f' % (mean_predict, var_predict))
length = real_last_one.shape[0]
real_last_one = np.reshape(real_last_one, length)
real_last_two = np.reshape(real_last_two, length)
predict_last_one = np.reshape(predict_last_one, length)
# Calculate sign accuracy
real_diff = real_last_one - real_last_two
predict_diff = predict_last_one - real_last_two
real_diff_sign = np.sign(real_diff)
predict_diff_sign = np.sign(predict_diff)
print('real_diff: ', real_diff[-20:])
print('predict_diff: ', predict_diff[-20:])
num_correct_sign = np.count_nonzero(
np.equal(real_diff_sign, predict_diff_sign))
sign_accuracy = num_correct_sign / real_diff_sign.shape[0]
statistcs['sign_accuracy'] = sign_accuracy
print('Sign Accuracy = %.10f' % sign_accuracy)