def predict(model_name, model_file_name, layers, input_features):
# Restore the well-trained model
if model_name == 'dbn':
network = dbn.DBN(layers=layers, batch_size=100)
else:
return -1, 'Invalid model'
with tf.Session() as sess:
tf_saver = tf.train.Saver()
tf_saver.restore(sess, model_file_path + model_file_name)
outputs = network.get_output(sess, input_features)
return 0, outputs
def test_main(testing_features):
# Restore the well-trained model
network = dbn.DBN(layers=layers, batch_size=1)
with tf.Session() as sess:
tf_saver = tf.train.Saver()
tf_saver.restore(sess, model_file_path + model_file_name)
# Get the ouput from test data
# for i in testing_features:
# outputs = network.get_output(sess, [i])[0]
# print outputs
outputs = network.get_output(sess, testing_features)
return outputs
def predict(model_name, model_file_name, layers, input_features, target_size):
img_width=input_features.shape[1]
img_height=input_features.shape[2]
input_features = input_features.reshape(len(input_features), len(input_features[0].flatten()))
# Restore the well-trained model
if model_name == 'dbn':
layers = map(int, layers.strip().split(','))
layers = [input_features.shape[1]] + layers + [target_size]
network = dbn.DBN(layers=layers, batch_size=100)
elif model_name == 'cnn':
conv_layers, hid_layers = layers.strip().split('#')
conv_layers = map(int, conv_layers.strip().split(','))
hid_layers = map(int, hid_layers.strip().split(','))
network = cnn.CNN(img_width=img_width, img_height=img_height, conv_layers=conv_layers, hidden_layers=hid_layers, batch_size=128)
else:
return -1, 'Invalid model'
with tf.Session() as sess:
tf_saver = tf.train.Saver()
tf_saver.restore(sess, model_file_path + model_file_name)
outputs = network.get_output(sess, input_features)
return 0, outputs
print 'The resource path is not existed, create a new one.'
os.makedirs(model_path)
if not os.path.exists(res_path):
print 'The result path is not existed, create a new one.'
os.makedirs(res_path)
# Training
# _, input_size = training_features.shape
# _, output_size = training_labels.shape
print '------ [ %s ] ------' % arrow.now()
print 'Create an instance of the neural network.'
if model == 'dbn':
layers = map(int, layers.strip().split(','))
layers = [features.shape[1]] + layers + [labels.shape[1]]
network = dbn.DBN(layers=layers,
iters=1000,
batch_size=100,
mu=LEARNING_RATE) #.0001)
elif model == 'cnn':
conv_layers, hid_layers = layers.strip().split('#')
conv_layers = map(int, conv_layers.strip().split(','))
hid_layers = map(int, hid_layers.strip().split(','))
network = cnn.CNN(img_width=img_width,
img_height=img_height,
conv_layers=conv_layers,
hidden_layers=hid_layers,
learning_rate=LEARNING_RATE,
training_iters=20000,
batch_size=128,
display_step=10)
with tf.Session() as sess:
print 'The number of training data:\t%d' % len(features)
print 'A example of the feature data:\t%s' % features[0]
print 'A example of the label data:\t%s' % labels[0]
# Divide the raw data into the training part and testing part
start_test = end_train = int(
float(len(features)) / float(TRAIN_TEST_RATIO + 1) * TRAIN_TEST_RATIO)
training_features = np.array(features[0:end_train])
training_labels = np.array(labels[0:end_train])
testing_features = np.array(features[start_test:len(features)])
testing_labels = np.array(labels[start_test:len(labels)])
# Check path
if not os.path.exists(model_path):
os.makedirs(model_path)
if not os.path.exists(res_path):
os.makedirs(res_path)
# Training
# _, input_size = training_features.shape
# _, output_size = training_labels.shape
network = dbn.DBN(layers=layers, iters=50, batch_size=100, mu=.0005)
with tf.Session() as sess:
tr, test = network.train(sess, training_features, training_labels,
testing_features, testing_labels)
np.savetxt(res_path + 'training_result.txt', tr)
np.savetxt(res_path + 'testing_result.txt', test)
tf_saver = tf.train.Saver()
tf_saver.save(sess, model_path + model_file_name)