-
Notifications
You must be signed in to change notification settings - Fork 0
/
nn_ds.py
192 lines (156 loc) · 6.05 KB
/
nn_ds.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
'''
GPU command:
THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python train_model.py
'''
# from common import GENRES
from keras.callbacks import Callback
from keras.utils import np_utils, to_categorical
from keras.models import Model, model_from_yaml
from keras.optimizers import RMSprop, Adam
from keras import backend as K
from keras.layers import Input, Dense, Lambda, Dropout, Activation, LSTM, \
TimeDistributed, Convolution1D, MaxPooling1D, Conv1D
from sklearn.model_selection import train_test_split
import numpy as np
import random
import cPickle
from optparse import OptionParser
from sys import stderr, argv
import sys, os
import lib.io
import lib.viz
import lib.cl
SEED = 42
N_LAYERS = 3 #3
FILTER_LENGTH = 5 #5
CONV_FILTER_COUNT = 256 #256
CONV_FILTER_STRIDES = 2
LSTM_COUNT = 256 #256
BATCH_SIZE = 32
EPOCH_COUNT = 600
CLASS_COUNT = 10
def predict(model, x):
x = np.expand_dims(x, axis=2)
print("x shape in prediction: %d, %d, %d"%x.shape)
return model.predict(x)
def load_model(model_filename, weights_filename):
with open(model_filename, 'r') as f:
model = model_from_yaml(f.read())
model.load_weights(weights_filename)
# self.pred_fun = get_layer_output_function(model, 'output_realtime')
print 'Model loaded'
return model
def train_model(x_train, y_train, x_val, y_val):
print 'Building model...'
B = K.backend()
if B=='tensorflow':
K.set_image_dim_ordering('tf')
x_train = np.expand_dims(x_train, axis=2)
x_val = np.expand_dims(x_val, axis=2)
print("x_train shape: %d, %d, %d"%x_train.shape)
#print x_train
y_cat = to_categorical(y_train)
y_val = to_categorical(y_val)
n_features = x_train.shape[1]
# input_shape = (n_features, None)
# input_shape = (None, n_features)
input_shape = (n_features, 1)
model_input = Input(input_shape, name='input')
layer = model_input
for i in range(N_LAYERS):
layer = Conv1D(
filters = CONV_FILTER_COUNT,
kernel_size = FILTER_LENGTH,
name='convolution_' + str(i + 1)
)(layer)
layer = Activation('relu')(layer)
layer = MaxPooling1D(2)(layer)
#layer = Dropout(0.25)(layer)
layer = Dropout(0.5)(layer)
layer = LSTM(LSTM_COUNT, return_sequences=True)(layer)
layer = Dropout(0.5)(layer)
layer = TimeDistributed(Dense(CLASS_COUNT))(layer)
layer = Activation('softmax', name='output_realtime')(layer)
time_distributed_merge_layer = Lambda(
function=lambda x: K.mean(x, axis=1),
output_shape=lambda shape: (shape[0],) + shape[2:],
name='output_merged'
)
model_output = time_distributed_merge_layer(layer)
model = Model(model_input, model_output)
opt = RMSprop(lr=0.00001)
#opt = Adam(lr=0.00001, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0)
model.compile(
loss='categorical_crossentropy',
optimizer=opt,
metrics=['accuracy']
)
print 'Training...'
# history = model.fit(x=x_train, y=y_cat, batch_size=BATCH_SIZE, epochs=EPOCH_COUNT,
# validation_data=(x_val, y_val), verbose=2)
return model
if __name__ == '__main__':
parser = OptionParser()
parser.add_option('-s', '--save_path', dest='model_path',
default=os.path.join(os.path.dirname(__file__),
'models'),
help='path to the output model', metavar='MODEL_PATH')
parser.add_option('-l', '--load_path', dest='load_path',
default=None,
help='path to the load model',
metavar='LOAD_PATH')
parser.add_option('-o', '--output_path', dest='output_path',
default='output',
help='path to save results',
metavar='OUTPUT_PATH')
options, args = parser.parse_args()
input_filename_x = 'train_data.csv'
input_filename_y = 'train_labels.csv'
test_input_filename = 'test_data.csv'
model_filename = 'model.yaml'
weights_filename = 'weights.h5'
io = lib.io.IO()
viz = lib.viz.Viz()
cl = lib.cl.CL(io, viz)
# Read data
print "Reading train data..."
X, y = io.read_data(input_filename_x, input_filename_y)
y = io.shift_v(y, shift=-1)
print "Reading test data..."
test_x = io.read_data(test_input_filename, None)
print "There are " + str(len(X)) + " samples in the train set."
print "There are " + str(len(test_x)) + " samples in the test set."
test_x = np.matrix(test_x)
test_ids = range(1, len(test_x)+1)
# load from file
if options.load_path is not None:
model = load_model(options.load_path+'/'+model_filename, options.load_path+'/'+weights_filename)
# train
else:
# Remove outliers
X,y = cl.lof(np.matrix(X), np.matrix(y))
# Suffle
X, y = io.shuffle(X, y)
# Pick val and train set
val_ids, val_x, val_y = io.pick_set(X, y, 526)
train_ids, train_x, train_y = io.pick_set(X, y, 3400)
# train
model = train_model(train_x, train_y, val_x, val_y)
# Save model
#full_model_filename = options.model_path + '/'+model_filename
#full_weights_filename = options.model_path + '/'+weights_filename
#with open(full_model_filename, 'w') as f:
# f.write(model.to_yaml())
# model.save_weights(full_weights_filename)
# Print metrics
#viz.plot_nn_perf(history, options.output_path+'/nn_perf.png')
# visualize model
viz.plot_model(model, options.output_path+'/model.png')
# viz.plot_all_feature_maps(model, test_x[:12], options.output_path+'/')
# predict
pred_proba = predict(model, test_x)
pred_class = np.argmax(pred_proba, axis=1)
pred_class = io.shift_v(pred_class, shift=1)
# Output
io.write_classes(options.output_path+'/classes_result.csv', test_ids, pred_class)
io.write_probabilities(options.output_path+'/probabilities_result.csv', test_ids, pred_proba)