/
dnn.py
219 lines (175 loc) · 8.16 KB
/
dnn.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
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
import numpy as np
from ..exception import UnSupportException, ParameterException
from .base import Model
from ..util import ParamValidator
class DNN(Model):
pv = ParamValidator(
{
"input_shape": {"type": int},
"shape": {"type": list},
"activations": {"type": list},
"eta": {"type": [int, float]},
"threshold": {"type": [int, float]},
"softmax": {"type": bool},
"max_epochs": {"type": int},
"regularization": {"type": [int, float]},
"minibatch_size": {"type": int},
"momentum": {"type": [int, float], "range": (0.0, 1.0)},
"decay_power": {"type": [int, float]},
"verbose": {"type": bool},
}
)
def __init__(self, shape, activations, eta=0.5, threshold=1e-5, softmax=False, max_epochs=20,
regularization=0, minibatch_size=20, momentum=0.9, decay_power=0.2, verbose=False):
Model.__init__(self)
if not len(shape) == len(activations):
raise ParameterException("activations must equal to number od layers.")
self.shape = self.pv("shape", shape)
self.depth = len(self.shape)
self.activity_levels = [np.mat([0])] * self.depth
self.outputs = [np.mat(np.mat([0]))] * (self.depth + 1)
self.deltas = [np.mat(np.mat([0]))] * self.depth
self.eta = self.pv("eta", eta)
self.effective_eta = self.eta
self.threshold = self.pv("threshold", threshold)
self.max_epochs = self.pv("max_epochs", max_epochs)
self.regularization = self.pv("regularization", regularization)
self.is_softmax = self.pv("softmax", softmax)
self.verbose = self.pv("verbose", verbose)
self.minibatch_size = self.pv("minibatch_size", minibatch_size)
self.momentum = self.pv("momentum", momentum)
self.decay_power = self.pv("decay_power", decay_power)
self.iterations = 0
self.epochs = 0
self.epoch_loss = []
self.activations = self.pv("activations", activations)
self.activation_func = []
self.activation_func_diff = []
for f in activations:
f = f.lower()
if f == "sigmoid":
self.activation_func.append(self.sigmoid)
self.activation_func_diff.append(self.sigmoid_diff)
elif f == "identity":
self.activation_func.append(self.identity)
self.activation_func_diff.append(self.identity_diff)
elif f == "relu":
self.activation_func.append(self.relu)
self.activation_func_diff.append(self.relu_diff)
elif f == "tanh":
self.activation_func.append(self.tanh)
self.activation_func_diff.append(self.tanh_diff)
else:
raise UnSupportException("activation function {:s}".format(f))
self.weights = [np.mat(np.mat([0]))] * self.depth
self.biases = [np.mat(np.mat([0]))] * self.depth
self.acc_weights_delta = [np.mat(np.mat([0]))] * self.depth
self.acc_biases_delta = [np.mat(np.mat([0]))] * self.depth
self.input_weights_initialized = False
for idx in np.arange(1, len(shape)):
self.weights[idx] = np.mat(np.random.random((shape[idx], shape[idx - 1])) / 200)
self.biases[idx] = np.mat(np.random.random((shape[idx], 1)) / 200)
def compute(self, x):
result = x
for idx in np.arange(0, self.depth):
self.outputs[idx] = result
al = self.weights[idx] * result + self.biases[idx]
self.activity_levels[idx] = al
result = np.mat(self.activation_func[idx](al))
self.outputs[self.depth] = result
return self.softmax(result) if self.is_softmax else result
def predict(self, features):
if features.shape[0] == 0 or features.shape[1] == 0:
raise ParameterException("data is empty.")
return self.compute(np.mat(features).T).T.A
def bp(self, d):
tmp = d.T
for idx in np.arange(0, self.depth)[::-1]:
delta = np.multiply(tmp, self.activation_func_diff[idx](self.outputs[idx + 1]).T)
self.deltas[idx] = delta
tmp = delta * self.weights[idx]
def update(self):
for idx in np.arange(0, self.depth):
# current gradient
weights_grad = -self.deltas[idx].T * self.outputs[idx].T / self.deltas[idx].shape[0] + \
self.regularization * self.weights[idx]
biases_grad = -np.mean(self.deltas[idx].T, axis=1) # + self.regularization * self.biases[idx]
# accumulated delta
self.acc_weights_delta[idx] = self.acc_weights_delta[
idx] * self.momentum - self.effective_eta * weights_grad
self.acc_biases_delta[idx] = self.acc_biases_delta[idx] * self.momentum - self.effective_eta * biases_grad
self.weights[idx] = self.weights[idx] + self.acc_weights_delta[idx]
self.biases[idx] = self.biases[idx] + self.acc_biases_delta[idx]
def train(self, features, response):
if features.shape[0] == 0 or features.shape[1] == 0 or features.shape[0] != response.shape[0] or response.shape[
1] == 0:
raise ParameterException("features or response is empty or number of instances is not equal")
x = np.mat(features)
y = np.mat(response)
loss = []
self.epoch_loss = []
self.iterations = 0
self.epochs = 0
start = 0
train_set_size = x.shape[0]
if not self.input_weights_initialized:
self.weights[0] = np.mat(np.random.random((self.shape[0], x.shape[1])) / 200)
self.biases[0] = np.mat(np.random.random((self.shape[0], 1)) / 200)
self.input_weights_initialized = True
while True:
end = start + self.minibatch_size
minibatch_x = x[start:end].T
minibatch_y = y[start:end].T
start = (start + self.minibatch_size) % train_set_size
yp = self.compute(minibatch_x)
d = minibatch_y - yp
self.bp(d)
self.update()
if self.is_softmax:
loss.append(np.mean(-np.sum(np.multiply(minibatch_y, np.log(yp + 1e-300)), axis=0)))
else:
loss.append(np.mean(np.sqrt(np.sum(np.power(d, 2), axis=0))))
self.iterations += 1
# decay the learning rate
self.effective_eta = self.eta / np.power(self.iterations, self.decay_power)
if self.iterations % train_set_size == 0:
self.epochs += 1
mean_e = np.mean(loss)
self.epoch_loss.append(mean_e)
loss = []
if self.verbose:
print("epoch: {:d}. mean loss: {:.6f}. learning rate: {:.8f}".format(self.epochs, mean_e,
self.effective_eta))
if self.epochs >= self.max_epochs or mean_e < self.threshold:
break
@staticmethod
def sigmoid(x):
return 1.0 / (1.0 + np.power(np.e, np.where(-x > 1e2, 1e2, -x)))
@staticmethod
def sigmoid_diff(x):
return np.multiply(x, (1 - x))
@staticmethod
def relu(x):
# return x if x > 0 else 0.0
return np.where(x > 0, x, 0.0)
@staticmethod
def relu_diff(x):
return np.where(x > 0, 1.0, 0.0)
@staticmethod
def identity(x):
return x
@staticmethod
def identity_diff(x):
return np.ones(x.shape)
@staticmethod
def tanh(x):
exp = 2 * np.where(x > 1e2, 1e2, x)
return (np.power(np.e, exp) - 1) / (np.power(np.e, exp) + 1)
@staticmethod
def tanh_diff(x):
return 1 - np.multiply(x, x)
@staticmethod
def softmax(x):
x[x > 1e2] = 1e2
ep = np.power(np.e, x)
return ep / np.sum(ep, axis=0)