-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathunrolling_alg_v2.py
274 lines (220 loc) · 8.43 KB
/
unrolling_alg_v2.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
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
from torch.autograd import Variable
import sklearn
import numpy as np
import matplotlib.pyplot as plt
import torch.nn.functional as F
import torch.utils.data as Data
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error, mean_absolute_error
import torch
import torch.nn as nn
from torch.optim import SGD
import csv
def generate_data(n):
num = 0
theta1_data = []
theta2_data = []
y_data = []
while num<n:
epsilon = 0.1
theta11 = round(np.random.random(1)[0], 2)
theta12 = round((1-theta11)*(1-np.random.random(1)[0]),2)
theta21 = round(np.random.random(1)[0],2)
theta22 = round((1 - theta21)*(1 - np.random.random(1)[0]),2)
c1 = round(10*np.random.random(1)[0],2)
c2 = round(10*np.random.random(1)[0],2)
err = 1
p1 = 1
p2 = 1
while err>10e-5:
tp1 = p1
tp2 = p2
p1 = c1*p1**theta11*p2**theta12
p2 = c2*p1**theta21*p2**theta22
err = np.abs(np.log(p1) - np.log(tp1)) + np.abs(np.log(p2) - np.log(tp2))
if p1<10e2 and p1>10e-2:
num = num +1
theta1_data.append([float(theta11),float(theta12),float(np.log(c1))])
theta2_data.append([float(theta21),float(theta22),float(np.log(c2))])
y_data.append([float(np.log(p1)),float(np.log(p2))])
return theta1_data, theta2_data, y_data
class MyLinear1(nn.Module):
def __init__(self):
super().__init__()
self.hidden1_1 = nn.Linear(in_features=3, out_features=100, bias=True)
self.hidden2_1 = nn.Linear(3, 100)
self.hidden1_2 = nn.Linear(100, 1)
self.hidden2_2 = nn.Linear(100, 1)
def forward(self, x):
x1 = x[:,0:3]
x2 = x[:,3:6]
batch_size = len(x1)
b = torch.tensor([[1] * batch_size]).T
l1_1 = F.relu6(self.hidden1_1(x1))
l2_1 = F.relu6(self.hidden2_1(x2))
l1_2 = F.relu6(self.hidden1_2(l1_1))
l2_2 = F.relu6(self.hidden2_2(l2_1))
l_out = torch.cat((l1_2, l2_2, b), 1)
x1_2_l_out = torch.cat((x, l_out), 1)
return x1_2_l_out
class MyLinear2(nn.Module):
def __init__(self):
super().__init__()
self.hidden1_1 = nn.Linear(in_features=3, out_features=100, bias=True)
self.hidden2_1 = nn.Linear(3, 100)
self.hidden1_2 = nn.Linear(100, 1)
self.hidden2_2 = nn.Linear(100, 1)
def forward(self, x1_2_l_out):
x1 = x1_2_l_out[:, 0:3]
x2 = x1_2_l_out[:, 3:6]
l_out = x1_2_l_out[:, 6:9]
batch_size = len(x1)
b = torch.tensor([[1] * batch_size]).T
l1_1 = F.relu6(self.hidden1_1(x1*l_out))
l2_1 = F.relu6(self.hidden2_1(x2*l_out))
# l1_1 = F.relu6(self.hidden1_1( l_out))
# l2_1 = F.relu6(self.hidden2_1( l_out))
l1_2 = F.relu6(self.hidden1_2(l1_1))
l2_2 = F.relu6(self.hidden2_2(l2_1))
l_out = torch.cat((l1_2, l2_2, b), 1)
x1_2_l_out = torch.cat((x1,x2, l_out), 1)
return x1_2_l_out
class MyLinear3(nn.Module):
def __init__(self):
super().__init__()
self.hidden1_1 = nn.Linear(in_features=3, out_features=100, bias=True)
self.hidden2_1 = nn.Linear(3, 100)
self.hidden1_2 = nn.Linear(100, 1)
self.hidden2_2 = nn.Linear(100, 1)
def forward(self, x1_2_l_out):
x1 = x1_2_l_out[:, 0:3]
x2 = x1_2_l_out[:, 3:6]
l_out = x1_2_l_out[:, 6:9]
l1_1 = F.relu6(self.hidden1_1(x1 * l_out))
l2_1 = F.relu6(self.hidden2_1(x2 * l_out))
# l1_1 = F.relu6(self.hidden1_1(l_out))
# l2_1 = F.relu6(self.hidden2_1(l_out))
l1_2 = self.hidden1_2(l1_1)
l2_2 = self.hidden2_2(l2_1)
l_out = torch.cat((l1_2, l2_2), 1)
return l_out
def training_pro2(train_loader):
net = []
layer_num = 1
for i in range(layer_num):
net.append(MyLinear2())
my_model = nn.Sequential(MyLinear1( ), *net, MyLinear3( ))
learning_rate = 10e-4
optimizer = SGD(my_model.parameters(), lr=learning_rate, weight_decay=10e-4)
epoch_num = 100
# test_data
# theta1_data_test, theta2_data_test, y_data_test = data_read(path = "test_data.csv")
theta1_data_test, theta2_data_test, y_data_test = generate_data(1)
y_pred1 = []
y_true1 = [ y_data_test[0][0]] * epoch_num
# print("y_true:", y_data_test[0])
print("y_true1:", y_true1)
y_pred2 = []
y_true2 = [y_data_test[0][1]] * epoch_num
print("y_true2:", y_true2)
test_x1t = torch.tensor(theta1_data_test)
test_x2t = torch.tensor(theta2_data_test)
for epoch in range(epoch_num):
print("epoch:", epoch)
train_loss = 0
train_num = 0
for step,(b_x1,b_x2,b_y) in enumerate(train_loader):
# print("step:", step)
x = torch.cat((b_x1,b_x2),1)
output = my_model(x)
# print("output:", output)
# print("b_y:", b_y)
# loss = loss_func(output,b_y)
loss = custom_mse(output, b_y)
# print("loss:", loss)
optimizer.zero_grad()
loss.backward()
optimizer.step()
# print("loss:", loss)
# print("loss.item:", loss.item())
train_loss += loss.item()
train_num += train_loss
print("train_loss:", train_loss / train_num)
# single_predict = mlpreg(s_feat_st)
test_x = torch.cat((test_x1t, test_x2t), 1)
single_predict = my_model(test_x)
p = single_predict.data.numpy()
# print("p:", p)
y_pred1.append( p[0][0])
y_pred2.append( p[0][1])
# print("y_pred:", y_pred)
plt.plot(y_pred1, label='y1_pred', c="g", marker="o", linewidth=2)
plt.plot(y_true1, label='y1_true', c="red", marker="o", linewidth=2)
plt.plot(y_pred2, label='y2_pred', c="blue", linewidth=2)
plt.plot(y_true2, label='y2_true', c="red", linewidth=2)
plt.xlabel(r'$x$')
plt.ylabel(r'$y$')
plt.legend()
plt.show()
def custom_mse(predicted, target):
total_mse = 0
# print("predicted:", predicted)
for i in range(target.shape[1]):
# print("predicted[i]:", predicted.T[i])
total_mse+=nn.MSELoss()(predicted.T[i], target.T[i])
return total_mse
def training_data_generate():
data_num = 1000
theta1_data, theta2_data, y_data = generate_data(data_num)
print("len:", len(theta1_data))
train_x1t = torch.tensor(theta1_data)
train_x2t = torch.tensor(theta2_data)
train_yt = torch.tensor(y_data)
with open("training_data.csv", "w") as csvfile:
writer = csv.writer(csvfile)
for i in range(data_num):
writer.writerow(theta1_data[i] + theta2_data[i] + y_data[i])
# print("theta1_data:",theta1_data)
return train_x1t, train_x2t, train_yt
def test_data_generate():
data_num = 1
theta1_data, theta2_data, y_data = generate_data(data_num)
print("len:", len(theta1_data))
train_x1t = torch.tensor(theta1_data)
train_x2t = torch.tensor(theta2_data)
train_yt = torch.tensor(y_data)
with open("test_data.csv", "w") as csvfile:
writer = csv.writer(csvfile)
for i in range(data_num):
writer.writerow(theta1_data[i] + theta2_data[i] + y_data[i])
# print("theta1_data:",theta1_data)
return train_x1t, train_x2t, train_yt
def data_read(path = "training_data.csv"):
csvFile = open(path, "r")
reader = csv.reader(csvFile)
train_x1t = []
train_x2t = []
train_yt = []
for item in reader:
train_x1t.append([float(x) for x in item[:3]])
train_x2t.append([float(x) for x in item[3:6]])
train_yt.append([float(x) for x in item[-2:]])
csvFile.close()
print("train_x1t:", train_x1t)
print("train_x2t:", train_x2t)
print("train_yt:", train_yt)
return torch.tensor(train_x1t), torch.tensor(train_x2t), torch.tensor(train_yt)
if __name__ == '__main__':
data_num = 1000
theta1_data, theta2_data, y_data = generate_data(data_num)
print("len:", len(theta1_data))
train_x1t = torch.tensor(theta1_data)
train_x2t = torch.tensor(theta2_data)
train_yt = torch.tensor(y_data)
# train_x1t, train_x2t, train_yt = data_read()
train_data = Data.TensorDataset(train_x1t, train_x2t, train_yt)
train_loader = Data.DataLoader(dataset=train_data, batch_size=50, shuffle=True, num_workers=0)
#
# training_progress(train_loader)
training_pro2(train_loader)