-
Notifications
You must be signed in to change notification settings - Fork 0
/
SMNN.py
242 lines (200 loc) · 9.22 KB
/
SMNN.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
import numpy as np
import pandas as pd
from sklearn.preprocessing import StandardScaler
import glob
from matplotlib import pyplot as plt
from keras.layers import Dense
from keras.wrappers.scikit_learn import KerasRegressor
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import KFold
from sklearn.pipeline import Pipeline
from tkinter import Tk
from tkinter.filedialog import askdirectory
import os
import time
from keras.layers import Conv2D, Add, SeparableConv2D, BatchNormalization, Activation, Flatten, MaxPooling2D, Input, AveragePooling2D, Reshape, Dropout, LeakyReLU, UpSampling2D
from sklearn.metrics import r2_score, accuracy_score
from keras.models import Model, load_model
from keras import backend as K
from keras.optimizers import adam
start_time = time.time()
#define useful parameters
features = 1011
group = 8
copies = 32
activation = 0.25
dropout = 0.3
test = False
patience = 4
def f7(seq):
seen = set()
seen_add = seen.add
return [x for x in seq if not (x in seen or seen_add(x))]
# import data
#filename = askdirectory()
filename = "/home/ragan/Downloads/chip_test/chip18"
file_list = sorted(glob.glob(filename + '/*.txt'))
list_array = sorted(glob.glob(filename + '/*.txt'))
#validation data
#filename_validation = askdirectory()
filename_validation = "/home/ragan/Downloads/chip_test/chip18"
file_list_validation = sorted(glob.glob(filename_validation + '/*.txt'))
list_array_validation = sorted(glob.glob(filename_validation + '/*.txt'))
#load data
X_validation_new = np.load("/home/ragan/Downloads/chip_test/chip15/matrix/x_val.npy")
y_validation_new = np.load("/home/ragan/Downloads/chip_test/chip15/matrix/y_val.npy")
X_validation_tensor = np.load("/home/ragan/Downloads/chip_test/chip15/matrix/x_val.npy")
y_validation_bundle = np.load("/home/ragan/Downloads/chip_test/chip15/matrix/y_val.npy")
def coeff_determination(y_true, y_pred):
SS_res = K.sum(K.square( y_true-y_pred ))
SS_tot = K.sum(K.square( y_true - K.mean(y_true) ) )
return ( 1 - SS_res/(SS_tot + K.epsilon()) )
print("data import --- %s seconds ---" % (time.time() - start_time))
#neural net
input = Input(shape=(group, group, features))
x = Dropout(dropout)(input)
x = Dense(512, kernel_initializer='normal', activation='linear')(input)
x = LeakyReLU(alpha=activation)(x)
x = Dropout(dropout)(x)
x = Dense(256, kernel_initializer='normal', activation='linear')(x)
x = LeakyReLU(alpha=activation)(x)
x = Dropout(dropout)(x)
x = Dense(128, kernel_initializer='normal', activation='linear')(x)
x = LeakyReLU(alpha=activation)(x)
x = Dropout(dropout)(x)
x = Conv2D(32, kernel_size=(3,3), strides=(1,1), activation='linear', padding='same')(x)
x = LeakyReLU(alpha=activation)(x)
x = Conv2D(32, kernel_size=(3, 3), strides=(1, 1), activation='linear', padding='same')(x)
x = LeakyReLU(alpha=activation)(x)
x = MaxPooling2D(pool_size=(2,2))(x)
x = Conv2D(64, kernel_size=(3,3), strides=(1,1), activation='linear', padding='same')(x)
x = LeakyReLU(alpha=activation)(x)
x = Conv2D(64, kernel_size=(3, 3), strides=(1, 1), activation='linear', padding='same')(x)
x = LeakyReLU(alpha=activation)(x)
x = SeparableConv2D(256, (2, 2), padding='same', use_bias=False)(x)
x = Activation('relu')(x)
x = SeparableConv2D(256, (2, 2), padding='same', use_bias=False)(x)
x = Activation('relu')(x)
for i in range(8):
residual = x
x = Activation('relu')(x)
x = SeparableConv2D(256, (2, 2), padding='same', use_bias=False)(x)
x = Activation('relu')(x)
x = SeparableConv2D(256, (2, 2), padding='same', use_bias=False)(x)
x = Activation('relu')(x)
x = SeparableConv2D(256, (2, 2), padding='same', use_bias=False)(x)
x = Add()([x, residual])
x = Activation('relu')(x)
x = MaxPooling2D(pool_size=(2,2))(x)
x = Flatten()(x)
x = Dense(16, kernel_initializer='normal', activation='linear')(x)
x = LeakyReLU(alpha=activation)(x)
x = Dropout(dropout)(x)
output = Dense(1, kernel_initializer='normal')(x)
model = Model(inputs=input, outputs=output)
#optimizer = adam(lr=1e-4)
model.compile(loss='mean_squared_error', optimizer="adam", metrics=['acc', coeff_determination])
model.summary()
#Train model, early stopping, history
best_loss_validation = 10000000
best_acc_validation = -1
loss_history_train = []
acc_history_train = []
r2_history_train = []
loss_history_validation = []
acc_history_validation = []
r2_history_validation = []
for i in range(1000):
for j in range(copies):
loss_list_train = []
acc_list_train = []
r2_list_train = []
loss_list_validation = []
acc_list_validation = []
r2_list_validation = []
history_train = model.train_on_batch(np.load(f"/home/ragan/Downloads/chip_test/chip14/matrix/x_{j}.npy"), np.load(f"/home/ragan/Downloads/chip_test/chip14/matrix/y_{j}.npy"))
history_validation = model.test_on_batch(X_validation_tensor, y_validation_bundle)
loss_list_train = np.append(loss_list_train, history_train[0])
acc_list_train = np.append(acc_list_train, history_train[1])
r2_list_train = np.append(r2_list_train, history_train[2])
loss_list_validation = np.append(loss_list_validation, history_validation[0])
acc_list_validation = np.append(acc_list_validation, history_validation[1])
r2_list_validation = np.append(r2_list_validation, history_validation[2])
mean_loss_train = sum(loss_list_train) / len(loss_list_train)
loss_history_train = np.append(loss_history_train, mean_loss_train)
mean_acc_train = sum(acc_list_train) / len(acc_list_train)
acc_history_train = np.append(acc_history_train, mean_acc_train)
mean_r2_train = sum(r2_list_train) / len(r2_list_train)
r2_history_train = np.append(r2_history_train, mean_r2_train)
mean_loss_validation = sum(loss_list_validation) / len(loss_list_validation)
loss_history_validation = np.append(loss_history_validation, mean_loss_validation)
mean_acc_validation = sum(acc_list_validation) / len(acc_list_validation)
acc_history_validation = np.append(acc_history_validation, mean_acc_validation)
mean_r2_validation = sum(r2_list_validation) / len(r2_list_validation)
r2_history_validation = np.append(r2_history_validation, mean_r2_validation)
if mean_loss_validation < best_loss_validation:
best_loss_validation = mean_loss_validation
rounds_without_improvement = 0
model.save("/home/ragan/Downloads/r_800_new/results/best_model.h5")
print(mean_loss_validation, mean_acc_validation, mean_r2_validation)
else:
rounds_without_improvement +=1
print("No Improvement")
if rounds_without_improvement == patience:
break
print("training NN--- %s seconds ---" % (time.time() - start_time))
#prepare validation data for errorplot
model = load_model("/home/ragan/Downloads/r_800_new/results/best_model.h5", custom_objects={'coeff_determination': coeff_determination})
predictions_validation = model.predict(X_validation_new)
loss_validation = model.evaluate(X_validation_new, y_validation_new)
print(loss_validation)
guess_validation = np.column_stack((predictions_validation, y_validation_new))
e_one_validation = []
e_validation = []
m_one_validation = []
m_validation = []
for i in range(0, len(list_array_validation)):
e_one_validation = np.std([predictions_validation for (predictions_validation,y_validation_new) in guess_validation if y_validation_new == int(os.path.basename(file_list_validation[i])[3:5])])
e_validation = np.append(e_validation, e_one_validation)
m_one_validation = np.mean([predictions_validation for (predictions_validation, y_validation_new) in guess_validation if y_validation_new == int(os.path.basename(file_list_validation[i])[3:5])])
m_validation = np.append(m_validation, m_one_validation)
real_validation = []
for i in range(0, len(list_array_validation)):
real_validation = np.append(real_validation, int(os.path.basename(file_list_validation[i])[3:5]))
for i in range(0, len(list_array_validation)):
print(m_validation[i],e_validation[i])
validation_r2 = format(r2_score(y_validation_new, predictions_validation), ".2f")
print("validation --- %s seconds ---" % (time.time() - start_time))
#np.save("/home/ragan/Downloads/chip_test/chip4/results/lowest_loss", loss_validation[0])
#plot error vs epoch
plt.subplot(311)
plt.plot(loss_history_train)
plt.plot(loss_history_validation)
plt.title('Model Accuracy/Loss')
plt.ylabel('Loss')
plt.subplot(312)
plt.plot(acc_history_train)
plt.plot(acc_history_validation)
plt.ylabel('Accuracy')
plt.subplot(313)
plt.plot(r2_history_train)
plt.plot(r2_history_validation)
plt.ylim(-1,1)
plt.ylabel('r2')
plt.xlabel('epoch')
#plot predictions
plt.figure(num=None, figsize=(8, 6), dpi=80, facecolor='w', edgecolor='k')
x_labels = []
for i in range(0, len(list_array_validation)):
x_labels = np.append(x_labels, os.path.basename(file_list_validation[i])[-10:-3])
labels = f7(x_labels)
plt.xticks(np.unique(real_validation), labels)
plt.yticks(np.unique(real_validation), labels)
plt.ylabel('Predicted Concentration', fontsize=16)
plt.xlabel('Real Concentration', fontsize=16)
plt.title(f'Methylene Blue', fontsize=20)
plt.margins(tight=False)
plt.errorbar(real_validation, m_validation, yerr=e_validation, fmt='o', label=f'Validation Data r^2: {validation_r2}')
plt.plot((0,1,2,3,4,5,6,7,8,9), label='Ideal Fit')
legend = plt.legend(loc='upper left', shadow=True, fontsize=14)
plt.show()