-
Notifications
You must be signed in to change notification settings - Fork 2
/
regression.py
143 lines (98 loc) · 3.63 KB
/
regression.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
import argparse
import time
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
# lr scheduler
from clr_callback import CyclicLR
# Make NumPy printouts easier to read.
np.set_printoptions(precision=3, suppress=True)
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior() # Enable tf v1 behavior as in v2 a lot have changed
# import dl_utils as utils
import newton_cg as es
print(tf.__version__)
parser = argparse.ArgumentParser(description='Keras regression case',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--optimizer', type=int, default=0, help='0 for ncg, 1 for adam, 2 for sgd')
args = parser.parse_args()
data = pd.read_csv("birth_rate.csv")
data.head()
print(data)
def sigmoid(x):
z = np.exp(-x)
sig = 1 / (1 + z)
return sig
# Split data/labels
data_X = np.array(data['Birth rate'])
data_X = np.array(range(10))*0.1-0.5# np.array(data['Birth rate'])
# data_X = np.array(data['Birth rate'])
#data_X = np.array(range(10))*0.1-0.5# np.array(data['Birth rate'])
data_Y = sigmoid(data_X) #p.array(data['Life expectancy'])
#data_Y = np.array(data['Life expectancy'])
print(data_X)
print(data_Y)
print(len(data))
#X = tf.placeholder(tf.float32, name='X')
#Y = tf.placeholder(tf.float32, name='Y')
#w = tf.get_variable('weight', initializer=tf.constant(0.1))
#b = tf.get_variable('bias', initializer=tf.constant(0.1))
#Y_hat = tf.add(tf.multiply(X,w), b)
#loss = (Y-Y_hat)*(Y-Y_hat) # tf.keras.losses.mean_squared_error(Y,Y_hat)
#loss = huber_loss(Y,Y_hat)
## Define gradient descent as the optimizer to minimise the loss
#optimizer = tf.train.GradientDescentOptimizer(0.001).minimize(loss)
##optimizer = es.EHNewtonOptimizer(0.001).minimize(loss)
#model = keras.Model(inputs=[X], outputs=Y, name=f'model{i}')
#model = tf.keras.Sequential([
# X,
# layers.Dense(units=1)
#])
clr_trig2 = CyclicLR(mode='exp_range', base_lr=0.00001, max_lr=0.0001, step_size=1, gamma=0.5) #gamma=0.9994)
model = tf.keras.Sequential()
model.add(tf.keras.layers.Dense(1,activation=tf.nn.sigmoid ))
model.build((None, 1))
print(model.get_weights())
if args.optimizer<1:
model.compile(optimizer=es.EHNewtonOptimizer(10), loss='mse')
elif args.optimizer<2:
model.compile(optimizer=tf.train.AdamOptimizer(1), loss='mse')
else:
model.compile(optimizer=tf.train.GradientDescentOptimizer(1), loss='mse')
sess = tf.Session()
#sess.run(tf.global_variables_initializer())
model.summary()
model.evaluate(data_X, data_Y)
#model.compile(optimizer=tf.train.GradientDescentOptimizer(1), loss='mse')
for i in range(10):
print(model.get_weights())
# adapt lr
#model.fit(data_X, data_Y, epochs=1, callbacks=[clr_trig2])
model.fit(data_X, data_Y, epochs=1)
print(model.trainable_variables)
print(model.get_weights())
#sess = tf.Session()
history = []
start = time.time()
# Graph variable initialization
sess.run(tf.global_variables_initializer())
# Open stream for tensorboard
#writer = tf.summary.FileWriter(logdir, sess.graph)
# Start training
#for i in range(50):
# total_loss = 0.0
# for x in range(len(data)):
# _, l = sess.run([optimizer,loss], feed_dict={X: data_X[x], Y:data_Y[x]})
# total_loss += l
# if (i) % 10 == 0:
# dw, db = sess.run([w,b])
# y_hat = data_X * dw + db
# history.append(y_hat)
# print('Epoch {0}: {1}, w: {2}, b: {3}'.format(i, total_loss/len(data), dw, db))
# print('dw: %f, db: %f\n' %(dw, db))
#writer.close()
#print('Train Time: %f seconds' %(time.time() - start))