-
Notifications
You must be signed in to change notification settings - Fork 0
/
mnist_bn0.py
89 lines (79 loc) · 3.56 KB
/
mnist_bn0.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
#!/usr/bin/python
#!coding=utf-8
import tensorflow as tf
import tensorflow.contrib.layers as tcl
from tensorflow.examples.tutorials.mnist import input_data
import sys
"""
TF version:1.3.1; python version: 2.7.9; GPU P40
This file aims to compare the batch_norm layer's effect.
Really find something different!
Run this with mnist_bn0.py
My code is not clean, but I will keep going!
STEP=49, loss is 5.37903255939, acc is 0.82360001497
STEP=99, loss is 1.79672749698, acc is 0.93960001111
[TEST]step=99, || loss=20.4746034718, acc=0.168099994585
STEP=149, loss is 1.64993903732, acc is 0.95000000596
STEP=199, loss is 1.5875512296, acc is 0.954400001764
[TEST]step=199, || loss=30.0848628616, acc=0.256899989545
"""
mnist = input_data.read_data_sets("MNIST_data", one_hot=True)
x = tf.placeholder(dtype=tf.float32, shape=[None, 784])
y = tf.placeholder(dtype=tf.float32, shape=[None, 10])
phase = tf.placeholder(dtype=tf.bool)
drop = tf.placeholder(dtype=tf.float32)
input_x = tf.reshape(x, [-1, 28, 28, 1])
with tf.variable_scope('mnist'):
a = tcl.conv2d(input_x, 16, [3, 3], [1, 1], padding='SAME')
#a_bn = tf.layers.batch_normalization(a, training=phase)
#a_relu = tf.nn.relu(a_bn)
b = tcl.conv2d(input_x, 16, [5, 5], [1, 1], padding='SAME')
#b_bn = tf.layers.batch_normalization(b, training=phase)
#b_relu = tf.nn.relu(b_bn)
c = tf.concat([a, b], axis=3)
a2 = tcl.conv2d(c, 64, [3, 3], [1, 1], padding='SAME')
#a2_bn = tf.layers.batch_normalization(a2, training=phase)
#a2_relu = tf.nn.relu(a2_bn)
b2 = tcl.conv2d(c, 64, [5, 5], [1, 1], padding='SAME')
#b2_bn = tf.layers.batch_normalization(b2, training=phase)
#b2_relu = tf.nn.relu(b2_bn)
c2 = tf.concat([a2, b2], axis=3)
print(c2.shape)
a3 = tcl.conv2d(c2, 128, [3, 3], [1, 1], padding="SAME")
#a3_bn = tf.layers.batch_normalization(a3, training=phase)
#a3_relu = tf.nn.relu(a3_bn)
b3 = tcl.conv2d(c2, 128, [5, 5], [1, 1], padding='SAME')
#b3_bn = tf.layers.batch_normalization(b3, training=phase)
#b3_relu = tf.nn.relu(b3_bn)
c3 = tf.concat([a3, b3], axis=3)
c3_flat = tcl.flatten(c3)
print(c3_flat.shape)
fc1 = tcl.fully_connected(c3_flat, 256)
fc1_drop = tf.nn.dropout(fc1, keep_prob=drop)
out = tcl.fully_connected(fc1_drop, 10, activation_fn=None)
loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=out, labels=y))
with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
train_op = tf.train.AdamOptimizer().minimize(loss)
correct = tf.equal(tf.argmax(out, 1), tf.argmax(y, 1))
acc = tf.reduce_mean(tf.cast(correct, dtype=tf.float32))
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
train_acc, train_loss = 0.0, 0.0
for i in range(10000):
batch_x, batch_y = mnist.train.next_batch(100)
_, _loss, _acc = sess.run([train_op, loss, acc], feed_dict={x: batch_x, y: batch_y, phase:True, drop:0.5})
train_acc += _acc
train_loss += _loss
if (i+1) % 50 == 0:
print('STEP={}, loss is {}, acc is {} '.format(i, train_loss/50, train_acc/50))
train_acc = 0.0
train_loss = 0.0
sys.stdout.flush()
if (i+1) % 100 == 0:
accuracy, val_loss = 0.0, 0.0
for j in range(100):
_x, _y = mnist.test.next_batch(100)
right, x_loss = sess.run([acc, loss], feed_dict={x:_x, y:_y, phase:False, drop:1.0})
accuracy += right
val_loss += x_loss
print('[TEST]step={}, || loss={}, acc={}'.format(i, val_loss/100, accuracy/100))