CNN: Deep Network Example

 

 

tf.nn.conv2dexpress

這個函數的功能是:給定4維的input和filter,計算出一個2維的卷積結果。函數的定義爲:apache

def conv2d(input, filter, strides, padding, use_cudnn_on_gpu=None,
           data_format=None, name=None):

前幾個參數分別是input, filter, strides, padding, use_cudnn_on_gpu, …下面來一一解釋
input:待卷積的數據。格式要求爲一個張量,[batch, in_height, in_width, in_channels].
分別表示 批次數,圖像高度,寬度,輸入通道數。
filter: 卷積核。格式要求爲[filter_height, filter_width, in_channels, out_channels].
分別表示 卷積核的高度,寬度,輸入通道數,輸出通道數。
strides :一個長爲4的list. 表示每次卷積之後卷積窗口在input中滑動的距離
padding :有SAME和VALID兩種選項,表示是否要保留圖像邊上那一圈不徹底卷積的部分。若是是SAME,則保留
use_cudnn_on_gpu :是否使用cudnn加速。默認是Trueapp

 

tf.nn.max_pool
進行最大值池化操做,而avg_pool 則進行平均值池化操做.函數的定義爲:less

def max_pool(value, ksize, strides, padding, data_format="NHWC", name=None):

value: 一個4D張量,格式爲[batch, height, width, channels],與conv2d中input格式同樣
ksize: 長爲4的list,表示池化窗口的尺寸
strides: 池化窗口的滑動值,與conv2d中的同樣
padding: 與conv2d中用法同樣。ide

# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""A very simple MNIST classifier.
See extensive documentation at
http://tensorflow.org/tutorials/mnist/beginners/index.md
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

# Import data
from tensorflow.examples.tutorials.mnist import input_data

import tensorflow as tf

flags = tf.app.flags
FLAGS = flags.FLAGS
flags.DEFINE_string('data_dir', '/tmp/data/', 'Directory for storing data') # 第一次啓動會下載文本資料,放在/tmp/data文件夾下

print(FLAGS.data_dir)
mnist = input_data.read_data_sets(FLAGS.data_dir, one_hot=True)

def weight_variable(shape):
    initial = tf.truncated_normal(shape, stddev=0.1) # 變量的初始值爲截斷正太分佈
    return tf.Variable(initial)

def bias_variable(shape):
    initial = tf.constant(0.1, shape=shape)
    return tf.Variable(initial)

def conv2d(x, W):
    """
    tf.nn.conv2d功能:給定4維的input和filter,計算出一個2維的卷積結果
    前幾個參數分別是input, filter, strides, padding, use_cudnn_on_gpu, ...
    input   的格式要求爲一個張量,[batch, in_height, in_width, in_channels],批次數,圖像高度,圖像寬度,通道數
    filter  的格式爲[filter_height, filter_width, in_channels, out_channels],濾波器高度,寬度,輸入通道數,輸出通道數
    strides 一個長爲4的list. 表示每次卷積之後在input中滑動的距離
    padding 有SAME和VALID兩種選項,表示是否要保留不徹底卷積的部分。若是是SAME,則保留
    use_cudnn_on_gpu 是否使用cudnn加速。默認是True
    """
    return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool_2x2(x):
    """
    tf.nn.max_pool 進行最大值池化操做,而avg_pool 則進行平均值池化操做
    幾個參數分別是:value, ksize, strides, padding,
    value:  一個4D張量,格式爲[batch, height, width, channels],與conv2d中input格式同樣
    ksize:  長爲4的list,表示池化窗口的尺寸
    strides: 窗口的滑動值,與conv2d中的同樣
    padding: 與conv2d中用法同樣。
    """
    return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],
                          strides=[1, 2, 2, 1], padding='SAME')

sess = tf.InteractiveSession()

x = tf.placeholder(tf.float32, [None, 784])
x_image = tf.reshape(x, [-1,28,28,1]) #將輸入按照 conv2d中input的格式來reshape,reshape

"""
# 第一層
# 卷積核(filter)的尺寸是5*5, 通道數爲1,輸出通道爲32,即feature map 數目爲32
# 又由於strides=[1,1,1,1] 因此單個通道的輸出尺寸應該跟輸入圖像同樣。即總的卷積輸出應該爲?*28*28*32
# 也就是單個通道輸出爲28*28,共有32個通道,共有?個批次
# 在池化階段,ksize=[1,2,2,1] 那麼卷積結果通過池化之後的結果,其尺寸應該是?*14*14*32
"""
W_conv1 = weight_variable([5, 5, 1, 32])  # 卷積是在每一個5*5的patch中算出32個特徵,分別是patch大小,輸入通道數目,輸出通道數目
b_conv1 = bias_variable([32])
h_conv1 = tf.nn.elu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)

"""
# 第二層
# 卷積核5*5,輸入通道爲32,輸出通道爲64。
# 卷積前圖像的尺寸爲 ?*14*14*32, 卷積後爲?*14*14*64
# 池化後,輸出的圖像尺寸爲?*7*7*64
"""
W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])
h_conv2 = tf.nn.elu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)

# 第三層 是個全鏈接層,輸入維數7*7*64, 輸出維數爲1024
W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])
h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.elu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
keep_prob = tf.placeholder(tf.float32) # 這裏使用了drop out,即隨機安排一些cell輸出值爲0,能夠防止過擬合
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

# 第四層,輸入1024維,輸出10維,也就是具體的0~9分類
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])
y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2) # 使用softmax做爲多分類激活函數
y_ = tf.placeholder(tf.float32, [None, 10])

cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(y_conv), reduction_indices=[1])) # 損失函數,交叉熵
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy) # 使用adam優化
correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1)) # 計算準確度
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
sess.run(tf.initialize_all_variables()) # 變量初始化
for i in range(20000):
    batch = mnist.train.next_batch(50)
    if i%100 == 0:
        # print(batch[1].shape)
        train_accuracy = accuracy.eval(feed_dict={
            x:batch[0], y_: batch[1], keep_prob: 1.0})
        print("step %d, training accuracy %g"%(i, train_accuracy))
    train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})

print("test accuracy %g"%accuracy.eval(feed_dict={
    x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))
相關文章
相關標籤/搜索