前情概要
上一次我们跟随着莫烦的视频用tensorflow写了一个简单的神经网络并将结果可视化了,详见
暑假补强-2:用tensorflow写简单的神经网络并将结果可视化(跟着莫烦python的视频学习)
接下来,我们对原有的代码进行一下改动,得到用tensorboard画出酷酷的整个神经网络的结构图以及参数变化图
完整代码
import tensorflow as tf
import numpy as np
def add_layer(input_data, in_size, out_size, n_layer, activation_function=None):
layer_name = 'layer' + str(n_layer)
with tf.name_scope(layer_name):
Weights = tf.Variable(tf.random_normal([in_size, out_size]), name='W')
tf.summary.histogram('Weights', Weights)
biases = tf.Variable(tf.zeros([1, out_size]) + 0.1, name='b')
tf.summary.histogram('biases', biases)
with tf.name_scope('Wx_plus_b'):
Wx_plus_b = tf.matmul(input_data, Weights) + biases
if activation_function is None:
outputs = Wx_plus_b
else:
outputs = activation_function(Wx_plus_b)
tf.summary.histogram('outputs', outputs)
return outputs
x_data = np.linspace(-1, 1, 300, dtype=np.float32)[:,np.newaxis]
noise = np.random.normal(0, 0.05, x_data.shape).astype(np.float32)
y_data = np.square(x_data) - 0.5 + noise
with tf.name_scope("inputs"):
xs = tf.placeholder(tf.float32, [None, 1], name='x_input')
ys = tf.placeholder(tf.float32, [None, 1], name='y_input')
l1 = add_layer(xs, 1, 10, 1,activation_function=tf.nn.relu)
prediction = add_layer(l1, 10, 1, 2,activation_function=None)
with tf.name_scope("loss"):
loss = tf.reduce_mean(tf.reduce_sum(tf.square(ys-prediction, name='pingfang'), reduction_indices=[1], name="reduce_sum"), name="reduce_mean")
tf.summary.scalar('loss', loss)
with tf.name_scope('train'):
train_step = tf.train.ProximalGradientDescentOptimizer(0.1).minimize(loss)
sess = tf.Session()
merged = tf.summary.merge_all()
writer = tf.summary.FileWriter('graph/', sess.graph)
init = tf.global_variables_initializer()
sess.run(init)
for i in range(1000):
sess.run(train_step, feed_dict={xs : x_data, ys : y_data})
if i % 50 == 0:
result = sess.run(merged, feed_dict={xs: x_data, ys: y_data})
writer.add_summary(result, i)
代码剖析
- 首先 看一下add_layer这个部分
def add_layer(input_data, in_size, out_size, n_layer, activation_function=None):
layer_name = 'layer' + str(n_layer)
with tf.name_scope(layer_name):
Weights = tf.Variable(tf.random_normal([in_size, out_size]), name='W')
tf.summary.histogram('Weights', Weights)
biases = tf.Variable(tf.zeros([1, out_size]) + 0.1, name='b')
tf.summary.histogram('biases', biases)
with tf.name_scope('Wx_plus_b'):
Wx_plus_b = tf.matmul(input_data, Weights) + biases
if activation_function is None:
outputs = Wx_plus_b
else:
outputs = activation_function(Wx_plus_b)
tf.summary.histogram('outputs', outputs)
return outputs
这个函数的参数里新加了一个n_layer,是为了区分不同的层,
如图
给每一层起了名字以后,在这里面的参数都属于这一层
代码中,把Weights命名成W,biases命名成b,并且对Weights,biases和outputs的变化画进了一个直方图里
- 再把loss的变化反映到scalars图中
with tf.name_scope("loss"):
loss = tf.reduce_mean(tf.reduce_sum(tf.square(ys-prediction, name='pingfang'), reduction_indices=[1], name="reduce_sum"), name="reduce_mean")
tf.summary.scalar('loss', loss)
这个loss的变化看起来还有点怪异,中间暴涨的原因可能是学习步长太长,突然偏离了正确值。
- 建立会话,将所有的summary整合到一起,并将结果写入到指定目录下
sess = tf.Session()
merged = tf.summary.merge_all() # 整合
writer = tf.summary.FileWriter('graph/', sess.graph) # 写入到graph目录下
init = tf.global_variables_initializer()
sess.run(init)
- 训练
add_summary()中的i很重要!i对应的是global_step参数,如果没有i,那么scalars画出来的图就只是一条直线。
for i in range(1000):
sess.run(train_step, feed_dict={xs:x_data, ys:y_data})
# 每50次将结果整合画图
if i % 50: == 0:
result = sess.run(merged, feed_dict{xs:x_data, ys:y_data})
writer.add_summary(result, i)
获得结果
将代码运行一次以后,在graph目录下就能得到一个文件。
此时,我们要从命令行窗口运行这个文件
- 进入conda环境
-
进入graph所在目录
-
输入并复制地址
输入tensorboard --logdir=graph --host localhost --port 8088
-
用谷歌浏览器打开
http://localhost:8088/
成果