# coding=utf-8 from __future__ import print_function import tensorflow as tf import numpy as np import matplotlib.pyplot as plt import os os.environ[‘TF_CPP_MIN_LOG_LEVEL‘] = ‘2‘ # ### 添加神经层 def add_layer(inputs, in_size, out_size, activation_function=None): with tf.name_scope(‘layer‘): # 使用with tf.name_scope定义图层,并指定在可视化图层中的显示名称 with tf.name_scope(‘weights‘): # 定义图层并指定名称,注意这里是上一图层的子图层 Weights = tf.Variable(tf.random_normal([in_size, out_size]), name=‘W‘) # 参数name指定名称 with tf.name_scope(‘biases‘): # 定义图层并指定名称 biases = tf.Variable(tf.zeros([1, out_size]) + 0.1, name=‘b‘) # 参数name指定名称 with tf.name_scope(‘Wx_plus_b‘): # 定义图层并指定名称 Wx_plus_b = tf.matmul(inputs, Weights) + biases if activation_function is None: outputs = Wx_plus_b else: outputs = activation_function(Wx_plus_b) return outputs # ### 构建数据 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 # ### 搭建网络 with tf.name_scope(‘inputs‘): # 定义图层并指定名称 xs = tf.placeholder(tf.float32, [None, 1], name=‘x_input‘) # 指定名称为x_input,也就是在可视化图层中的显示名称 ys = tf.placeholder(tf.float32, [None, 1], name=‘y_input‘) # 指定名称为y_input h1 = add_layer(xs, 1, 10, activation_function=tf.nn.relu) # 隐藏层 prediction = add_layer(h1, 10, 1, activation_function=None) # 输出层 with tf.name_scope(‘loss‘): # 定义图层并指定名称 loss = tf.reduce_mean(tf.reduce_sum(tf.square(ys - prediction), reduction_indices=[1])) with tf.name_scope(‘train‘): # 定义图层并指定名称 train_step = tf.train.GradientDescentOptimizer(0.1).minimize(loss) sess = tf.Session() writer = tf.summary.FileWriter("logs/", sess.graph) # 创建FileWriter对象和event文件,指定event文件的存放目录 init = tf.global_variables_initializer() sess.run(init) # ### 结果可视化 fig = plt.figure() ax = fig.add_subplot(1, 1, 1) ax.scatter(x_data, y_data) plt.ion() plt.show() # ### 训练 for i in range(1001): sess.run(train_step, feed_dict={xs: x_data, ys: y_data}) if i % 50 == 0: result = sess.run(loss, feed_dict={xs: x_data, ys: y_data}) print("Steps:{} Loss:{}".format(i, result)) try: ax.lines.remove(lines[0]) except Exception: pass prediction_value = sess.run(prediction, feed_dict={xs: x_data}) lines = ax.plot(x_data, prediction_value, ‘r-‘, lw=5) plt.pause(0.2) # ### TensorBoard # TensorFlow自带的可视化工具: # - 可视化学习:https://www.tensorflow.org/guide/summaries_and_tensorboard # - 图的直观展示:https://www.tensorflow.org/guide/graph_viz; # - 直方图信息中心:https://www.tensorflow.org/guide/tensorboard_histograms # 能够以直观的流程图的方式,清楚展示出整个神经网络的结构和框架,便于理解模型和发现问题; # 使用tf.name_scope()形成图层,图层名字就是方法里的参数; # # ### 启动TensorBoard # 使用命令“tensorboard --logdir=path/to/log-directory”(或者“python -m tensorboard.main”); # 参数logdir指向FileWriter将数据序列化的目录,建议在logdir上一级目录执行此命令; # TensorBoard运行后,在浏览器输入“localhost:6006”即可查看TensorBoard;
原文:https://www.cnblogs.com/anliven/p/10424915.html