2018.05.30 Dropout程序精读_plot dropout-优快云博客

本文链接：https://blog.youkuaiyun.com/o0haidee0o/article/details/80516547

本文通过一个具体的实验案例，展示了如何利用Dropout技术来防止神经网络训练过程中的过拟合现象。实验中设置了两种不同的网络结构进行对比，一种为普通的全连接层网络，另一种则加入了Dropout层。通过可视化训练过程中的损失变化，直观地说明了Dropout的有效性。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

生成的实验数据：

实验结果：

实验代码：

先附链接吧：

https://download.youkuaiyun.com/download/o0haidee0o/10448563

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

#tf.set_random_seed(1)
#np.random.seed(1)#这两句没有也没毛病，所以暂时没搞懂干嘛的

# Hyper parameters超参数
N_SAMPLES = 20#20个样本
N_HIDDEN = 300#300个hidden layer
LR = 0.01#learning rate=0.01

# training data
x = np.linspace(-1, 1, N_SAMPLES)[:, np.newaxis]
#np.linspace(linear space 线性等分向量):产生20个[-1,1]线性等分的随机数
##print(x)
y = x + 0.3*np.random.randn(N_SAMPLES)[:, np.newaxis]#正态分布随机20个数*0.3+x

# test data x还是training data, y由于是随机数，所以有微小变化
test_x = x.copy()
test_y = test_x + 0.3*np.random.randn(N_SAMPLES)[:, np.newaxis]

# show data
plt.ion()   # something about plotting打开交互模式
plt.figure()
plt.scatter(x, y, c='magenta', s=50, alpha=0.5, label='train')#magenta 洋红
plt.scatter(test_x, test_y, c='cyan', s=50, alpha=0.5, label='test')#cyan 亮蓝色
plt.legend(loc='upper left')
plt.ylim((-2.5, 2.5))#ylim = y limited 限制y轴显示区域
##plt.show()

# tf placeholders
tf_x = tf.placeholder(tf.float32, [None, 1])
tf_y = tf.placeholder(tf.float32, [None, 1])
tf_is_training = tf.placeholder(tf.bool, None)
# to control dropout when training and testing，training需要dropout，但testing不需要dropout

# overfitting net
o1 = tf.layers.dense(tf_x, N_HIDDEN, tf.nn.relu)
o2 = tf.layers.dense(o1, N_HIDDEN, tf.nn.relu)
o_out = tf.layers.dense(o2, 1)
o_loss = tf.losses.mean_squared_error(tf_y, o_out)
o_train = tf.train.AdamOptimizer(LR).minimize(o_loss)#LR=learning rate

# dropout net
d1 = tf.layers.dense(tf_x, N_HIDDEN, tf.nn.relu)
d1 = tf.layers.dropout(d1, rate=0.5, training=tf_is_training)   # drop out 50% of inputs
d2 = tf.layers.dense(d1, N_HIDDEN, tf.nn.relu)
d2 = tf.layers.dropout(d2, rate=0.5, training=tf_is_training)   # drop out 50% of inputs
d_out = tf.layers.dense(d2, 1)
d_loss = tf.losses.mean_squared_error(tf_y, d_out)
d_train = tf.train.AdamOptimizer(LR).minimize(d_loss)

sess = tf.Session()
sess.run(tf.global_variables_initializer())#虽然没有Variables但是这句话还是要有，不然就一群错误，唉

##plt.ion()   # something about plotting打开交互模式
plt.figure()
for t in range(500):
    sess.run([o_train, d_train], {tf_x: x, tf_y: y, tf_is_training: True})  # train, set is_training=True

    if t % 10 == 0:
        # plotting
        plt.cla()
        o_loss_, d_loss_, o_out_, d_out_ = sess.run(
            [o_loss, d_loss, o_out, d_out], {tf_x: test_x, tf_y: test_y, tf_is_training: False} )
        
        plt.scatter(x, y, c='magenta', s=50, alpha=0.3, label='train')
        plt.scatter(test_x, test_y, c='cyan', s=50, alpha=0.3, label='test')
        plt.plot(test_x, o_out_, 'r-', lw=3, label='overfitting')
        plt.plot(test_x, d_out_, 'b--', lw=3, label='dropout(50%)')
        plt.text(-0.5, -1.6, 'overfitting loss=%.4f' % o_loss_, fontdict={'size': 20,'color':'red'});
        plt.text(-0.5, -2, 'dropout loss=%.4f' % d_loss_, fontdict={'size': 20, 'color': 'blue'})
        plt.legend(loc='upper left')
        plt.ylim((-2.5, 2.5))
        plt.pause(0.1)

plt.ioff()
plt.show()