tensorflow与感知机

TensorFlow 支持自动求导，可以使用 TensorFlow 优化器来计算和使用梯度。它使用梯度自动更新用变量定义的张量。本文将使用 TensorFlow 优化器来训练网络。将使用  Contrib（层）来定义神经网络层，可以用来添加各种层到神经网络模型，如添加构建块。这里使用的一个方法是 tf.contrib.layers.fully_connected及使用 TensorFlow 自带的优化器来计算和使用梯度。

1.Tensorflow多层感知机实现MNIST分类

import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'
import tensorflow as tf
import tensorflow.contrib.layers as layers
from tensorflow.examples.tutorials.mnist import input_data

mnist = input_data.read_data_sets('./data/mnist_data', one_hot=True)

batch_size = 200
eta = 0.001
max_epoch = 50
n_hidden = 30
n_classes = 10
n_input = 784

def multilayer_perceptron(x):
    fc1 = layers.fully_connected(x, n_hidden, activation_fn=tf.nn.relu)
    out = layers.fully_connected(fc1, n_classes, activation_fn=None)
    return out

x = tf.placeholder(tf.float32, [None, n_input])
y = tf.placeholder(tf.float32, [None, n_classes])
y_hat = multilayer_perceptron(x)

loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=y_hat, labels=y))
train = tf.train.AdamOptimizer(learning_rate=eta).minimize(loss)
correct_prediction = tf.equal(tf.argmax(y_hat,1), tf.argmax(y,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, dtype=tf.float32))
init = tf.global_variables_initializer()


with tf.Session() as sess:
    sess.run(init)
    for epoch in range(max_epoch):
        epoch_loss = 0.0
        batch_steps = int(mnist.train.num_examples/batch_size)
        for i in range(batch_steps):
            batch_x, batch_y = mnist.train.next_batch(batch_size)
            _, c = sess.run([train,loss], feed_dict={x:batch_x, y:batch_y})
        epoch_loss += c / batch_steps
        accur = sess.run(accuracy,feed_dict={x:mnist.test.images, y:mnist.test.labels}) 
        print('Epoch %02d, Loss = %.6f, Accuracy = %.6f' %(epoch, epoch_loss, accur))

            

2.tensorflow2.0与图片分类

from __future__ import absolute_import, division, print_function, unicode_literals
import tensorflow as tf
from tensorflow import keras
import numpy as np
import matplotlib.pyplot as plt

fashion_mnist = keras.datasets.fashion_mnist
(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()  # 得到4个Numpy Array
class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat', 
               'Sandal',      'Shirt',   'Sneaker',  'Bag',   'Ankle boot']

//数据的预处理：
#原始数据中图片的每个像素由[0, 255]区间上的整数表示。为了更好的训练模型，需要将所有的值都标准化到区间[0, 1]
train_images = train_images / 255.0
test_images = test_images / 255.0

//模型建立
model = keras.Sequential([
     keras.layers.Flatten(input_shape=(28, 28)),
     keras.layers.Dense(128, activation='relu'),  #该层具有128个神经单元，激活函数为Relu，输出为一个长度为128的向量
     keras.layers.Dense(10)  #该层具有10个神经单元，未设置激活函数，输出为一个长度为10的向量，也是该网络的输出层.
 ])

//模型编译：
model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

//模型训练：
model.fit(train_images, train_labels, batch_size=32, epochs=10)

为了做预测，在模型的后面添加一个softmax层，将logits转换成概率，这样可解释性更好：

probability_model = tf.keras.Sequential([model, tf.keras.layers.Softmax()])

对部分结果进行可视化：

predictions = probability_model.predict(test_images)

def plot_image(i, predictions_array, true_label, img):
  predictions_array, true_label, img = predictions_array, true_label[i], img[i]
  plt.grid(False)
  plt.xticks([])
  plt.yticks([])

  plt.imshow(img, cmap=plt.cm.binary)

  predicted_label = np.argmax(predictions_array)
  if predicted_label == true_label:
    color = 'blue'
  else:
    color = 'red'

  plt.xlabel("{} {:2.0f}% ({})".format(class_names[predicted_label],
                                100*np.max(predictions_array),
                                class_names[true_label]),
                                color=color)

def plot_value_array(i, predictions_array, true_label):
  predictions_array, true_label = predictions_array, true_label[i]
  plt.grid(False)
  plt.xticks(range(10))
  plt.yticks([])
  thisplot = plt.bar(range(10), predictions_array, color="#777777")
  plt.ylim([0, 1])
  predicted_label = np.argmax(predictions_array)

  thisplot[predicted_label].set_color('red')
  thisplot[true_label].set_color('blue')

画出前15个样本：

num_rows = 5
num_cols = 3
num_images = num_rows*num_cols
plt.figure(figsize=(2*2*num_cols, 2*num_rows))
for i in range(num_images):
  plt.subplot(num_rows, 2*num_cols, 2*i+1)
  plot_image(i, predictions[i], test_labels, test_images)
  plt.subplot(num_rows, 2*num_cols, 2*i+2)
  plot_value_array(i, predictions[i], test_labels)
plt.tight_layout()
plt.savefig('demo_15_img2.png', dpi=100)
plt.show()