Tensorflow安装速成教程

Linux平台安装tensorflow

sudo apt install python3 sudo apt install python3-pip

pip3 install tensorflow --upgrade

MacOS安装tensorflow

brew install python3

pip3 install tensorflow --upgrade

Windows平台安装Tensorflow

conda create -n tensorflow pip activate tensorflow pip install --ignore-installed --upgrade tensorflow

import tensorflow as tf a = tf.constant(1) b = tf.constant(2) c = a * b sess = tf.Session() print(sess.run(c))

Tensorflow顾名思义，是一些Tensor张量的流组成的运算。

Tensor("mul_1:0", shape=(), dtype=int32)

tf.multiply(X, w)

import tensorflow as tf import numpy as np trX = np.linspace(-1, 1, 101) trY = 2 * trX + np.random.randn(*trX.shape) * 0.33 # 创建一些线性值附近的随机值 X = tf.placeholder("float") Y = tf.placeholder("float") def model(X, w):    return tf.multiply(X, w) # X*w线性求值，非常简单 w = tf.Variable(0.0, name="weights") y_model = model(X, w) cost = tf.square(Y - y_model) # 用平方误差做为优化目标 train_op = tf.train.GradientDescentOptimizer(0.01).minimize(cost) # 梯度下降优化 # 开始创建Session干活！ with tf.Session() as sess:    # 首先需要初始化全局变量，这是Tensorflow的要求    tf.global_variables_initializer().run()    for i in range(100):        for (x, y) in zip(trX, trY):            sess.run(train_op, feed_dict={X: x, Y: y})    print(sess.run(w))

import tensorflow as tf import numpy as np from tensorflow.examples.tutorials.mnist import input_data def init_weights(shape):    return tf.Variable(tf.random_normal(shape, stddev=0.01)) def model(X, w):    return tf.matmul(X, w) # 模型还是矩阵乘法 mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels X = tf.placeholder("float", [None, 784]) Y = tf.placeholder("float", [None, 10]) w = init_weights([784, 10]) py_x = model(X, w) cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y)) # 计算误差 train_op = tf.train.GradientDescentOptimizer(0.05).minimize(cost) # construct optimizer predict_op = tf.argmax(py_x, 1) with tf.Session() as sess:    tf.global_variables_initializer().run()    for i in range(100):        for start, end in zip(range(0, len(trX), 128), range(128, len(trX)+1, 128)):            sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end]})        print(i, np.mean(np.argmax(teY, axis=1) ==                         sess.run(predict_op, feed_dict={X: teX})))

h = tf.nn.sigmoid(tf.matmul(X, w_h))    return tf.matmul(h, w_o)

import tensorflow as tf import numpy as np from tensorflow.examples.tutorials.mnist import input_data # 所有连接随机生成权值 def init_weights(shape):    return tf.Variable(tf.random_normal(shape, stddev=0.01)) def model(X, w_h, w_o):    h = tf.nn.sigmoid(tf.matmul(X, w_h))    return tf.matmul(h, w_o) mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels X = tf.placeholder("float", [None, 784]) Y = tf.placeholder("float", [None, 10]) w_h = init_weights([784, 625]) w_o = init_weights([625, 10]) py_x = model(X, w_h, w_o) cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y)) # 计算误差损失 train_op = tf.train.GradientDescentOptimizer(0.05).minimize(cost) # construct an optimizer predict_op = tf.argmax(py_x, 1) with tf.Session() as sess:    tf.global_variables_initializer().run()    for i in range(100):        for start, end in zip(range(0, len(trX), 128), range(128, len(trX)+1, 128)):            sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end]})        print(i, np.mean(np.argmax(teY, axis=1) ==                         sess.run(predict_op, feed_dict={X: teX})))

X = tf.nn.dropout(X, p_keep_input)    h = tf.nn.relu(tf.matmul(X, w_h))    h = tf.nn.dropout(h, p_keep_hidden)    h2 = tf.nn.relu(tf.matmul(h, w_h2))    h2 = tf.nn.dropout(h2, p_keep_hidden)    return tf.matmul(h2, w_o)

import tensorflow as tf import numpy as np from tensorflow.examples.tutorials.mnist import input_data def init_weights(shape):    return tf.Variable(tf.random_normal(shape, stddev=0.01)) def model(X, w_h, w_h2, w_o, p_keep_input, p_keep_hidden):    X = tf.nn.dropout(X, p_keep_input)    h = tf.nn.relu(tf.matmul(X, w_h))    h = tf.nn.dropout(h, p_keep_hidden)    h2 = tf.nn.relu(tf.matmul(h, w_h2))    h2 = tf.nn.dropout(h2, p_keep_hidden)    return tf.matmul(h2, w_o) mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels X = tf.placeholder("float", [None, 784]) Y = tf.placeholder("float", [None, 10]) w_h = init_weights([784, 625]) w_h2 = init_weights([625, 625]) w_o = init_weights([625, 10]) p_keep_input = tf.placeholder("float") p_keep_hidden = tf.placeholder("float") py_x = model(X, w_h, w_h2, w_o, p_keep_input, p_keep_hidden) cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y)) train_op = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cost) predict_op = tf.argmax(py_x, 1) with tf.Session() as sess:    # you need to initialize all variables    tf.global_variables_initializer().run()    for i in range(100):        for start, end in zip(range(0, len(trX), 128), range(128, len(trX)+1, 128)):            sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end],                                          p_keep_input: 0.8, p_keep_hidden: 0.5})        print(i, np.mean(np.argmax(teY, axis=1) ==                         sess.run(predict_op, feed_dict={X: teX,                                                         p_keep_input: 1.0,                                                         p_keep_hidden: 1.0})))

import tensorflow as tf import numpy as np from tensorflow.examples.tutorials.mnist import input_data batch_size = 128 test_size = 256 def init_weights(shape):    return tf.Variable(tf.random_normal(shape, stddev=0.01)) def model(X, w, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden):    l1a = tf.nn.relu(tf.nn.conv2d(X, w,                       # l1a shape=(?, 28, 28, 32)                        strides=[1, 1, 1, 1], padding='SAME'))    l1 = tf.nn.max_pool(l1a, ksize=[1, 2, 2, 1],              # l1 shape=(?, 14, 14, 32)                        strides=[1, 2, 2, 1], padding='SAME')    l1 = tf.nn.dropout(l1, p_keep_conv)    l2a = tf.nn.relu(tf.nn.conv2d(l1, w2,                     # l2a shape=(?, 14, 14, 64)                        strides=[1, 1, 1, 1], padding='SAME'))    l2 = tf.nn.max_pool(l2a, ksize=[1, 2, 2, 1],              # l2 shape=(?, 7, 7, 64)                        strides=[1, 2, 2, 1], padding='SAME')    l2 = tf.nn.dropout(l2, p_keep_conv)    l3a = tf.nn.relu(tf.nn.conv2d(l2, w3,                     # l3a shape=(?, 7, 7, 128)                        strides=[1, 1, 1, 1], padding='SAME'))    l3 = tf.nn.max_pool(l3a, ksize=[1, 2, 2, 1],              # l3 shape=(?, 4, 4, 128)                        strides=[1, 2, 2, 1], padding='SAME')    l3 = tf.reshape(l3, [-1, w4.get_shape().as_list()[0]])    # reshape to (?, 2048)    l3 = tf.nn.dropout(l3, p_keep_conv)    l4 = tf.nn.relu(tf.matmul(l3, w4))    l4 = tf.nn.dropout(l4, p_keep_hidden)    pyx = tf.matmul(l4, w_o)    return pyx mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) trX, trY, teX, teY = mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels trX = trX.reshape(-1, 28, 28, 1)  # 28x28x1 input img teX = teX.reshape(-1, 28, 28, 1)  # 28x28x1 input img X = tf.placeholder("float", [None, 28, 28, 1]) Y = tf.placeholder("float", [None, 10]) w = init_weights([3, 3, 1, 32])       # 3x3x1 conv, 32 outputs w2 = init_weights([3, 3, 32, 64])     # 3x3x32 conv, 64 outputs w3 = init_weights([3, 3, 64, 128])    # 3x3x32 conv, 128 outputs w4 = init_weights([128 * 4 * 4, 625]) # FC 128 * 4 * 4 inputs, 625 outputs w_o = init_weights([625, 10])         # FC 625 inputs, 10 outputs (labels) p_keep_conv = tf.placeholder("float") p_keep_hidden = tf.placeholder("float") py_x = model(X, w, w2, w3, w4, w_o, p_keep_conv, p_keep_hidden) cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=py_x, labels=Y)) train_op = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cost) predict_op = tf.argmax(py_x, 1) with tf.Session() as sess:    # you need to initialize all variables    tf.global_variables_initializer().run()    for i in range(100):        training_batch = zip(range(0, len(trX), batch_size),                             range(batch_size, len(trX)+1, batch_size))        for start, end in training_batch:            sess.run(train_op, feed_dict={X: trX[start:end], Y: trY[start:end],                                          p_keep_conv: 0.8, p_keep_hidden: 0.5})        test_indices = np.arange(len(teX)) # Get A Test Batch        np.random.shuffle(test_indices)        test_indices = test_indices[0:test_size]        print(i, np.mean(np.argmax(teY[test_indices], axis=1) ==                         sess.run(predict_op, feed_dict={X: teX[test_indices],                                                         p_keep_conv: 1.0,                                                         p_keep_hidden: 1.0})))

0 0.95703125 1 0.9921875 2 0.9921875 3 0.98046875 4 0.97265625 5 0.98828125 6 0.99609375

7 0.99609375 8 0.99609375 9 0.98828125 10 0.98828125 11 0.9921875 12 0.98046875 13 0.99609375 14 0.9921875 15 0.99609375 16 1.0