使用Estimators、Experiment高级API
from __future__ import division, print_function, absolute_import
# Import MNIST data，MNIST数据集导入
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/tmp/data/", one_hot=False)
import tensorflow as tf
import matplotlib.pyplot as plt
import numpy as np
# In[2]:
# Training Parameters，超参数
learning_rate = 0.001 #学习率
num_steps = 2000 # 训练步数
batch_size = 128 # 训练数据批的大小
# Network Parameters，网络参数
num_input = 784 # MNIST数据输入 (img shape: 28*28)
num_classes = 10 # MNIST所有类别 (0-9 digits)
dropout = 0.75 # Dropout, probability to keep units，保留神经元相应的概率
# In[3]:
# Create the neural network，创建深度神经网络
def conv_net(x_dict, n_classes, dropout, reuse, is_training):
# Define a scope for reusing the variables，确定命名空间
with tf.variable_scope('ConvNet', reuse=reuse):
# TF Estimator类型的输入为像素
x = x_dict['images']
# MNIST数据输入格式为一位向量，包含784个特征 (28*28像素)
# 用reshape函数改变形状以匹配图像的尺寸 [高 x 宽 x 通道数]
# 输入张量的尺度为四维: [(每一)批数据的数目, 高，宽，通道数]
x = tf.reshape(x, shape=[-1, 28, 28, 1])
# 卷积层，32个卷积核，尺寸为5x5
conv1 = tf.layers.conv2d(x, 32, 5, activation=tf.nn.relu)
# 最大池化层，步长为2，无需学习任何参量
conv1 = tf.layers.max_pooling2d(conv1, 2, 2)
# 卷积层，32个卷积核，尺寸为5x5
conv2 = tf.layers.conv2d(conv1, 64, 3, activation=tf.nn.relu)
# 最大池化层，步长为2，无需学习任何参量
conv2 = tf.layers.max_pooling2d(conv2, 2, 2)
# 展开特征为一维向量，以输入全连接层
fc1 = tf.contrib.layers.flatten(conv2)
# 全连接层 展开成1024 维度矩阵
fc1 = tf.layers.dense(fc1, 1024)
# 应用Dropout (训练时打开，测试时关闭)
fc1 = tf.layers.dropout(fc1, rate=dropout, training=is_training)
# 输出层，预测类别
out = tf.layers.dense(fc1, n_classes)
return out
# In[4]:
# 确定模型功能 (参照TF Estimator模版) 参数分别为输入特征、标签、
def model_fn(features, labels, mode):
# 构建神经网络
# 因为dropout在训练与测试时的特性不一，我们此处为训练和测试过程创建两个独立但共享权值的计算图
logits_train = conv_net(features, num_classes, dropout, reuse=False, is_training=True)
logits_test = conv_net(features, num_classes, dropout, reuse=True, is_training=False)
# 预测 axis = 1的时候返回每一行最大值的位置索引
#tf.argmax 计算正确答案对应的类别编号
pred_classes = tf.argmax(logits_test, axis=1)
#计算非线性激励
pred_probas = tf.nn.softmax(logits_test)
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(mode, predictions=pred_classes)
# 确定误差函数与优化器
#tf.nn.sparse_softmax_cross_entropy_with_logits 计算交叉熵
#tf.reduce_mean 计算交叉熵平均值
loss_op = tf.reduce_mean(tf.nn.sparse_softmax_cross_entropy_with_logits(
logits=logits_train, labels=tf.cast(labels, dtype=tf.int32)))
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate)
train_op = optimizer.minimize(loss_op, global_step=tf.train.get_global_step())
# 评估模型精确度
acc_op = tf.metrics.accuracy(labels=labels, predictions=pred_classes)
# TF Estimators需要返回EstimatorSpec
estim_specs = tf.estimator.EstimatorSpec(
mode=mode,
predictions=pred_classes,
loss=loss_op,
train_op=train_op,
eval_metric_ops={'accuracy': acc_op})
return estim_specs
# In[5]:
# 构建Estimator
model = tf.estimator.Estimator(model_fn)
# In[6]:
# 确定训练输入函数
input_fn = tf.estimator.inputs.numpy_input_fn(
x={'images': mnist.train.images}, y=mnist.train.labels,
batch_size=batch_size, num_epochs=None, shuffle=True)
# 开始训练模型
model.train(input_fn, steps=num_steps)
# In[7]:
# 评判模型
# 确定评判用输入函数
input_fn = tf.estimator.inputs.numpy_input_fn(
x={'images': mnist.test.images}, y=mnist.test.labels,
batch_size=batch_size, shuffle=False)
model.evaluate(input_fn)
# In[8]:
# 预测单个图像 循环图片个数
n_images = 10
# 从数据集得到测试图像 获取前10张图片
test_images = mnist.test.images[:n_images]
# 准备输入数据
input_fn = tf.estimator.inputs.numpy_input_fn(
x={'images': test_images}, shuffle=False)
# 用训练好的模型预测图片类别
preds = list(model.predict(input_fn))
# 可视化显示
for i in range(n_images):
plt.imshow(np.reshape(test_images[i], [28, 28]), cmap='gray')
plt.show()
print("Model prediction:", preds[i])
原生版Tensorflow训练模型
from __future__ import division, print_function, absolute_import
import tensorflow as tf
# Import MNIST data，MNIST数据集导入
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/tmp/data/", one_hot=True)
# In[2]:
# Hyper-parameters，超参数
learning_rate = 0.001
num_steps = 500
batch_size = 128
display_step = 10
# Network Parameters，网络参数
num_input = 784 # MNIST数据输入 (img shape: 28*28)
num_classes = 10 # MNIST所有类别 (0-9 digits)
dropout = 0.75 # Dropout, probability to keep units，保留神经元相应的概率
# tf Graph input，TensorFlow图结构输入
X = tf.placeholder(tf.float32, [None, num_input])
Y = tf.placeholder(tf.float32, [None, num_classes])
keep_prob = tf.placeholder(tf.float32) # dropout (keep probability)，保留i
# In[3]:
# Create some wrappers for simplicity，创建基础卷积函数，简化写法
def conv2d(x, W, b, strides=1):
# Conv2D wrapper, with bias and relu activation，卷积层，包含bias与非线性relu激励
x = tf.nn.conv2d(x, W, strides=[1, strides, strides, 1], padding='SAME')
x = tf.nn.bias_add(x, b)
return tf.nn.relu(x)
def maxpool2d(x, k=2):
# MaxPool2D wrapper，最大池化层
return tf.nn.max_pool(x, ksize=[1, k, k, 1], strides=[1, k, k, 1],
padding='SAME')
# Create model，创建模型
def conv_net(x, weights, biases, dropout):
# MNIST数据为维度为1，长度为784 (28*28 像素)的
# Reshape to match picture format [Height x Width x Channel]
# Tensor input become 4-D: [Batch Size, Height, Width, Channel]
x = tf.reshape(x, shape=[-1, 28, 28, 1])
# Convolution Layer，卷积层
conv1 = conv2d(x, weights['wc1'], biases['bc1'])
# Max Pooling (down-sampling)，最大池化层／下采样
conv1 = maxpool2d(conv1, k=2)
# Convolution Layer，卷积层
conv2 = conv2d(conv1, weights['wc2'], biases['bc2'])
# Max Pooling (down-sampling)，最大池化层／下采样
conv2 = maxpool2d(conv2, k=2)
# Fully connected layer，全连接网络
# Reshape conv2 output to fit fully connected layer input，调整conv2层输出的结果以符合全连接层的需求
fc1 = tf.reshape(conv2, [-1, weights['wd1'].get_shape().as_list()[0]])
fc1 = tf.add(tf.matmul(fc1, weights['wd1']), biases['bd1'])
fc1 = tf.nn.relu(fc1)
# Apply Dropout，应用dropout
fc1 = tf.nn.dropout(fc1, dropout)
# Output, class prediction，最后输出预测
out = tf.add(tf.matmul(fc1, weights['out']), biases['out'])
return out
# In[4]:
# Store layers weight & bias 存储每一层的权值和全差
weights = {
# 5x5 conv, 1 input, 32 outputs
'wc1': tf.Variable(tf.random_normal([5, 5, 1, 32])),
# 5x5 conv, 32 inputs, 64 outputs
'wc2': tf.Variable(tf.random_normal([5, 5, 32, 64])),
# fully connected, 7*7*64 inputs, 1024 outputs
'wd1': tf.Variable(tf.random_normal([7*7*64, 1024])),
# 1024 inputs, 10 outputs (class prediction)
'out': tf.Variable(tf.random_normal([1024, num_classes]))
}
biases = {
'bc1': tf.Variable(tf.random_normal([32])),
'bc2': tf.Variable(tf.random_normal([64])),
'bd1': tf.Variable(tf.random_normal([1024])),
'out': tf.Variable(tf.random_normal([num_classes]))
}
# Construct model，构建模型
logits = conv_net(X, weights, biases, keep_prob)
prediction = tf.nn.softmax(logits)
# Define loss and optimizer，定义误差函数与优化器
loss_op = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(
logits=logits, labels=Y))
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate)
train_op = optimizer.minimize(loss_op)
# Evaluate model，评估模型
correct_pred = tf.equal(tf.argmax(prediction, 1), tf.argmax(Y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32))
# Initialize the variables (i.e. assign their default value)，初始化图结构所有变量
init = tf.global_variables_initializer()
# In[5]:
# Start training，开始训练
with tf.Session() as sess:
# Run the initializer，初始化
sess.run(init)
for step in range(1, num_steps+1):
batch_x, batch_y = mnist.train.next_batch(batch_size)
# Run optimization op (backprop)，优化
sess.run(train_op, feed_dict={X: batch_x, Y: batch_y, keep_prob: dropout})
if step % display_step == 0 or step == 1:
# Calculate batch loss and accuracy
loss, acc = sess.run([loss_op, accuracy], feed_dict={X: batch_x,
Y: batch_y,
keep_prob: 1.0})
print("Step " + str(step) + ", Minibatch Loss= " + "{:.4f}".format(loss) + ", Training Accuracy= " + "{:.3f}".format(acc))
print("Optimization Finished!")
# Calculate accuracy for 256 MNIST test images，以每256个测试图像为例，
print("Testing Accuracy:", sess.run(accuracy, feed_dict={X: mnist.test.images[:256],
Y: mnist.test.labels[:256],
keep_prob: 1.0}))
