WARNING:tensorflow:From D:\ProgramData\Anaconda3\lib\site-packages\tensorflow\contrib\learn\python\learn\datasets\mnist.py:290: DataSet.__init__ (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.

Instructions for updating:

Please use alternatives such as official/mnist/dataset.py from tensorflow/models.

Traceback (most recent call last):

  File "D:/code/PycharmProject/mnist_testdemo2/mnist/convolutional.py", line 12, in <module>

    y, variables = model.convolutional(x, keep_prob)

  File "D:\code\PycharmProject\mnist_testdemo2\mnist\model.py", line 24, in convolutional

    W_conv1 = weight_variable([5, 5, 1, 32])

  File "D:\code\PycharmProject\mnist_testdemo2\mnist\model.py", line 17, in weight_variable

    initial = tf.truncated.normal(shape, stddev=0.1)

AttributeError: module 'tensorflow' has no attribute 'truncated'

------------------------------------------------------

我的model文件代码：

import tensorflow as tf

# 线性模型 Y=W*x + b
def regressions(x):
    W = tf.Variable(tf.zeros([784, 10]), name="W")
    b = tf.Variable(tf.zeros([10]), name='b')
    y = tf.nn.softmax(tf.matmul(x, W) + b)
    return y, [W,b]

# 卷积模型
def convolutional(x, keep_prob):
    def conv2d(x, W):
        return tf.nn.conv2d([1, 1, 1, 1], padding='SAME')
    def max_pool_2x2(x):
        return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1])
    def weight_variable(shape):
        initial = tf.truncated.normal(shape, stddev=0.1)
        return tf.Variable(initial)
    def bias_variable(shape):
        initial = tf.constant(0.1, shape=shape)
        return tf.Variable(initial)

    x_image = tf.reshape(x, [-1, 28, 28, 1])
    W_conv1 = weight_variable([5, 5, 1, 32])
    b_conv1 = bias_variable([32])
    h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
    h_pool1 = max_pool_2x2(h_conv1)

    W_conv2 = weight_variable([5, 5, 32, 64])
    b_conv2 = bias_variable([64])
    h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
    h_pool2 = max_pool_2x2(h_conv2)

    # full connection
    W_fc1 = weight_variable([7 * 7 * 64, 1024])
    b_fc1 = bias_variable([1024])
    h_pool2_flat = tf.reshape(h_pool2, [-1, 7 * 7 * 64])
    h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

    h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

    W_fc2 = weight_variable([1024, 10])
    b_fc2 = bias_variable([10])
    y = tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)

    return y, [W_conv1, b_conv1, W_conv2, b_conv2, W_fc1, b_fc1, W_fc2, b_fc2]

我的convolutional代码：

import os
import model
import tensorflow as tf
import input_data

data = input_data.read_data_sets('MNIST_data', one_hot=True)

#model
with tf.variable_scope("convolutional"):
    x = tf.placeholder(tf.float32, [None, 784], name='x')
    keep_prob = tf.placeholder(tf.float32)
    y, variables = model.convolutional(x, keep_prob)

#train
y_ = tf.placeholder(tf.float32, [None, 10], name='y')
cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

saver = tf.train.Saver(variables)

with tf.Session() as sess:
    merged_summary_op = tf.summary.merge_all()
    summay_writer = tf.summary.FileWriter('./tem/mnist_log/1', sess.graph)
    summay_writer.add_graph(sess.graph)
    sess.run(tf.lobal_variables_initializer())

    #最好做两万次训练
    for i in range(2000):
        batch = data.train.next_batch(50)
        if(i % 100 == 0):
            train_accuracy = accuracy.eval(feed_dict={x: batch[0], y_: batch[1], keep_prob: 1.0})
            print("step %d, training accuracy %g" % (i, train_accuracy))
        sess.run(train_step, feed_dict={x: batch[0], y_: batch[1], keep_prob:0.5})

    print(sess.run(accuracy, feed_dict={x: data.test.images, y_: data.test.labels, keep_prob: 1.0}))
    path = saver.save(sess, os.path.join(os.path.dirname(__file__), 'data', 'convalutional.ckpt', write_meta_graph=False, write_state=False))
    print("Saved:", path)