Question

我想用我刚刚安装的TF实现MNIST的简单逻辑回归，并希望用TensorBoard监控minibatch-SGD的进度。

我首先没有编译的张量板，并且在testset上获得了0.9166的准确度。

然而，当我添加tensorboard以查看发生了什么时，我甚至无法再编译它了：

the placeholders must be fed with dtype float但我的所有数组都是带有dtype float的np数组！

如果你能在我的代码中指出那些令人惊讶的问题：

  # -*- coding: utf-8 -*-
    """
    Created on Thu Jan 14 13:06:44 2016

    @author: me
    """
    #from tensorflow.examples.tutorials.mnist import input_data
    #mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)

    import tensorflow as tf
    import os
    import random
    import numpy as np
    from array import array
    import struct
    import matplotlib.pyplot as plt
    import time

   #I first placed the decompressed -ubyte files from mnist on the path indicated
    os.chdir('/home/me/Bureau/Step1/')

    with open("train-labels.idx1-ubyte") as file:
        magic, size = struct.unpack(">II",file.read(8))
        train_labels_data=np.asarray(array("B",file.read()))
    with open("t10k-labels.idx1-ubyte") as file:
        magic, size = struct.unpack(">II",file.read(8))
        test_labels_data=np.asarray(array("B",file.read()))
    with open("train-images.idx3-ubyte") as file:
        magic, size, rows, cols =struct.unpack(">IIII",file.read(16))
        train_images_data=np.reshape(np.asarray(array("B",file.read())),(size,rows,cols))
    with open("t10k-images.idx3-ubyte") as file:
        magic, size, rows, cols =struct.unpack(">IIII",file.read(16))
        test_images_data=np.reshape(np.asarray(array("B",file.read())),(size,rows,cols))

    for i in range(10):
       plt.imshow(train_images_data[i,:])
       plt.show()
       print(train_labels_data[i])

    train_images=np.reshape(train_images_data,(60000,28*28)).astype(np.float32)*1/255
    test_images=np.reshape(test_images_data,(10000,28*28)).astype(np.float32)*1/255

    train_labels=np.zeros((60000,10),dtype=np.float32)
    test_labels=np.zeros((10000,10),dtype=np.float32)

    for i in range(60000):
        a=train_labels_data[i]
        train_labels[i,a]=1.

    for j in range(10000):
        b=test_labels_data[j]
        test_labels[j,b]=1.

    sess=tf.Session()

    x=tf.placeholder(tf.float32, [None, 784],name="x-input")
    W=tf.Variable(tf.zeros([784, 10]),name="weights")
    b=tf.Variable(tf.zeros([10]),name="bias")

    with tf.name_scope("Wx_b") as scope:
        y=tf.nn.softmax(tf.matmul(x,W) + b)

    w_hist=tf.histogram_summary("weights",W)
    b_hist=tf.histogram_summary("bias",b)
    y_hist=tf.histogram_summary("y",y)


    y_ =tf.placeholder(tf.float32, [None, 10], name="y-input")


    with tf.name_scope("xent") as scope:
        cross_entropy= -tf.reduce_sum(y_*tf.log(y))
        ce_summ=tf.scalar_summary("cross_entropy", cross_entropy)

    with tf.name_scope("train") as scope:
        train_step=tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)

    with tf.name_scope("test") as scope:
        correct_prediction =tf.equal(tf.argmax(y,1), tf.argmax(y_,1))
        accuracy=tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
        accuracy_summary=tf.scalar_summary("accuracy",accuracy)

    merged=tf.merge_all_summaries()
    writer=tf.train.SummaryWriter("/tmp/mnist_logs",sess.graph_def)

    init=tf.initialize_all_variables()
    sess.run(init)

    for i in range(1000):
        if i % 10 == 0:
            feed={x:test_images, y_: test_labels}
            result=sess.run([merged, accuracy],feed_dict=feed)
            summary_str=result[0]
            acc=result[1]
            writer.add_summary(summary_str, i)
            print("Accuracy at step %s: %s" % (i,acc))
        else:
            index=np.random.randint(60000-1,size=100)
            batch_xs, batch_ys = train_images[index,:], train_labels[index]
            sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})

    print(sess.run(accuracy, feed_dict={x: train_images, y_: train_labels}))

它发生的行是在要合并的Feed中，但是当我提供与我喂养train_step完全相同的方式时，我感到很茫然......

Answer 1

事实证明，当我重新打开一个新的spyder并启动它运行的程序时，你不能一遍又一遍地运行相同的脚本！头脑=吹

占位符x-input TensorFlow

1 个答案: