Question

我很想使用tensorflow并且正在用于学校项目。在这里，我试图建立一个房屋标识符，我在excel表上做了一些数据，把它变成了一个csv文件，我正在测试是否可以读取数据。读取数据但是当我进行矩阵乘法时它会产生多个错误并且表示....＆＃34; ValueError：Shape必须是等级2，但是对于＆＃39; MatMul＆＃39; （op：＆＃39; MatMul＆＃39;）输入形状：[]，[1,1]。＆＃34;非常感谢你！

import tensorflow as tf
import os
dir_path = os.path.dirname(os.path.realpath(__file__))
filename = dir_path+ "\House Price Data .csv"
w1=tf.Variable(tf.zeros([1,1]))
w2=tf.Variable(tf.zeros([1,1])) #Feature 1's weight
w3=tf.Variable(tf.zeros([1,1])) #Feature 1's weight
b=tf.Variable(tf.zeros([1])) #bias for various features
x1= tf.placeholder(tf.float32,[None, 1])
x2= tf.placeholder(tf.float32,[None, 1])
x3= tf.placeholder(tf.float32,[None, 1])
Y= tf.placeholder(tf.float32,[None, 1])
y_=tf.placeholder(tf.float32,[None,1])
with tf.Session() as sess:
    sess.run( tf.global_variables_initializer())
    with open(filename) as inf:
        # Skip header
        next(inf)
        for line in inf:
            # Read data, using python, into our features
            housenumber, x1, x2, x3, y_ = line.strip().split(",")
            x1 = float(x1)
            product = tf.matmul(x1, w1)
            y = product + b

Answer 1

@Aaron是对的，当您从csv文件加载数据时，您将覆盖变量。

您需要将加载的值保存到单独的变量中，例如_x1而不是x1，然后使用feed_dict将值提供给占位符。由于x1的形状为[None,1]，因此您需要将字符串标量_x1转换为具有相同形状的浮点数，在这种情况下为[1,1]。< / p>

import tensorflow as tf
import os
dir_path = os.path.dirname(os.path.realpath(__file__))
filename = dir_path+ "\House Price Data .csv"
w1=tf.Variable(tf.zeros([1,1]))
b=tf.Variable(tf.zeros([1])) #bias for various features
x1= tf.placeholder(tf.float32,[None, 1])

y_pred = tf.matmul(x1, w1) + b

with tf.Session() as sess:
    sess.run( tf.global_variables_initializer())
    with open(filename) as inf:
        # Skip header
        next(inf)
        for line in inf:
            # Read data, using python, into our features
            housenumber, _x1, _x2, _x3, _y_ = line.strip().split(",")
            sess.run(y_pred, feed_dict={x1:[[float(_x1)]]})

矩阵乘法不工作 - Tensorflow

1 个答案: