Tensorflow InvalidArgumentError（回溯见上文）：不兼容的形状：[10000,10]与[10000]

Question

我在尝试我的第一次尝试tensorflow——我试图通过获取一个10维的输入向量X并输出一个标量Y来学习一个线性回归器。具体地说，我试图使用基于闭合形式梯度的解决方案。

我收到了以下错误，我不确定我做错了什么。任何能给我指明正确方向的东西都将不胜感激！

PS C:\Users\Dave\Documents\School\Deep Learning\Assignment_1> python test1.py
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\stream_executor\dso_loader.cc:128] successfu
lly opened CUDA library cublas64_80.dll locally
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\stream_executor\dso_loader.cc:128] successfu
lly opened CUDA library cudnn64_5.dll locally
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\stream_executor\dso_loader.cc:128] successfu
lly opened CUDA library cufft64_80.dll locally
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\stream_executor\dso_loader.cc:128] successfu
lly opened CUDA library nvcuda.dll locally
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\stream_executor\dso_loader.cc:128] successfu
lly opened CUDA library curand64_80.dll locally
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:885] F
ound device 0 with properties:
name: GeForce GTX 1080
major: 6 minor: 1 memoryClockRate (GHz) 1.86
pciBusID 0000:01:00.0
Total memory: 8.00GiB
Free memory: 6.63GiB
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:906] D
MA: 0
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:916] 0
:   Y
I c:\tf_jenkins\home\workspace\release-win\device\gpu\os\windows\tensorflow\core\common_runtime\gpu\gpu_device.cc:975] C
reating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1080, pci bus id: 0000:01:00.0)
Traceback (most recent call last):
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 10
    return fn(*args)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 10
    status, run_metadata)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\contextlib.py", line 66, in __exit__
    next(self.gen)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\errors_impl.py",
n_on_not_ok_status
    pywrap_tensorflow.TF_GetCode(status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [10000,10] vs. [10000]
         [[Node: sub = Sub[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/gpu:0"](Add, _recv_Placeholder_1_0/_7)]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "test1.py", line 43, in <module>
    c = sess.run(cost, feed_dict={X: train_X, Y: train_Y})
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 76
    run_metadata_ptr)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 96
    feed_dict_string, options, run_metadata)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 10
    target_list, options, run_metadata)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\client\session.py", line 10
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [10000,10] vs. [10000]
         [[Node: sub = Sub[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/gpu:0"](Add, _recv_Placeholder_1_0/_7)]]

Caused by op 'sub', defined at:
  File "test1.py", line 25, in <module>
    cost = tf.reduce_sum(tf.pow(pred-Y, 2))/(2*n_samples)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\ops\math_ops.py", line 814,
    return func(x, y, name=name)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\ops\gen_math_ops.py", line
    result = _op_def_lib.apply_op("Sub", x=x, y=y, name=name)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\op_def_library.py
    op_def=op_def)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\ops.py", line 224
    original_op=self._default_original_op, op_def=op_def)
  File "C:\Users\Dave\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\python\framework\ops.py", line 112
    self._traceback = _extract_stack()

InvalidArgumentError (see above for traceback): Incompatible shapes: [10000,10] vs. [10000]
         [[Node: sub = Sub[T=DT_FLOAT, _device="/job:localhost/replica:0/task:0/gpu:0"](Add, _recv_Placeholder_1_0/_7)]]

这是我的代码：

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
rng = np.random

#from IPython import get_ipython
#get_ipython().run_line_magic('matplotlib', 'inline')

learning_rate = 0.01
training_epochs = 1000
display_step = 50

train_X = np.loadtxt('data.txt', usecols=[0,1,2,3,4,5,6,7,8,9])
train_Y = np.loadtxt('data.txt', usecols=[10])
n_samples = train_X.shape[0]
X = tf.placeholder(tf.float32)
Y = tf.placeholder(tf.float32)
W = tf.Variable(rng.randn(), name = "weight")
b = tf.Variable(rng.randn(), name = "bias")

#build the model
pred = tf.add(tf.mul(X,W), b)

#mean squared error
cost = tf.reduce_sum(tf.pow(pred-Y, 2))/(2*n_samples)

#gradient descent 
optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)

#initialize the variables
init = tf.global_variables_initializer()
#launch the graph
with tf.Session() as sess:
    sess.run(init)

    #fit training data
    for epoch in range(training_epochs):
        for (x, y) in zip(train_X, train_Y):
            sess.run(optimizer, feed_dict = {X: x, Y: y})

        #display logs
        if (epoch+1) % display_step ==0:
            c = sess.run(cost, feed_dict={X: train_X, Y: train_Y})
           # print "Epoch:", '%04d' % (epoch+1), "cost=", "{:.9f".format(c), \
                #"W=", sess.run(W), "b=" sess.run(b)

    #print "Optimization done"
    training_cost = sess.run(cost, feed_dict={X: train_X, Y: train_Y})
    #print "Training cost=", training_cost, "W=", sess.run(W), "b=" sess.run(b), '\n'

    #display graphically
    plt.plot(train_X, train_Y, 'ro', label = 'Orig data')
    plt.plot(train_X, sess.run(W) * train_X + sess.run(b), label = 'Fitted Line')
    plt.legend()
    plt.show()

Answer 1

我不知道tensorflow，我也不确定你的代码到底是怎么回事，所以我试图根据numpy的行为做出一个合理的猜测。我本来想加上这个作为评论，但时间太长了。

加载训练数据时，train_X具有10列（这是一个2d数组）的形状，而train_Y具有单个列（这是一个2d数组）的形状。这两个形状不能一起广播，因此pred中的Y和pred-Y具有不兼容的形状。为此，您需要将train_X转置，或者将train_Y转换为一个形状数组(10000,1)以使它们兼容。前者可以通过将unpack=True传递给np.loadtxt来实现；后者则使用类似train_Y = train_Y[:,None]的内容（至少我怀疑unpack=True在这种情况下不会有帮助，但无论如何值得一试）。

但是如果你要转换你的数组，你需要小心你的训练循环仍然有效。目前，您的(10000,10)形数组相当于一个长度为10000的10个列表的列表，而您的(10000,)形数组相当于一个长度为10000的列表。这些可以很好地组合在一起。如果您进行转置，例如train_X以使广播工作，则需要修改此循环：

    for (x, y) in zip(train_X.T, train_Y):
        sess.run(optimizer, feed_dict = {X: x, Y: y})

事后看来，我的意思可能是您需要将train_X向后移位，以确保train_X和train_Y的第一个维度与压缩匹配。

换位也可能影响后面的步骤，例如绘图。如果有任何奇怪的输出或错误，您需要调回。或者更好的方法：只对tensorflow特定的操作进行转置（但是我不熟悉这一部分，所以我不知道是否以及如何按照惯用方式完成）。