TypeError：参数大小错误答案

【问题标题】：TypeError: argument has wrong sizeTypeError：参数大小错误
【发布时间】：2018-05-07 10:00:23
【问题描述】：

我正在尝试根据大小索引矩阵。我收到 TypeError: argument has wrong size in images[i,:] = img[ind[i]*rows*cols:(ind[i]+1)*rows*cols] 循环，不确定那里发生了什么。但是labels[i] = lbl[ind[i]] 工作正常。

def read(digits, target_index):
    rows=28
    cols=28
    #lbl = array("b", reformatData['target'])

    if target_index == -1:
        data = dataset[:, :-1]

        target = dataset[:, -1]
    elif target_index == 0:
        data   = dataset[:, 1:]
        print "data size: ", data.size,"length", type(data)
        #output: data size:  791840 length <type 'numpy.ndarray'>

        target = dataset[:, 0]
        print "data size: ", target.size,"length", type(target) #output data size:  1010 length <type 'numpy.ndarray'>


    lbl =np.asarray(target.flatten())
    print "length labell", lbl.size, "length", type(lbl) # output length labell 1010 length <type 'numpy.ndarray'>

    img=np.asarray(data[0].flatten())
    print "length of image: ",img.size, "length", type(img) #output length of image:  784 length <type 'numpy.ndarray'>



    ind = [ k for k in xrange(lbl.size) if lbl[k] in digits]
    print "index length ", len(ind), "length", type(ind) #output index length  1010 length <type 'list'>

    images =  matrix(0, (len(ind), rows*cols))
    print "images length ", images.size, "images type", type(images) #output: images length  (1010, 784) images type <type 'cvxopt.base.matrix'>
    labels = matrix(0, (len(ind), 1))
    print "lables length ", labels.size, "labels type", type(labels) #output: lables length  (1010, 1) labels type <type 'cvxopt.base.matrix'>


    for i in xrange(len(ind)):
        images[i,:] = img[ind[i]*rows*cols:(ind[i]+1)*rows*cols] #output: see error below



        labels[i] = lbl[ind[i]]
    return images, labels

print read(digits=digits1, target_index=target_index)

Traceback (most recent call last):
  File "svm.py", line 79, in <module>
    print read(digits=digits1, target_index=target_index)
  File "svm.py", line 74, in read
    images[i,:] = img[ind[i]*rows*cols:(ind[i]+1)*rows*cols]
TypeError: argument has wrong size

样本数据：

3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 41 146 146 48 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 129 253 253 253 250 163 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 133 253 253 253 253 253 253 229 70 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 101 253 252 145 102 107 237 253 247 128 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 181 253 167 0 0 0 61 235 253 253 163 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 255 253 43 0 0 0 0 58 193 253 253 164 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 187 253 32 0 0 0 0 0 55 236 253 253 86 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 146 253 32 0 100 190 87 87 87 147 253 253 123 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 94 253 78 40 248 253 253 253 253 253 253 253 223 84 15 0 0 0 0 0 0 0 0 0 0 0 0 0 14 92 12 35 240 253 253 253 253 253 253 253 253 253 244 89 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 75 161 179 253 253 253 253 253 253 253 253 253 209 43 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 16 16 39 38 16 16 145 243 253 253 185 48 0 0 0 0 0 0 0 0 0 0 0 0 0 20 58 0 0 0 0 0 0 0 0 58 209 253 253 183 0 0 0 0 0 0 0 0 0 0 0 0 77 221 247 79 0 0 0 0 0 0 0 0 13 219 253 240 72 0 0 0 0 0 0 0 0 0 0 0 90 247 253 252 57 0 0 0 0 0 0 0 0 53 251 253 191 0 0 0 0 0 0 0 0 0 0 0 0 116 253 253 59 0 0 0 0 0 0 0 0 99 252 253 145 0 0 0 0 0 0 0 0 0 0 0 0 14 188 253 221 158 38 0 0 0 0 111 211 246 253 253 145 0 0 0 0 0 0 0 0 0 0 0 0 0 12 221 246 253 251 249 249 249 249 253 253 253 253 200 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 95 183 228 253 253 253 253 253 253 195 124 23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 24 37 138 74 126 88 37 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

解决方案：

经过一系列的调试，我终于找到了一个解决方案，将数组包中的标签转换为数组：img=array("B", data)。在此之前，我通过data = ([int(i) for i in (dataset[:, 1:].flatten())]) 将 numpy.ndarray 扁平化到列表中：四个小时的调试帮助了我。可能不是很有效，但现在可以使用。在这里为尝试解决 scikit 库中的分类问题的人发布解决方案。

def read(digits, target_index):
    rows=18
    cols=15
    #lbl = array("b", reformatData['target'])

    if target_index == -1:
        data = dataset[:, :-1]

        target = dataset[:, -1]
    elif target_index == 0:
        data   = dataset[:, 1:]
        #print (data[0].flatten(), type(data[0].flatten()),data[0].flatten().size)
        #data=(list(data[0].flatten()))
        data = ([int(i) for i in (dataset[:, 1:].flatten())])

        #print ("data size: ", data.size,"length", type(data))
        #target = dataset[:, 0]
        #print ("data size: ", target.size,"length", type(target) )
        lbl=([int(i) for i in (dataset[:, 0].flatten())])
        #print lbl

    #lbl =np.asarray(target[:].flatten())
    lbl=array("b", lbl)
    #print(type(lbl)), len(lbl)
    #print ("length labell", lbl.size, "length", type(lbl) )

    img=array("B", data)
    #print(type(img)), len(img)
    #print ("length of image: ",img.size, "length", type(img) )

    ind = [ k for k in range(len(lbl)) if lbl[k] in digits]
    print ("index length ", len(ind), "length", type(ind) )

    images =  matrix(0, (len(ind), rows*cols))
    print ("images length ", images.size, "images type", type(images))
    labels = matrix(0, (len(ind), 1))
    print ("lables length ", labels.size, "labels type", type(labels))


    for i in xrange(len(ind)):
        images[i,:] = img[ind[i]*rows*cols:(ind[i]+1)*rows*cols]

        labels[i]= lbl[ind[i]]

    return images, labels

print read(digits=[3], target_index=0)

【问题讨论】：

每当您看到IndexError 或TypeError 时添加print ... 无处不在。请考虑切换到 Python 3
打击乐，谢谢。尝试了 - print + python27 和 python3。两者都给出了完全相同的 TypeError: argument has wrong size
啊抱歉，我的意思是在每一步通过print <var> 检查参数以及它们发生了什么，那么问题就很清楚了
您能否添加一些示例数据，以便我们重新创建问题？

标签： python numpy scikit-learn svm

【解决方案1】：

经过一系列的调试，我终于找到了一个解决方案，将数组包中的标签转换为数组：img=array("B", data)。在此之前，我通过data = ([int(i) for i in (dataset[:, 1:].flatten())]) 将 numpy.ndarray 扁平化到列表中：四个小时的调试帮助了我。可能效率不高，但目前有效。在这里为尝试解决 scikit 库中的分类问题的人发布解决方案。

def read(digits, target_index):
    rows=18
    cols=15
    #lbl = array("b", reformatData['target'])

    if target_index == -1:
        data = dataset[:, :-1]

        target = dataset[:, -1]
    elif target_index == 0:
        data   = dataset[:, 1:]
        #print (data[0].flatten(), type(data[0].flatten()),data[0].flatten().size)
        #data=(list(data[0].flatten()))
        data = ([int(i) for i in (dataset[:, 1:].flatten())])

        #print ("data size: ", data.size,"length", type(data))
        #target = dataset[:, 0]
        #print ("data size: ", target.size,"length", type(target) )
        lbl=([int(i) for i in (dataset[:, 0].flatten())])
        #print lbl

    #lbl =np.asarray(target[:].flatten())
    lbl=array("b", lbl)
    #print(type(lbl)), len(lbl)
    #print ("length labell", lbl.size, "length", type(lbl) )

    img=array("B", data)
    #print(type(img)), len(img)
    #print ("length of image: ",img.size, "length", type(img) )

    ind = [ k for k in range(len(lbl)) if lbl[k] in digits]
    print ("index length ", len(ind), "length", type(ind) )

    images =  matrix(0, (len(ind), rows*cols))
    print ("images length ", images.size, "images type", type(images))
    labels = matrix(0, (len(ind), 1))
    print ("lables length ", labels.size, "labels type", type(labels))


    for i in xrange(len(ind)):
        images[i,:] = img[ind[i]*rows*cols:(ind[i]+1)*rows*cols]

        labels[i]= lbl[ind[i]]

    return images, labels

print read(digits=[3], target_index=0)

【讨论】：