TypeError: Expected int32, got list containing Tensors of type '_Message' instead.

来源:互联网 发布:c语言写学生管理系统 编辑:程序博客网 时间:2024/06/06 00:08

代码:

def convolve(img, W):

    # The W matrix is only 2D
    # But conv2d will need a tensor which is 4d:
    # height x width x n_input x n_output
    if len(W.get_shape()) == 2:
        dims = W.get_shape().as_list() + [1, 1]
        W = tf.reshape(W, dims)


    if len(img.get_shape()) == 2:
        # num x height x width x channels
        dims = [1] + img.get_shape().as_list() + [1]
        img = tf.reshape(img, dims)
    elif len(img.get_shape()) == 3:
        dims = [1] + img.get_shape().as_list()
        img = tf.reshape(img, dims)
        # if the image is 3 channels, then our convolution
        # kernel needs to be repeated for each input channel
        W = tf.concat(2, [W, W, W])


    # Stride is how many values to skip for the dimensions of
    # num, height, width, channels
    convolved = tf.nn.conv2d(img, W,
                             strides=[1, 1, 1, 1], padding='SAME')
    return convolved


x = tf.placeholder(tf.float32, shape=img.shape)

out = convolve(x, gabor())
 报错:

TypeError: Expected int32, got list containing Tensors of type '_Message' instead.


解决办法:

tf.concat([W, W, W],2)

阅读全文
0 0
原创粉丝点击