问不知道代码错误来自哪里？

以下代码出现：AttributeError: 'tuple' object has no attribute 'ndims'

但是调试不出哪一行出现问题

import tensorflow as tf

import numpy as np

from random import choice

#1、产生数据,定义参数，输入层

#定义参数

reuse=tf.AUTO_REUSE

Im = 100

In = tf.random_uniform([1], minval=0, maxval=0.1, dtype=tf.float32)

Um = 220

W0 = 50

A_list = [46.70,28.01,20.01,15.56]

index = choice([0,1,2,3])

A = A_list[index]

Fai = np.random.random((180,180))

t = np.arange(0,100,0.0001)

T = np.arange(0,0.1,0.0001)

M = tf.reciprocal(np.multiply(2,T))

#基波电流信号

x0 = np.multiply(Um,np.sin(np.multiply(W0,t)))

def Harmonic_signal(t,W0,Fai,a):

if a==None:

#在电网基波信号上叠加正数次的谐波信号的稳态电网畸变信号

x1 = np.add(np.multiply(Im,np.sin(W0,t)),np.multiply(In,np.sin(np.add(np.multiply(A,W0,t),Fai))))

return x1

else:

#振荡谐波稳态电网畸变信号

x1 = np.add(np.multiply(Im,np.sin(W0,t)),np.multiply(In,np.sin(np.add(np.multiply(A,W0,t),Fai)),np.multiply(M,np.exp(np.divide(-t,T,dtype = tf.float32)))))

return x1

#x1 = Harmonic_signal(t,W0,Fai,None)

#瞬时电压暂升或暂降，q大于1，为暂升，q小于1，为暂降

def Voltage_flicker(t,x0,q):

t1 = tf.get_variable("t1",dtype=tf.float32,initializer=tf.random_uniform_initializer(minval=0,maxval=t,dtype=tf.float32))

x2 = np.select([t<t1, t>np.add(t1,6), True], [x0, x0,np.multiply(q,x0)])

return x2

#x2 = Voltage_flicker(t,x0,q)

#2，运行参数设定

training_rate = 0.001

training_iters = 100000

batch_size = 128

display_step = 10

n_input = 28

n_steps = 28

n_hidden = 128

n_classes = 5

#中间隐藏单元

def Model(net,n_hidden,

data_format='channels_last',

kernel_regularizer=None,

bias_initializer=tf.zeros_initializer(),

activation = tf.nn.relu,

scope = None,

reuse = None):

for i in range(n_hidden):

with tf.variable_scope('block_1d',net,reuse = reuse):

#tf.layters.conv1d(inputs, filters, kernel_size, strides=1, padding='valid', data_format='channels_last', dilation_rate=1, activation=None, use_bias=True，...)

with tf.variable_scope("BRANCH_0"):

branch_0 = tf.layers.conv1d(net,1,1,strides=1,padding="VALID",name = "Conv1d_0a_1x1")

branch_0 = tf.layers.conv1d(net,128,3,strides=1,padding="SAME",name = "Conv1d_1a_1x3")

with tf.variable_scope("BRANCH_1"):

branch_1 = tf.layers.conv1d(net,1,1,strides=1,padding="VALID",name = "Conv1d_0b_1x1")

branch_1 = tf.layers.conv1d(net,64,5,strides=1,padding="SAME",name = "Conv1d_1b_1x5")

with tf.variable_scope("BRANCH_2"):

branch_2 = tf.layers.conv1d(net,1,1,strides=1,padding="VALID",name = "Conv1d_0c_1x1")

branch_2 = tf.layers.conv1d(net,32,7,strides=1,padding="SAME",name = "Conv1d_1c_1x7")

link = tf.concat(1,[branch_0,branch_1,branch_2])

link = tf.multiply(link,link)

return Model(link)

x = tf.placeholder("float",[None,n_steps,n_input])

y = tf.placeholder("float",[None,n_classes])

out0 = Model(x0,n_hidden)

out1 = Model(Harmonic_signal(t,W0,Fai,None))

out2 = Model(Harmonic_signal(t,W0,Fai,1))

out3 = Model(Voltage_flicker(t,x0,0.5))

out4 = Model(Voltage_flicker(t,x0,1))

#3，输出层

out = [out0,out1,out2,out3,out4]

cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits = out,labels = y))

optimizer = tf.train.AdamOptimizer(learning_rate = training_rate).minimize(cost)

correct_pred = tf.equal(tf.argmax(out,1),tf.argmax(y,1))

accuaracy = tf.reduce_mean(tf.cast(correct_pred,tf.float32))

init = tf.glorot_normal_initializer()

#启动会话

with tf.Session() as sess:

sess = tf.Session(config=tf.ConfigProto(gpu_options = True))

sess.run(init)

step = 1

while step * batch_size < training_iters:

batch_x,batch_y = tf.train.next_batch(batch_size)

batch_x = batch_x.reshape(batch_size,n_steps,n_input)

sess.run(optimizer,feed_dict = {x:batch_x,y:batch_y})

if step % display_step == 0:

acc = sess.run(accuaracy,feed_dict = {x:batch_x,y:batch_y})

loss = sess.run(accuaracy,feed_dict = {x:batch_x,y:batch_y})

print("iter" + str(step * batch_size) + ", Minibatch Loss =" + "\{:.6f}".format(loss) + ",Training Accuracy=" + "\{:.5f}".format(acc))

step += 1