ValueError: Variable rnn/basic_rnn_cell/kernel already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope?

详细错误信息

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-3-7b6c61e32b09> in <module>129         plt.show()130
--> 131 prediction()<ipython-input-3-7b6c61e32b09> in prediction()107 def prediction():108     global batch_size
--> 109     pred,_=lstm(batch_size)      #预测时只输入[1,time_step,input_size]的测试数据110     saver=tf.train.Saver(tf.global_variables())111     with tf.Session() as sess:<ipython-input-3-7b6c61e32b09> in lstm(batch)62         #dynamic_rnn 用于创建由RNNCell细胞指定的循环神经网络，对inputs进行动态展开63         #output_rnn是记录lstm每个输出节点的结果，final_states是最后一个cell的结果
---> 64     output_rnn,final_states=tf.nn.dynamic_rnn(cell, input_rnn,initial_state=init_state, dtype=tf.float32)65     #函数的作用是将tensor变换为参数shape的形式。66     output=tf.reshape(output_rnn,[-1,rnn_unit])...........
.......................ValueError: Tensor("rnn/Const:0", shape=(1,), dtype=int32) must be from the same graph as Tensor("ExpandDims:0", shape=(1,), dtype=int32).

方案一：

使用如下方法对原始网络结构图进行重置：

tf.reset_default_graph()

方案二：

将模型建模部分与预测部分分离

以jupyter为例，那么在模型训练的时候，使用一个kernel并把模型文件保存在对应的目录中；

另外打开一个jupyter notebook，然后构建预测函数进行预测分析OK。

# 训练过程代码，在第一个jupyter notebook中；

import pandas as pd  #数据分析包
import numpy as np   #提供多维数组对象的库
import matplotlib.pyplot as plt  #画图的库
# import tensorflow as tf          #深度学习框架
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()#导入数据
f=open(r'600196.csv')
df=pd.read_csv(f)     #读入股票数据
data=np.array(df['target'])   #获取最高价序列
data=data[::-1]      #反转，使数据按照日期先后顺序排列
#以折线图展示data
plt.figure()
plt.plot(data)
plt.show()
normalize_data=(data-np.mean(data))/np.std(data)  #标准化
normalize_data=normalize_data[:,np.newaxis]       #增加维度#生成训练集
#设置常量
time_step=10      #时间步
rnn_unit=10       #hidden layer units
batch_size=60     #每一批次训练多少个样例
input_size=1      #输入层维度
output_size=1     #输出层维度
lr=0.0006         #学习率
train_x,train_y=[],[]   #训练集
for i in range(len(normalize_data)-time_step-1):x=normalize_data[i:i+time_step]y=normalize_data[i+1:i+time_step+1]train_x.append(x.tolist())  #将数组转化成列表train_y.append(y.tolist())#定义神经网络变量
X=tf.placeholder(tf.float32, [None,time_step,input_size])    #每批次输入网络的tensor/定义placeholder
Y=tf.placeholder(tf.float32, [None,time_step,output_size])   #每批次tensor对应的标签
#输入层、输出层权重、偏置
weights={'in':tf.Variable(tf.random_normal([input_size,rnn_unit])),'out':tf.Variable(tf.random_normal([rnn_unit,1]))}
biases={'in':tf.Variable(tf.constant(0.1,shape=[rnn_unit,])),'out':tf.Variable(tf.constant(0.1,shape=[1,]))}#定义神经网络变量
def lstm(batch):      #参数：输入网络批次数目w_in=weights['in']b_in=biases['in']input=tf.reshape(X,[-1,input_size])  #需要将tensor转成2维进行计算，计算后的结果作为隐藏层的输入input_rnn=tf.matmul(input,w_in)+b_in #表示矩阵乘法input_rnn=tf.reshape(input_rnn,[-1,time_step,rnn_unit])  #将tensor转成3维，作为lstm cell的输入cell=tf.nn.rnn_cell.BasicLSTMCell(rnn_unit)  #定义单个基本的LSTM单元init_state=cell.zero_state(batch,dtype=tf.float32)  #这个函数用于返回全0的state tensor#dynamic_rnn 用于创建由RNNCell细胞指定的循环神经网络，对inputs进行动态展开#output_rnn是记录lstm每个输出节点的结果，final_states是最后一个cell的结果output_rnn,final_states=tf.nn.dynamic_rnn(cell, input_rnn,initial_state=init_state, dtype=tf.float32)#函数的作用是将tensor变换为参数shape的形式。output=tf.reshape(output_rnn,[-1,rnn_unit])w_out=weights['out']b_out=biases['out']pred=tf.matmul(output,w_out)+b_out #表示矩阵乘法return pred,final_states#训练模型
def train_lstm():global batch_sizepred,_=lstm(batch_size) #调用的构建的lstm变量#损失函数 平均平方误差(MSE)loss=tf.reduce_mean(tf.square(tf.reshape(pred,[-1])-tf.reshape(Y, [-1])))#实现梯度下降算法的优化器，优化损失函数train_op=tf.train.AdamOptimizer(lr).minimize(loss)#保存和恢复模型的方法；方法返回checkpoint文件的路径。可以直接传给restore() 进行调用saver=tf.train.Saver(tf.global_variables())with tf.Session() as sess:sess.run(tf.global_variables_initializer())#重复训练10000次for i in range(10):step=0start=0end=start+batch_sizewhile(end<len(train_x)):_,loss_=sess.run([train_op,loss],feed_dict={X:train_x[start:end],Y:train_y[start:end]})start+=batch_sizeend=start+batch_size#每10步保存一次参数if step%10==0:print(i,step,loss_)print("保存模型：",saver.save(sess,'stock.model'))step+=1train_lstm()
tf.reset_default_graph()

# 预测部分，在另外一个jupyter notebook中；

import pandas as pd  #数据分析包
import numpy as np   #提供多维数组对象的库
import matplotlib.pyplot as plt  #画图的库
# import tensorflow as tf          #深度学习框架
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()#导入数据
f=open(r'600196.csv')
df=pd.read_csv(f)     #读入股票数据
data=np.array(df['target'])   #获取最高价序列
data=data[::-1]      #反转，使数据按照日期先后顺序排列
#以折线图展示data
plt.figure()
plt.plot(data)
plt.show()
normalize_data=(data-np.mean(data))/np.std(data)  #标准化
normalize_data=normalize_data[:,np.newaxis]       #增加维度#生成训练集
#设置常量
time_step=10      #时间步
rnn_unit=10       #hidden layer units
batch_size=60     #每一批次训练多少个样例
input_size=1      #输入层维度
output_size=1     #输出层维度
lr=0.0006         #学习率
train_x,train_y=[],[]   #训练集
for i in range(len(normalize_data)-time_step-1):x=normalize_data[i:i+time_step]y=normalize_data[i+1:i+time_step+1]train_x.append(x.tolist())  #将数组转化成列表train_y.append(y.tolist())#定义神经网络变量
X=tf.placeholder(tf.float32, [None,time_step,input_size])    #每批次输入网络的tensor/定义placeholder
Y=tf.placeholder(tf.float32, [None,time_step,output_size])   #每批次tensor对应的标签
#输入层、输出层权重、偏置
weights={'in':tf.Variable(tf.random_normal([input_size,rnn_unit])),'out':tf.Variable(tf.random_normal([rnn_unit,1]))}
biases={'in':tf.Variable(tf.constant(0.1,shape=[rnn_unit,])),'out':tf.Variable(tf.constant(0.1,shape=[1,]))}#定义神经网络变量
def lstm(batch):      #参数：输入网络批次数目w_in=weights['in']b_in=biases['in']input=tf.reshape(X,[-1,input_size])  #需要将tensor转成2维进行计算，计算后的结果作为隐藏层的输入input_rnn=tf.matmul(input,w_in)+b_in #表示矩阵乘法input_rnn=tf.reshape(input_rnn,[-1,time_step,rnn_unit])  #将tensor转成3维，作为lstm cell的输入cell=tf.nn.rnn_cell.BasicLSTMCell(rnn_unit)  #定义单个基本的LSTM单元init_state=cell.zero_state(batch,dtype=tf.float32)  #这个函数用于返回全0的state tensor#dynamic_rnn 用于创建由RNNCell细胞指定的循环神经网络，对inputs进行动态展开#output_rnn是记录lstm每个输出节点的结果，final_states是最后一个cell的结果output_rnn,final_states=tf.nn.dynamic_rnn(cell, input_rnn,initial_state=init_state, dtype=tf.float32)#函数的作用是将tensor变换为参数shape的形式。output=tf.reshape(output_rnn,[-1,rnn_unit])w_out=weights['out']b_out=biases['out']pred=tf.matmul(output,w_out)+b_out #表示矩阵乘法return pred,final_states#预测模型
def prediction():pred,_=lstm(1)      #预测时只输入[1,time_step,input_size]的测试数据saver=tf.train.Saver(tf.global_variables())with tf.Session() as sess:#参数恢复module_file = tf.train.latest_checkpoint('.')saver.restore(sess, module_file)#取训练集最后一行为测试样本。shape=[1,time_step,input_size]prev_seq=train_x[-1]predict=[]#得到之后100个预测结果for i in range(100):next_seq=sess.run(pred,feed_dict={X:[prev_seq]})predict.append(next_seq[-1])#每次得到最后一个时间步的预测结果，与之前的数据加在一起，形成新的测试样本prev_seq=np.vstack((prev_seq[1:],next_seq[-1]))#以折线图表示结果plt.figure()plt.plot(list(range(len(normalize_data))), normalize_data, color='b')plt.plot(list(range(len(normalize_data), len(normalize_data) + len(predict))), predict, color='r')plt.show()prediction()

参考：ValueError: Variable rnn/basic_rnn_cell/kernel already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope?

参考：Tensorflow: Variable rnn/basic_lstm_cell/kernel already exists, disallowed

ValueError: Variable rnn/basic_rnn_cell/kernel already exists, disallowed. Did you mean to set reuse相关推荐

ValueError: Variable rnn/basic_lstm_cell/kernel already exists, disallowed.
ValueError: Variable rnn/basic_lstm_cell/kernel already exists, disallowed. 用spyder运行mnist_lstm.py 第 ...
ValueError: Variable eval_net/l1/w1 already exists, disallowed. Did you mean to set reuse=True or...
当在spyder中运行代码时报错如下: ValueError: Variable eval_net/l1/w1 already exists, disallowed. Did you mean to ...
Variable Generator/dense/kernel already exists, disallowed.
Jupyter notebook运行下面程序:https://github.com/MorvanZhou/Tensorflow-Tutorial/blob/master/tutorial-conten ...
tf1.x error:Variable generator/encoder_image/conv2d/kernel already exists, disallowed
今天在用tf1.x进行重复多次计算的时候,出现问题Variable generator/encoder_image/conv2d/kernel already exists, disallowed 错 ...
Variable W already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE……
Variable W already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarS ...
Variable conv1/weights already exists, disallowed. Did you mean to set reuse=True...tensorflow报错
在训练神经网络模型使用tensorflow时,training模块训练多次会出现报错如: 1,Variable conv1/weights already exists, disallowed. Di ...
【报错及解决】Variable w already exists, disallowed. Did you mean to set reuse=True in VarScope? Originally
一.问题描述在构建tensorflow计算图的时候,如下代码运行报错 x = tf.placeholder(tf.float32,[None,3072]) #[None] y = tf.placeh ...
tensorflow Variable already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUS
问题描述使用with tf.Session() as sess:后下次循环进入再次使用 with tf.Session() 问题原因 tf.graph() 未清空解决在withtf.get_d ...
...weights already exists, disallowed. Did you mean to set reuse=True...（解决办法）
model.inference(train_batch, BATCH_SIZE, N_CLASSES)这个里面应该有with tf.variable_scope('layer1'):这样的代码,改为w ...

ValueError: Variable rnn/basic_rnn_cell/kernel already exists, disallowed. Did you mean to set reuse

ValueError: Variable rnn/basic_rnn_cell/kernel already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope?

详细错误信息

方案一：

方案二：

ValueError: Variable rnn/basic_rnn_cell/kernel already exists, disallowed. Did you mean to set reuse相关推荐

最新文章

热门文章