python给我的解决方案是什么"ValueError: setting an array element with a sequence."

Question

我正在运行下面的代码，但它给了我一个关于数组的错误。我试图找到一个解决方案，并以某种方式理解这个问题，但我无法解决这个问题。这是我的代码：

import tensorflow as tf
import pandas as pa
import numpy as np

iris = pa.read_csv("iris.csv", names = ['F1', 'F2', 'F3', 'F4', 'class'])
print(iris.head(5))
iris['class'].value_counts()
#mapping data
A1 = np.asarray([1,0,0])
A2 = np.asarray([0,1,0])
A3 = np.asarray([0,0,1])
Irises = {'Iris-setosa' : A1, 'two' : A2, 'Iris-virginica' : A3}
iris['class'] = iris['class'].map(Irises)

#Mjesanje podataka 
iris = iris.iloc[np.random.permutation(len(iris))]
print(iris.head(10))
iris = iris.reset_index(drop=True)
print(iris.head(10))
#splitting data into training and testing
x_train = iris.ix[0:100,['F1', 'F2', 'F3', 'F4']]
y_train = iris.ix[0:100,['class']]
x_test = iris.ix[101:, ['F1', 'F2', 'F3', 'F4']]
y_test = iris.ix[101:, ['class']]

print(x_train.tail(5))
print(y_train.tail(5))
print(x_test.tail(5))
print(y_test.tail(5))
n_nodes_hl1 = 150
n_nodes_hl2 = 150

n_classes = 3 # U ovom slucaju tri, 1-> Iris-setosa, Iris-versicolo, Iris-virginica
batch_size = 50 # Da li ima neko optimalno rijesenje koliko uzeti?
x = tf.placeholder('float', shape = [None, 4]) # 4 featrues 
y = tf.placeholder('float', shape = [None, n_classes]) # 3 classes 

def neural_network_model(data):
hidden_layer_1 = {'weights': tf.Variable(tf.random_normal([4, n_nodes_hl1])),
'biases':tf.Variable(tf.random_normal([n_nodes_hl1]))}
hidden_layer_2 = {'weights': tf.Variable(tf.random_normal([n_nodes_hl1, n_nodes_hl2])),
'biases':tf.Variable(tf.random_normal([n_nodes_hl2]))}
output_layer = {'weights': tf.Variable(tf.random_normal([n_nodes_hl2, n_classes])),
'biases': tf.Variable(tf.random_normal([n_classes]))}

l1 = tf.add(tf.matmul(data, hidden_layer_1['weights']), hidden_layer_1['biases']) #(input_data * weights) + biases
l1 = tf.nn.relu(l1) #activation function, im using rectified 
l2 = tf.add(tf.matmul(l1, hidden_layer_2['weights']), hidden_layer_2['biases'])
l2 = tf.nn.relu(l2)
output_layer = tf.matmul(l2, output_layer['weights'] + output_layer['biases'])
return output_layer

def train_neural_network(x):
prediction = neural_network_model(x)
cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(prediction, y)) #loss
optimizer = tf.train.GradientDescentOptimizer(0.1).minimize(cross_entropy)
#koliko puta ce ici back
hm_epoch = 10
with tf.Session() as sess:
sess.run(tf.initialize_all_variables()) 
for step in range(hm_epoch):                
_, c = sess.run([optimizer, cross_entropy], feed_dict={x: x_train, y:[t for t in y_train.as_matrix()]})
print(c)
correct = tf.equal(tf.argmax(prediction, 1), tf.argmax(y, 1))
accuracy = tf.reduce_mean(tf.cast(correct, tf.float32))
#prediction = sess.run(accuracy, feed_dict=(x: x_test, y:[t for t in y_test.as_matrix()]))
#print(prediction)
train_neural_network(x)

我得到这个错误：

Traceback(最后一次调用)：文件"NeuralNet.py"，第92行，在train_neural_network(x)文件"NeuralNet.py"，第83行，在train_neurl_network中_，c=sess.run([优化器，cross_entropy]，feed_dict=｛x:x_train，y:[t for t in y_train.as_matrix()]｝)文件"/home/jusuf/anaconda3/lib/python3.5/site packages/tensorflow/python/client/session.py"，第717行，在runrun_metadata_ptr)文件中"/home/jusuf/anaconda3/lib/python3.5/site packages/tensorflow/python/client/session.py"，在_run np_val＝np.asary(subfeed_val，dtype=subfeed_type)文件"/home/jusuf/anaconda3/lib/python3.5/site packages/numpy/core/number.py"，行482，在阵列中return array(a，dtype，copy=False，order=order)ValueError：用序列设置数组元素。

Answer 1

产生此错误的一个操作是将列表分配给数组元素：

In [498]: x=np.zeros(3)
In [499]: x
Out[499]: array([ 0.,  0.,  0.])
In [500]: x[0] = [1,2,3]
....
ValueError: setting an array element with a sequence.

由于错误在np.asarray(subfeed_val, dtype=subfeed_dtype)语句中，因此它更有可能正在执行以下操作：

In [502]: np.array([[1,2,3],[1,2]], dtype=int)
ValueError: setting an array element with a sequence.

这仍然是试图将一系列数字放入一个插槽的问题。

进一步查看错误堆栈，错误位于：

sess.run([optimizer, cross_entropy], feed_dict={x: x_train, y:[t for t in y_train.as_matrix()]})

我认为这与分配给c无关。

sess.run是一个tensorflow函数，我对此一无所知。

====================

正确格式化的错误堆栈是

Traceback (most recent call last): 
File "NeuralNet.py", line 92, in train_neural_network(x) 
File "NeuralNet.py", line 83, in train_neural_network
_, c = sess.run([optimizer, cross_entropy], feed_dict={x: x_train, y:[t for t in y_train.as_matrix()]}) 
File "/home/jusuf/anaconda3/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 717, in    
runrun_metadata_ptr) 
File "/home/jusuf/anaconda3/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 888, in 
_run np_val = np.asarray(subfeed_val, dtype=subfeed_dtype) 
File "/home/jusuf/anaconda3/lib/python3.5/site-packages/numpy/core/numeric.py", line 482, in 
asarray return array(a, dtype, copy=False, order=order) 
ValueError: setting an array element with a sequence.

我建议查看tensorflow文档，并确保对该函数的输入是正确的。关注允许的类型，如果是数组，则关注维度、形状和数据类型。

Answer 2

错误消息为；

ValueError: setting an array element with a sequence.

解释如下：您正试图用序列设置数组元素。好吧，错误在哪里，见下文；

c = sess.run([optimizer, cross_entropy]

告诉我什么是"c"。它可能是一个浮点，整数或其他什么。但我确信它不是一个数组。这就是您收到上述异常的原因。

但如果你想打印出这个数组，你可以直接打印；

print(sess.run([optimizer, cross_entropy])而不是运行print(c)

就我从您的代码中看到的，您在任何地方都没有使用"c"。