Python Hopfield Network:训练网络 - 权重错误

时间:2014-03-05 19:17:58

标签: python connection iteration neural-network weighted-average

我是编程的新手,我目前正在尝试训练我的Hopfield网络时遇到一些简单的问题但是在尝试计算连接的重量时我一直遇到这个错误。也许我不理解如何“训练”网络,或者我错过了某个地方的某个步骤。但我已经在节点类下定义了函数:

    def update_weight(self):
    for i in self.incoming_connections:
        i.weight += (2*self.activation - 1)*(2*i.sender.activation-1)

哪个应该是正确的,但是当我更新重量,然后输入,然后激活(位于末尾)。我得到一个错误,说“不支持的操作数类型”我的更新权重函数,我不明白。有人可以帮我看看我的问题似乎是什么?

# 
#                               Preparations 
# 

import random
import math
import pygame
nodes=[]
training=[]
NUMNODES=16

# 
#                                   Node Class
# 

class Node(object): 

    def __init__(self,name=None): 
        self.name=name 
        self.activation_threshold=1.0
        self.net_input=0.0
        self.outgoing_connections=[] 
        self.incoming_connections=[]
        self.activation=None

    def __str__(self):
        return self.name

    def addconnection(self,sender,weight=0.0):
        self.incoming_connections.append(Connection(sender,self,weight)) 

    def update_input(self): 
        self.net_input=0.0
        for conn in self.incoming_connections: 
            self.net_input += conn.weight * conn.sender.activation 
        print 'Updated Input for node', str(self), 'is', self.net_input 

    def update_activation(self):
        if self.net_input > self.activation_threshold:
            self.activation = 1.0
            print 'Node', str(self), 'is activated : ', self.activation
        elif self.net_input <= self.activation_threshold:
            self.activation = 0.0
            print 'Node', str(self), 'is not activated : ', self.activation

    def update_training(self):
        Node = random.choice(nodes)

    def update_weight(self):
        for i in self.incoming_connections:
            i.weight += (2*self.activation - 1)*(2*i.sender.activation-1)
            print 'Weight is now set'

# 
#                                   Connection Class
# 

class Connection(object): 
    def __init__(self, sender, reciever, weight): 
        self.weight=weight 
        self.sender=sender 
        self.reciever=reciever

    def __str__(self):
        string = "Connection from " + str(self.sender) + " to " + str(self.reciever) + ", weight = " + str(self.weight)
        return string

# 
#                                 Other Programs 
# 

def set_activations(act_vector): 
    for i in xrange(len(act_vector)): 
        nodes[i].activation = act_vector[i] 

for i in xrange(NUMNODES): 
    nodes.append(Node(str(i)))

for i in xrange(NUMNODES):#go thru all the nodes calling them i 
    for j in xrange(NUMNODES):#go thru all the nodes calling them j 
        if i!=j:#as long as i and j are not the same 
            nodes[i].addconnection(nodes[j])#connects the nodes together

#
#                                         Training Patterns
#

train1=(1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0)
training.append(train1)
train2=(1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0)
training.append(train2)
train3=(1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0)
training.append(train3)
train4=(1.0,0.0,1.0,0.0,1.0,0.0,1.0,0.0,1.0,0.0,1.0,0.0,1.0,0.0,1.0,0.0)
training.append(train4)

set_activations=(train1)

#
#                                        Running 10 Iterations
#

for i in xrange(10):
    print '                     *********** Iteration', str(i+1), '***********'
    for thing in nodes:
        thing.update_weight()
    for thing in nodes:
        thing.update_input()
    for thing in nodes:
        thing.update_activation()

out_file=open('output.txt','w')
out_file.close()

1 个答案:

答案 0 :(得分:0)

这些activation属性中的一个可能是None

    i.weight += (2*self.activation - 1)*(2*i.sender.activation-1)

这是一件好事(与静默失败相反),因为它表示某个节点尚未在某处正确设置的错误。

如果您发布实际的追溯,即使它对您没有意义,也会有所帮助。

修改

看起来这是一个错误

set_activations=(train1)

你应该打电话给set_activations(train1)吗?