- #1
saminator910
- 96
- 1
I'm trying to make a neural network in python, but I'm having a lot of trouble. Specifically after I have the network set up, what weights to initially assign to each neuron's inputs, the threshold, and evolving the networks to do what you want.
Here is my neuron class.
I am wondering what I should be setting my initial input weights to be? nothing seems to be working...
Here is my neuron class.
Code:
class Neuron:
def __init__(self, inputs):
self.k = inputs
self.t = random.gauss(0,10)
self.b = 1
self.x_avg = inputs**-1
self.x = []
for a in range(0,inputs):
#self.x.append(self.x_avg)
self.x.append(random.gauss(0,10))
def clear(self):
self.x = []
self.b = 1
self.x_avg = 0
def out(self,*insa):
y = []
ins=insa[0]
for a in ins:
y.append(a*self.x[ins.index(a)])
sums = sum(y)-self.t
out = (1+e**-sums)**-1
return out
#def update(self, x_index):
#def update(perchange):
def reset(self):
self.t = random.gauss(0,10)
self.x =[]
for a in range(0,self.k):
#self.x.append(self.x_avg)
self.x.append(random.gauss(0,10))
#def update(self, branch, val):
I am wondering what I should be setting my initial input weights to be? nothing seems to be working...