-
Notifications
You must be signed in to change notification settings - Fork 27
/
vectorized_costNN.py
71 lines (50 loc) · 1.86 KB
/
vectorized_costNN.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
# -*- coding: utf-8 -*-
"""
Created on Fri May 27 12:03:15 2016
@author: hossam
"""
import numpy as np
import neurolab as nl
import time
#import warnings
#warnings.filterwarnings("ignore")
def costNN(x,inputs,outputs,net):
trainInput=inputs
trainOutput=outputs
numInputs=np.shape(trainInput)[1] #number of inputs
#number of hidden neurons
HiddenNeurons = net.layers[0].np['b'][:].shape[0]
popSize = len(x)
######################################
split1=HiddenNeurons*numInputs
split2=split1+HiddenNeurons
split3=split2+HiddenNeurons
# input_w = 3X8 (HiddenNeurons*numInputs)
input_w =x[:, 0:split1].reshape(popSize, HiddenNeurons,numInputs)
# layer_w = 1 X 3 (HiddenNeurons)
layer_w=x[:, split1:split2].reshape(popSize, 1,HiddenNeurons)
# input_bias = hiddenNeurons
input_bias=x[:, split2:split3].reshape(popSize, 1,HiddenNeurons)
#input_bias = np.array([0.4747,-1.2475,-1.2470])
# bias_2 = 1
bias_2 =x[:, split3:split3+1]
nets = np.array([net] * popSize)
nets = np.array(list(map(updateLayers, nets, input_w, layer_w, input_bias, bias_2)))
'''
net.layers[0].np['w'][:] = input_w
net.layers[1].np['w'][:] = layer_w
net.layers[0].np['b'][:] = input_bias
net.layers[1].np['w'][:] = bias_2
'''
pred = np.array([net.sim(trainInput).reshape(len(trainOutput)) for net in nets])
#pred=net.sim(trainInput).reshape(len(trainOutput))
trainOutputs = np.array([trainOutput] * popSize)
mse = ((pred - trainOutputs) ** 2).mean(axis=1)
return mse
def updateLayers(net, input_w, layer_w, input_bias, bias_2):
newNet = net.copy()
newNet.layers[0].np['w'][:] = input_w
newNet.layers[1].np['w'][:] = layer_w
newNet.layers[0].np['b'][:] = input_bias
newNet.layers[1].np['b'][:] = bias_2
return newNet