04-RED-NEURONAL-1

import numpy as np
import scipy as sc
import matplotlib.pyplot as plt
from sklearn.datasets import make_circles

n = 500
p = 2

X,Y = make_circles(n_samples=n, factor=0.5, noise=0.05)

Y = Y[:, np.newaxis]


plt.scatter(X[Y[:, 0] == 0, 0], X[Y[:, 0] == 0, 1], c="skyblue")
plt.scatter(X[Y[:, 0] == 1, 0], X[Y[:, 0] == 1, 1], c="salmon")

plt.axis("equal")
plt.show()

X.shape

(500, 2)

Y.shape

(500, 1)

# CLASE DE LA CAPA DE LA RED

class neural_layer():
    
    def __init__(self, n_conn, n_neur, act_f):
        
        self.act_f = act_f
        
        self.b = np.random.rand(1, n_neur) * 2 - 1
        self.W = np.random.rand(n_conn, n_neur) * 2 -1

# FUNCION DE ACTIVACIÓN

sigm = (lambda x: 1 / (1 + np.e ** (-x)),
        lambda x: x * (1 - x))

relu = lambda x: np.maximum(0, x)

_x = np.linspace(-5, 5, 100)
plt.plot(_x, sigm[0](_x))

[<matplotlib.lines.Line2D at 0x1a18e13990>]

l0 = neural_layer(p, 4, sigm)
l1 = neural_layer(4, 8, sigm)

def create_nn(topology, act_f):
    
    nn = []
    
    for l, layer in enumerate(topology[:-1]):
        
        nn.append(neural_layer(topology[1], topology[l+1], act_f))
        
    return nn

topology = [p, 2, 2, 2, 2, 1]

neural_net = create_nn(topology, sigm)


def train(neural_net, X, Y, l2_cost, lr=0.5, train=True):
    
    out = [(None, X)]
    
    for l, layer in enumerate(neural_net):
        
        z = out[-1][1] @ neural_net[l].W + neural_net[l].b
        
        a = neural_net[l].act_f[0](z)
        
        out.append((z, a))
        
    print(l2_cost[0](out[-1][1], Y))
    
    if train:
        
        deltas = []
        
        for l in reversed(range(0, len(neural_net))):
            
            z = out[l+1][0]
            a = out[l+1][1]
                        
            if l == len(neural_net) - 1:
                deltas.insert(0, l2_cost[1](a, Y) * neural_net[l].act_f[1](a))
            else:
                deltas.insert(0, deltas[0] @ _W.T * neural_net[l].act_f[1](a))
                
            _W = neural_net[l].W 
        
            neural_net[l].b = neural_net[l].b - np.mean(deltas[0], axis=0, keepdims=True) * lr
            
            neural_net[l].W = neural_net[l].W - out[l][1].T @ deltas[0] * lr

            
    return out[-1][1]

train(neural_net, X, Y, l2_cost, 0.5)

0.2653412083117705





array([[0.37616783],
       [0.37613184],
       [0.37614159],
       [0.37613825],
       [0.37616555],
       [0.37612217],
       [0.37612483],
       [0.37614193],
       [0.37612939],
       [0.37616386],
       [0.37611881],
       [0.37614457],
       [0.37616113],
       [0.37612409],
       [0.37614938],
       [0.37615185],
       [0.37613517],
       [0.37615149],
       [0.37615682],
       [0.37613212],
       [0.37615313],
       [0.37614649],
       [0.37613029],
       [0.3761464 ],
       [0.37616104],
       [0.37616889],
       [0.37616569],
       [0.37615057],
       [0.37611912],
       [0.37614661],
       [0.37615321],
       [0.37616688],
       [0.37613438],
       [0.376166  ],
       [0.37613281],
       [0.37613893],
       [0.37616918],
       [0.37615394],
       [0.37613291],
       [0.37613137],
       [0.37614072],
       [0.37614927],
       [0.37615042],
       [0.37616633],
       [0.37611004],
       [0.37615074],
       [0.37613648],
       [0.37615931],
       [0.37615896],
       [0.37614583],
       [0.37615545],
       [0.37616041],
       [0.3761561 ],
       [0.37614   ],
       [0.37614801],
       [0.37614373],
       [0.37615218],
       [0.37613922],
       [0.3761282 ],
       [0.37615009],
       [0.37614267],
       [0.37611572],
       [0.37615898],
       [0.37613267],
       [0.37616768],
       [0.37614669],
       [0.37612807],
       [0.3761333 ],
       [0.37610977],
       [0.37613584],
       [0.37615055],
       [0.37613284],
       [0.37611862],
       [0.37613918],
       [0.37611036],
       [0.37614682],
       [0.37611204],
       [0.37612632],
       [0.37611455],
       [0.37612403],
       [0.37614644],
       [0.3761259 ],
       [0.37615066],
       [0.37611413],
       [0.37613788],
       [0.37615064],
       [0.37614199],
       [0.37612417],
       [0.37613447],
       [0.37611915],
       [0.37611787],
       [0.37614015],
       [0.37613193],
       [0.37613272],
       [0.37614782],
       [0.37616842],
       [0.37615213],
       [0.37614894],
       [0.37616271],
       [0.37614796],
       [0.37610987],
       [0.376129  ],
       [0.3761445 ],
       [0.37614856],
       [0.37613286],
       [0.37612573],
       [0.37616884],
       [0.37612853],
       [0.37611036],
       [0.37613858],
       [0.37613709],
       [0.37616586],
       [0.37614579],
       [0.37612589],
       [0.37615882],
       [0.37614984],
       [0.37616307],
       [0.37616771],
       [0.37615214],
       [0.37612508],
       [0.37615381],
       [0.37612729],
       [0.37612826],
       [0.37611104],
       [0.37616468],
       [0.37612668],
       [0.37612423],
       [0.3761226 ],
       [0.37613331],
       [0.37614683],
       [0.37613586],
       [0.37613362],
       [0.37613297],
       [0.37614683],
       [0.37615312],
       [0.37616574],
       [0.37612467],
       [0.37615333],
       [0.37613576],
       [0.37615714],
       [0.37615059],
       [0.37616875],
       [0.37613365],
       [0.37611385],
       [0.37612321],
       [0.37613181],
       [0.37616679],
       [0.37613288],
       [0.37613672],
       [0.3761172 ],
       [0.37611057],
       [0.37612016],
       [0.37615238],
       [0.37612734],
       [0.37615393],
       [0.37615578],
       [0.37616243],
       [0.37612652],
       [0.37615325],
       [0.37615352],
       [0.37612417],
       [0.37614662],
       [0.37611362],
       [0.37615148],
       [0.37611451],
       [0.37610958],
       [0.3761198 ],
       [0.37612884],
       [0.37611461],
       [0.37615316],
       [0.3761354 ],
       [0.37612961],
       [0.37613226],
       [0.37611096],
       [0.37611566],
       [0.37614167],
       [0.37614247],
       [0.37615191],
       [0.3761121 ],
       [0.37614357],
       [0.37610856],
       [0.3761128 ],
       [0.3761652 ],
       [0.37616669],
       [0.37614519],
       [0.37612811],
       [0.37611365],
       [0.37612065],
       [0.37613507],
       [0.37615428],
       [0.37614616],
       [0.37614109],
       [0.37615979],
       [0.37613679],
       [0.37614616],
       [0.37612094],
       [0.37613493],
       [0.37610931],
       [0.37616323],
       [0.37612367],
       [0.37614359],
       [0.37613126],
       [0.37612607],
       [0.37615491],
       [0.37613992],
       [0.37613775],
       [0.37615364],
       [0.37614055],
       [0.3761548 ],
       [0.37612386],
       [0.37614453],
       [0.37612631],
       [0.37616505],
       [0.37611273],
       [0.37614564],
       [0.37612877],
       [0.37611178],
       [0.37612623],
       [0.37615043],
       [0.37615502],
       [0.37615316],
       [0.37611624],
       [0.37613542],
       [0.37612601],
       [0.37611101],
       [0.37612367],
       [0.37616357],
       [0.37612654],
       [0.37613434],
       [0.37612969],
       [0.37614126],
       [0.37611745],
       [0.37612934],
       [0.37614778],
       [0.37615277],
       [0.37616606],
       [0.37612787],
       [0.37614507],
       [0.3761487 ],
       [0.37614513],
       [0.3761278 ],
       [0.3761146 ],
       [0.37616677],
       [0.37616244],
       [0.37615508],
       [0.37615593],
       [0.37611153],
       [0.37611369],
       [0.37611106],
       [0.37614234],
       [0.37615091],
       [0.37612708],
       [0.37611989],
       [0.3761523 ],
       [0.37611222],
       [0.37613112],
       [0.37616866],
       [0.37615318],
       [0.37615649],
       [0.37611531],
       [0.37614683],
       [0.3761523 ],
       [0.37615242],
       [0.37613708],
       [0.37614725],
       [0.37611752],
       [0.37612905],
       [0.37616838],
       [0.37614697],
       [0.37611292],
       [0.37611691],
       [0.37615635],
       [0.37615263],
       [0.37614055],
       [0.37612446],
       [0.37611407],
       [0.376137  ],
       [0.37616222],
       [0.37611825],
       [0.37615825],
       [0.37616298],
       [0.37612392],
       [0.37615173],
       [0.37614172],
       [0.37616732],
       [0.37613858],
       [0.37614951],
       [0.37612767],
       [0.37615464],
       [0.37615383],
       [0.37613627],
       [0.37612637],
       [0.37612585],
       [0.37616814],
       [0.3761282 ],
       [0.37613465],
       [0.37614725],
       [0.3761407 ],
       [0.37614174],
       [0.37616082],
       [0.37614531],
       [0.37615489],
       [0.37610976],
       [0.3761666 ],
       [0.37612011],
       [0.3761294 ],
       [0.37612462],
       [0.37612767],
       [0.37612541],
       [0.37616201],
       [0.37612002],
       [0.37612799],
       [0.37614105],
       [0.37614617],
       [0.37611541],
       [0.37614697],
       [0.37614719],
       [0.37613335],
       [0.37612466],
       [0.37613958],
       [0.37616684],
       [0.3761329 ],
       [0.37612786],
       [0.37615195],
       [0.37612804],
       [0.3761663 ],
       [0.37615412],
       [0.37611004],
       [0.37615583],
       [0.37611406],
       [0.37612615],
       [0.37615151],
       [0.37613863],
       [0.37612496],
       [0.37614455],
       [0.37613063],
       [0.37614942],
       [0.37612946],
       [0.37615782],
       [0.37610991],
       [0.37612662],
       [0.37612947],
       [0.37614948],
       [0.37614606],
       [0.37613746],
       [0.37616362],
       [0.37612897],
       [0.37612707],
       [0.37612537],
       [0.37614177],
       [0.37613174],
       [0.3761613 ],
       [0.37615418],
       [0.37612856],
       [0.37615135],
       [0.37613187],
       [0.376125  ],
       [0.3761621 ],
       [0.37613843],
       [0.37615704],
       [0.37612691],
       [0.3761213 ],
       [0.37614797],
       [0.37611135],
       [0.37615304],
       [0.37612388],
       [0.37611365],
       [0.37613557],
       [0.37615045],
       [0.37615522],
       [0.37611944],
       [0.37611646],
       [0.37612457],
       [0.37610994],
       [0.37615487],
       [0.37615186],
       [0.37613223],
       [0.37613441],
       [0.37614157],
       [0.37615167],
       [0.37614201],
       [0.37614791],
       [0.37613822],
       [0.37614639],
       [0.37613356],
       [0.37615175],
       [0.37614028],
       [0.37615437],
       [0.37615173],
       [0.37616497],
       [0.3761245 ],
       [0.37611267],
       [0.37615251],
       [0.3761491 ],
       [0.37612896],
       [0.37613779],
       [0.37613594],
       [0.37614289],
       [0.37611112],
       [0.37612604],
       [0.37612986],
       [0.37615078],
       [0.37614659],
       [0.37610895],
       [0.37616904],
       [0.37616926],
       [0.37613455],
       [0.37614706],
       [0.37613026],
       [0.37611502],
       [0.37612805],
       [0.37616198],
       [0.37615314],
       [0.37613759],
       [0.37615413],
       [0.37616585],
       [0.37614069],
       [0.37612972],
       [0.37612963],
       [0.37611227],
       [0.37611062],
       [0.37612902],
       [0.37615512],
       [0.3761267 ],
       [0.37613021],
       [0.37615465],
       [0.37612699],
       [0.37612613],
       [0.3761562 ],
       [0.37612748],
       [0.37614528],
       [0.37615246],
       [0.37612777],
       [0.37614196],
       [0.37611647],
       [0.37615308],
       [0.3761262 ],
       [0.37613306],
       [0.37613772],
       [0.37614343],
       [0.37616724],
       [0.3761353 ],
       [0.37612802],
       [0.37616838],
       [0.37613192],
       [0.37612382],
       [0.37613892],
       [0.37614561],
       [0.37615178],
       [0.37616927],
       [0.37612816],
       [0.37612166],
       [0.37615607],
       [0.3761699 ],
       [0.37613435],
       [0.37616889],
       [0.37616089],
       [0.37612533],
       [0.37612394],
       [0.37615398],
       [0.37615043],
       [0.37615619],
       [0.37616442],
       [0.37614549],
       [0.37616774],
       [0.37614937],
       [0.37614073],
       [0.37612656],
       [0.37615359],
       [0.37612975],
       [0.37614227],
       [0.37614844],
       [0.37613216],
       [0.37614739],
       [0.37612889],
       [0.37613102],
       [0.37615034],
       [0.37616192],
       [0.37616696],
       [0.37611   ],
       [0.37614177],
       [0.37611399],
       [0.37610987],
       [0.37611778],
       [0.37616617],
       [0.37612772],
       [0.37612299],
       [0.37615599],
       [0.37615823],
       [0.3761303 ],
       [0.37613253],
       [0.3761129 ],
       [0.37613343],
       [0.37612415],
       [0.37611552],
       [0.37613917],
       [0.37614523],
       [0.37613512],
       [0.37614579],
       [0.37616397]])

paulpopaul / 04-red-neuronal-1 Goto Github PK

04-red-neuronal-1's Introduction

04-RED-NEURONAL-1

04-red-neuronal-1's People

Contributors

Watchers

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent