Day 3
從感知機(jī)到復(fù)雜神經(jīng)網(wǎng)絡(luò)
簡(jiǎn)單理解，感知機(jī)可看作是激活函數(shù)為階躍函數(shù)的“神經(jīng)元”习勤。而許多神經(jīng)元進(jìn)行串踪栋、并聯(lián)可以構(gòu)成復(fù)雜的神經(jīng)網(wǎng)絡(luò)，并隨著層數(shù)的加深擁有近乎無(wú)限的表達(dá)能力图毕。
激活函數(shù)
import numpy as np
def step_function(x: np.array):
y: np.array = x > 0
return y.astype(np.int)
def sigmoid(x):
return 1 / (1 + np.exp(-x))
def relu(x):
return np.maximum(0, x)
def identity_function(x):
return x
def softmax(a: np.array):
c = np.max(a)
exp_a = np.exp(a - c) # to avoid overflow
sum_exp_a = np.sum(exp_a)
y: np.array = exp_a / sum_exp_a
return y
多維數(shù)組
import numpy as np
a = np.random.rand(2, 5)
print(a)
a = a.reshape(5, 2)
print(a)
