03_Deeplearning_build a neural network

x = np.array([[200.0, 17.0]])
layer_1 = Dense(units=3,activation='sigmoid') #layer1,neurons:3,activation
#activation function is sigmoid
a1 = layer_1(x)   #a1 is a vector which has three elements

layer_2 = Dense(units=1,activation='sigmoid')#layer2:neurons:1,activation:sigmoid
a2 = layer_2(a1)   #a2 has a number

just as same as before

arrays and metrix

metrix:two []

tensorflow deal with metrix.

np.array([[]]):2D-array - metrix

np.array([]):1-d, number list

building a neuron network

the first method to creat the layer.

x = np.array([[200.0,17.0]]) #initialize x
layer_1 =Dense(units=3,activation='sigmoid') # creat layer_1
a1 = layer_1(x)								#get the layer_1 output
layer_2 = Dense(units=1, activation = 'sigmoid') #creat layer_2
a2 = layer_2(a1)                            #get the layer_2 output

the second one

layer_1 = Dense(units = 3,activation = "sigmoid")
layer_2 = Dense(units = 1, activation = "sigmoid")
model = Sequential([layer_1, layer_2]) #connect the two layers
x = np.array([[200.0,17.0],
			   [120.0,5.0],
			   [425.0,20.0],
			   [212.0,18.0]])    #initialize the x
y = np.array([1,0,0,1])
model.compile(...)
model.fit(x,y)			#get the data to practise
model.predict(x_new)

the third way

model = Sequential([
Dense(units=3,activation="sigmoid")
Dense(units=1,activation="sigmoid")
])
x = np.array([[200.0,17.0],
			   [120.0,5.0],
			   [425.0,20.0],
			   [212.0,18.0]])    #initialize the x
y = np.array([1,0,0,1])
model.compile(...)
model.fit(x,y)			#get the data to practise
model.predict(x_new)

Digit Classification model

layer_1 = Dense(units = 25,activation= "sigmoid")
layer_2 = Dense(units = 15,activation = "sigmoid")
layer_3 = Dense(units = 1,activation = "sigmoid")
model = Sequential([layer_1,layer_2,layer_3])
model.compile(...)
x = np.array([0...,245,...,17],
			  [0...,200,...,184])
y = np.array([1,0])
model.fit(x,y)
model.predict(x_new)