Forward prop in tensorflow——MLP

1.how to build layers

1. Manually（1x3matrix）

   x = np.arrat([200.0, 17.0])
   layer_1 = Dense(units=3, activation='sigmoid')
   a1 = layer_1(x)
   layer_2 = Dense(units=1, activation='sigmoid')
   a2 = layer_2(a1)
   #a1&a2 are stored in TensorFlow format. Run this to convert it to NumPy: a1.numpy().
   

2. Automatically

   layer_1 = Dense(units=3, activation='sigmoid')
   layer_2 = Dense(units=1, activation='sigmoid')
   model = Sequential([layer_1, layer_2])
   

3. Compile the model

   model.compile(...)
   x = np.array([...])
   y = np.array([...])
   model.fit(x,y)
   model.predict(x_test)
   

4. All

   model = Sequential([
     Dense(units=25, activation='sigmoid'),
     Dense(units=15, activation='sigmoid'),
     Dense(units=1, activation='sigmoid')])
   model.compile(...)
   x = np.array([...])
   y = np.array([...])
   model.fit(x,y)
   model.predict(x_test)
   

5. def function to return a for next layer

   def dense(a_in, W, b, g):  # Here W is 2by3 
       units = W.shape[1]       # Obtain columns in W(=number of neurons in this layer)
       a_out = np.zeros(units)
       for j in range(units):
         w = W[:,j]
         z = np.dot(w,a_in) + b[j]
         a_out[j] = g[z]
       return a_out
       # capital letters refer to a matrix
   

6. Sequential Structure

   def sequential(x):
     a1 = dense(x, W1, b1) 
     a2 = dense(x, W2, b2) 
     a3 = dense(x, W3, b3) 
     a4 = dense(x, W4, b4) 
     f_x = a4
     return f_x
   

7. example

# 构建网络层
import tensorflow as ts
from tensorflow.keras import Squential
  model = Sequential([
      Dense(units=25, activation='sigmoid'),
      Dense(units=15, activation='sigmoid'),
      Dense(units=1, activation='sigmoid')])
# 编译模型
from tensorflow.keras.losses import BinaryCrossentropy
  model.compile(loss=BinaryCrossentropy())
# 训练
model.fit(X,Y,epochs=100)