0

我创建了一个带有分类标签的以下连体神经网络

EXAMPLES=10000
FEATURES=30
LEFT=np.random.random((EXAMPLES,FEATURES))
RIGHT=np.random.random((EXAMPLES,FEATURES))
LABELS=[]
for i in range(EXAMPLES):
    LABELS.append(np.random.randint(0,2))
LABELS=np.asarray(LABELS)

LABELSOFTMAX=to_categorical(LABELS)

def cosine_distance(vecs):
    #I'm not sure about this function too
    y_true, y_pred = vecs
    y_true = K.l2_normalize(y_true, axis=-1)
    y_pred = K.l2_normalize(y_pred, axis=-1)
    return K.mean(1 - K.sum((y_true * y_pred), axis=-1))

def cosine_dist_output_shape(shapes):
    shape1, shape2 = shapes
    print((shape1[0], 1))
    return (shape1[0], 1)






inputShape=Input(shape=(FEATURES,))
left_input = Input(shape=(FEATURES,))
right_input = Input(shape=(FEATURES,))

首次实施

model = Sequential()
model.add(Dense(20, activation='relu', input_shape=(30,)))
model.add(BatchNormalization())
model.add(Dense(10, activation='relu'))



encoded_l = model(left_input)
encoded_r = model(right_input)

L1_Distance = Lambda(cosine_distance, output_shape=cosine_dist_output_shape)([encoded_l, encoded_r])
siamese_net = Model([left_input, right_input], L1_Distance)

siamese_net.summary()
siamese_net.compile(loss="binary_crossentropy",optimizer=Adam(lr=0.0001))
siamese_net.fit(x=[LEFT,RIGHT],y=LABELS,batch_size=64,epochs=100)

第一个模型总结

(None, 1)
Model: "model_28"
__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
input_132 (InputLayer)          (None, 30)           0                                            
__________________________________________________________________________________________________
input_133 (InputLayer)          (None, 30)           0                                            
__________________________________________________________________________________________________
sequential_44 (Sequential)      (None, 10)           910         input_132[0][0]                  
                                                                 input_133[0][0]                  
__________________________________________________________________________________________________
lambda_37 (Lambda)              (None, 1)            0           sequential_44[1][0]              
                                                                 sequential_44[2][0]              
==================================================================================================
Total params: 910
Trainable params: 870
Non-trainable params: 40

二次实施

model = Sequential()
model.add(Dense(20, activation='relu', input_shape=(30,)))
model.add(BatchNormalization())
model.add(Dense(10, activation='relu'))

#model.add(Dense(30, activation='relu'))

   
    
encoded_l = model(left_input)
encoded_r = model(right_input)


L1_Layer = Lambda(cosine_distance, output_shape=cosine_dist_output_shape)([encoded_l, encoded_r])
L1_Diatance = L1_layer([encoded_l, encoded_r])
prediction = Dense(2,activation='softmax')(L1_Diatance)

siamese_net = Model([left_input, right_input], prediction)
siamese_net.compile(loss="binary_crossentropy",optimizer=Adam(lr=0.001))
siamese_net.summary()

第二个模型总结

Model: "model_29"
__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
input_135 (InputLayer)          (None, 30)           0                                            
__________________________________________________________________________________________________
input_136 (InputLayer)          (None, 30)           0                                            
__________________________________________________________________________________________________
sequential_45 (Sequential)      (None, 10)           910         input_135[0][0]                  
                                                                 input_136[0][0]                  
__________________________________________________________________________________________________
lambda_19 (Lambda)              multiple             0           sequential_45[1][0]              
                                                                 sequential_45[2][0]              
__________________________________________________________________________________________________
dense_140 (Dense)               (None, 2)            22          lambda_19[10][0]                 
==================================================================================================
Total params: 932
Trainable params: 892
Non-trainable params: 40

所以我的问题是哪个是正确和更好的实现，因为两者都工作正常，它们与这些模型中的任何一个是否存在一些微妙的问题，因为我相信余弦相似层只会给出一个缩放器张量，这在这种情况下让我感到困惑？

4

0 回答 0