我想使用自己的binary_crossentropy而不是使用Keras库附带的那个。这是我的自定义功能:
import theano
from keras import backend as K
def elementwise_multiply(a, b): # a and b are tensors
c = a * b
return theano.function([a, b], c)
def custom_objective(y_true, y_pred):
first_log = K.log(y_pred)
first_log = elementwise_multiply(first_log, y_true)
second_log = K.log(1 - y_pred)
second_log = elementwise_multiply(second_log, (1 - y_true))
result = second_log + first_log
return K.mean(result, axis=-1)
注意:这是为了练习。我知道 T.nnet.binary_crossentropy(y_pred,y_true)
但是,当我编译模型时:
sgd = SGD(lr=0.001)
model.compile(loss = custom_objective, optimizer = sgd)
我收到此错误:
----------------------------------------------- ---------------------------- TypeError Traceback(最近一次调用 最后)in() 36 37 sgd = SGD(lr = 0.001) ---> 38 model.compile(loss = custom_objective,optimizer = sgd) 39#==============================================
C:\ Program Files(x86)\ Anaconda3 \ lib \ site-packages \ keras \ models.py in 编译(self,optimizer,loss,class_mode) 418否则: 419 mask =无 - > 420 train_loss = weighted_loss(self.y,self.y_train,self.weights,mask) 421 test_loss = weighted_loss(self.y,self.y_test,self.weights,mask) 422
C:\ Program Files(x86)\ Anaconda3 \ lib \ site-packages \ keras \ models.py in 加权(y_true,y_pred,权重,掩码) 80''' 81#score_array有ndim> = 2 ---> 82 score_array = fn(y_true,y_pred) 83如果面具不是无: 84#mask应该与score_array具有相同的形状
custom_objective中的(y_true,y_pred) 11 second_log = K.log(1 - K.clip(y_true,K.epsilon(),np.inf)) 12 second_log = elementwise_multiply(second_log,(1-y_true)) ---> 13 result = second_log + first_log 14 #result = np.multiply(result,y_pred) 15返回K.mean(结果,轴= -1)
TypeError:+:'Function'和不支持的操作数类型 '功能'
当我用内联函数替换elementwise_multiply时:
def custom_objective(y_true, y_pred):
first_log = K.log(y_pred)
first_log = first_log * y_true
second_log = K.log(1 - y_pred)
second_log = second_log * (1-y_true)
result = second_log + first_log
return K.mean(result, axis=-1)
模型编译但损失值 nan :
大纪元1/1 945/945 [==============================] - 62s - 损失:nan - acc:0.0011 - val_loss:nan - val_acc:0.0000e + 00
有人可以帮我这个吗?!
由于
答案 0 :(得分:4)
我发现了问题。我不得不将返回值乘以“-1”,因为我使用随机梯度下降(sgd)作为优化器而不是随机梯度上升!
这是代码,它就像一个魅力:
import theano
from keras import backend as K
def custom_objective(y_true, y_pred):
first_log = K.log(y_pred)
first_log = first_log * y_true
second_log = K.log(1 - y_pred)
second_log = second_log * (1 - y_true)
result = second_log + first_log
return (-1 * K.mean(result))