J'essaie d'écrire ma fonction de perte personnalisée: je veux appliquer categorical_crossentropy
Aux parties du vecteur d'entrée puis additionner.
Supposons que y_true, y_pred sont des vecteurs 1D.
Code:
def custom_loss(y_true, y_pred):
loss_sum= 0.0
for i in range(0,y_true.shape[0],dictionary_dims):
loss_sum+= keras.backend.categorical_crossentropy(y_true[i*dictionary_dims:(i+1)*dictionary_dims], y_pred[i*dictionary_dims:(i+1)*dictionary_dims])
return loss_sum
Mais je reçois une erreur:
for i in range(0,y_true.shape[0],dictionary_dims):
TypeError: __index__ returned non-int (type NoneType)
Alors, comment accéder à la forme des tenseurs d'entrée pour obtenir un sous-ensemble de tenseur?
pdate: A également essayé d'écrire directement la perte via tensorflow:
def custom_loss_tf(y_true, y_pred):
print('tf.shape(y_true)',tf.shape(y_true)) #
print('type(tf.shape(y_true))',type(tf.shape(y_true))) #
sys.exit()
loss_sum= 0.0
for i in range(0,y_true.shape[0],dictionary_dims):
loss_sum+= keras.backend.categorical_crossentropy(y_true[i*dictionary_dims:(i+1)*dictionary_dims], y_pred[i*dictionary_dims:(i+1)*dictionary_dims])
return loss_sum
Sortie:
tf.shape(y_true) Tensor("Shape:0", shape=(2,), dtype=int32)
type(tf.shape(y_true)) <class 'tensorflow.python.framework.ops.Tensor'>
Je ne sais pas ce que signifie shape=(2,)
, mais ce n'est pas ce que j'attends, car model.summary()
montre que la dernière couche est (None, 26)
:
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_1 (InputLayer) (None, 80, 120, 3) 0
_________________________________________________________________
conv2d_1 (Conv2D) (None, 80, 120, 32) 896
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 40, 60, 32) 0
_________________________________________________________________
activation_1 (Activation) (None, 40, 60, 32) 0
_________________________________________________________________
conv2d_2 (Conv2D) (None, 40, 60, 32) 9248
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 20, 30, 32) 0
_________________________________________________________________
activation_2 (Activation) (None, 20, 30, 32) 0
_________________________________________________________________
conv2d_3 (Conv2D) (None, 20, 30, 64) 18496
_________________________________________________________________
max_pooling2d_3 (MaxPooling2 (None, 10, 15, 64) 0
_________________________________________________________________
activation_3 (Activation) (None, 10, 15, 64) 0
_________________________________________________________________
conv2d_4 (Conv2D) (None, 10, 15, 64) 36928
_________________________________________________________________
max_pooling2d_4 (MaxPooling2 (None, 5, 7, 64) 0
_________________________________________________________________
activation_4 (Activation) (None, 5, 7, 64) 0
_________________________________________________________________
flatten_1 (Flatten) (None, 2240) 0
_________________________________________________________________
head (Dense) (None, 26) 58266
=================================================================
Deux choses ici:
int_shape
de keras.backend
.int_shape(y_true)[0]
vous renverra une taille de lot. Vous devez utiliser int_shape(y_true)[1]
.