# Loss decrease different between the pytorch and keras

hi,all:
I meet a problem when I rewrite the network form keras
in keras, i use
model.compile(optimizer=optimizer, loss=‘binary_crossentropy’, metrics=[‘accuracy’])
in keras,and use
Could you make sure to use the same `torch.nn.init` methods for your layers as in your Keras model?
`torch.nn.init.kaiming_normal` corresponds to Keras’ `he_normal`.