qq_41231776的博客

这是我在用keras做神经网络自定义损失函数loss时遇到的问题，代码如下：import tensorflow as tffrom sklearn import datasetsimport numpy as npfrom keras import backend as Kx_train = datasets.load_iris().datay_train = datasets.load_iris().targetnp.random.seed(116)np.random.shuffl