在Keras的一个回路中创建LAMDA层的加载模型

我有一个功能模型,它在喀拉斯的一个回路中创建了几个门 Lambda 层:

predictions = []
for ii, kk in enumerate(label_cols):
    slicer = Lambda(lambda x: x[:,:,:,ii:ii+1],
                    output_shape=gates_shape[:-2]+(1,),
                    name='slice_'+kk)
    gate_  = slicer(gates)
    ...
    prediction = Dense(n_classes[kk], activation=final_activation, name=kk)(x)
    predictions.append(prediction)

当我试图加载模型时,它会抱怨:

/anaconda3/lib/python3.6/site-packages/keras/layers/core.py in call(self, inputs, mask)
    691         if has_arg(self.function, 'mask'):
    692             arguments['mask'] = mask
--> 693         return self.function(inputs, **arguments)
    694 
    695     def compute_mask(self, inputs, mask=None):

/anaconda3/lib/python3.6/site-packages/keras/layers/core.py in <lambda>(x)
    334 print('gates_shape', gates_shape)
    335 for ii, kk in enumerate(prms.label_cols):
--> 336     slicer = Lambda(lambda x: x[:,:,:,ii:ii+1], 
    337                     output_shape=gates_shape[:-2]+(1,),
    338                     name='slice_'+kk)

NameError: name 'ii' is not defined

在我添加 兰姆达 . 正确的方法是什么?