python - 喀拉斯 : TypeError: 'AddNL' object has no attribute '__getitem__'

Question

我正在使用 Keras 功能 API 和 TensorFlow 后端来定义卷积层，并且我必须实现自定义合并层。

假设:

# All necessary libraries imported.
# image_shape = (28, 28, 3)
# 16 and (3,3) are kernel_count and kernel_size respectively.
def MyModel(image_shape):
    input = Input(shape=image_shape)
    conv1 = Conv2D(16, (3,3), kernel_initializer='he_normal')(input)
    conv2 = Conv2D(16, (3,3), kernel_initializer='he_normal')(input)
    conv3 = AddNL()[conv1, conv2]

我的自定义合并层AddNL在keras.layers.merge中实现，如下所示:

class AddNL(_Merge):
    def _merge_function(self, inputs):
        nK1 = inputs[0].shape[-1]
        nK2 = inputs[1].shape[-1]
        # <?, image,image, channel> Channel = Number of Kernels.
        k = nK1 if nK1 < nK2 else nK2
        output = input[0][...,0:k] + input[1][...,0:k] 
        return output

当我的代码到达该行时，抛出以下TypeError:

类型错误:“AddNL”对象没有属性“__getitem__”

Answer 1

有一个拼写错误。代码应如下所示:

conv3 = AddNL([conv1, conv2])