2

我想在 keras 的剩余块之间添加一个跳过连接。这是我当前的实现,它不起作用,因为张量具有不同的形状。

该函数如下所示:

def build_res_blocks(net, x_in, num_res_blocks, res_block, num_filters, res_block_expansion, kernel_size, scaling):
    net_next_in = net
    for i in range(num_res_blocks):
        net = res_block(net_next_in, num_filters, res_block_expansion, kernel_size, scaling)

        # net tensor shape: (None, None, 32)
        # x_in tensor shape: (None, None, 3)
        # Error here, net_next_in should be in the shape of (None, None, 32) to be fed into next layer
        net_next_in = Add()([net, x_in]) 

    return net

我得到的错误是:ValueError: Operands could not be broadcast together with shapes (None, None, 32) (None, None, 3)

我的问题是如何将这些张量添加或合并成正确的形状(无、无、32)。如果这不是正确的方法,你怎么能达到预期的结果?

编辑:

这是 res_block 的样子:

def res_block(x_in, num_filters, expansion, kernel_size, scaling):
    x = Conv2D(num_filters * expansion, kernel_size, padding='same')(x_in)
    x = Activation('relu')(x)
    x = Conv2D(num_filters, kernel_size, padding='same')(x)
    x = Add()([x_in, x])
return x
4

1 回答 1

3

你不能添加不同形状的张量。您可以使用keras.layers.Concatenate将它们连接起来,但这会给您留下一个 shape 的张量[None, None, 35]

或者,看看 Keras 中的 Resnet50实现。对于要添加的维度不同的情况,它们的残差块在快捷方式中具有 1x1xC 卷积。

于 2019-02-21T14:47:09.443 回答