Question

这是错误。您看到任何不匹配的形状吗？

A `Concatenate` layer requires inputs with matching shapes except for the concat axis. Got inputs shapes: [[(None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1)], [(None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1), (None, 1, 1)]]

TF是不是真的想说些什么？

Answer 1

的确，TF试图说形状还不错，但是有两个列表，它只需要一个列表。（文档和错误消息可能会更好。）我通过实验弄清楚了。形状不是问题。

修复它的代码：

  all_outputs_in_1_list = []
  all_outputs_in_1_list.extend(bias_outputs)
  all_outputs_in_1_list.extend(mm_outputs)
  x = concatenate(all_outputs_in_1_list)

由于需要匹配形状而导致层级错误，但是它们都已经匹配

1 个答案: