原因是因为keras在建立网格是, 将数据流定义为Layer和Tensor会产生不兼容的现象, 因此会有这种错误.
只需要将我们tensor要进行的运算用keras提供的Lambda封装一下就可以将Tensor类型转化为为Layer类型.
出现错误提示的代码:
def TotalModel(input_shape, noise_model, avo_model, verbose = 1):
noise_input = Input(input_shape)
noise_output = noise_model(noise_input)
avo_output = avo_model(noise_input)
total_output =avo_output + noise_output
total = Model(inputs = noise_input, outputs = total_output)
if verbose == 1:
print()
total.summary()
return total
修改为:
def TotalModel(input_shape, noise_model, avo_model, verbose = 1):
noise_input = Input(input_shape)
noise_output = noise_model(noise_input)
avo_output = avo_model(noise_input)
total_output = Lambda(lambda x: x[0] + x[1])([avo_output, noise_output])
total = Model(inputs = noise_input, outputs = total_output)
if verbose == 1:
print()
total.summary()
return total
参考链接:
https://blog.csdn.net/scythe666/article/details/79962836
https://codeday.me/bug/20190131/581332.html
AttributeError: 'Tensor' object has no attribute '_keras_history'
原文：https://www.cnblogs.com/daangzi96/p/11858426.html