tensorflow2.0:ValueError: Operands could not be broadcast together with shapes

最新推荐文章于 2024-07-18 13:58:46 发布

whieper

最新推荐文章于 2024-07-18 13:58:46 发布

阅读量1.3k

点赞数

分类专栏：问题

本文链接：https://blog.csdn.net/qq_42208267/article/details/105071107

版权

在使用TensorFlow 2.0时遇到ValueError，提示操作数无法进行广播配合，形状不匹配。此外，还可能出现TypeError，指出添加的层不是Layer类的实例，而是keras.layers.convolutional.Conv2D对象。问题可能源于使用了tensorflow.keras而非原生keras导致。解决方案是确保所有Sequential和Layers都是从tensorflow.keras导入的，因为TensorFlow 2.0内部的keras与原生keras存在差异。

摘要由CSDN通过智能技术生成

Traceback (most recent call last):
  File "F:/sosoon/PaddlePorject/tf2/CIRFA100resnet.py", line 134, in <module>
    model.build(input_shape=(None, 32, 32, 3))
  File "D:\setup\Anaconda\envs\paddle\lib\site-packages\tensorflow_core\python\keras\engine\network.py", line 673, in build
    self.call(x, **kwargs)
  File "F:/sosoon/PaddlePorject/tf2/CIRFA100resnet.py", line 65, in call
    x = self.layer2(x)
  File "D:\setup\Anaconda\envs\paddle\lib\site-packages\keras\backend\tensorflow_backend.py", line 75, in symbolic_fn_wrapper
    return func(*args, **kwargs)
  File "D:\setup\Anaconda\envs\paddle\lib\site-packages\keras\engine\base_layer.py", line 463, in __call__
    self.build(unpack_singleton(input_shapes))
  File "D:\setup\Anaconda\envs\paddle\lib\site-packages\keras\engine\sequential.py", line 222, in build
    x = layer(x)
  File "D:\setup\Anaconda\envs\paddle\lib\site-packages\keras\backend\tensorfl