multiinput GAN returning error ValueError: Graph disconnected:
multiinput GAN returning error ValueError: Graph disconnected:
整个周末都在尝试解决这个问题。我希望有人能提供帮助。
我有一个模型可以采用普通数组并在 GAN 中对其进行处理，它有效但一旦我将其更改为多输入，我开始得到：
ValueError: Graph disconnected:
我的原码：
# Build stacked GAN model
gan_input = Input(shape=Xtrain.shape[1])
H = generator(gan_input)
gd_input=Concatenate()([gan_input,H])
gan_V = discriminator(gd_input)
GAN = Model(gan_input, [gan_V,H])
GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
GAN.summary()
然后我修改为多输入：
gan_dataframe_input = Input(shape=Xtrain[1][:-2].shape) #new testing
numpy_input = Input(shape=Xtrain[1][-1].shape)
gan_input = layers.concatenate([gan_dataframe_input, numpy_input])
print(gan_input)
print(mergedLayer)
H = generator([gan_dataframe_input,numpy_input]) <<--two shapes being imputed
gd_input=Concatenate()([gan_input,H]) <<--merged layer + above two shapes being imputed
gan_V = discriminator(gd_input)
GAN = Model(gan_input, [gan_V,H]) <<--this line returns an error
GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
GAN.summary()
堆栈跟踪：
KerasTensor(type_spec=TensorSpec(shape=(None, 736), dtype=tf.float32, name=None), name='concatenate_28/concat:0', description="created by layer 'concatenate_28'")
KerasTensor(type_spec=TensorSpec(shape=(None, 736), dtype=tf.float32, name=None), name='concatenate_27/concat:0', description="created by layer 'concatenate_27'")
WARNING:tensorflow:Functional model inputs must come from `tf.keras.Input` (thus holding past layer metadata), they cannot be the output of a previous non-Input layer. Here, a tensor specified as input to "model_34" was not an Input tensor, it was generated by layer concatenate_28.
Note that input tensors are instantiated via `tensor = tf.keras.Input(shape)`.
The tensor that caused the issue was: concatenate_28/concat:0
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-94-ac83091846e6> in <module>()
69 gd_input=Concatenate()([gan_input,H])
70 gan_V = discriminator(gd_input)
---> 71 GAN = Model(gan_input, [gan_V,H])
72 GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
73 GAN.summary()
4 frames
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/functional.py in _map_graph_network(inputs, outputs)
988 'The following previous layers '
989 'were accessed without issue: ' +
--> 990 str(layers_with_complete_input))
991 for x in nest.flatten(node.outputs):
992 computable_tensors.add(id(x))
ValueError: Graph disconnected: cannot obtain value for tensor KerasTensor(type_spec=TensorSpec(shape=(None, 659), dtype=tf.float32, name='input_71'), name='input_71', description="created by layer 'input_71'") at layer "concatenate_28". The following previous layers were accessed without issue: []
奇怪的是看完整轨迹，在图层上打印数据后，数组中的项目数似乎没有对齐？ (659,) 是其中一个输入的大小，而另一个是 (77,)。我不确定我在这里做错了什么。有什么建议吗？
当您构建 multi-input/multi-output 模型时，您必须将模型输入和输出编译并提供为数组，而不是像您那样将它们连接起来。此外，模型的输入必须始终为 tf.keras.layers.Input。所以正确的代码是
gan_dataframe_input = Input(shape=Xtrain[1][:-2].shape) #new testing
numpy_input = Input(shape=Xtrain[1][-1].shape)
gan_input = layers.concatenate([gan_dataframe_input, numpy_input])
print(gan_input)
print(mergedLayer)
H = generator([gan_dataframe_input,numpy_input]) <<--two shapes being imputed
gd_input=Concatenate()([gan_input,H]) <<--merged layer + above two shapes being imputed
gan_V = discriminator(gd_input)
GAN = Model([gan_dataframe_input, numpy_input ], [gan_V,H]) <<--this line is modified
GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
GAN.summary()
整个周末都在尝试解决这个问题。我希望有人能提供帮助。
我有一个模型可以采用普通数组并在 GAN 中对其进行处理，它有效但一旦我将其更改为多输入，我开始得到：
ValueError: Graph disconnected:
我的原码：
# Build stacked GAN model
gan_input = Input(shape=Xtrain.shape[1])
H = generator(gan_input)
gd_input=Concatenate()([gan_input,H])
gan_V = discriminator(gd_input)
GAN = Model(gan_input, [gan_V,H])
GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
GAN.summary()
然后我修改为多输入：
gan_dataframe_input = Input(shape=Xtrain[1][:-2].shape) #new testing
numpy_input = Input(shape=Xtrain[1][-1].shape)
gan_input = layers.concatenate([gan_dataframe_input, numpy_input])
print(gan_input)
print(mergedLayer)
H = generator([gan_dataframe_input,numpy_input]) <<--two shapes being imputed
gd_input=Concatenate()([gan_input,H]) <<--merged layer + above two shapes being imputed
gan_V = discriminator(gd_input)
GAN = Model(gan_input, [gan_V,H]) <<--this line returns an error
GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
GAN.summary()
堆栈跟踪：
KerasTensor(type_spec=TensorSpec(shape=(None, 736), dtype=tf.float32, name=None), name='concatenate_28/concat:0', description="created by layer 'concatenate_28'")
KerasTensor(type_spec=TensorSpec(shape=(None, 736), dtype=tf.float32, name=None), name='concatenate_27/concat:0', description="created by layer 'concatenate_27'")
WARNING:tensorflow:Functional model inputs must come from `tf.keras.Input` (thus holding past layer metadata), they cannot be the output of a previous non-Input layer. Here, a tensor specified as input to "model_34" was not an Input tensor, it was generated by layer concatenate_28.
Note that input tensors are instantiated via `tensor = tf.keras.Input(shape)`.
The tensor that caused the issue was: concatenate_28/concat:0
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-94-ac83091846e6> in <module>()
69 gd_input=Concatenate()([gan_input,H])
70 gan_V = discriminator(gd_input)
---> 71 GAN = Model(gan_input, [gan_V,H])
72 GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
73 GAN.summary()
4 frames
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/functional.py in _map_graph_network(inputs, outputs)
988 'The following previous layers '
989 'were accessed without issue: ' +
--> 990 str(layers_with_complete_input))
991 for x in nest.flatten(node.outputs):
992 computable_tensors.add(id(x))
ValueError: Graph disconnected: cannot obtain value for tensor KerasTensor(type_spec=TensorSpec(shape=(None, 659), dtype=tf.float32, name='input_71'), name='input_71', description="created by layer 'input_71'") at layer "concatenate_28". The following previous layers were accessed without issue: []
奇怪的是看完整轨迹，在图层上打印数据后，数组中的项目数似乎没有对齐？ (659,) 是其中一个输入的大小，而另一个是 (77,)。我不确定我在这里做错了什么。有什么建议吗？
当您构建 multi-input/multi-output 模型时，您必须将模型输入和输出编译并提供为数组，而不是像您那样将它们连接起来。此外，模型的输入必须始终为 tf.keras.layers.Input。所以正确的代码是
gan_dataframe_input = Input(shape=Xtrain[1][:-2].shape) #new testing
numpy_input = Input(shape=Xtrain[1][-1].shape)
gan_input = layers.concatenate([gan_dataframe_input, numpy_input])
print(gan_input)
print(mergedLayer)
H = generator([gan_dataframe_input,numpy_input]) <<--two shapes being imputed
gd_input=Concatenate()([gan_input,H]) <<--merged layer + above two shapes being imputed
gan_V = discriminator(gd_input)
GAN = Model([gan_dataframe_input, numpy_input ], [gan_V,H]) <<--this line is modified
GAN.compile(loss=['categorical_crossentropy','mse'], optimizer=opt) #Complete GAN have both loss functions
GAN.summary()