All inputs to the layer should be tensors.

[aloreggia]

Hi,
I'm trying to use GCN for graph learning, but when I run your train.py script I get the following error:

Using local pooling filters...
Traceback (most recent call last):
File "train.py", line 55, in
H = GraphConvolution(16, support, activation='relu', kernel_regularizer=l2(5e-4))([H]+G)
File "/Users/aloreggia/Documents/anaconda2/lib/python2.7/site-packages/keras/engine/topology.py", line 575, in call
self.assert_input_compatibility(inputs)
File "/Users/aloreggia/Documents/anaconda2/lib/python2.7/site-packages/keras/engine/topology.py", line 448, in assert_input_compatibility
str(inputs) + '. All inputs to the layer '
ValueError: Layer graph_convolution_1 was called with an input that isn't a symbolic tensor. Received type: <class 'theano.sparse.basic.SparseVariable'>. Full input: [if{}.0, SparseVariable{csr,float32}]. All inputs to the layer should be tensors.

Can you help me solving this issue?

Thank you
Andrea

[tkipf]

Maybe this is a keras version issue. Can you try with v. 1.0.9?

[aloreggia]

Keras version is 2.2.0. I'm using the Anaconda environment.

[tkipf]

Thanks - it could be that 2.2.0 is not working due to some change in the API. Can you try v.1.0.9?

[aloreggia]

I think I've just get where the code fails with this version of Keras. Creating the Input layers
G = [Input(shape=(None, None), batch_shape=(None, None))]

with no specification about sparsity, everything works perfectly. Is this correct?

Thanks
Andrea

[tkipf]

Interesting, maybe the new keras version doesn't need the sparsity argument anymore for sparse inputs. Glad you could figure it out. Let's leave this issue open so others can find it in case they run into the same problem.

[tehreemnaqvi]

Hi,
I also got this error in keras.
Can anyone tell me how to fix it?
Thanks

[thisishardik]

I was able to solve this error. Probably my solution helps.
What you need to do is define a keras Input layer, for e.g.
inputs = tf.keras.Input(shape=(128, 128, 3)) to feed into the first layer of your model.
Use it like this -
conv_1 = tf.keras.layers.Conv2D(filters=filters, kernel_size=kernel_size, padding=padding)(inputs)

Hope this helps !!

[mrtrieuphong]

Thank you @thisishardik, it's work for me!

I was able to solve this error. Probably my solution helps. What you need to do is define a keras Input layer, for e.g. inputs = tf.keras.Input(shape=(128, 128, 3)) to feed into the first layer of your model. Use it like this - conv_1 = tf.keras.layers.Conv2D(filters=filters, kernel_size=kernel_size, padding=padding)(inputs)

Hope this helps !!