Merge of Two CNN models with Keras Concatenate API

[asad raza]

I have trained two CNN of the same architecture with Different Datasets. The first dataset shape is (414938,41) and the second dataset shape is (432302,19). I want to merge them models to create a new ensemble CNN model. I have used Keras API concatenate() but it gives a bellow error.

layers.concatenate([model.input,host_model.input])

ValueError: A `Concatenate` layer requires inputs with matching shapes except for the concat axis. Got inputs shapes: [(None, 41, 1), (None, 19, 1)]

Please there is any technique, code, or other Keras API  so that I can merge two CNN models of different shapes to develop a new ensemble CNN model.

[Lance Norskog]

Trim the final (1,) dimension from each, concatenate the final large axis, add the final (1,) dimension back on:

left = layers.Reshape((-1, 41))(model.input)

right = layers.Reshape((-1, 19))(host_model.input))

combined = layers.Concatenate([left, right])

final = layers.Reshape((-1, 41+19, 1))(combined)

--
You received this message because you are subscribed to the Google Groups "Keras-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keras-users...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/keras-users/5d01cedc-f722-4553-a36b-6a4a844da0d7n%40googlegroups.com.

--

Lance Norskog
lance....@gmail.com
Redwood City, CA

[asad raza]

I have applied that code but it gives the following error

ValueError: Attempt to convert a value (<tensorflow.python.keras.layers.merge.Concatenate object at 0x000001B7B31F3B70>) with an unsupported type (<class 'tensorflow.python.keras.layers.merge.Concatenate'>) to a Tensor.

[Lance Norskog]

It might be:

 layers.Concatenate()([left, right])

I have not used it lately.

[asad raza]

thanks. It now works.