Keras model having multiple inputs causes strange errors when fitting

Question

I am currently working on a encoder-decoder model using GRUs. It takes 2 inputs, encoder input and decoder input. There is only one output from the decoder. The model is:

encoder=tf.keras.layers.GRU(10,return_state=True)
_,state=encoder(encoder_input)

decoder_input=tf.keras.layers.Input(shape=(None,10))
decoder=tf.keras.layers.GRU(10,return_sequences=True)
decoder_output=decoder(decoder_input,initial_state=state)

model=tf.keras.models.Model(inputs=[encoder_input,decoder_input],outputs=decoder_output)

model.compile(optimizer='Adam',loss='MeanSquaredError',metrics=['Accuracy'])

When I try to fit the model with the following pseudocode: model.fit(x=[encoder_data,decoder_data],y=decoder_truth) , encoder_data , decoder_data and decoder_truth all being nested lists of lists and having shape (None,None,10) , and decoder_data and decoder_truth having the same shape

The code raises: ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type list).

Answer 1

Decoder_data and decoder_truth should be the same length as GRUs give one output for each input. Also, the number of time steps per batch should remain constant.