GenerateKeras produces empty list as output if Layers with multiple outputs are used

[umesh-timalsina]

GenerateKeras produces invalid model code if used on a model as shown in the figure below. Since, LSTM has multiple outputs, it is not able to assign any to the actual Model code.

The code generated is the following(Master Branch):

import keras
from keras.models import Model
from keras.layers import *


input2 = Input(shape=(None, 37), batch_shape=None, dtype=None, sparse=False, tensor=None)
input3 = Input(shape=(None, 37), batch_shape=None, dtype=None, sparse=False, tensor=None)
output, hidden_state, cell_state = LSTM(units=256, activation="tanh", recurrent_activation="hard_sigmoid", use_bias=True, kernel_initializer="glorot_uniform", recurrent_initializer="orthogonal", bias_initializer="zeros", unit_forget_bias=True, kernel_regularizer=None, recurrent_regularizer=None, bias_regularizer=None, kernel_constraint=None, recurrent_constraint=None, bias_constraint=None, dropout=0, recurrent_dropout=0, implementation=1, return_sequences=True, return_state=True, go_backwards=False, stateful=False, unroll=False)(inputs=input3)
output2, hidden_state2, cell_state2 = LSTM(units=256, activation="tanh", recurrent_activation="hard_sigmoid", use_bias=True, kernel_initializer="glorot_uniform", recurrent_initializer="orthogonal", bias_initializer="zeros", unit_forget_bias=True, kernel_regularizer=None, recurrent_regularizer=None, bias_regularizer=None, kernel_constraint=None, recurrent_constraint=None, bias_constraint=None, dropout=0, recurrent_dropout=0, implementation=1, return_sequences=False, return_state=True, go_backwards=False, stateful=False, unroll=False)(inputs=input2, initial_state=[hidden_state, cell_state])

custom_objects = {}


model = Model(inputs=[input3,input2], outputs=[])
result = model
model.custom_objects = custom_objects

[brollb]

Is there a test for this? (I thought you mentioned that there was one before but it might be nice to add it here as a comment.)