Added gold standard data.seq2seq with attention. Training works. Inference models have an error:

Traceback (most recent call last):
  File "seq2seq_attention.py", line 163, in <module>
    outputs=[decoder_out] + decoder_states )
  File "/home/sevajuri/anaconda3/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper
    return func(*args, **kwargs)
  File "/home/sevajuri/anaconda3/lib/python3.6/site-packages/keras/engine/topology.py", line 1811, in __init__
    str(layers_with_complete_input))
RuntimeError: Graph disconnected: cannot obtain value for tensor Tensor("input_1_1:0", shape=(?, 27), dtype=float32) at layer "input_1". The following previous layers were accessed without issue: ['input_2']