Error in Neural Network construction

[rodolfobolconte]

I trying to run the original code, but arrived in the follow error when I think occours in NN creation:

  0%|                                                                                                        | 0/209.4375 [00:00<?, ?it/s]
Traceback (most recent call last):
  File "trainer_hrlce.py", line 784, in <module>
    main()
  File "trainer_hrlce.py", line 730, in main
    one_fold(idx, _train_index, _dev_index)
  File "trainer_hrlce.py", line 591, in one_fold
    pred, pred2, pred3 = model(a.cuda(), a_len, b.cuda(), b_len, c.cuda(), c_len, emoji_a.cuda(), emoji_b.cuda(), emoji_c.cuda(), elmo_a, elmo_b, elmo_c)
  File "/home/rodolfo/anaconda3/envs/hrlce/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/rodolfo/avard/HRLCE/SemEval2019Task3/model/hrlce.py", line 144, in forward
    attention_layer=self.a_self_attention)
  File "/home/rodolfo/avard/HRLCE/SemEval2019Task3/model/hrlce.py", line 113, in lstm_forward
    packed_output, hidden = lstm(packed_input, hidden)
  File "/home/rodolfo/anaconda3/envs/hrlce/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/rodolfo/anaconda3/envs/hrlce/lib/python3.7/site-packages/torch/nn/modules/rnn.py", line 759, in forward
    self.check_forward_args(input, hx, batch_sizes)
  File "/home/rodolfo/anaconda3/envs/hrlce/lib/python3.7/site-packages/torch/nn/modules/rnn.py", line 684, in check_forward_args
    self.check_input(input, batch_sizes)
  File "/home/rodolfo/anaconda3/envs/hrlce/lib/python3.7/site-packages/torch/nn/modules/rnn.py", line 207, in check_input
    self.input_size, input.size(-1)))
RuntimeError: input.size(-1) must be equal to input_size. Expected 1324, got 556

Any help to fix this?