NRMS Issue on running

[tnnanh1010]

Traceback (most recent call last):
File "/content/NewsRecommendation/src/main.py", line 295, in
train(None, args)
File "/content/NewsRecommendation/src/main.py", line 91, in train
bz_loss, y_hat = model(log_ids, log_mask, input_ids, targets)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/dist-packages/torch/nn/modules/module.py", line 1736, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/dist-packages/torch/nn/modules/module.py", line 1747, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/content/NewsRecommendation/src/model/NRMS.py", line 87, in forward
candidate_news_vecs = self.news_encoder(candidate_news).reshape(-1, 1 + self.args.npratio, self.args.news_dim)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: shape '[-1, 5, 400]' is invalid for input of size 70400

Are there any modification with NRMS encoder