I get this issues:Dimension out of range (expected to be in range of [-2, 1], but got 2) about a-pytorch-tutorial-to-image-captioning

[Kevinskt]

I have solved this problem,
“Dimension out of range (expected to be in range of [-2, 1], but got 2)”
The reason for this issue is that scores was originally a three-dimensional tensor, but after the operation
scores = pack_padded_sequence(scores, decode_lengths, batch_first=True)[0],
it became a two-dimensional vector.

Here are two solutions:
1.Use the pad_packed_sequence function to restore it.
2.Before reducing dimensionality, use unpacked_scores = scores, then: _, preds = torch.max(unpacked_scores, dim=2).
Note: The original code used scores_copy, I replaced it to avoid errors.