Implement ATT-INPUT method for computing alignments from cross-attention

The HuggingFaceNmtEngine class currently implements the ATT-OUTPUT approach from this paper. The ATT-INPUT method would generate better quality alignments. In order to implement ATT-INPUT, the class would need to shift the attentions to the left one step. This can be done by not adding a 0 matrix at the beginning of the attentions. We would also need to change what layer the attentions are retrieved from (bottom layers). For ATT-INPUT, it is possible for the last token to not get aligned if the translation has hit the max generation length. This edge case should be handled properly.

Locate the HuggingFaceNmtEngine implementation and compare its current ATT-OUTPUT attention handling with the linked paper. Trace where the leading zero matrix and attention layer selection are applied, then determine how ATT-INPUT should handle translations that reach the maximum generation length without aligning the final token. Done means ATT-INPUT produces the intended alignments, including that edge case.