Seq2Seq Mode
Implemented a sequence-to-sequence model with an attention mechanism for machine translation using Multi30K machine translation dataset.
Compared the performance of greedy search, beam search, and nucleus sampling strategies used for decoding using BLEU score and translation sample.