msarmi9
/

multi30k

Model card Files Files and versions

Metrics Training metrics Community

Seq2seq + Attention

Pytorch implementation of Neural Machine Translation by Jointly Learning to Align and Translate. Trained on the Multi30k-de-en dataset with sentencepiece as the tokenizer.

Here's the attention heatmap of a random sample from the test set:

Downloads last month: -; Downloads are not tracked for this model. How to track

Space using msarmi9/multi30k 1

Evaluation results

Test BLEU on multi30k-de-en
self-reported

33.468