Pytorch implementation of English to German Neural machine translation model.
It is a basic Encoder-Decoder RNN with LSTMs implemented on pytorch. Heavily inspired from this tutorial.
Please check other branches of this repo as well. I have implemented several improved versions of this model.
Basic Encoder-Decoder RNN (LSTM) results - Training time: 5m 44.22s (15 epochs) | Testing loss: 2.921 | Testing ppl: 18.554