/pytorch-simple-seq2seq

Simple Seq2Seq with Attention in Pytorch

Primary LanguagePython

Pytorch Implementation of Simple Seq2Seq with Attention

A Seq2seq network to translate from French to English.

Pytorch version: 1.1.0

Reference:

Dataset:

To evaluate, we sample 5 sentences randomly, and network performance are as follows:

epoch input sentence target sentence output sentence
30 nous sommes deja en retard . we re already late . we re all proud of us .
30 on s occupe bien d eux . they are well looked after . they re jealous of us .
30 tu agis comme un enfant . you re acting like a child . you re acting like a child .
30 elles sont toutes normales . they re all normal . they re all dead .
30 e ne te donnerai aucun argent . i m not giving you any money . i m not giving you any more money .
70 nous sommes deja en retard . we re already late . we re all late .
70 on s occupe bien d eux . they are well looked after . she s a bit in agreement .
70 tu agis comme un enfant . you re acting like a child . you re acting like a child .
70 elles sont toutes normales . they re all normal . they re all mad .
70 e ne te donnerai aucun argent . i m not giving you any money . i m not giving you any money .
100 nous sommes deja en retard . we re already late . we re already late .
100 on s occupe bien d eux . they are well looked after . they are well looked after .
100 tu agis comme un enfant . you re acting like a child . you re acting like a child .
100 elles sont toutes normales . they re all normal . they re all normal .
100 e ne te donnerai aucun argent . i m not giving you any money . i m not giving you any money .