Multilingual Machine Translation Models

Overview

This app holds multilingual machine translation models developed using the M2M100 multilingual encoder-decoder (seq-to-seq) model trained for Many-to-Many multilingual translation.

The M2M100 model was introduced in this paper and first released in this repository; where "a true Many-to-Many multilingual translation model that can translate directly between any pair of 100 languages is created".

The models available here support forward translation from a source language to english and back translation from english to the source language.

Supported Languages

German	French	Spanish	Chinese
Russian	Arabic	Japanese	Indonesian
Turkish	Welsh	Italian	Dutch
Irish	Scottish-gaelic	Polish	Danish
Norwegian	Swedish	Finnish	Greek
Ukrainian	Belarusian	Bulgarian	Romanian
Latvian	Lithuanian	Croatian	Czech
Hungarian	Georgian	Slovak	Slovenian
Albanian	Serbian	Estonian	Georgian
Bosnian	Hebrew	Hindi	Korean
Swahili	Somali	Afrikaans	Amharic
Azerbaijani	Armenian	Bashkir	Bengali
Breton	Catalan	Cebuano	Persian
Fula	Western Frisian	Galician	Gujarati
Hausa	Haitian	Igbo	Iloko
Icelandic	Javanese	Kazakh	Khmer
Kannada	Luxembourgish	Ganda	Lingala
Lao	Malagasy	Macedonian	Malayalam
Mongolian	Marathi	Malay	Burmese
Nepali	Occitan	Oriya	Panjabi
Pashto	Sindhi	Sinhala	Swati
Sundanese	Tamil	Thai	Tagalog
Tswana	Urdu	Uzbek	Vietnamese
Wolof	Xhosa	Yiddish	Yoruba
Zulu