An initial impact analysis of @Wikimedia's self hosted machine translation service - MinT:
Supports 206 languages. Increased translated articles, with low deletion rate. Now the second most utilized translation service.
https://t.co/Ya6NIpDVvi
NLLB200, OpusMT, Softcatalà and IndicTrans2 are the current backend models used. This expands our machine translation support to large set of languages.
Wikimedia Foundation announced a new machine translation service to support more underserved languages. Named as MinT, this service uses multiple opensource neural machine translation models and provides a unified interface. https://t.co/B0DMtQlV3n
In July, we launched No Language Left Behind to more languages in @Wikimedia’s Content Translation tool that helps Wikipedia editors jumpstart article translation — a new report is already showing encouraging impact on usage & translation quality ➡️ https://t.co/OHAP22YLbZ
🧵/5
Our goal is to bring the sum of all knowledge to everyone, everywhere, in their language, at no cost.
On this #InternationalTranslationDay, let's talk about one of the tools that's helping us get there. https://t.co/izx14l8RKj
Proceedings of the 22nd Annual Conference of the European Association for Machine Translation is now available https://t.co/91iQ9c52n6
It has a project paper "OPUS-MT – Building open translation services for the World" authored by @TiedemannJoerg and me.
We added another machine translation system - OpusMT - an opensource neural machine translation system based on open corpus project(OPUS). The neural machine translation is powered by MarianNMT.
Currently available for English->Assamese on experimental basis.
The Wikipedia article on the Universal Declaration of Human Rights is still missing in several languages, maybe you can help translating it: https://t.co/kxxyy2PRZy #WikiForHumanRights
The 500,000th @Wikipedia article published using Content Translation is about the French football team Drouais, and it was translated from English to Vietnamese by user Vokaanhduy.
@FCDrouais
https://t.co/T6sCmuBoou
Half a million is an overwhelming number of #Wikipedia articles to write. That is precisely what has been done by the wonderful Wikipedia editors using the Content Translation tool. Thank you! Here is a video from when the 100000th article was written: https://t.co/oLZFvo6iPI
Wikipedia added half a million new articles using its article translation system in last 4 years, to its language editions.
My team at Wikimedia foundation is so proud and thanks every translators for this achievement.
@Wikimedia@WhatToTranslate