Firefox's Translation Engine is NOT Open Data
Автор: Quippd
Загружено: 2026-02-27
Просмотров: 64
Описание:
Mozilla staff said that the on-device translation models shipped in Firefox were trained on open data.
Unfortunately, that is a lie.
Mozilla states that they use OPUS as their main resource for translation datasets.
A look at Opus reveals that the datasets are not permissively licensed. Instead, they are scraped from copyrighted web sources, exactly like the big tech AI companies.
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке: