MihaiPopa-1/opus-mt-tatoeba-conlang

This is just a list of Tatoeba snapshots that I used to fine-tune Opus MT! Why? Because we need to like train translation models that support every single language on Earth. Today, we're starting with conlangs, and adding weird languages, and moving out only of Tatoeba and putting more data sources (for more languages)! Variants (I will update later) Variant File Used to Train Model Used Languages Supported Finetuned From opus-mt-en-jbo English-Lojban… See the full description on the dataset page: https://huggingface.co/datasets/MihaiPopa-1/opus-mt-tatoeba-conlang.