Let’s democratize NLP for all languages! https://abs.twimg.com/emoji/v2/... draggable="false" alt="🌎" title="Earth globe americas" aria-label="Emoji: Earth globe americas">https://abs.twimg.com/emoji/v2/... draggable="false" alt="🌎" title="Earth globe americas" aria-label="Emoji: Earth globe americas">https://abs.twimg.com/emoji/v2/... draggable="false" alt="🌎" title="Earth globe americas" aria-label="Emoji: Earth globe americas">

Today, with v2.9.1, we are releasing 1,008 machine translation models, covering ` of 140 different languages trained by @jorgtiedemann with @marian, ported by @sam_shleifer. Find your language here: https://huggingface.co/models?search=Helsinki-NLP%2Fopus-mt">https://huggingface.co/models... [1/4]

To celebrate, @Teven, our historian-in-residence, wrote up a history of MT in honor of the release: https://medium.com/huggingface/a-brief-history-of-machine-translation-paradigms-d5c09d8a5b7e">https://medium.com/huggingfa... [2/4]

Here& #39;s a colab for playing with the en-ROMANCE multilingual model https://colab.research.google.com/drive/1z9UtSETxVrDhYnH1eN9lyMv2g-YTFNrz?usp=sharing">https://colab.research.google.com/drive/1z9... [3 / 4]

and a gist of how some of the bilingual models translate the first part of this thread: https://gist.github.com/sshleifer/a5498e4d829a016b5875516d659ed50f">https://gist.github.com/sshleifer... [4/4]

Correction, models were trained with @marian_nmt !