A: Hi!

TL;DR supported languages for custom trained text model include: ar, bg, ca, cs, da, de, el, en, es, et, fa, fi, fr, fr-ca, gl, gu, he, hi, hr, hu, hy, id, it, ja, ka, ko, ku, lt, lv, mk, mn, mr, ms, my, nb, nl, pl, pt, pt-br, ro, ru, sk, sl, sq, sr, sv, th, tr, uk, ur, vi, zh-cn, zh-tw.

When you train a custom text model Cogniflow identifies the main language in the dataset and if it's not English it uses instead a cross lingual language model.

The best thing is that once the cross lingual model is trained you can use it with new texts not necessary in the language the model was trained! The predictions with the cross lingual model can be run on any text with any of the following 50+ languages: ar, bg, ca, cs, da, de, el, en, es, et, fa, fi, fr, fr-ca, gl, gu, he, hi, hr, hu, hy, id, it, ja, ka, ko, ku, lt, lv, mk, mn, mr, ms, my, nb, nl, pl, pt, pt-br, ro, ru, sk, sl, sq, sr, sv, th, tr, uk, ur, vi, zh-cn, zh-tw.

What features are NOT supported for other languages different from english?

Well, some pretrained text models are specific for English, these are:

- pretrained emotion detection model works for English only

- pretrained tocix language identification works for English only

- pretrained named entity recognition (NER) works for English only

About pretrained text models, other languages than english are supported in these models:

- pretrained translation model (ES/IT/PT/FR to English)

- pretrained Sentiment Analysis (without neutral)

- pretrained Sentiment Analysis (with neutral)

- pretrained model for language identification in written text

For audio models, the speech2text models are only available for English or Spanish. New languages are planed to be added in the future but not in the short term.

Finally, COR visual models are optimized for Latin languages and English, but they could work with other languages too (decreasing a little bit their performance).

Hope this could be helpful!

Happy training