API, auto-detect, NLLB-200

[MeinDeutschkurs]

Hello everyone,
I love the translation feature. I have some questions:

1. is it possible to call it over API?
2. is it possible to autodetect the source language?
3. is it possible to use NLLB-200 anyhow?

Thank you so much,
Johannes

[ParisNeo]

Anything is possible. It is just a matter of having time to do it :)

I'm working on the lollms project which takes all my time that's why I'm not advancing this one that much.
The LoLLMs's artbot allows generating images from any language. It creafts the prompt for you. So you can describe what you want to do in any language and it will craft an english prompt. It is still in its beginnings so many features are not yet implemented.

[MeinDeutschkurs]

Anything is possible. It is just a matter of having time to do it :)

I'm working on the lollms project which takes all my time that's why I'm not advancing this one that much. The LoLLMs's artbot allows generating images from any language. It creafts the prompt for you. So you can describe what you want to do in any language and it will craft an english prompt. It is still in its beginnings so many features are not yet implemented.

Thx for reply. I have not much experience but...

I currently experiment with NLLB-200 to autodetect the language, and to translate it then to a specified output language, like "eng_Latn" or "deu_Latn". Works fine. I also try to autodetect if a content is not suitable for all the audience. For this I let comment gpt4all's model the English translated version of the prompt. It's outstanding and saves money in comparison to gpt-3.5-turbo/text-davinci-003. But it needs lots of resources. So this does not work for all.

I hope to get into grado and fastAPI at the weekend. I love the queue-system that is supported by AUTOMATIC1111. I hope that I can fuse my testings to it.

[OfficiallyCrazy]

Any updates on this? I owuld love to have this over API.