Hi.
Anyone got any experience with using (a set of) local LLMs for practicing a new language? (Spanish, not Python). Curious about experiences and knowledge gained.
And, in the extension of that thought, what would be required ‘scaffolding’ around a set of LLMs to be able to:
- assess a student’s current proficiency
- set up some kind of study guide
- provide assignments (vocab training, writing prompts, reading comprehension, speaking exercises, listening exercises)
- evaluate responses to assignments
- give feedback on responses
- keep track of progress over time and adjust assignments accordingly
I *assume* something like this would require multiple LLMs, in order to handle Text To Speech and Automatic Speech Recognition. Is whisper (for example) useful for evaluating (and give feedback on) pronunciation?
The most multi-lingual capable model I’m aware of is OpenBuddy 70b. I use it as a foreign language tutor, and it does an ok job. I constantly check it against google translate, and it hasn’t let me down yet, but ymmv. I don’t use it a ton.
I think the problem is that, in general, technology hasn’t been the best at foreign language translations. Google Translate is SOTA in that realm, and it’s not perfect. I’m not sure I’d trust it for doing this in a real production sense, but I do trust it enough to help me learn just enough to get by.
So with that said, you could likely get halfway far mixing any LLM with a handful of tools. For example- SillyTavern I believe has a Google Translate module built in. You could use Google to do the translations. Then, having multiple speech to text/text to speech modules, one for each language, might give you that flexibility of input and output.
Essentially, I would imagine that 90% of the work will be developing tooling around any decent LLM, regardless of its language abilities, and then using external tooling to support that. I could be wrong, though.