Multilingual support
#21
by
Fedoration
- opened
What languages can this solution be used for? Is there any support for Russian?
this model was fine-tuned based on Qwen2.5-VL-3B, which originally supports multiple languages, not sure if nanonets also use multiple languages data in their fine-tuning?
Hi.
I've tried it on Russian documents and even handwriting. Worked well. Удачи.
still lack a lot on arabic, did well on french though
What is the best model on arabic?
and will it be worth it doing the training from the start on a bit bigger model like 7B or 32B, on public datasets(i hope there is some..)