Thanks for the excellent work!
I'm just wondering did you use some x-to-en labeled dataset to distill large-v3 model to make it support multilingual? or you just used en dataset you mentioned.
Thanks!
· Sign up or log in to comment