Languages supported and context length?

#2
by abpani1994 - opened

Is GLiNER2’s context length of 2,048 based on the number of tokens processed by gliner.data_processor.words_splitter(), or does it use a normal split(' ') word count?
I can see that the vocab size is 250k does that mean it can support 100+ languages like BGE M3?

abpani1994 changed discussion title from Languages supported to Languages supported?
abpani1994 changed discussion title from Languages supported? to Languages supported and context length?

Sign up or log in to comment