xueleecs opened this issue 10 months ago · 1 comments
Thank you so much~
I think it is the sequence that processed by tokenizer