zhangruoceng

Results 1 comments of zhangruoceng

@haotian-liu In my experiment, tokenizer set "use_fast=True" works , with transformers==4.34.1 and tokenizers==0.14.1. But don't know why mismatch when set "use_fast=False"