(fix): align quantized models tokenizers with un-quantized ones (#4) ff136ed verified ybelkada commited on 6 days ago