File size: 300 Bytes
751936e
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10

import os
from transformers import LlamaTokenizer

CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
TOKENIZER_DIR = os.path.join(CURRENT_DIR, "tokenizer")

tokenizer = LlamaTokenizer.from_pretrained(TOKENIZER_DIR)

tokenizer.comments = "比chinese_llama词典多一个`[PAD]`,请勿混用"