ghj
Browse files- tokenizeConfig.py +2 -1
tokenizeConfig.py
CHANGED
@@ -1,7 +1,8 @@
|
|
|
|
1 |
from tokenizers import Tokenizer, models, pre_tokenizers, trainers, decoders
|
2 |
import json
|
3 |
|
4 |
-
class OBITokenizer:
|
5 |
def __init__(self):
|
6 |
# Initialize a BPE model for tokenization
|
7 |
bpe_model = models.BPE()
|
|
|
1 |
+
from transformers import PreTrainedTokenizer
|
2 |
from tokenizers import Tokenizer, models, pre_tokenizers, trainers, decoders
|
3 |
import json
|
4 |
|
5 |
+
class OBITokenizer(PreTrainedTokenizer):
|
6 |
def __init__(self):
|
7 |
# Initialize a BPE model for tokenization
|
8 |
bpe_model = models.BPE()
|