kskshr commited on
Commit
285acff
1 Parent(s): 1127213

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -63,7 +63,7 @@ def preprocess(text):
63
 
64
  from transformers import BertTokenizer, RobertaModel
65
  tokenizer = BertTokenizer.from_pretrained('ku-accms/roberta-base-japanese-ssuw')
66
- model = BertModel.from_pretrained("ku-accms/roberta-base-japanese-ssuw")
67
  text = "京都大学で自然言語処理を専攻する。"
68
  encoded_input = tokenizer(preprocess(text), return_tensors='pt')
69
  output = model(**encoded_input)
@@ -73,7 +73,7 @@ output = model(**encoded_input)
73
  We used a Japanese Wikipedia dump (as of 20230101, 3.3GB) and a Japanese portion of CC100 (70GB).
74
 
75
  ## Training procedure
76
- We first segmented the texts into words by KyTea and then tokenized the words into subwords using WordPiece with a vocabulary size of 32,000. We pre-trained the BERT model using [transformers](https://github.com/huggingface/transformers) library. The training took about 7 days using 4 NVIDIA A100-SXM4-80GB GPUs.
77
 
78
  The following hyperparameters were used for the pre-training.
79
 
 
63
 
64
  from transformers import BertTokenizer, RobertaModel
65
  tokenizer = BertTokenizer.from_pretrained('ku-accms/roberta-base-japanese-ssuw')
66
+ model = RobertaModel.from_pretrained("ku-accms/roberta-base-japanese-ssuw")
67
  text = "京都大学で自然言語処理を専攻する。"
68
  encoded_input = tokenizer(preprocess(text), return_tensors='pt')
69
  output = model(**encoded_input)
 
73
  We used a Japanese Wikipedia dump (as of 20230101, 3.3GB) and a Japanese portion of CC100 (70GB).
74
 
75
  ## Training procedure
76
+ We first segmented the texts into words by KyTea and then tokenized the words into subwords using WordPiece with a vocabulary size of 32,000. We pre-trained the RoBERTa model using [transformers](https://github.com/huggingface/transformers) library. The training took about 7 days using 4 NVIDIA A100-SXM4-80GB GPUs.
77
 
78
  The following hyperparameters were used for the pre-training.
79