Update README.md

README.md

@@ -1,3 +1,88 @@
 ---
 license: cc-by-nc-sa-4.0
+datasets:
+- wikipedia
+- cc100
+language:
+- ja
+library_name: transformers
+pipeline_tag: fill-mask
 ---
+
+BERT-base (Vaporetto + BPE)
+===
+
+## How to load the tokenizer
+Please download the dictionary file for Vaporetto + BPE from [our GitHub repository](https://github.com/hitachi-nlp/compare-ja-tokenizer/blob/public/data/dict/vaporetto_bpe.json).
+Then you can load the tokenizer by specifying the path of the dictionary file to `dict_path`.
+
+```python
+from typing import Optional
+
+from tokenizers import Tokenizer, NormalizedString, PreTokenizedString
+from tokenizers.processors import BertProcessing
+from tokenizers.pre_tokenizers import PreTokenizer
+from transformers import PreTrainedTokenizerFast
+
+import vaporetto
+import textspan
+
+class VaporettoPreTokenizer:
+    def __init__(self, unidic_path: str):
+        with open(unidic_path, 'rb') as fp:
+            model = fp.read()
+        self.tokenizer = vaporetto.Vaporetto(model, predict_tags=False)
+    
+    def tokenize(self, sequence: str) -> list[str]:
+        tokens = self.tokenizer.tokenize(sequence)
+        return [token.surface() for token in tokens]
+    
+    def custom_split(self, i: int, normalized_string: NormalizedString) -> list[NormalizedString]:
+        text = str(normalized_string)
+        tokens = self.tokenize(text)
+        tokens_spans = textspan.get_original_spans(tokens, text)
+        return [normalized_string[st:ed] for cahr_spans in tokens_spans for st,ed in cahr_spans]
+    
+    def pre_tokenize(self, pretok: PreTokenizedString):
+        pretok.split(self.custom_split)
+
+# load a pre-tokenizer
+pre_tokenizer = VaporettoPreTokenizer("/path/to/bccwj-suw+unidic+tag.model.zst")
+
+# load a tokenizer
+dict_path = /path/to/vaporetto_bpe.json
+tokenizer = Tokenizer.from_file(dict_path)
+tokenizer.post_processor = BertProcessing(
+    cls=("[CLS]", tokenizer.token_to_id('[CLS]')),
+    sep=("[SEP]", tokenizer.token_to_id('[SEP]'))
+)
+
+# convert to PreTrainedTokenizerFast
+tokenizer = PreTrainedTokenizerFast(
+    tokenizer_object=tokenizer,
+    unk_token='[UNK]',
+    cls_token='[CLS]',
+    sep_token='[SEP]',
+    pad_token='[PAD]',
+    mask_token='[MASK]'
+)
+
+# set a pre-tokenizer
+tokenizer._tokenizer.pre_tokenizer = PreTokenizer.custom(pre_tokenizer)
+```
+
+```python
+# Test
+test_str = "こんにちは。私は形態素解析器について研究をしています。"
+tokenizer.convert_ids_to_tokens(tokenizer(test_str).input_ids)
+# -> ['[CLS]','こん','にち','は','。','私','は','形態','素','解析','器','に','つい','て','研究','を','し','て','い','ます','。','[SEP]']
+```
+
+## How to load the model
+```python
+from transformers import AutoModelForMaskedLM
+model = AutoModelForMaskedLM.from_pretrained("hitachi-nlp/bert-base_vaporetto-bpe")
+```
+
+
+**See [our repository](https://github.com/hitachi-nlp/compare-ja-tokenizer) for more details!**