First model version
Browse files- added_tokens.json +7 -0
- special_tokens_map.json +14 -0
- tokenizer_config.json +24 -0
- vocab.txt +0 -0
added_tokens.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</s>": 21130,
|
| 3 |
+
"<mask>": 21132,
|
| 4 |
+
"<pad>": 21128,
|
| 5 |
+
"<s>": 21129,
|
| 6 |
+
"<unk>": 21131
|
| 7 |
+
}
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<s>",
|
| 4 |
+
"<pad>",
|
| 5 |
+
"</s>",
|
| 6 |
+
"<unk>",
|
| 7 |
+
"<mask>"
|
| 8 |
+
],
|
| 9 |
+
"cls_token": "[CLS]",
|
| 10 |
+
"mask_token": "[MASK]",
|
| 11 |
+
"pad_token": "<pad>",
|
| 12 |
+
"sep_token": "[SEP]",
|
| 13 |
+
"unk_token": "[UNK]"
|
| 14 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<s>",
|
| 4 |
+
"<pad>",
|
| 5 |
+
"</s>",
|
| 6 |
+
"<unk>",
|
| 7 |
+
"<mask>"
|
| 8 |
+
],
|
| 9 |
+
"cls_token": "[CLS]",
|
| 10 |
+
"do_basic_tokenize": true,
|
| 11 |
+
"do_lower_case": false,
|
| 12 |
+
"mask_token": "[MASK]",
|
| 13 |
+
"max_len": 512,
|
| 14 |
+
"model_max_length": 512,
|
| 15 |
+
"name_or_path": "bert-base-chinese",
|
| 16 |
+
"never_split": null,
|
| 17 |
+
"pad_token": "<pad>",
|
| 18 |
+
"sep_token": "[SEP]",
|
| 19 |
+
"special_tokens_map_file": null,
|
| 20 |
+
"strip_accents": null,
|
| 21 |
+
"tokenize_chinese_chars": true,
|
| 22 |
+
"tokenizer_class": "BertTokenizer",
|
| 23 |
+
"unk_token": "[UNK]"
|
| 24 |
+
}
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|