Commit
·
8aa0d8b
1
Parent(s):
2151e84
Upload tokenizer (#4)
Browse files- Upload tokenizer (5e552f66bbdfdd24a9e0701085768861dd57bd47)
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +48 -0
- vocab.txt +0 -0
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"clean_up_tokenization_spaces": true,
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"do_basic_tokenize": true,
|
| 5 |
+
"do_lower_case": false,
|
| 6 |
+
"mask_token": "[MASK]",
|
| 7 |
+
"max_len": 512,
|
| 8 |
+
"model_max_length": 512,
|
| 9 |
+
"never_split": [
|
| 10 |
+
"+ك",
|
| 11 |
+
"+كما",
|
| 12 |
+
"ك+",
|
| 13 |
+
"+وا",
|
| 14 |
+
"+ين",
|
| 15 |
+
"و+",
|
| 16 |
+
"+كن",
|
| 17 |
+
"+ان",
|
| 18 |
+
"+هم",
|
| 19 |
+
"+ة",
|
| 20 |
+
"[بريد]",
|
| 21 |
+
"لل+",
|
| 22 |
+
"+ي",
|
| 23 |
+
"+ت",
|
| 24 |
+
"+ن",
|
| 25 |
+
"س+",
|
| 26 |
+
"ل+",
|
| 27 |
+
"[مستخدم]",
|
| 28 |
+
"+كم",
|
| 29 |
+
"+ا",
|
| 30 |
+
"ب+",
|
| 31 |
+
"ف+",
|
| 32 |
+
"+نا",
|
| 33 |
+
"+ها",
|
| 34 |
+
"+ون",
|
| 35 |
+
"+هما",
|
| 36 |
+
"ال+",
|
| 37 |
+
"+ه",
|
| 38 |
+
"+هن",
|
| 39 |
+
"+ات",
|
| 40 |
+
"[رابط]"
|
| 41 |
+
],
|
| 42 |
+
"pad_token": "[PAD]",
|
| 43 |
+
"sep_token": "[SEP]",
|
| 44 |
+
"strip_accents": null,
|
| 45 |
+
"tokenize_chinese_chars": true,
|
| 46 |
+
"tokenizer_class": "BertTokenizer",
|
| 47 |
+
"unk_token": "[UNK]"
|
| 48 |
+
}
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|