Xenova HF staff commited on
Commit
08f491e
1 Parent(s): 468799b

Upload tokenizer.json

Browse files
Files changed (1) hide show
  1. tokenizer.json +130 -0
tokenizer.json ADDED
@@ -0,0 +1,130 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": null,
4
+ "padding": null,
5
+ "added_tokens": [
6
+ {
7
+ "id": 72,
8
+ "content": "<unk>",
9
+ "single_word": false,
10
+ "lstrip": false,
11
+ "rstrip": false,
12
+ "normalized": false,
13
+ "special": true
14
+ }
15
+ ],
16
+ "normalizer": {
17
+ "type": "Sequence",
18
+ "normalizers": [
19
+ {
20
+ "type": "Lowercase"
21
+ },
22
+ {
23
+ "type": "Replace",
24
+ "pattern": {
25
+ "Regex": "[^\u092b41\\-\u0905\u09078 0\u091b\u0928\u090f\u0910\u093c\u0937\u093f\u0901\u091a2\u0911\u0925\u092d\u0940\u200d\u09453\u091e\u0948\u0941\u0920\u0902\u0949\u0909_\u0908\u0903\u0939\u0927\u0932\u0930\u0938\u092c\u0916\u0923'`\u0935\u0918\u092a\u0917\u0922\u092f\u0947\u094d\u093e\u0906\u0921\u091c\u091d\u0936\u0914\u094b\u0926\u0943\u094c\u090a\u0942\u0913\u091f\u0924\u0915\u092e]"
26
+ },
27
+ "content": ""
28
+ },
29
+ {
30
+ "type": "Strip",
31
+ "strip_left": true,
32
+ "strip_right": true
33
+ },
34
+ {
35
+ "type": "Replace",
36
+ "pattern": {
37
+ "Regex": "(?=.)|(?<!^)$"
38
+ },
39
+ "content": "\u092b"
40
+ }
41
+ ]
42
+ },
43
+ "pre_tokenizer": {
44
+ "type": "Split",
45
+ "pattern": {
46
+ "Regex": ""
47
+ },
48
+ "behavior": "Isolated",
49
+ "invert": false
50
+ },
51
+ "post_processor": null,
52
+ "decoder": null,
53
+ "model": {
54
+ "vocab": {
55
+ "\u092b": 0,
56
+ "4": 1,
57
+ "1": 2,
58
+ "-": 3,
59
+ "\u0905": 4,
60
+ "\u0907": 5,
61
+ "8": 6,
62
+ " ": 7,
63
+ "0": 8,
64
+ "\u091b": 9,
65
+ "\u0928": 10,
66
+ "\u090f": 11,
67
+ "\u0910": 12,
68
+ "\u093c": 13,
69
+ "\u0937": 14,
70
+ "\u093f": 15,
71
+ "\u0901": 16,
72
+ "\u091a": 17,
73
+ "2": 18,
74
+ "\u0911": 19,
75
+ "\u0925": 20,
76
+ "\u092d": 21,
77
+ "\u0940": 22,
78
+ "\u200d": 23,
79
+ "\u0945": 24,
80
+ "3": 25,
81
+ "\u091e": 26,
82
+ "\u0948": 27,
83
+ "\u0941": 28,
84
+ "\u0920": 29,
85
+ "\u0902": 30,
86
+ "\u0949": 31,
87
+ "\u0909": 32,
88
+ "_": 33,
89
+ "\u0908": 34,
90
+ "\u0903": 35,
91
+ "\u0939": 36,
92
+ "\u0927": 37,
93
+ "\u0932": 38,
94
+ "\u0930": 39,
95
+ "\u0938": 40,
96
+ "\u092c": 41,
97
+ "\u0916": 42,
98
+ "\u0923": 43,
99
+ "'": 44,
100
+ "`": 45,
101
+ "\u0935": 46,
102
+ "\u0918": 47,
103
+ "\u092a": 48,
104
+ "\u0917": 49,
105
+ "\u0922": 50,
106
+ "\u092f": 51,
107
+ "\u0947": 52,
108
+ "\u094d": 53,
109
+ "\u093e": 54,
110
+ "\u0906": 55,
111
+ "\u0921": 56,
112
+ "\u091c": 57,
113
+ "\u091d": 58,
114
+ "\u0936": 59,
115
+ "\u0914": 60,
116
+ "\u094b": 61,
117
+ "\u0926": 62,
118
+ "\u0943": 63,
119
+ "\u094c": 64,
120
+ "\u090a": 65,
121
+ "\u0942": 66,
122
+ "\u0913": 67,
123
+ "\u091f": 68,
124
+ "\u0924": 69,
125
+ "\u0915": 70,
126
+ "\u092e": 71,
127
+ "<unk>": 72
128
+ }
129
+ }
130
+ }