sanchit-gandhi HF staff commited on
Commit
b520ace
1 Parent(s): 1fd560e

Upload tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +4 -0
  2. tokenizer_config.json +12 -0
  3. vocab.json +66 -0
special_tokens_map.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "pad_token": "क",
3
+ "unk_token": "<unk>"
4
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_blank": true,
3
+ "clean_up_tokenization_spaces": true,
4
+ "is_uroman": false,
5
+ "language": "new",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "normalize": true,
8
+ "pad_token": "क",
9
+ "phonemize": false,
10
+ "tokenizer_class": "VitsTokenizer",
11
+ "unk_token": "<unk>"
12
+ }
vocab.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ " ": 61,
3
+ "'": 34,
4
+ "-": 43,
5
+ "_": 29,
6
+ "ँ": 8,
7
+ "ं": 44,
8
+ "ः": 30,
9
+ "अ": 22,
10
+ "आ": 52,
11
+ "इ": 50,
12
+ "ई": 13,
13
+ "उ": 17,
14
+ "ऊ": 48,
15
+ "ए": 26,
16
+ "ओ": 16,
17
+ "क": 0,
18
+ "ख": 36,
19
+ "ग": 28,
20
+ "घ": 63,
21
+ "ङ": 27,
22
+ "च": 56,
23
+ "छ": 7,
24
+ "ज": 14,
25
+ "झ": 42,
26
+ "ञ": 19,
27
+ "ट": 10,
28
+ "ठ": 9,
29
+ "ड": 51,
30
+ "ढ": 4,
31
+ "ण": 35,
32
+ "त": 58,
33
+ "थ": 18,
34
+ "द": 21,
35
+ "ध": 54,
36
+ "न": 20,
37
+ "प": 46,
38
+ "फ": 47,
39
+ "ब": 11,
40
+ "भ": 49,
41
+ "म": 6,
42
+ "य": 45,
43
+ "र": 24,
44
+ "ल": 60,
45
+ "व": 41,
46
+ "श": 31,
47
+ "ष": 62,
48
+ "स": 1,
49
+ "ह": 23,
50
+ "ा": 15,
51
+ "ि": 53,
52
+ "ी": 59,
53
+ "ु": 57,
54
+ "ू": 38,
55
+ "ृ": 40,
56
+ "े": 5,
57
+ "ै": 32,
58
+ "ो": 3,
59
+ "ौ": 55,
60
+ "्": 39,
61
+ "०": 33,
62
+ "१": 12,
63
+ "२": 2,
64
+ "६": 37,
65
+ "‍": 25
66
+ }