kingabzpro commited on
Commit
3e8f7b7
·
1 Parent(s): c0de802

Upload tokenizer

Browse files
Files changed (2) hide show
  1. added_tokens.json +2 -2
  2. vocab.json +67 -93
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 94,
3
- "<s>": 93
4
  }
 
1
  {
2
+ "</s>": 69,
3
+ "<s>": 68
4
  }
vocab.json CHANGED
@@ -1,96 +1,70 @@
1
  {
2
- "[PAD]": 93,
3
- "[UNK]": 92,
4
- "a": 1,
5
- "b": 2,
6
- "c": 3,
7
- "d": 4,
8
- "e": 5,
9
- "f": 6,
10
- "g": 7,
11
- "h": 8,
12
- "i": 9,
13
- "j": 10,
14
- "k": 11,
15
- "l": 12,
16
- "m": 13,
17
- "n": 14,
18
- "o": 15,
19
- "p": 16,
20
- "r": 17,
21
- "s": 18,
22
- "t": 19,
23
- "u": 20,
24
- "v": 21,
25
- "w": 22,
26
- "x": 23,
27
- "y": 24,
28
- "z": 25,
29
  "|": 0,
30
- "ँ": 27,
31
- "ं": 28,
32
- "ः": 29,
33
- "अ": 30,
34
- "आ": 31,
35
- "इ": 32,
36
- "ई": 33,
37
- "उ": 34,
38
- "ऊ": 35,
39
- "ऋ": 36,
40
- "ए": 37,
41
- "ऐ": 38,
42
- "ऑ": 39,
43
- "ओ": 40,
44
- "औ": 41,
45
- "क": 42,
46
- "ख": 43,
47
- "ग": 44,
48
- "घ": 45,
49
- "च": 46,
50
- "छ": 47,
51
- "ज": 48,
52
- "झ": 49,
53
- "ञ": 50,
54
- "ट": 51,
55
- "ठ": 52,
56
- "ड": 53,
57
- "ढ": 54,
58
- "ण": 55,
59
- "त": 56,
60
- "थ": 57,
61
- "द": 58,
62
- "ध": 59,
63
- "न": 60,
64
- "प": 61,
65
- "फ": 62,
66
- "ब": 63,
67
- "भ": 64,
68
- "म": 65,
69
- "य": 66,
70
- "र": 67,
71
- "ल": 68,
72
- "व": 69,
73
- "श": 70,
74
- "ष": 71,
75
- "स": 72,
76
- "ह": 73,
77
- "़": 74,
78
- "ा": 75,
79
- "ि": 76,
80
- "ी": 77,
81
- "ु": 78,
82
- "ू": 79,
83
- "ृ": 80,
84
- "े": 81,
85
- "ै": 82,
86
- "ॉ": 83,
87
- "ो": 84,
88
- "ौ": 85,
89
- "्": 86,
90
- "ज़": 87,
91
- "ड़": 88,
92
- "ढ़": 89,
93
- "फ़": 90,
94
- "।": 91,
95
- "’": 92
96
  }
 
1
  {
2
+ "[PAD]": 67,
3
+ "[UNK]": 66,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  "|": 0,
5
+ "ँ": 1,
6
+ "ं": 2,
7
+ "ः": 3,
8
+ "अ": 4,
9
+ "आ": 5,
10
+ "इ": 6,
11
+ "ई": 7,
12
+ "उ": 8,
13
+ "ऊ": 9,
14
+ "ऋ": 10,
15
+ "ए": 11,
16
+ "ऐ": 12,
17
+ "ऑ": 13,
18
+ "ओ": 14,
19
+ "औ": 15,
20
+ "क": 16,
21
+ "ख": 17,
22
+ "ग": 18,
23
+ "घ": 19,
24
+ "च": 20,
25
+ "छ": 21,
26
+ "ज": 22,
27
+ "झ": 23,
28
+ "ञ": 24,
29
+ "ट": 25,
30
+ "ठ": 26,
31
+ "ड": 27,
32
+ "ढ": 28,
33
+ "ण": 29,
34
+ "त": 30,
35
+ "थ": 31,
36
+ "द": 32,
37
+ "ध": 33,
38
+ "न": 34,
39
+ "प": 35,
40
+ "फ": 36,
41
+ "ब": 37,
42
+ "भ": 38,
43
+ "म": 39,
44
+ "य": 40,
45
+ "र": 41,
46
+ "ल": 42,
47
+ "व": 43,
48
+ "श": 44,
49
+ "ष": 45,
50
+ "स": 46,
51
+ "ह": 47,
52
+ "़": 48,
53
+ "ा": 49,
54
+ "ि": 50,
55
+ "ी": 51,
56
+ "ु": 52,
57
+ "ू": 53,
58
+ "ृ": 54,
59
+ "े": 55,
60
+ "ै": 56,
61
+ "ॉ": 57,
62
+ "ो": 58,
63
+ "ौ": 59,
64
+ "्": 60,
65
+ "ज़": 61,
66
+ "ड़": 62,
67
+ "ढ़": 63,
68
+ "फ़": 64,
69
+ "।": 65
 
70
  }