Hugging Face's logo

shishirml
/
wav2vec2-large-mms-1b-Nepali-colab-Shishir

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
wav2vec2
Inference Endpoints
Model card Files Metrics Community
1
shishirml commited on
Commit
d8f33c2
•
1 Parent(s): a70a8be

Upload tokenizer

Browse files
Files changed (2) hide show
  1. added_tokens.json +2 -2
  2. vocab.json +84 -93
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 96,
3
- "<s>": 95
4
  }
 
1
  {
2
+ "</s>": 87,
3
+ "<s>": 86
4
  }
vocab.json CHANGED
@@ -1,99 +1,90 @@
1
  {
2
  "npi": {
3
- "[PAD]": 94,
4
- "[UNK]": 93,
5
  "\\": 1,
6
- "a": 2,
7
- "b": 3,
8
- "c": 4,
9
- "e": 5,
10
- "f": 6,
11
- "k": 7,
12
- "o": 8,
13
  "|": 0,
14
- "ँ": 9,
15
- "ं": 10,
16
- "ः": 11,
17
- "अ": 12,
18
- "आ": 13,
19
- "इ": 14,
20
- "ई": 15,
21
- "उ": 16,
22
- "ऊ": 17,
23
- "ऋ": 18,
24
- "ए": 19,
25
- "ऐ": 20,
26
- "ऑ": 21,
27
- "ओ": 22,
28
- "औ": 23,
29
- "क": 24,
30
- "ख": 25,
31
- "ग": 26,
32
- "घ": 27,
33
- "ङ": 28,
34
- "च": 29,
35
- "छ": 30,
36
- "ज": 31,
37
- "झ": 32,
38
- "ञ": 33,
39
- "ट": 34,
40
- "ठ": 35,
41
- "ड": 36,
42
- "ढ": 37,
43
- "ण": 38,
44
- "त": 39,
45
- "थ": 40,
46
- "द": 41,
47
- "ध": 42,
48
- "न": 43,
49
- "प": 44,
50
- "फ": 45,
51
- "ब": 46,
52
- "भ": 47,
53
- "म": 48,
54
- "य": 49,
55
- "र": 50,
56
- "ऱ": 51,
57
- "ल": 52,
58
- "व": 53,
59
- "श": 54,
60
- "ष": 55,
61
- "स": 56,
62
- "ह": 57,
63
- "़": 58,
64
- "ा": 59,
65
- "ि": 60,
66
- "ी": 61,
67
- "ु": 62,
68
- "ू": 63,
69
- "ृ": 64,
70
- "े": 65,
71
- "ै": 66,
72
- "ॉ": 67,
73
- "ॊ": 68,
74
- "ो": 69,
75
- "ौ": 70,
76
- "्": 71,
77
- "ॐ": 72,
78
- "॑": 73,
79
- "ॠ": 74,
80
- "।": 75,
81
- "०": 76,
82
- "१": 77,
83
- "२": 78,
84
- "३": 79,
85
- "४": 80,
86
- "५": 81,
87
- "६": 82,
88
- "७": 83,
89
- "८": 84,
90
- "९": 85,
91
- "॰": 86,
92
- "​": 87,
93
- "‌": 88,
94
- "‍": 89,
95
- "‎": 90,
96
- "‏": 91,
97
- " ": 92
98
  }
99
  }
 
1
  {
2
  "npi": {
3
+ "[PAD]": 85,
4
+ "[UNK]": 84,
5
  "\\": 1,
 
 
 
 
 
 
 
6
  "|": 0,
7
+ "ँ": 2,
8
+ "ं": 3,
9
+ "ः": 4,
10
+ "अ": 5,
11
+ "आ": 6,
12
+ "इ": 7,
13
+ "ई": 8,
14
+ "उ": 9,
15
+ "ऊ": 10,
16
+ "ऋ": 11,
17
+ "ए": 12,
18
+ "ऐ": 13,
19
+ "ओ": 14,
20
+ "औ": 15,
21
+ "क": 16,
22
+ "ख": 17,
23
+ "ग": 18,
24
+ "घ": 19,
25
+ "ङ": 20,
26
+ "च": 21,
27
+ "छ": 22,
28
+ "ज": 23,
29
+ "झ": 24,
30
+ "ञ": 25,
31
+ "ट": 26,
32
+ "ठ": 27,
33
+ "ड": 28,
34
+ "ढ": 29,
35
+ "ण": 30,
36
+ "त": 31,
37
+ "थ": 32,
38
+ "द": 33,
39
+ "ध": 34,
40
+ "न": 35,
41
+ "प": 36,
42
+ "फ": 37,
43
+ "ब": 38,
44
+ "भ": 39,
45
+ "म": 40,
46
+ "य": 41,
47
+ "र": 42,
48
+ "ऱ": 43,
49
+ "ल": 44,
50
+ "व": 45,
51
+ "श": 46,
52
+ "ष": 47,
53
+ "स": 48,
54
+ "ह": 49,
55
+ "़": 50,
56
+ "ा": 51,
57
+ "ि": 52,
58
+ "ी": 53,
59
+ "ु": 54,
60
+ "ू": 55,
61
+ "ृ": 56,
62
+ "े": 57,
63
+ "ै": 58,
64
+ "ॉ": 59,
65
+ "ॊ": 60,
66
+ "ो": 61,
67
+ "ौ": 62,
68
+ "्": 63,
69
+ "॑": 64,
70
+ "ॠ": 65,
71
+ "।": 66,
72
+ "०": 67,
73
+ "१": 68,
74
+ "२": 69,
75
+ "३": 70,
76
+ "४": 71,
77
+ "५": 72,
78
+ "६": 73,
79
+ "७": 74,
80
+ "८": 75,
81
+ "९": 76,
82
+ "॰": 77,
83
+ "​": 78,
84
+ "‌": 79,
85
+ "‍": 80,
86
+ "‎": 81,
87
+ "‏": 82,
88
+ " ": 83
 
 
89
  }
90
  }