Hugging Face's logo

pius-code
/
asha_twi_adapter

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Model card Files
xet
 Community
pius-code commited on
Commit
f3aa9a8
·
verified ·
1 Parent(s): 46d6cb5

Upload tokenizer

Browse files
Files changed (3) hide show
  1. added_tokens.json +2 -2
  2. tokenizer_config.json +4 -4
  3. vocab.json +68 -41
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 43,
3
- "<s>": 42
4
  }
 
1
  {
2
+ "</s>": 70,
3
+ "<s>": 69
4
  }
tokenizer_config.json CHANGED
@@ -8,7 +8,7 @@
8
  "single_word": false,
9
  "special": true
10
  },
11
- "40": {
12
  "content": "[UNK]",
13
  "lstrip": true,
14
  "normalized": false,
@@ -16,7 +16,7 @@
16
  "single_word": false,
17
  "special": false
18
  },
19
- "41": {
20
  "content": "[PAD]",
21
  "lstrip": true,
22
  "normalized": false,
@@ -24,7 +24,7 @@
24
  "single_word": false,
25
  "special": false
26
  },
27
- "42": {
28
  "content": "<s>",
29
  "lstrip": false,
30
  "normalized": false,
@@ -32,7 +32,7 @@
32
  "single_word": false,
33
  "special": true
34
  },
35
- "43": {
36
  "content": "</s>",
37
  "lstrip": false,
38
  "normalized": false,
 
8
  "single_word": false,
9
  "special": true
10
  },
11
+ "67": {
12
  "content": "[UNK]",
13
  "lstrip": true,
14
  "normalized": false,
 
16
  "single_word": false,
17
  "special": false
18
  },
19
+ "68": {
20
  "content": "[PAD]",
21
  "lstrip": true,
22
  "normalized": false,
 
24
  "single_word": false,
25
  "special": false
26
  },
27
+ "69": {
28
  "content": "<s>",
29
  "lstrip": false,
30
  "normalized": false,
 
32
  "single_word": false,
33
  "special": true
34
  },
35
+ "70": {
36
  "content": "</s>",
37
  "lstrip": false,
38
  "normalized": false,
vocab.json CHANGED
@@ -1,46 +1,73 @@
1
  {
2
  "aka": {
3
- "[PAD]": 41,
4
- "[UNK]": 40,
5
- "a": 1,
6
- "b": 2,
7
- "c": 3,
8
- "d": 4,
9
- "e": 5,
10
- "f": 6,
11
- "g": 7,
12
- "h": 8,
13
- "i": 9,
14
- "j": 10,
15
- "k": 11,
16
- "l": 12,
17
- "m": 13,
18
- "n": 14,
19
- "o": 15,
20
- "p": 16,
21
- "r": 17,
22
- "s": 18,
23
- "t": 19,
24
- "u": 20,
25
- "v": 21,
26
- "w": 22,
27
- "y": 23,
28
- "z": 24,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
29
  "|": 0,
30
- "·": 25,
31
- "á": 26,
32
- "ã": 27,
33
- "õ": 28,
34
- "ĩ": 29,
35
- "ũ": 30,
36
- "ɔ": 31,
37
- "ɛ": 32,
38
- "ʹ": 33,
39
- "́": 34,
40
- "—": 35,
41
- "‘": 36,
42
- "’": 37,
43
- "“": 38,
44
- "•": 39
 
 
45
  }
46
  }
 
1
  {
2
  "aka": {
3
+ "!": 1,
4
+ "'": 2,
5
+ ",": 3,
6
+ ".": 4,
7
+ ":": 5,
8
+ "?": 6,
9
+ "A": 7,
10
+ "B": 8,
11
+ "C": 9,
12
+ "D": 10,
13
+ "E": 11,
14
+ "F": 12,
15
+ "G": 13,
16
+ "H": 14,
17
+ "I": 15,
18
+ "J": 16,
19
+ "K": 17,
20
+ "M": 18,
21
+ "N": 19,
22
+ "O": 20,
23
+ "P": 21,
24
+ "R": 22,
25
+ "S": 23,
26
+ "T": 24,
27
+ "W": 25,
28
+ "Y": 26,
29
+ "[PAD]": 68,
30
+ "[UNK]": 67,
31
+ "a": 27,
32
+ "b": 28,
33
+ "c": 29,
34
+ "d": 30,
35
+ "e": 31,
36
+ "f": 32,
37
+ "g": 33,
38
+ "h": 34,
39
+ "i": 35,
40
+ "k": 36,
41
+ "l": 37,
42
+ "m": 38,
43
+ "n": 39,
44
+ "o": 40,
45
+ "p": 41,
46
+ "r": 42,
47
+ "s": 43,
48
+ "t": 44,
49
+ "u": 45,
50
+ "v": 46,
51
+ "w": 47,
52
+ "y": 48,
53
+ "z": 49,
54
  "|": 0,
55
+ "·": 50,
56
+ "á": 51,
57
+ "ã": 52,
58
+ "õ": 53,
59
+ "ĩ": 54,
60
+ "ũ": 55,
61
+ "Ɔ": 56,
62
+ "Ɛ": 57,
63
+ "ɔ": 58,
64
+ "ɛ": 59,
65
+ "ʹ": 60,
66
+ "́": 61,
67
+ "—": 62,
68
+ "‘": 63,
69
+ "’": 64,
70
+ "“": 65,
71
+ "•": 66
72
  }
73
  }