Hugging Face's logo

Thunpitcha
/
wav2vec2-base-thai-5-google-colab

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Thunpitcha commited on
Commit
c84e3e4
•
1 Parent(s): da9264e

Upload tokenizer

Browse files
Files changed (1) hide show
  1. vocab.json +85 -85
vocab.json CHANGED
@@ -1,90 +1,90 @@
1
  {
2
- "'": 57,
3
  "[PAD]": 87,
4
  "[UNK]": 86,
5
- "_": 49,
6
- "a": 76,
7
- "c": 34,
8
- "e": 60,
9
- "h": 31,
10
- "i": 10,
11
- "j": 22,
12
- "k": 32,
13
- "l": 48,
14
- "m": 85,
15
- "n": 59,
16
- "o": 16,
17
- "r": 55,
18
- "s": 79,
19
- "t": 27,
20
- "y": 13,
21
- "|": 3,
22
- "~": 84,
23
- "ก": 6,
24
- "ข": 5,
25
- "ค": 41,
26
- "ฆ": 66,
27
- "ง": 78,
28
- "จ": 33,
29
- "ฉ": 26,
30
- "ช": 64,
31
- "ซ": 46,
32
- "ฌ": 29,
33
- "ญ": 56,
34
- "ฎ": 39,
35
- "ฏ": 53,
36
- "ฐ": 35,
37
- "ฑ": 69,
38
- "ฒ": 74,
39
- "ณ": 37,
40
- "ด": 43,
41
- "ต": 4,
42
- "ถ": 25,
43
- "ท": 21,
44
- "ธ": 71,
45
- "น": 61,
46
- "บ": 14,
47
- "ป": 72,
48
- "ผ": 68,
49
- "ฝ": 9,
50
- "พ": 62,
51
- "ฟ": 80,
52
- "ภ": 28,
53
- "ม": 12,
54
- "ย": 75,
55
- "ร": 54,
56
- "ฤ": 52,
57
- "ล": 0,
58
- "ว": 58,
59
- "ศ": 7,
60
- "ษ": 47,
61
- "ส": 17,
62
  "ห": 20,
63
- "ฬ": 19,
64
- "อ": 50,
65
- "ฮ": 51,
66
- "ะ": 42,
67
- "ั": 38,
68
- "า": 11,
69
- "ำ": 65,
70
- "ิ": 18,
71
- "ี": 40,
72
- "ึ": 67,
73
- "ื": 2,
74
- "ุ": 15,
75
- "ู": 73,
76
- "เ": 45,
77
- "แ": 8,
78
- "โ": 77,
79
- "ใ": 83,
80
- "ไ": 70,
81
- "ๅ": 63,
82
- "ๆ": 30,
83
- "็": 82,
84
- "่": 36,
85
- "้": 23,
86
- "๊": 24,
87
- "๋": 44,
88
- "์": 81,
89
- "’": 1
90
  }
 
1
  {
2
+ "'": 41,
3
  "[PAD]": 87,
4
  "[UNK]": 86,
5
+ "_": 30,
6
+ "a": 4,
7
+ "c": 39,
8
+ "e": 5,
9
+ "h": 80,
10
+ "i": 40,
11
+ "j": 61,
12
+ "k": 81,
13
+ "l": 63,
14
+ "m": 16,
15
+ "n": 70,
16
+ "o": 44,
17
+ "r": 31,
18
+ "s": 11,
19
+ "t": 45,
20
+ "y": 21,
21
+ "|": 29,
22
+ "~": 23,
23
+ "ก": 3,
24
+ "ข": 58,
25
+ "ค": 36,
26
+ "ฆ": 12,
27
+ "ง": 37,
28
+ "จ": 38,
29
+ "ฉ": 75,
30
+ "ช": 35,
31
+ "ซ": 8,
32
+ "ฌ": 74,
33
+ "ญ": 73,
34
+ "ฎ": 13,
35
+ "ฏ": 48,
36
+ "ฐ": 85,
37
+ "ฑ": 60,
38
+ "ฒ": 82,
39
+ "ณ": 34,
40
+ "ด": 57,
41
+ "ต": 28,
42
+ "ถ": 72,
43
+ "ท": 51,
44
+ "ธ": 76,
45
+ "น": 6,
46
+ "บ": 42,
47
+ "ป": 78,
48
+ "ผ": 46,
49
+ "ฝ": 69,
50
+ "พ": 1,
51
+ "ฟ": 67,
52
+ "ภ": 68,
53
+ "ม": 77,
54
+ "ย": 26,
55
+ "ร": 79,
56
+ "ฤ": 71,
57
+ "ล": 65,
58
+ "ว": 54,
59
+ "ศ": 2,
60
+ "ษ": 64,
61
+ "ส": 56,
62
  "ห": 20,
63
+ "ฬ": 55,
64
+ "อ": 15,
65
+ "ฮ": 49,
66
+ "ะ": 83,
67
+ "ั": 47,
68
+ "า": 10,
69
+ "ำ": 19,
70
+ "ิ": 27,
71
+ "ี": 0,
72
+ "ึ": 33,
73
+ "ื": 53,
74
+ "ุ": 43,
75
+ "ู": 52,
76
+ "เ": 17,
77
+ "แ": 32,
78
+ "โ": 9,
79
+ "ใ": 59,
80
+ "ไ": 18,
81
+ "ๅ": 66,
82
+ "ๆ": 14,
83
+ "็": 62,
84
+ "่": 50,
85
+ "้": 22,
86
+ "๊": 7,
87
+ "๋": 25,
88
+ "์": 84,
89
+ "’": 24
90
  }