Pongsathorn commited on
Commit
f485141
·
1 Parent(s): 82bb052

Upload 4 files

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +6 -0
  2. tokenizer_config.json +13 -0
  3. vocab.json +96 -0
special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "pad_token": "[PAD]",
5
+ "unk_token": "[UNK]"
6
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "do_lower_case": false,
5
+ "eos_token": "</s>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "[PAD]",
8
+ "processor_class": "Wav2Vec2Processor",
9
+ "replace_word_delimiter_char": " ",
10
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
11
+ "unk_token": "[UNK]",
12
+ "word_delimiter_token": "|"
13
+ }
vocab.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "[PAD]": 93,
3
+ "[UNK]": 92,
4
+ "a": 68,
5
+ "b": 84,
6
+ "c": 5,
7
+ "e": 28,
8
+ "f": 78,
9
+ "g": 82,
10
+ "h": 49,
11
+ "i": 76,
12
+ "j": 10,
13
+ "k": 17,
14
+ "l": 77,
15
+ "m": 44,
16
+ "n": 26,
17
+ "o": 13,
18
+ "p": 83,
19
+ "r": 52,
20
+ "s": 8,
21
+ "t": 72,
22
+ "u": 59,
23
+ "w": 29,
24
+ "x": 90,
25
+ "y": 47,
26
+ "|": 31,
27
+ "ก": 58,
28
+ "ข": 0,
29
+ "ฃ": 1,
30
+ "ค": 40,
31
+ "ฆ": 87,
32
+ "ง": 86,
33
+ "จ": 45,
34
+ "ฉ": 60,
35
+ "ช": 69,
36
+ "ซ": 79,
37
+ "ฌ": 53,
38
+ "ญ": 6,
39
+ "ฎ": 61,
40
+ "ฏ": 3,
41
+ "ฐ": 65,
42
+ "ฑ": 11,
43
+ "ฒ": 39,
44
+ "ณ": 91,
45
+ "ด": 75,
46
+ "ต": 85,
47
+ "ถ": 81,
48
+ "ท": 62,
49
+ "ธ": 25,
50
+ "น": 63,
51
+ "บ": 20,
52
+ "ป": 18,
53
+ "ผ": 64,
54
+ "ฝ": 42,
55
+ "พ": 50,
56
+ "ฟ": 36,
57
+ "ภ": 24,
58
+ "ม": 73,
59
+ "ย": 19,
60
+ "ร": 70,
61
+ "ฤ": 23,
62
+ "ล": 35,
63
+ "ว": 9,
64
+ "ศ": 32,
65
+ "ษ": 46,
66
+ "ส": 55,
67
+ "ห": 37,
68
+ "ฬ": 89,
69
+ "อ": 51,
70
+ "ฮ": 30,
71
+ "ฯ": 38,
72
+ "ะ": 43,
73
+ "ั": 7,
74
+ "า": 88,
75
+ "ำ": 34,
76
+ "ิ": 48,
77
+ "ี": 33,
78
+ "ึ": 74,
79
+ "ื": 71,
80
+ "ุ": 16,
81
+ "ู": 22,
82
+ "เ": 15,
83
+ "แ": 27,
84
+ "โ": 67,
85
+ "ใ": 41,
86
+ "ไ": 56,
87
+ "ๅ": 12,
88
+ "ๆ": 66,
89
+ "็": 54,
90
+ "่": 57,
91
+ "้": 80,
92
+ "๊": 14,
93
+ "๋": 2,
94
+ "์": 4,
95
+ "ํ": 21
96
+ }