ad019el commited on
Commit
322b957
1 Parent(s): c332e4d

add tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +6 -0
  2. tokenizer_config.json +10 -0
  3. vocab.json +68 -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 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "do_lower_case": false,
4
+ "eos_token": "</s>",
5
+ "pad_token": "[PAD]",
6
+ "replace_word_delimiter_char": " ",
7
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
8
+ "unk_token": "[UNK]",
9
+ "word_delimiter_token": "|"
10
+ }
vocab.json ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "&": 5,
3
+ "/": 29,
4
+ "0": 34,
5
+ "1": 38,
6
+ "2": 0,
7
+ "7": 43,
8
+ "9": 56,
9
+ "[PAD]": 65,
10
+ "[UNK]": 64,
11
+ "_": 13,
12
+ "a": 35,
13
+ "c": 58,
14
+ "e": 31,
15
+ "f": 26,
16
+ "j": 49,
17
+ "l": 40,
18
+ "m": 27,
19
+ "o": 41,
20
+ "q": 46,
21
+ "r": 12,
22
+ "t": 50,
23
+ "u": 15,
24
+ "w": 42,
25
+ "|": 14,
26
+ "ء": 23,
27
+ "آ": 28,
28
+ "أ": 4,
29
+ "ؤ": 52,
30
+ "إ": 10,
31
+ "ئ": 21,
32
+ "ا": 19,
33
+ "ب": 47,
34
+ "ة": 3,
35
+ "ت": 62,
36
+ "ث": 48,
37
+ "ج": 37,
38
+ "ح": 9,
39
+ "خ": 45,
40
+ "د": 32,
41
+ "ذ": 18,
42
+ "ر": 11,
43
+ "ز": 25,
44
+ "س": 1,
45
+ "ش": 39,
46
+ "ص": 2,
47
+ "ض": 7,
48
+ "ط": 54,
49
+ "ظ": 55,
50
+ "ع": 16,
51
+ "غ": 51,
52
+ "ف": 53,
53
+ "ق": 60,
54
+ "ك": 63,
55
+ "ل": 6,
56
+ "م": 36,
57
+ "ن": 33,
58
+ "ه": 30,
59
+ "و": 61,
60
+ "ى": 24,
61
+ "ي": 20,
62
+ "ٱ": 17,
63
+ "چ": 57,
64
+ "ڤ": 44,
65
+ "ک": 8,
66
+ "ی": 22,
67
+ "’": 59
68
+ }