YassineBenlaria commited on
Commit
cf07fec
1 Parent(s): 28772cd

Upload tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +6 -0
  2. tokenizer_config.json +13 -0
  3. vocab.json +46 -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
+ "replace_word_delimiter_char": " ",
9
+ "target_lang": null,
10
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
11
+ "unk_token": "[UNK]",
12
+ "word_delimiter_token": "|"
13
+ }
vocab.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "[PAD]": 43,
3
+ "[UNK]": 42,
4
+ "|": 13,
5
+ "ء": 35,
6
+ "آ": 3,
7
+ "أ": 17,
8
+ "ؤ": 14,
9
+ "إ": 31,
10
+ "ئ": 41,
11
+ "ا": 10,
12
+ "ب": 40,
13
+ "ة": 16,
14
+ "ت": 21,
15
+ "ث": 30,
16
+ "ج": 12,
17
+ "ح": 25,
18
+ "خ": 33,
19
+ "د": 29,
20
+ "ذ": 19,
21
+ "ر": 39,
22
+ "ز": 1,
23
+ "س": 32,
24
+ "ش": 28,
25
+ "ص": 38,
26
+ "ض": 7,
27
+ "ط": 8,
28
+ "ظ": 34,
29
+ "ع": 24,
30
+ "غ": 11,
31
+ "ف": 22,
32
+ "ق": 23,
33
+ "ك": 0,
34
+ "ل": 2,
35
+ "م": 36,
36
+ "ن": 5,
37
+ "ه": 6,
38
+ "و": 9,
39
+ "ى": 20,
40
+ "ي": 26,
41
+ "ٱ": 4,
42
+ "چ": 27,
43
+ "ڤ": 18,
44
+ "ک": 15,
45
+ "ی": 37
46
+ }