Hugging Face's logo

bouim
/
hubert-large-arabic-darija

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
hubert
Generated from Trainer
Model card Files
xet
 Metrics Community
bouim commited on
Commit
3ea7411
·
1 Parent(s): 13cf31e

Upload tokenizer

Browse files
Files changed (2) hide show
  1. added_tokens.json +2 -2
  2. vocab.json +54 -41
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 42,
3
- "<s>": 41
4
  }
 
1
  {
2
+ "</s>": 55,
3
+ "<s>": 54
4
  }
vocab.json CHANGED
@@ -1,43 +1,56 @@
1
  {
2
- "3": 10,
3
- "[PAD]": 40,
4
- "[UNK]": 39,
5
- "x": 25,
6
- "y": 9,
7
- "|": 33,
8
- "؛": 16,
9
- "ء": 34,
10
- "أ": 35,
11
- "ؤ": 28,
12
- "ئ": 20,
13
- "ا": 30,
14
- "ب": 18,
15
- "ة": 12,
16
- "ت": 27,
17
- "ج": 21,
18
- "ح": 32,
19
- "خ": 6,
20
- "د": 7,
21
- "ذ": 22,
22
- "ر": 15,
23
- "ز": 23,
24
- "س": 11,
25
- "ش": 38,
26
- "ص": 14,
27
- "ض": 4,
28
- "ط": 8,
29
- "ظ": 31,
30
- "ع": 3,
31
- "غ": 37,
32
- "ف": 0,
33
- "ق": 5,
34
- "ك": 13,
35
- "ل": 29,
36
- "م": 19,
37
- "ن": 26,
38
- "ه": 24,
39
- "و": 2,
40
- "ى": 1,
41
- "ي": 36,
42
- "💜": 17
 
 
 
 
 
 
 
 
 
 
 
 
 
43
  }
 
1
  {
2
+ "(": 14,
3
+ ")": 24,
4
+ "0": 49,
5
+ "1": 47,
6
+ "2": 31,
7
+ "3": 37,
8
+ "5": 25,
9
+ "6": 18,
10
+ "8": 46,
11
+ "[PAD]": 53,
12
+ "[UNK]": 52,
13
+ "x": 45,
14
+ "y": 12,
15
+ "|": 32,
16
+ "°": 42,
17
+ "؛": 3,
18
+ "ء": 36,
19
+ "آ": 40,
20
+ "أ": 26,
21
+ "ؤ": 15,
22
+ "إ": 7,
23
+ "ئ": 48,
24
+ "ا": 20,
25
+ "ب": 28,
26
+ "ة": 11,
27
+ "ت": 19,
28
+ "ث": 51,
29
+ "ج": 13,
30
+ "ح": 22,
31
+ "خ": 50,
32
+ "د": 44,
33
+ "ذ": 38,
34
+ "ر": 35,
35
+ "ز": 29,
36
+ "س": 33,
37
+ "ش": 34,
38
+ "ص": 30,
39
+ "ض": 8,
40
+ "ط": 43,
41
+ "ظ": 5,
42
+ "ع": 2,
43
+ "غ": 16,
44
+ "ف": 41,
45
+ "ق": 17,
46
+ "ك": 21,
47
+ "ل": 39,
48
+ "م": 4,
49
+ "ن": 9,
50
+ "ه": 0,
51
+ "و": 1,
52
+ "ى": 23,
53
+ "ي": 27,
54
+ "💜": 10,
55
+ "😛": 6
56
  }