khaoulaoub commited on
Commit
1318b42
1 Parent(s): 62cd51f

Upload tokenizer

Browse files
Files changed (1) hide show
  1. vocab.json +30 -30
vocab.json CHANGED
@@ -1,36 +1,36 @@
1
  {
2
  "[PAD]": 33,
3
  "[UNK]": 32,
4
- "|": 25,
5
- "ء": 6,
6
- "ا": 15,
7
- "ب": 14,
8
- "ة": 1,
9
- "ت": 17,
10
- "ث": 3,
11
- "ج": 29,
12
- "ح": 21,
13
- "خ": 27,
14
- "د": 31,
15
- "ذ": 10,
16
- "ر": 2,
17
- "ز": 7,
18
- "س": 19,
19
- "ش": 4,
20
- "ص": 8,
21
- "ض": 26,
22
- "ط": 20,
23
- "ظ": 12,
24
- "ع": 9,
25
- "غ": 30,
26
- "ف": 23,
27
- "ق": 16,
28
- "ك": 13,
29
- "ل": 18,
30
  "م": 5,
31
- "ن": 22,
32
  "ه": 11,
33
- "و": 0,
34
- "ى": 28,
35
- "ي": 24
36
  }
 
1
  {
2
  "[PAD]": 33,
3
  "[UNK]": 32,
4
+ "|": 21,
5
+ "ء": 15,
6
+ "ا": 14,
7
+ "ب": 30,
8
+ "ة": 0,
9
+ "ت": 8,
10
+ "ث": 12,
11
+ "ج": 7,
12
+ "ح": 31,
13
+ "خ": 23,
14
+ "د": 3,
15
+ "ذ": 24,
16
+ "ر": 1,
17
+ "ز": 13,
18
+ "س": 28,
19
+ "ش": 2,
20
+ "ص": 25,
21
+ "ض": 22,
22
+ "ط": 16,
23
+ "ظ": 4,
24
+ "ع": 17,
25
+ "غ": 29,
26
+ "ف": 26,
27
+ "ق": 9,
28
+ "ك": 20,
29
+ "ل": 6,
30
  "م": 5,
31
+ "ن": 10,
32
  "ه": 11,
33
+ "و": 18,
34
+ "ى": 19,
35
+ "ي": 27
36
  }