FarmerlineML commited on
Commit
d5b0b05
verified
1 Parent(s): c20caea

update vocab.json for twi language

Browse files

refined vocab.json for improved conciseness, removing numerals as MMS models struggle with accurate transcription.

Files changed (1) hide show
  1. vocab.json +28 -31
vocab.json CHANGED
@@ -1,32 +1,29 @@
1
  {
2
- " ": 8,
3
- "'": 5,
4
- "-": 3,
5
- "2": 23,
6
- "3": 7,
7
- "_": 6,
8
- "a": 0,
9
- "b": 29,
10
- "d": 28,
11
- "e": 24,
12
- "f": 15,
13
- "g": 17,
14
- "h": 20,
15
- "i": 26,
16
- "k": 19,
17
- "l": 21,
18
- "m": 10,
19
- "n": 4,
20
- "o": 16,
21
- "p": 9,
22
- "r": 25,
23
- "s": 22,
24
- "t": 2,
25
- "u": 18,
26
- "w": 12,
27
- "y": 13,
28
- "": 11,
29
- "蓴": 27,
30
- "蓻": 14,
31
- "始": 1
32
- }
 
1
  {
2
+ "'": 1,
3
+ "[PAD]": 26,
4
+ "[UNK]": 25,
5
+ "a": 2,
6
+ "b": 3,
7
+ "d": 4,
8
+ "e": 5,
9
+ "f": 6,
10
+ "g": 7,
11
+ "h": 8,
12
+ "i": 9,
13
+ "k": 10,
14
+ "l": 11,
15
+ "m": 12,
16
+ "n": 13,
17
+ "o": 14,
18
+ "p": 15,
19
+ "r": 16,
20
+ "s": 17,
21
+ "t": 18,
22
+ "u": 19,
23
+ "w": 20,
24
+ "y": 21,
25
+ "|": 0,
26
+ "": 22,
27
+ "": 23,
28
+ "": 24
29
+ }