sanchit-gandhi commited on
Commit
ad132b0
·
1 Parent(s): 14744c5

Upload tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +4 -0
  2. tokenizer_config.json +12 -0
  3. vocab.json +36 -0
special_tokens_map.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "pad_token": "d",
3
+ "unk_token": "<unk>"
4
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_blank": true,
3
+ "clean_up_tokenization_spaces": true,
4
+ "is_uroman": false,
5
+ "language": "hau",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "normalize": true,
8
+ "pad_token": "d",
9
+ "phonemize": false,
10
+ "tokenizer_class": "VitsTokenizer",
11
+ "unk_token": "<unk>"
12
+ }
vocab.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ " ": 33,
3
+ "'": 26,
4
+ "-": 20,
5
+ "6": 21,
6
+ "_": 1,
7
+ "a": 9,
8
+ "b": 17,
9
+ "c": 14,
10
+ "d": 0,
11
+ "e": 22,
12
+ "f": 24,
13
+ "g": 4,
14
+ "h": 28,
15
+ "i": 12,
16
+ "j": 16,
17
+ "k": 23,
18
+ "l": 19,
19
+ "m": 30,
20
+ "n": 15,
21
+ "o": 6,
22
+ "r": 8,
23
+ "s": 3,
24
+ "t": 5,
25
+ "u": 10,
26
+ "w": 27,
27
+ "y": 7,
28
+ "z": 29,
29
+ "ā": 11,
30
+ "ă": 25,
31
+ "ū": 13,
32
+ "ƙ": 31,
33
+ "ɓ": 18,
34
+ "ɗ": 2,
35
+ "ˈ": 32
36
+ }