Matthijs commited on
Commit
4355803
1 Parent(s): 9bf7685

Upload tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +4 -0
  2. tokenizer_config.json +9 -0
  3. vocab.json +37 -0
special_tokens_map.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "pad_token": "<pad>",
3
+ "unk_token": "<unk>"
4
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_blank": true,
3
+ "clean_up_tokenization_spaces": true,
4
+ "language": null,
5
+ "model_max_length": 1000000000000000019884624838656,
6
+ "pad_token": "<pad>",
7
+ "tokenizer_class": "VitsMmsTokenizer",
8
+ "unk_token": "<unk>"
9
+ }
vocab.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ " ": 30,
3
+ "'": 21,
4
+ "-": 27,
5
+ "6": 25,
6
+ "<pad>": 33,
7
+ "<unk>": 34,
8
+ "_": 0,
9
+ "a": 22,
10
+ "b": 3,
11
+ "c": 6,
12
+ "d": 18,
13
+ "e": 2,
14
+ "f": 8,
15
+ "g": 13,
16
+ "h": 12,
17
+ "i": 10,
18
+ "j": 5,
19
+ "k": 16,
20
+ "l": 14,
21
+ "m": 15,
22
+ "n": 23,
23
+ "o": 26,
24
+ "p": 28,
25
+ "q": 11,
26
+ "r": 7,
27
+ "s": 20,
28
+ "t": 1,
29
+ "u": 29,
30
+ "v": 31,
31
+ "w": 9,
32
+ "x": 24,
33
+ "y": 17,
34
+ "z": 32,
35
+ "ā": 19,
36
+ "ō": 4
37
+ }