Matthijs commited on
Commit
5ffdbaf
1 Parent(s): e353c1d

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 +29 -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 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ " ": 21,
3
+ "'": 15,
4
+ "-": 11,
5
+ "<pad>": 25,
6
+ "<unk>": 26,
7
+ "_": 1,
8
+ "a": 7,
9
+ "b": 13,
10
+ "c": 14,
11
+ "d": 18,
12
+ "e": 22,
13
+ "g": 19,
14
+ "h": 8,
15
+ "i": 9,
16
+ "j": 10,
17
+ "k": 12,
18
+ "l": 17,
19
+ "m": 23,
20
+ "n": 16,
21
+ "o": 5,
22
+ "p": 24,
23
+ "r": 20,
24
+ "s": 4,
25
+ "t": 2,
26
+ "u": 0,
27
+ "w": 3,
28
+ "y": 6
29
+ }