Matthijs commited on
Commit
486380b
1 Parent(s): 05408b2

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 +49 -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 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ " ": 0,
3
+ "-": 5,
4
+ "0": 44,
5
+ "1": 7,
6
+ "2": 2,
7
+ "3": 13,
8
+ "4": 11,
9
+ "5": 31,
10
+ "6": 30,
11
+ "7": 42,
12
+ "8": 8,
13
+ "<pad>": 45,
14
+ "<unk>": 46,
15
+ "_": 27,
16
+ "a": 9,
17
+ "b": 22,
18
+ "c": 33,
19
+ "d": 21,
20
+ "e": 38,
21
+ "f": 6,
22
+ "g": 4,
23
+ "h": 10,
24
+ "i": 34,
25
+ "j": 28,
26
+ "k": 37,
27
+ "l": 17,
28
+ "m": 15,
29
+ "n": 18,
30
+ "o": 25,
31
+ "p": 36,
32
+ "q": 3,
33
+ "r": 14,
34
+ "s": 29,
35
+ "t": 19,
36
+ "u": 26,
37
+ "v": 1,
38
+ "w": 40,
39
+ "x": 43,
40
+ "y": 23,
41
+ "z": 41,
42
+ "ß": 24,
43
+ "ä": 16,
44
+ "ë": 20,
45
+ "ï": 32,
46
+ "ö": 12,
47
+ "ü": 35,
48
+ "–": 39
49
+ }