Matthijs commited on
Commit
c1510a3
1 Parent(s): 7737068

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 +57 -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 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ " ": 3,
3
+ "'": 20,
4
+ "-": 45,
5
+ "<pad>": 53,
6
+ "<unk>": 54,
7
+ "_": 47,
8
+ "a": 52,
9
+ "b": 13,
10
+ "c": 39,
11
+ "d": 36,
12
+ "e": 30,
13
+ "f": 22,
14
+ "g": 49,
15
+ "h": 5,
16
+ "i": 7,
17
+ "j": 43,
18
+ "k": 9,
19
+ "l": 44,
20
+ "m": 27,
21
+ "n": 33,
22
+ "o": 35,
23
+ "p": 14,
24
+ "r": 48,
25
+ "s": 25,
26
+ "t": 46,
27
+ "u": 17,
28
+ "w": 19,
29
+ "y": 24,
30
+ "á": 10,
31
+ "â": 37,
32
+ "é": 41,
33
+ "ê": 1,
34
+ "í": 15,
35
+ "î": 26,
36
+ "ó": 23,
37
+ "ô": 38,
38
+ "ú": 34,
39
+ "û": 51,
40
+ "ě": 6,
41
+ "ń": 18,
42
+ "ǎ": 40,
43
+ "ǐ": 12,
44
+ "ǒ": 2,
45
+ "ǔ": 16,
46
+ "ɔ": 0,
47
+ "ɛ": 28,
48
+ "ɩ": 8,
49
+ "ʋ": 31,
50
+ "́": 42,
51
+ "̂": 4,
52
+ "̌": 11,
53
+ "έ": 29,
54
+ "ί": 21,
55
+ "ε": 50,
56
+ "ḿ": 32
57
+ }