Martha-987 commited on
Commit
9934b36
1 Parent(s): 18e8c28

Upload tokenizer

Browse files
Files changed (3) hide show
  1. special_tokens_map.json +6 -0
  2. tokenizer_config.json +10 -0
  3. vocab.json +65 -0
special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "pad_token": "[PAD]",
5
+ "unk_token": "[UNK]"
6
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "do_lower_case": false,
4
+ "eos_token": "</s>",
5
+ "pad_token": "[PAD]",
6
+ "replace_word_delimiter_char": " ",
7
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
8
+ "unk_token": "[UNK]",
9
+ "word_delimiter_token": "|"
10
+ }
vocab.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "!": 59,
3
+ "\"": 14,
4
+ "'": 9,
5
+ ",": 56,
6
+ "-": 5,
7
+ ".": 54,
8
+ ":": 47,
9
+ ";": 4,
10
+ "?": 32,
11
+ "A": 40,
12
+ "B": 3,
13
+ "C": 26,
14
+ "D": 46,
15
+ "E": 19,
16
+ "F": 20,
17
+ "G": 11,
18
+ "H": 39,
19
+ "I": 34,
20
+ "J": 18,
21
+ "K": 30,
22
+ "L": 31,
23
+ "M": 23,
24
+ "N": 38,
25
+ "O": 13,
26
+ "P": 10,
27
+ "Q": 52,
28
+ "R": 0,
29
+ "S": 7,
30
+ "T": 25,
31
+ "U": 44,
32
+ "V": 22,
33
+ "W": 37,
34
+ "X": 27,
35
+ "Y": 12,
36
+ "[PAD]": 62,
37
+ "[UNK]": 61,
38
+ "a": 41,
39
+ "b": 53,
40
+ "c": 6,
41
+ "d": 55,
42
+ "e": 45,
43
+ "f": 29,
44
+ "g": 17,
45
+ "h": 49,
46
+ "i": 15,
47
+ "j": 57,
48
+ "k": 33,
49
+ "l": 42,
50
+ "m": 36,
51
+ "n": 21,
52
+ "o": 58,
53
+ "p": 35,
54
+ "q": 2,
55
+ "r": 1,
56
+ "s": 28,
57
+ "t": 51,
58
+ "u": 48,
59
+ "v": 8,
60
+ "w": 16,
61
+ "x": 43,
62
+ "y": 24,
63
+ "z": 50,
64
+ "|": 60
65
+ }