paruwka commited on
Commit
75d04f9
1 Parent(s): a69266b

Upload tokenizer

Browse files
Files changed (2) hide show
  1. README.md +1 -1
  2. vocab.json +66 -66
README.md CHANGED
@@ -1,8 +1,8 @@
1
  ---
2
  license: apache-2.0
3
- base_model: facebook/wav2vec2-large-xlsr-53
4
  tags:
5
  - generated_from_trainer
 
6
  metrics:
7
  - wer
8
  model-index:
 
1
  ---
2
  license: apache-2.0
 
3
  tags:
4
  - generated_from_trainer
5
+ base_model: facebook/wav2vec2-large-xlsr-53
6
  metrics:
7
  - wer
8
  model-index:
vocab.json CHANGED
@@ -1,70 +1,70 @@
1
  {
2
- "&": 28,
3
- "'": 3,
4
- "(": 13,
5
- ")": 49,
6
- "=": 54,
7
  "[PAD]": 67,
8
  "[UNK]": 66,
9
- "a": 38,
10
- "b": 8,
11
- "c": 43,
12
- "d": 21,
13
- "e": 62,
14
- "f": 57,
15
- "g": 52,
16
- "h": 44,
17
- "i": 39,
18
- "j": 60,
19
- "k": 48,
20
- "l": 19,
21
- "m": 26,
22
- "n": 15,
23
- "o": 11,
24
- "p": 17,
25
- "q": 23,
26
- "r": 14,
27
- "s": 63,
28
- "t": 2,
29
- "u": 41,
30
- "v": 30,
31
- "w": 51,
32
- "x": 20,
33
- "y": 36,
34
- "z": 18,
35
- "|": 42,
36
- "«": 46,
37
- "»": 29,
38
- "ß": 53,
39
- "à": 59,
40
- "á": 32,
41
- "â": 55,
42
- "ç": 24,
43
- "è": 27,
44
- "é": 16,
45
- "ê": 6,
46
- "ë": 5,
47
- "î": 7,
48
- "ï": 45,
49
- "ó": 56,
50
- "ô": 33,
51
- "ö": 61,
52
- "ø": 40,
53
- "ù": 58,
54
- "ú": 9,
55
- "û": 47,
56
- "ü": 34,
57
- "ć": 4,
58
- "ē": 0,
59
- "ō": 22,
60
- "œ": 10,
61
- "š": 31,
62
- "ū": 37,
63
- "̀": 65,
64
- "́": 35,
65
- "̂": 12,
66
- "̧": 64,
67
- "—": 25,
68
- "’": 50,
69
- "…": 1
70
  }
 
1
  {
2
+ "&": 11,
3
+ "'": 57,
4
+ "(": 51,
5
+ ")": 44,
6
+ "=": 3,
7
  "[PAD]": 67,
8
  "[UNK]": 66,
9
+ "a": 58,
10
+ "b": 50,
11
+ "c": 4,
12
+ "d": 40,
13
+ "e": 65,
14
+ "f": 53,
15
+ "g": 5,
16
+ "h": 46,
17
+ "i": 42,
18
+ "j": 12,
19
+ "k": 47,
20
+ "l": 14,
21
+ "m": 23,
22
+ "n": 1,
23
+ "o": 28,
24
+ "p": 49,
25
+ "q": 2,
26
+ "r": 52,
27
+ "s": 20,
28
+ "t": 63,
29
+ "u": 16,
30
+ "v": 35,
31
+ "w": 60,
32
+ "x": 34,
33
+ "y": 48,
34
+ "z": 59,
35
+ "|": 64,
36
+ "«": 39,
37
+ "»": 55,
38
+ "ß": 62,
39
+ "à": 10,
40
+ "á": 7,
41
+ "â": 15,
42
+ "ç": 56,
43
+ "è": 6,
44
+ "é": 33,
45
+ "ê": 45,
46
+ "ë": 29,
47
+ "î": 21,
48
+ "ï": 22,
49
+ "ó": 24,
50
+ "ô": 38,
51
+ "ö": 36,
52
+ "ø": 26,
53
+ "ù": 27,
54
+ "ú": 31,
55
+ "û": 0,
56
+ "ü": 9,
57
+ "ć": 8,
58
+ "ē": 37,
59
+ "ō": 32,
60
+ "œ": 61,
61
+ "š": 13,
62
+ "ū": 41,
63
+ "̀": 19,
64
+ "́": 30,
65
+ "̂": 25,
66
+ "̧": 54,
67
+ "—": 43,
68
+ "’": 18,
69
+ "…": 17
70
  }