Hugging Face's logo

tuanio
/
wav2vec2-processor-vietnamese-phonemes

Model card Files Community
tuanio commited on
Commit
fd16f1b
•
1 Parent(s): 65f5882

Upload processor

Browse files
Files changed (4) hide show
  1. preprocessor_config.json +10 -0
  2. special_tokens_map.json +1 -0
  3. tokenizer_config.json +14 -0
  4. vocab.json +126 -0
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": false,
9
+ "sampling_rate": 16000
10
+ }
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {}
tokenizer_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": null,
3
+ "clean_up_tokenization_spaces": true,
4
+ "do_lower_case": false,
5
+ "eos_token": null,
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": null,
8
+ "processor_class": "Wav2Vec2Processor",
9
+ "replace_word_delimiter_char": " ",
10
+ "target_lang": null,
11
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
12
+ "unk_token": null,
13
+ "word_delimiter_token": "|"
14
+ }
vocab.json ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "<pad>": 0,
3
+ "a-0": 1,
4
+ "a-1": 2,
5
+ "a-2": 3,
6
+ "a-3": 4,
7
+ "a-4": 5,
8
+ "a-5": 6,
9
+ "aː-0": 7,
10
+ "aː-1": 8,
11
+ "aː-2": 9,
12
+ "aː-3": 10,
13
+ "aː-4": 11,
14
+ "aː-5": 12,
15
+ "e-0": 13,
16
+ "e-1": 14,
17
+ "e-2": 15,
18
+ "e-3": 16,
19
+ "e-4": 17,
20
+ "e-5": 18,
21
+ "eaː-0": 19,
22
+ "eaː-1": 20,
23
+ "eaː-2": 21,
24
+ "eaː-3": 22,
25
+ "eaː-4": 23,
26
+ "eaː-5": 24,
27
+ "f": 25,
28
+ "h": 26,
29
+ "i-0": 27,
30
+ "i-1": 28,
31
+ "i-2": 29,
32
+ "i-3": 30,
33
+ "i-4": 31,
34
+ "i-5": 32,
35
+ "iz": 33,
36
+ "iə-0": 34,
37
+ "iə-1": 35,
38
+ "iə-2": 36,
39
+ "iə-3": 37,
40
+ "iə-4": 38,
41
+ "iə-5": 39,
42
+ "k": 40,
43
+ "kpz": 41,
44
+ "kz": 42,
45
+ "k̟z": 43,
46
+ "l": 44,
47
+ "m": 45,
48
+ "mz": 46,
49
+ "n": 47,
50
+ "nz": 48,
51
+ "o-0": 49,
52
+ "o-1": 50,
53
+ "o-2": 51,
54
+ "o-3": 52,
55
+ "o-4": 53,
56
+ "o-5": 54,
57
+ "p": 55,
58
+ "pz": 56,
59
+ "s": 57,
60
+ "t": 58,
61
+ "tz": 59,
62
+ "tʰ": 60,
63
+ "t͡ɕ": 61,
64
+ "u-0": 62,
65
+ "u-1": 63,
66
+ "u-2": 64,
67
+ "u-3": 65,
68
+ "u-4": 66,
69
+ "u-5": 67,
70
+ "uz": 68,
71
+ "uə-0": 69,
72
+ "uə-1": 70,
73
+ "uə-2": 71,
74
+ "uə-3": 72,
75
+ "uə-4": 73,
76
+ "uə-5": 74,
77
+ "v": 75,
78
+ "w": 76,
79
+ "x": 77,
80
+ "z": 78,
81
+ "|": 124,
82
+ "ŋ": 79,
83
+ "ŋmz": 80,
84
+ "ŋz": 81,
85
+ "ŋ̟z": 82,
86
+ "ɓ": 83,
87
+ "ɔ-0": 84,
88
+ "ɔ-1": 85,
89
+ "ɔ-2": 86,
90
+ "ɔ-3": 87,
91
+ "ɔ-4": 88,
92
+ "ɔ-5": 89,
93
+ "ɗ": 90,
94
+ "ə-0": 91,
95
+ "ə-1": 92,
96
+ "ə-2": 93,
97
+ "ə-3": 94,
98
+ "ə-4": 95,
99
+ "ə-5": 96,
100
+ "əː-0": 97,
101
+ "əː-1": 98,
102
+ "əː-2": 99,
103
+ "əː-3": 100,
104
+ "əː-4": 101,
105
+ "əː-5": 102,
106
+ "ɛ-0": 103,
107
+ "ɛ-1": 104,
108
+ "ɛ-2": 105,
109
+ "ɛ-3": 106,
110
+ "ɛ-4": 107,
111
+ "ɛ-5": 108,
112
+ "ɣ": 109,
113
+ "ɨ-0": 110,
114
+ "ɨ-1": 111,
115
+ "ɨ-2": 112,
116
+ "ɨ-3": 113,
117
+ "ɨ-4": 114,
118
+ "ɨ-5": 115,
119
+ "ɨə-0": 116,
120
+ "ɨə-1": 117,
121
+ "ɨə-2": 118,
122
+ "ɨə-3": 119,
123
+ "ɨə-4": 120,
124
+ "ɨə-5": 121,
125
+ "ɲ": 122
126
+ }