tuanio commited on
Commit
14f5a9c
1 Parent(s): 2aff10d

Upload folder using huggingface_hub (#1)

Browse files

- a7ccadfe5b823ddddc443afba9e4e3c27791e0b4b01fd33335286851e5d3b612 (e53c97eee644d4be5c7c1404a7c209664b6085c2)

Files changed (3) hide show
  1. special_tokens_map.json +6 -0
  2. tokenizer_config.json +14 -0
  3. vocab.json +50 -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 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "do_phonemize": false,
5
+ "eos_token": "</s>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "<pad>",
8
+ "phone_delimiter_token": " ",
9
+ "phonemizer_backend": "espeak",
10
+ "phonemizer_lang": "en-us",
11
+ "tokenizer_class": "Wav2Vec2PhonemeCTCTokenizer",
12
+ "unk_token": "<unk>",
13
+ "word_delimiter_token": null
14
+ }
vocab.json ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</s>": 1,
3
+ "<pad>": 3,
4
+ "<s>": 0,
5
+ "<unk>": 2,
6
+ "aɪ": 26,
7
+ "aʊ": 45,
8
+ "b": 27,
9
+ "d": 18,
10
+ "eɪ": 17,
11
+ "f": 21,
12
+ "h": 40,
13
+ "i": 12,
14
+ "iə": 44,
15
+ "j": 29,
16
+ "k": 25,
17
+ "l": 33,
18
+ "m": 16,
19
+ "n": 11,
20
+ "o": 42,
21
+ "oʊ": 22,
22
+ "p": 7,
23
+ "s": 13,
24
+ "t": 4,
25
+ "uː": 24,
26
+ "v": 14,
27
+ "w": 9,
28
+ "z": 36,
29
+ "æ": 6,
30
+ "ð": 23,
31
+ "ŋ": 34,
32
+ "ɑ": 31,
33
+ "ɑː": 41,
34
+ "ɔ": 39,
35
+ "ɔɪ": 47,
36
+ "ɔː": 38,
37
+ "ə": 15,
38
+ "ɚ": 8,
39
+ "ɛ": 10,
40
+ "ɡ": 28,
41
+ "ɪ": 20,
42
+ "ɹ": 32,
43
+ "ɾ": 43,
44
+ "ʃ": 5,
45
+ "ʊ": 37,
46
+ "ʊɹ": 30,
47
+ "ʌ": 35,
48
+ "ʒ": 19,
49
+ "θ": 46
50
+ }