roviso commited on
Commit
7324298
1 Parent(s): 71673b8

Upload tokenizer

Browse files
Files changed (4) hide show
  1. added_tokens.json +4 -0
  2. special_tokens_map.json +22 -0
  3. tokenizer_config.json +13 -0
  4. vocab.json +181 -0
added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 177,
3
+ "<s>": 176
4
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": true,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ {
11
+ "content": "</s>",
12
+ "lstrip": false,
13
+ "normalized": true,
14
+ "rstrip": false,
15
+ "single_word": false
16
+ }
17
+ ],
18
+ "bos_token": "<s>",
19
+ "eos_token": "</s>",
20
+ "pad_token": "[PAD]",
21
+ "unk_token": "[UNK]"
22
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "do_lower_case": false,
5
+ "eos_token": "</s>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "[PAD]",
8
+ "replace_word_delimiter_char": " ",
9
+ "target_lang": "npi",
10
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
11
+ "unk_token": "[UNK]",
12
+ "word_delimiter_token": "|"
13
+ }
vocab.json ADDED
@@ -0,0 +1,181 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "npi": {
3
+ "\n": 0,
4
+ "!": 2,
5
+ "\"": 3,
6
+ "#": 4,
7
+ "$": 5,
8
+ "%": 6,
9
+ "&": 7,
10
+ "'": 8,
11
+ "(": 9,
12
+ ")": 10,
13
+ "+": 11,
14
+ ",": 12,
15
+ "-": 13,
16
+ ".": 14,
17
+ "/": 15,
18
+ "0": 16,
19
+ "1": 17,
20
+ "2": 18,
21
+ "3": 19,
22
+ "4": 20,
23
+ "5": 21,
24
+ "6": 22,
25
+ "7": 23,
26
+ "8": 24,
27
+ "9": 25,
28
+ ":": 26,
29
+ ";": 27,
30
+ "=": 28,
31
+ "?": 29,
32
+ "A": 30,
33
+ "B": 31,
34
+ "C": 32,
35
+ "D": 33,
36
+ "E": 34,
37
+ "F": 35,
38
+ "G": 36,
39
+ "H": 37,
40
+ "I": 38,
41
+ "J": 39,
42
+ "K": 40,
43
+ "L": 41,
44
+ "M": 42,
45
+ "N": 43,
46
+ "O": 44,
47
+ "P": 45,
48
+ "Q": 46,
49
+ "R": 47,
50
+ "S": 48,
51
+ "T": 49,
52
+ "U": 50,
53
+ "V": 51,
54
+ "W": 52,
55
+ "Y": 53,
56
+ "Z": 54,
57
+ "[": 55,
58
+ "[PAD]": 176,
59
+ "[UNK]": 175,
60
+ "]": 56,
61
+ "_": 57,
62
+ "`": 58,
63
+ "a": 59,
64
+ "b": 60,
65
+ "c": 61,
66
+ "d": 62,
67
+ "e": 63,
68
+ "f": 64,
69
+ "g": 65,
70
+ "h": 66,
71
+ "i": 67,
72
+ "j": 68,
73
+ "k": 69,
74
+ "l": 70,
75
+ "m": 71,
76
+ "n": 72,
77
+ "o": 73,
78
+ "p": 74,
79
+ "q": 75,
80
+ "r": 76,
81
+ "s": 77,
82
+ "t": 78,
83
+ "u": 79,
84
+ "v": 80,
85
+ "w": 81,
86
+ "x": 82,
87
+ "y": 83,
88
+ "z": 84,
89
+ "|": 1,
90
+ "¥": 86,
91
+ "«": 87,
92
+ "®": 88,
93
+ "·": 89,
94
+ "Ï": 90,
95
+ "è": 91,
96
+ "é": 92,
97
+ "÷": 93,
98
+ "ँ": 94,
99
+ "ं": 95,
100
+ "ः": 96,
101
+ "अ": 97,
102
+ "आ": 98,
103
+ "इ": 99,
104
+ "ई": 100,
105
+ "उ": 101,
106
+ "ऊ": 102,
107
+ "ऋ": 103,
108
+ "ए": 104,
109
+ "ऐ": 105,
110
+ "ओ": 106,
111
+ "औ": 107,
112
+ "क": 108,
113
+ "ख": 109,
114
+ "ग": 110,
115
+ "घ": 111,
116
+ "ङ": 112,
117
+ "च": 113,
118
+ "छ": 114,
119
+ "ज": 115,
120
+ "झ": 116,
121
+ "ञ": 117,
122
+ "ट": 118,
123
+ "ठ": 119,
124
+ "ड": 120,
125
+ "ढ": 121,
126
+ "ण": 122,
127
+ "त": 123,
128
+ "थ": 124,
129
+ "द": 125,
130
+ "ध": 126,
131
+ "न": 127,
132
+ "प": 128,
133
+ "फ": 129,
134
+ "ब": 130,
135
+ "भ": 131,
136
+ "म": 132,
137
+ "य": 133,
138
+ "र": 134,
139
+ "ल": 135,
140
+ "व": 136,
141
+ "श": 137,
142
+ "ष": 138,
143
+ "स": 139,
144
+ "ह": 140,
145
+ "़": 141,
146
+ "ा": 142,
147
+ "ि": 143,
148
+ "ी": 144,
149
+ "ु": 145,
150
+ "ू": 146,
151
+ "ृ": 147,
152
+ "े": 148,
153
+ "ै": 149,
154
+ "ो": 150,
155
+ "ौ": 151,
156
+ "्": 152,
157
+ "ॐ": 153,
158
+ "।": 154,
159
+ "०": 155,
160
+ "१": 156,
161
+ "२": 157,
162
+ "३": 158,
163
+ "४": 159,
164
+ "५": 160,
165
+ "६": 161,
166
+ "७": 162,
167
+ "८": 163,
168
+ "९": 164,
169
+ "​": 165,
170
+ "‌": 166,
171
+ "‍": 167,
172
+ "–": 168,
173
+ "—": 169,
174
+ "‘": 170,
175
+ "’": 171,
176
+ "“": 172,
177
+ "”": 173,
178
+ "…": 174,
179
+ " ": 175
180
+ }
181
+ }