sanchit-gandhi
commited on
Commit
·
2ecf5f1
1
Parent(s):
1970a92
Upload tokenizer
Browse files- special_tokens_map.json +4 -0
- tokenizer_config.json +12 -0
- vocab.json +48 -0
special_tokens_map.json
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"pad_token": "ч",
|
3 |
+
"unk_token": "<unk>"
|
4 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_blank": true,
|
3 |
+
"clean_up_tokenization_spaces": true,
|
4 |
+
"is_uroman": false,
|
5 |
+
"language": "kjh",
|
6 |
+
"model_max_length": 1000000000000000019884624838656,
|
7 |
+
"normalize": true,
|
8 |
+
"pad_token": "ч",
|
9 |
+
"phonemize": false,
|
10 |
+
"tokenizer_class": "VitsTokenizer",
|
11 |
+
"unk_token": "<unk>"
|
12 |
+
}
|
vocab.json
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
" ": 20,
|
3 |
+
"-": 40,
|
4 |
+
"0": 1,
|
5 |
+
"1": 28,
|
6 |
+
"2": 43,
|
7 |
+
"4": 19,
|
8 |
+
"6": 44,
|
9 |
+
"_": 9,
|
10 |
+
"́": 21,
|
11 |
+
"а": 8,
|
12 |
+
"б": 24,
|
13 |
+
"в": 39,
|
14 |
+
"г": 26,
|
15 |
+
"д": 38,
|
16 |
+
"е": 14,
|
17 |
+
"з": 7,
|
18 |
+
"и": 11,
|
19 |
+
"й": 33,
|
20 |
+
"к": 17,
|
21 |
+
"л": 15,
|
22 |
+
"м": 13,
|
23 |
+
"н": 27,
|
24 |
+
"о": 29,
|
25 |
+
"п": 41,
|
26 |
+
"р": 2,
|
27 |
+
"с": 12,
|
28 |
+
"т": 22,
|
29 |
+
"у": 35,
|
30 |
+
"ф": 16,
|
31 |
+
"х": 10,
|
32 |
+
"ц": 6,
|
33 |
+
"ч": 0,
|
34 |
+
"ш": 30,
|
35 |
+
"ы": 5,
|
36 |
+
"ь": 4,
|
37 |
+
"э": 23,
|
38 |
+
"ю": 36,
|
39 |
+
"я": 34,
|
40 |
+
"ё": 37,
|
41 |
+
"і": 3,
|
42 |
+
"ғ": 45,
|
43 |
+
"ң": 18,
|
44 |
+
"ҷ": 32,
|
45 |
+
"ӧ": 25,
|
46 |
+
"ӱ": 42,
|
47 |
+
"–": 31
|
48 |
+
}
|