Upload tokenizer
Browse files- special_tokens_map.json +16 -0
- tokenizer_config.json +30 -0
- vocab.json +45 -0
special_tokens_map.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"pad_token": {
|
3 |
+
"content": "<pad>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"unk_token": {
|
10 |
+
"content": "<unk>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
}
|
16 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_blank": true,
|
3 |
+
"added_tokens_decoder": {
|
4 |
+
"41": {
|
5 |
+
"content": "<unk>",
|
6 |
+
"lstrip": false,
|
7 |
+
"normalized": false,
|
8 |
+
"rstrip": false,
|
9 |
+
"single_word": false,
|
10 |
+
"special": true
|
11 |
+
},
|
12 |
+
"42": {
|
13 |
+
"content": "<pad>",
|
14 |
+
"lstrip": false,
|
15 |
+
"normalized": false,
|
16 |
+
"rstrip": false,
|
17 |
+
"single_word": false,
|
18 |
+
"special": true
|
19 |
+
}
|
20 |
+
},
|
21 |
+
"clean_up_tokenization_spaces": false,
|
22 |
+
"is_uroman": false,
|
23 |
+
"language": "rus",
|
24 |
+
"model_max_length": 2048,
|
25 |
+
"normalize": false,
|
26 |
+
"pad_token": "<pad>",
|
27 |
+
"phonemize": false,
|
28 |
+
"tokenizer_class": "VitsTokenizer",
|
29 |
+
"unk_token": "<unk>"
|
30 |
+
}
|
vocab.json
ADDED
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
" ": 0,
|
3 |
+
"!": 37,
|
4 |
+
"+": 40,
|
5 |
+
",": 39,
|
6 |
+
".": 36,
|
7 |
+
":": 2,
|
8 |
+
"<pad>": 42,
|
9 |
+
"<unk>": 41,
|
10 |
+
"?": 38,
|
11 |
+
"_": 1,
|
12 |
+
"а": 3,
|
13 |
+
"б": 4,
|
14 |
+
"в": 5,
|
15 |
+
"г": 6,
|
16 |
+
"д": 7,
|
17 |
+
"е": 8,
|
18 |
+
"ж": 10,
|
19 |
+
"з": 11,
|
20 |
+
"и": 12,
|
21 |
+
"й": 13,
|
22 |
+
"к": 14,
|
23 |
+
"л": 15,
|
24 |
+
"м": 16,
|
25 |
+
"н": 17,
|
26 |
+
"о": 18,
|
27 |
+
"п": 19,
|
28 |
+
"р": 20,
|
29 |
+
"с": 21,
|
30 |
+
"т": 22,
|
31 |
+
"у": 23,
|
32 |
+
"ф": 24,
|
33 |
+
"х": 25,
|
34 |
+
"ц": 26,
|
35 |
+
"ч": 27,
|
36 |
+
"ш": 28,
|
37 |
+
"щ": 29,
|
38 |
+
"ъ": 30,
|
39 |
+
"ы": 31,
|
40 |
+
"ь": 32,
|
41 |
+
"э": 33,
|
42 |
+
"ю": 34,
|
43 |
+
"я": 35,
|
44 |
+
"ё": 9
|
45 |
+
}
|