YassineBenlaria
commited on
Commit
•
49ab6bf
1
Parent(s):
841f38f
Upload tokenizer
Browse files- special_tokens_map.json +6 -0
- tokenizer_config.json +13 -0
- vocab.json +45 -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 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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": null,
|
10 |
+
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
11 |
+
"unk_token": "[UNK]",
|
12 |
+
"word_delimiter_token": "|"
|
13 |
+
}
|
vocab.json
ADDED
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"[PAD]": 42,
|
3 |
+
"[UNK]": 41,
|
4 |
+
"|": 11,
|
5 |
+
"ء": 1,
|
6 |
+
"آ": 39,
|
7 |
+
"أ": 17,
|
8 |
+
"ؤ": 36,
|
9 |
+
"إ": 35,
|
10 |
+
"ئ": 6,
|
11 |
+
"ا": 20,
|
12 |
+
"ب": 5,
|
13 |
+
"ة": 34,
|
14 |
+
"ت": 4,
|
15 |
+
"ث": 26,
|
16 |
+
"ج": 29,
|
17 |
+
"ح": 40,
|
18 |
+
"خ": 22,
|
19 |
+
"د": 31,
|
20 |
+
"ذ": 2,
|
21 |
+
"ر": 15,
|
22 |
+
"ز": 21,
|
23 |
+
"س": 10,
|
24 |
+
"ش": 38,
|
25 |
+
"ص": 19,
|
26 |
+
"ض": 14,
|
27 |
+
"ط": 7,
|
28 |
+
"ظ": 24,
|
29 |
+
"ع": 9,
|
30 |
+
"غ": 16,
|
31 |
+
"ف": 32,
|
32 |
+
"ق": 33,
|
33 |
+
"ك": 8,
|
34 |
+
"ل": 27,
|
35 |
+
"م": 28,
|
36 |
+
"ن": 37,
|
37 |
+
"ه": 30,
|
38 |
+
"و": 0,
|
39 |
+
"ى": 3,
|
40 |
+
"ي": 13,
|
41 |
+
"ٱ": 25,
|
42 |
+
"چ": 12,
|
43 |
+
"ڤ": 18,
|
44 |
+
"ۚ": 23
|
45 |
+
}
|