Upload tokenizer
Browse files- special_tokens_map.json +6 -0
- tokenizer_config.json +14 -0
- vocab.json +98 -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 |
+
"cache_dir": "/tmp/",
|
4 |
+
"clean_up_tokenization_spaces": true,
|
5 |
+
"do_lower_case": false,
|
6 |
+
"eos_token": "</s>",
|
7 |
+
"model_max_length": 1000000000000000019884624838656,
|
8 |
+
"pad_token": "[PAD]",
|
9 |
+
"replace_word_delimiter_char": " ",
|
10 |
+
"target_lang": "ara",
|
11 |
+
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
12 |
+
"unk_token": "[UNK]",
|
13 |
+
"word_delimiter_token": "|"
|
14 |
+
}
|
vocab.json
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"ara": {
|
3 |
+
"#": 1,
|
4 |
+
"'": 2,
|
5 |
+
"(": 3,
|
6 |
+
")": 4,
|
7 |
+
"-": 5,
|
8 |
+
".": 6,
|
9 |
+
"A": 7,
|
10 |
+
"C": 8,
|
11 |
+
"D": 9,
|
12 |
+
"E": 10,
|
13 |
+
"F": 11,
|
14 |
+
"G": 12,
|
15 |
+
"H": 13,
|
16 |
+
"I": 14,
|
17 |
+
"L": 15,
|
18 |
+
"M": 16,
|
19 |
+
"N": 17,
|
20 |
+
"O": 18,
|
21 |
+
"P": 19,
|
22 |
+
"R": 20,
|
23 |
+
"S": 21,
|
24 |
+
"T": 22,
|
25 |
+
"U": 23,
|
26 |
+
"V": 24,
|
27 |
+
"Y": 25,
|
28 |
+
"[PAD]": 93,
|
29 |
+
"[UNK]": 92,
|
30 |
+
"_": 26,
|
31 |
+
"a": 27,
|
32 |
+
"b": 28,
|
33 |
+
"c": 29,
|
34 |
+
"d": 30,
|
35 |
+
"e": 31,
|
36 |
+
"f": 32,
|
37 |
+
"g": 33,
|
38 |
+
"h": 34,
|
39 |
+
"i": 35,
|
40 |
+
"k": 36,
|
41 |
+
"l": 37,
|
42 |
+
"m": 38,
|
43 |
+
"n": 39,
|
44 |
+
"o": 40,
|
45 |
+
"p": 41,
|
46 |
+
"r": 42,
|
47 |
+
"s": 43,
|
48 |
+
"t": 44,
|
49 |
+
"u": 45,
|
50 |
+
"v": 46,
|
51 |
+
"w": 47,
|
52 |
+
"y": 48,
|
53 |
+
"z": 49,
|
54 |
+
"|": 0,
|
55 |
+
"،": 50,
|
56 |
+
"؟": 51,
|
57 |
+
"ء": 52,
|
58 |
+
"آ": 53,
|
59 |
+
"أ": 54,
|
60 |
+
"ؤ": 55,
|
61 |
+
"إ": 56,
|
62 |
+
"ئ": 57,
|
63 |
+
"ا": 58,
|
64 |
+
"ب": 59,
|
65 |
+
"ة": 60,
|
66 |
+
"ت": 61,
|
67 |
+
"ث": 62,
|
68 |
+
"ج": 63,
|
69 |
+
"ح": 64,
|
70 |
+
"خ": 65,
|
71 |
+
"د": 66,
|
72 |
+
"ذ": 67,
|
73 |
+
"ر": 68,
|
74 |
+
"ز": 69,
|
75 |
+
"س": 70,
|
76 |
+
"ش": 71,
|
77 |
+
"ص": 72,
|
78 |
+
"ض": 73,
|
79 |
+
"ط": 74,
|
80 |
+
"ظ": 75,
|
81 |
+
"ع": 76,
|
82 |
+
"غ": 77,
|
83 |
+
"ـ": 78,
|
84 |
+
"ف": 79,
|
85 |
+
"ق": 80,
|
86 |
+
"ك": 81,
|
87 |
+
"ل": 82,
|
88 |
+
"م": 83,
|
89 |
+
"ن": 84,
|
90 |
+
"ه": 85,
|
91 |
+
"و": 86,
|
92 |
+
"ى": 87,
|
93 |
+
"ي": 88,
|
94 |
+
"ً": 89,
|
95 |
+
"ُ": 90,
|
96 |
+
"ّ": 91
|
97 |
+
}
|
98 |
+
}
|