marziye-A's picture
Upload tokenizer
1e56595
raw
history blame contribute delete
No virus
1.57 kB
{
"#": 1,
"&": 2,
"(": 3,
")": 4,
"[PAD]": 127,
"[UNK]": 126,
"_": 5,
"a": 6,
"b": 7,
"c": 8,
"d": 9,
"e": 10,
"f": 11,
"g": 12,
"h": 13,
"i": 14,
"k": 15,
"l": 16,
"m": 17,
"n": 18,
"o": 19,
"p": 20,
"q": 21,
"r": 22,
"s": 23,
"t": 24,
"u": 25,
"v": 26,
"w": 27,
"x": 28,
"y": 29,
"z": 30,
"|": 0,
"«": 31,
"»": 32,
"ā": 33,
"š": 34,
"؛": 35,
"؟": 36,
"ء": 37,
"آ": 38,
"أ": 39,
"ؤ": 40,
"ئ": 41,
"ا": 42,
"ب": 43,
"ة": 44,
"ت": 45,
"ث": 46,
"ج": 47,
"ح": 48,
"خ": 49,
"د": 50,
"ذ": 51,
"ر": 52,
"ز": 53,
"س": 54,
"ش": 55,
"ص": 56,
"ض": 57,
"ط": 58,
"ظ": 59,
"ع": 60,
"غ": 61,
"ـ": 62,
"ف": 63,
"ق": 64,
"ك": 65,
"ل": 66,
"م": 67,
"ن": 68,
"ه": 69,
"و": 70,
"ى": 71,
"ي": 72,
"ً": 73,
"ٌ": 74,
"َ": 75,
"ُ": 76,
"ِ": 77,
"ّ": 78,
"ْ": 79,
"ٔ": 80,
"٬": 81,
"پ": 82,
"چ": 83,
"ژ": 84,
"ک": 85,
"گ": 86,
"ۀ": 87,
"ی": 88,
"ے": 89,
"ە": 90,
"–": 91,
"…": 92,
"ﭘ": 93,
"ﮐ": 94,
"ﮔ": 95,
"ﯽ": 96,
"ﯾ": 97,
"ﯿ": 98,
"ﺍ": 99,
"ﺎ": 100,
"ﺑ": 101,
"ﺒ": 102,
"ﺖ": 103,
"ﺘ": 104,
"ﺧ": 105,
"ﺩ": 106,
"ﺪ": 107,
"ﺭ": 108,
"ﺮ": 109,
"ﺱ": 110,
"ﺴ": 111,
"ﺷ": 112,
"ﺸ": 113,
"ﻀ": 114,
"ﻋ": 115,
"ﻌ": 116,
"ﻟ": 117,
"ﻡ": 118,
"ﻢ": 119,
"ﻤ": 120,
"ﻥ": 121,
"ﻧ": 122,
"ﻪ": 123,
"ﻭ": 124,
"ﻮ": 125
}