abdulhade's picture
Upload tokenizer
a9e65f2 verified
{
"<|af|>": 50327,
"<|am|>": 50334,
"<|as|>": 50350,
"<|az|>": 50304,
"<|ba|>": 50355,
"<|be|>": 50330,
"<|bg|>": 50292,
"<|bn|>": 50302,
"<|bo|>": 50347,
"<|br|>": 50309,
"<|bs|>": 50315,
"<|cy|>": 50297,
"<|endoftext|>": 50106,
"<|et|>": 50307,
"<|eu|>": 50310,
"<|fa|>": 50300,
"<|fo|>": 50338,
"<|gl|>": 50319,
"<|gu|>": 50333,
"<|haw|>": 50352,
"<|ha|>": 50354,
"<|hr|>": 50291,
"<|ht|>": 50339,
"<|hu|>": 50286,
"<|hy|>": 50312,
"<|is|>": 50311,
"<|jw|>": 50356,
"<|ka|>": 50329,
"<|kk|>": 50316,
"<|km|>": 50323,
"<|kn|>": 50306,
"<|la|>": 50294,
"<|lb|>": 50345,
"<|ln|>": 50353,
"<|lo|>": 50336,
"<|lt|>": 50293,
"<|lv|>": 50301,
"<|mg|>": 50349,
"<|mi|>": 50295,
"<|mk|>": 50308,
"<|ml|>": 50296,
"<|mn|>": 50314,
"<|mr|>": 50320,
"<|mt|>": 50343,
"<|my|>": 50346,
"<|ne|>": 50313,
"<|nn|>": 50342,
"<|nocaptions|>": 50362,
"<|notimestamps|>": 50363,
"<|no|>": 50288,
"<|oc|>": 50328,
"<|pa|>": 50321,
"<|ps|>": 50340,
"<|sa|>": 50344,
"<|sd|>": 50332,
"<|si|>": 50322,
"<|sk|>": 50298,
"<|sl|>": 50305,
"<|sn|>": 50324,
"<|so|>": 50326,
"<|sq|>": 50317,
"<|sr|>": 50303,
"<|startoflm|>": 50360,
"<|startofprev|>": 50361,
"<|su|>": 50357,
"<|sw|>": 50318,
"<|ta|>": 50287,
"<|te|>": 50299,
"<|tg|>": 50331,
"<|th|>": 50289,
"<|tk|>": 50341,
"<|tl|>": 50348,
"<|transcribe|>": 50359,
"<|translate|>": 50358,
"<|tt|>": 50351,
"<|ur|>": 50290,
"<|uz|>": 50337,
"<|yi|>": 50335,
"<|yo|>": 50325,
"<|¡|>": 50107,
"<|¢|>": 50108,
"<|£|>": 50109,
"<|¤|>": 50110,
"<|¥|>": 50111,
"<|¦|>": 50112,
"<|§|>": 50113,
"<|¨|>": 50114,
"<|©|>": 50115,
"<|ª|>": 50116,
"<|«|>": 50117,
"<|¬|>": 50118,
"<|®|>": 50119,
"<|¯|>": 50120,
"<|°|>": 50121,
"<|±|>": 50122,
"<|²|>": 50123,
"<|³|>": 50124,
"<|´|>": 50125,
"<|µ|>": 50126,
"<|¶|>": 50127,
"<|·|>": 50128,
"<|¸|>": 50129,
"<|¹|>": 50130,
"<|º|>": 50131,
"<|»|>": 50132,
"<|¼|>": 50133,
"<|½|>": 50134,
"<|¾|>": 50135,
"<|¿|>": 50136,
"<|À|>": 50137,
"<|Á|>": 50138,
"<|Â|>": 50139,
"<|Ã|>": 50140,
"<|Ä|>": 50141,
"<|Å|>": 50142,
"<|Æ|>": 50143,
"<|Ç|>": 50144,
"<|È|>": 50145,
"<|É|>": 50146,
"<|Ê|>": 50147,
"<|Ë|>": 50148,
"<|Ì|>": 50149,
"<|Í|>": 50150,
"<|Î|>": 50151,
"<|Ï|>": 50152,
"<|Ð|>": 50153,
"<|Ñ|>": 50154,
"<|Ò|>": 50155,
"<|Ó|>": 50156,
"<|Ô|>": 50157,
"<|Õ|>": 50158,
"<|Ö|>": 50159,
"<|×|>": 50160,
"<|Ø|>": 50161,
"<|ا|>": 50271,
"<|اÙĨ|>": 50279,
"<|ت|>": 50278,
"<|د|>": 50281,
"<|ر|>": 50275,
"<|س|>": 50285,
"<|Ù|>": 50162,
"<|Ùħ|>": 50282,
"<|ÙĨ|>": 50274,
"<|ÙĪ|>": 50273,
"<|Ú|>": 50163,
"<|Ú©|>": 50276,
"<|Û|>": 50164,
"<|ÛĨ|>": 50284,
"<|ÛĮ|>": 50270,
"<|Ûİ|>": 50280,
"<|Ûķ|>": 50269,
"<|Ü|>": 50165,
"<|Ý|>": 50166,
"<|Þ|>": 50167,
"<|ß|>": 50168,
"<|à|>": 50169,
"<|á|>": 50170,
"<|â|>": 50171,
"<|ã|>": 50172,
"<|ä|>": 50173,
"<|å|>": 50174,
"<|æ|>": 50175,
"<|ç|>": 50176,
"<|è|>": 50177,
"<|é|>": 50178,
"<|ê|>": 50179,
"<|ë|>": 50180,
"<|ì|>": 50181,
"<|í|>": 50182,
"<|î|>": 50183,
"<|ï|>": 50184,
"<|ð|>": 50185,
"<|ñ|>": 50186,
"<|ò|>": 50187,
"<|ó|>": 50188,
"<|ô|>": 50189,
"<|õ|>": 50190,
"<|ö|>": 50191,
"<|÷|>": 50192,
"<|ø|>": 50193,
"<|ù|>": 50194,
"<|ú|>": 50195,
"<|û|>": 50196,
"<|ü|>": 50197,
"<|ý|>": 50198,
"<|þ|>": 50199,
"<|ÿ|>": 50200,
"<|Ā|>": 50201,
"<|ā|>": 50202,
"<|Ă|>": 50203,
"<|ă|>": 50204,
"<|Ą|>": 50205,
"<|ą|>": 50206,
"<|Ć|>": 50207,
"<|ć|>": 50208,
"<|Ĉ|>": 50209,
"<|ĉ|>": 50210,
"<|Ċ|>": 50211,
"<|ċ|>": 50212,
"<|Č|>": 50213,
"<|č|>": 50214,
"<|Ď|>": 50215,
"<|ď|>": 50216,
"<|Đ|>": 50217,
"<|đ|>": 50218,
"<|Ē|>": 50219,
"<|ē|>": 50220,
"<|Ĕ|>": 50221,
"<|ĕ|>": 50222,
"<|Ė|>": 50223,
"<|ė|>": 50224,
"<|Ę|>": 50225,
"<|ę|>": 50226,
"<|Ě|>": 50227,
"<|ě|>": 50228,
"<|Ĝ|>": 50229,
"<|ĝ|>": 50230,
"<|Ğ|>": 50231,
"<|ğ|>": 50232,
"<|Ġ|>": 50233,
"<|ĠØ|>": 50272,
"<|Ġب|>": 50283,
"<|ĠÙ|>": 50277,
"<|ġ|>": 50234,
"<|Ģ|>": 50235,
"<|ģ|>": 50236,
"<|Ĥ|>": 50237,
"<|ĥ|>": 50238,
"<|Ħ|>": 50239,
"<|ħ|>": 50240,
"<|Ĩ|>": 50241,
"<|ĩ|>": 50242,
"<|Ī|>": 50243,
"<|ī|>": 50244,
"<|Ĭ|>": 50245,
"<|ĭ|>": 50246,
"<|Į|>": 50247,
"<|į|>": 50248,
"<|İ|>": 50249,
"<|ı|>": 50250,
"<|IJ|>": 50251,
"<|ij|>": 50252,
"<|Ĵ|>": 50253,
"<|ĵ|>": 50254,
"<|Ķ|>": 50255,
"<|ķ|>": 50256,
"<|ĸ|>": 50257,
"<|Ĺ|>": 50258,
"<|ĺ|>": 50259,
"<|Ļ|>": 50260,
"<|ļ|>": 50261,
"<|Ľ|>": 50262,
"<|ľ|>": 50263,
"<|Ŀ|>": 50264,
"<|ŀ|>": 50265,
"<|Ł|>": 50266,
"<|ł|>": 50267,
"<|Ń|>": 50268
}