{ "merges": [ [ "Ġ", "Ġ" ], [ "ĠĠ", "ĠĠ" ], [ "ã", "ģ" ], [ "ã", "Ĥ" ], [ "Ċ", "ĠĠĠĠ" ], [ "ã", "ĥ" ], [ "ĊĠĠĠĠ", "ĠĠĠĠ" ], [ "ã", "Ģ" ], [ "ãģ", "®" ], [ "`", "`" ], [ "o", "r" ], [ "a", "t" ], [ "e", "n" ], [ "ãĢ", "Ĥ" ], [ "o", "n" ], [ "i", "n" ], [ "-", "-" ], [ "e", "r" ], [ "Ġ", "=" ], [ "e", "l" ], [ "ãĤ", "ĭ" ], [ "a", "l" ], [ "ãĤ", "Ĵ" ], [ "l", "o" ], [ "a", "r" ], [ "ĊĠĠĠĠĠĠĠĠ", "ĠĠĠĠ" ], [ "ãģ", "«" ], [ "i", "t" ], [ "\"", "\"" ], [ "s", "t" ], [ "e", "t" ], [ "ãĢ", "ģ" ], [ "e", "s" ], [ "ãģ", "Ļ" ], [ "i", "on" ], [ "--", "--" ], [ "o", "d" ], [ "c", "h" ], [ "ãģ", "§" ], [ "ãĥ", "¼" ], [ "ãģ", "¯" ], [ "Ċ", "ĊĠĠĠĠ" ], [ "d", "_" ], [ "e", "_" ], [ "r", "a" ], [ "ãģĻ", "ãĤĭ" ], [ "Ġ", "ãģ®" ], [ "ãģ", "¨" ], [ "ãģ", "ª" ], [ "ãģ", "Ĺ" ], [ "å", "Ī" ], [ "Ġ", "n" ], [ "Ċ", "Ċ" ], [ "å", "®" ], [ ")", "ãĢĤ" ], [ "ãĥ", "Ī" ], [ "l", "a" ], [ "l", "e" ], [ "ãģ", "Į" ], [ "ãģ", "Ł" ], [ "å", "Ĳ" ], [ "ãĥ", "³" ], [ "s", ":" ], [ "s", "el" ], [ "sel", "f" ], [ "u", "r" ], [ "ãģ", "Ħ" ], [ "\"\"", "\"" ], [ "a", "x" ], [ "æ", "ķ" ], [ "Ġ", "N" ], [ "e", "d" ], [ "r", "o" ], [ "ãĤ", "¹" ], [ "ãĤ", "¯" ], [ "en", "t" ], [ "d", "e" ], [ "i", "z" ], [ "i", "d" ], [ "on", "e" ], [ "Ġ", "``" ], [ "t", "o" ], [ "ãĥ", "«" ], [ "æķ", "°" ], [ "ä", "½" ], [ "u", "t" ], [ "ãĥ", "ĥ" ], [ "s", "e" ], [ "ä", "¸" ], [ "c", "o" ], [ "p", "o" ], [ "e", "x" ], [ "ãģ", "¦" ], [ "Ġ", "-" ], [ "a", "n" ], [ "lo", "g" ], [ "å", "ħ" ], [ "c", "t" ], [ "Ġ", "f" ], [ "self", "." ], [ "m", "od" ], [ "in", "g" ], [ "or", "m" ], [ "Ġ", "in" ], [ "a", "m" ], [ "è", "¨" ], [ "å", "Ń" ], [ "----", "----" ], [ "ĠN", "one" ], [ "c", "e" ], [ "å", "ı" ], [ "u", "n" ], [ "_", "b" ], [ "æ", "ľ" ], [ "lo", "at" ], [ "mod", "el" ], [ "m", "p" ], [ "æ", "ĸ" ], [ "ãĤ", "¤" ], [ "0", "0" ], [ "k", "en" ], [ "p", "." ], [ "ãģ®", "å" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "ĠĠĠĠ" ], [ "et", "ur" ], [ "etur", "n" ], [ "ãĤ", "ģ" ], [ "i", "f" ], [ "Ġ", "t" ], [ "st", "r" ], [ "ĊĠĠĠĠ", "\"\"\"" ], [ "Ġ", "1" ], [ "d", "i" ], [ "Ġ", "(" ], [ "r", "e" ], [ "s", "_" ], [ "ãĤ", "Į" ], [ "s", "or" ], [ "å®", "ļ" ], [ "at", "ion" ], [ "\"", "," ], [ "(", ")" ], [ "å", "¤" ], [ "å", "½" ], [ "_", "s" ], [ "Ġ", "S" ], [ "l", "i" ], [ "a", "s" ], [ "it", "ion" ], [ "ãĥ", "©" ], [ "Ġ", "b" ], [ "v", "al" ], [ "ãĥĥ", "ãĥ" ], [ "ç", "Ķ" ], [ "l", "en" ], [ "Ģ", "¤" ], [ "e", "(" ], [ "å", "ī" ], [ "ãģ", "¿" ], [ "ãģ", "Ĩ" ], [ "ãĥ", "»" ], [ "Ġ", "i" ], [ "m", "ax" ], [ "ex", "t" ], [ "Ġ-", ">" ], [ "_", "f" ], [ "t", "i" ], [ "Ġ", "+" ], [ "s", "ition" ], [ "q", "u" ], [ "¯", "¾" ], [ "to", "ken" ], [ "i", "g" ], [ "Ġ", "T" ], [ "o", "l" ], [ "åĲ", "Ī" ], [ "en", "sor" ], [ "è", "¦" ], [ "Ġ", "ãĤĴ" ], [ "p", "ar" ], [ "d_", "model" ], [ "de", "f" ], [ "t", "h" ], [ "a", "d" ], [ ")", ")" ], [ "ãģ", "Ĥ" ], [ "c", "od" ], [ "åĪ", "Ĺ" ], [ "or", "ch" ], [ "ãģĻãĤĭ", "ãĢĤ" ], [ "(", "``" ], [ "ãģ", "ĭ" ], [ "å", "¯¾" ], [ "å", "Ĭ" ], [ "_", "n" ], [ "å", "Į" ], [ "è", "¡" ], [ "è", "ª" ], [ "ãĤ", "¿" ], [ "å", "Ģ¤" ], [ "ãĥ", "ª" ], [ "Ġ", "/" ], [ "c", "a" ], [ "ãĤ", "ī" ], [ "ãģª", "ãģĦ" ], [ "åĪ", "Ĩ" ], [ "i", "v" ], [ "ãģĹ", "ãģ¦" ], [ "Ġ", "2" ], [ "ãĤ", "Ĭ" ], [ "h", "a" ], [ "m", "b" ], [ "o", "t" ], [ "ãģ", "ĵ" ], [ "ãĥ¼", "ãĥ" ], [ "ãģ", "¾" ], [ "\\", "\\" ], [ "la", "b" ], [ "r", "eturn" ], [ "u", "m" ], [ "Ġf", "loat" ], [ "val", "u" ], [ "åĮ", "ĸ" ], [ "e", "p" ], [ "er", "_" ], [ "å®", "Ł" ], [ "å", "¾" ], [ "æĸ", "ĩ" ], [ "Ġ", "s" ], [ "orch", "." ], [ "ä", "»" ], [ "Ġn", "p." ], [ "_", "c" ], [ "ãģ", "į" ], [ "ãĥ", "Ĩ" ], [ "par", "am" ], [ "Ġ", "self." ], [ "t", "ent" ], [ "ãģ", "ķ" ], [ "æ", "Į" ], [ "tent", "ion" ], [ "y", "t" ], [ "ç", "®" ], [ "at", "h" ], [ "w", "ar" ], [ "æ", "¸" ], [ "ç®", "Ĺ" ], [ "èª", "ŀ" ], [ "po", "sition" ], [ "¦", "ç" ], [ "ãĥ¼", "ãĤ¯" ], [ "p", "e" ], [ "u", "e" ], [ "ĊĊ", "Ċ" ], [ "c", "k" ], [ "un", "k" ], [ "]", "," ], [ "s", ")" ], [ "Ġ", "ãģ«" ], [ "ĠT", "ensor" ], [ "(", "n" ], [ "et", "a" ], [ "Ġ", "str" ], [ "ł", "´" ], [ ".", "s" ], [ "s", "," ], [ "é", "ĸ" ], [ "Ġ", "*" ], [ ".", "0" ], [ "Ġ", "0" ], [ "ç", "µ" ], [ "u", "l" ], [ "_", "t" ], [ "g", "ra" ], [ "é", "ĩ" ], [ "ł´", "åĲĪ" ], [ "iz", "er" ], [ "ãģŁ", "ãĤģ" ], [ "Ġ", "m" ], [ "Ġ", "A" ], [ "lab", "el" ], [ "å", "ĩ" ], [ "Ġ", "p" ], [ "Ġ", "ãĥ" ], [ "_", "h" ], [ "ä¸", "Ģ" ], [ "Ġin", "t" ], [ "¿", "Ĵ" ], [ "åŃ", "¦ç" ], [ "r", "es" ], [ "åŃ¦ç", "¿Ĵ" ], [ "é", "ł" ], [ "max", "_" ], [ "Ġ", "_" ], [ "a", "ct" ], [ "½", "®" ], [ "å½", "¢" ], [ "æ", "Ń" ], [ "ç", "½®" ], [ "or", "t" ], [ "ãĥ", "ĩ" ], [ "ä½", "¿" ], [ "al", "p" ], [ "alp", "ha" ], [ "ch", "unk" ], [ "mp", "ort" ], [ "Ċ", "ĊĠĠĠĠĠĠĠĠ" ], [ "e", "ad" ], [ "h", "t" ], [ "ä½", "į" ], [ "Ġ", "å" ], [ "å", "°" ], [ ".", "." ], [ "at", "iv" ], [ "Ġ", ":" ], [ "Ġ", "ãģ§" ], [ "f", "or" ], [ "y", "_" ], [ "ãĥĪ", "ãĥ¼ãĤ¯" ], [ "åŃ", "Ĺ" ], [ "ra", "n" ], [ "st", "ep" ], [ "u", "p" ], [ "ãĤ", "Ī" ], [ "ãĥ©", "ãĥ" ], [ "ä", "º" ], [ "æ", "ı" ], [ "r", "g" ], [ "æŃ", "£" ], [ "è¨", "Ī" ], [ "é", "Ģ" ], [ "åĩ", "º" ], [ "Ġ", "|" ], [ "it", "_" ], [ "æĸĩ", "åŃĹ" ], [ "A", "rg" ], [ "Arg", "s:" ], [ "_", "p" ], [ "iz", "e" ], [ "æ", "¬" ], [ "æľ", "Ģ" ], [ "f", "loat" ], [ "_", "id" ], [ "ig", "ht" ], [ "ãĤĭ", "ãĢĤ" ], [ "Ġ", "ç" ], [ "i", "r" ], [ "li", "st" ], [ "t", "ext" ], [ "åĲ", "Į" ], [ "é", "ģ" ], [ "ĊĠĠĠĠĠĠĠĠ", "self." ], [ "Ġ", "ãģ¨" ], [ ")", "``" ], [ "--------", "--------" ], [ "è¡", "Į" ], [ "qu", "en" ], [ "ãĥ³", "ãĤ" ], [ "v", "er" ], [ "ãĤ", "Ĥ" ], [ "Ġ", "x" ], [ "_", "w" ], [ "çĶ", "¨" ], [ "Ġ", "\"" ], [ "Ġ", "ãģ¯" ], [ "ãģĹ", "ãģŁ" ], [ "æ¬", "¡" ], [ "ãģ", "Ī" ], [ "å", "ł´åĲĪ" ], [ "±", "¤" ], [ "s", "." ], [ "valu", "es" ], [ "ãģĭ", "ãĤī" ], [ "å", "¿" ], [ "Ġt", "orch." ], [ "f", "f" ], [ "ro", "m" ], [ "è¦", "ı" ], [ "ĊĊĊ", "def" ], [ "(", "x" ], [ "e", "st" ], [ "ãģ", "©" ], [ "ãĥ", "ĳ" ], [ "ĊĊĠĠĠĠ", "Args:" ], [ "_", "m" ], [ "ativ", "e_" ], [ "e", "ight" ], [ "y", "er" ], [ "ç", "Ĭ" ], [ "çĬ", "¶" ], [ "Ġ", "d" ], [ "e", "y" ], [ "i", "s" ], [ "s", "s" ], [ "ãģ", "¤" ], [ "ãĤ", "·" ], [ "n", "orm" ], [ "ãĥĪãĥ¼ãĤ¯", "ãĥ³" ], [ "ä½į", "ç½®" ], [ "åħ", "ĥ" ], [ "ĊĠĠĠĠ", "\"" ], [ ")", "ãĤĴ" ], [ "_", "in" ], [ "åī", "²" ], [ "ãĤ¤", "ãĤ" ], [ "R", "eturn" ], [ "Return", "s:" ], [ "d", "(" ], [ "f", "rom" ], [ "ç", "Ľ" ], [ "è", "£" ], [ "é", "Ļ" ], [ "as", "k" ], [ "ãģ®", "æ" ], [ "ãĤ¹", "ãĥĪ" ], [ "è", "¿" ], [ "è", "ĩ" ], [ "é", "ķ" ], [ "éķ", "·" ], [ "Ġ", "c" ], [ "Ġ", "co" ], [ "g", "th" ], [ "len", "gth" ], [ "or", "y" ], [ "po", "ut" ], [ "r", "at" ], [ "ra", "in" ], [ "ro", "pout" ], [ "t", "y" ], [ "ãĤ", "Ń" ], [ "åħ", "¨" ], [ "è¨Ī", "ç®Ĺ" ], [ "ĠNone", ":" ], [ "Ġi", "mport" ], [ ")", ":" ], [ "ar", "ra" ], [ "arra", "y" ], [ "i", "l" ], [ "r", "ue" ], [ "t", "tention" ], [ "ãĥ", "Ĳ" ], [ "ãĥ¼", "ãĤ¿" ], [ "(", "00" ], [ "P", "o" ], [ "i", "as" ], [ "s", "s_" ], [ "ç", "ķ" ], [ "Ġ", "[" ], [ "_", "co" ], [ "ãģ", "ĳ" ], [ "ãĤ", "³" ], [ "å", "±¤" ], [ "Ġ", "L" ], [ ".", "," ], [ "=", "\"" ], [ "at", "t" ], [ "mp", "ut" ], [ "u", "b" ], [ "³", "»" ], [ "ãģ", "ı" ], [ "åī", "į" ], [ "İ", "ĩ" ], [ ")", "," ], [ "mb", "ol" ], [ "n", "." ], [ "v", "i" ], [ "y", "mbol" ], [ "ãģ", "Ŀ" ], [ "ãĥ©ãĥ", "¡" ], [ "å½¢", "çĬ¶" ], [ "åĬ", "Ľ" ], [ "æ", "Ī" ], [ "éĩ", "į" ], [ "ĊĠĠĠĠ", "return" ], [ "er", "y" ], [ "in", "e" ], [ "ãĥĥãĥ", "Ĺ" ], [ "ĠS", "_" ], [ "u", "s" ], [ "³»", "åĪĹ" ], [ "Ġ", "log" ], [ "_b", "yt" ], [ "c", "al" ], [ "c", "or" ], [ "in", "t" ], [ "æ", "Ŀ" ], [ "ç", "©" ], [ "Ġ", "C" ], [ "F", "al" ], [ "a", "g" ], [ "a", "p" ], [ "mput", "e_" ], [ "ãĤ", "µ" ], [ "ãĥ", "£" ], [ "ä", "¾" ], [ "ä½¿", "ãģĨ" ], [ "æ¬¡", "åħĥ" ], [ "Ġ", "ãģĮ" ], [ "Ġf", "or" ], [ "Ĥ", "ç" ], [ "Fal", "se" ], [ "[", "str" ], [ "o", "p" ], [ "ãģ", "Ķ" ], [ "å", "¼" ], [ "ç", "İĩ" ], [ "å", "ĭ" ], [ "å®Ł", "è£" ], [ "å®Łè£", "ħ" ], [ "æ", "³" ], [ "æĪ", "Ĳ" ], [ "éĸ", "ĵ" ], [ "ĊĊĠĠĠĠ", "Returns:" ], [ "Ġ", "d_" ], [ "Ġint", "," ], [ "N", "one" ], [ "es", ":" ], [ "iz", "ation" ], [ "Ġ", "\\\\" ], [ "Ġ", "d_model" ], [ "_h", "ead" ], [ "e", "m" ], [ "i", "k" ], [ "i", "p" ], [ "ti", "m" ], [ "ãģĶ", "ãģ¨" ], [ "ãĤ", "£" ], [ "ãĤ¤", "ãĥĪ" ], [ "åĪĨ", "åī²" ], [ "æĸ", "¹" ], [ "ļ", "Ħ" ], [ "a", "ir" ], [ "ed", "i" ], [ "ty", "pe" ], [ "ãĥ«", "ãĢĤ" ], [ "å", "·" ], [ "å", "Ľ" ], [ "Ġ", "B" ], [ "Ġ", "list" ], [ "ĠNone", "," ], [ "Ġ``", "(" ], [ "Ġi", "s" ], [ "b", "eta" ], [ "ca", "b" ], [ "it", "le" ], [ "o", "cab" ], [ "u", "al" ], [ "ãĤ", "¸" ], [ "ãĤ", "ı" ], [ "ç", "ļĦ" ], [ ")", "ãģ®" ], [ "o", "ut" ], [ "å½", "Ļ" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "ĠĠĠĠ" ], [ "Ġ", "di" ], [ "a", "b" ], [ "d", "ing" ], [ "de", "vi" ], [ "devi", "ce" ], [ "s", "(" ], [ "v", "ocab" ], [ "ä¸", "Ĭ" ], [ "å", "Ĩ" ], [ "åĬ", "ł" ], [ "èªŀ", "å½Ļ" ], [ "-", "1" ], [ "ar", "n" ], [ "e", ":" ], [ "e", "d_" ], [ "el", "ative_" ], [ "len", "(" ], [ "r", "c" ], [ "r", "elative_" ], [ "ãĥ", "¢" ], [ "æĮ", "ĩ" ], [ "ĊĠĠĠĠĠĠĠĠ", "\"\"\"" ], [ "di", "ent" ], [ "lo", "t" ], [ "ãĤ", "¢" ], [ "ãĤĴ", "æ" ], [ "ãĥ³ãĤ", "°" ], [ "ãĥĳ", "ãĥ©ãĥ¡" ], [ "é", "ħ" ], [ "éħ", "į" ], [ "P", "E" ], [ "_", "rat" ], [ "li", "p" ], [ "æŃ£", "è¦ı" ], [ "Ġ", "V" ], [ "cod", "er" ], [ "gra", "dient" ], [ "o", "ken" ], [ "p", "l" ], [ "ãĢģ", "``" ], [ "ç", "³»åĪĹ" ], [ "ç", "·" ], [ "è", "¾" ], [ "ĊĠĠĠĠ", "ax" ], [ "ĊĠĠĠĠ", "if" ], [ "i", "es" ], [ "p", "ro" ], [ "¯", "Ķ" ], [ "ãģ", "£" ], [ "ãģ§", "ãģĤ" ], [ "\"", ")" ], [ "*", "*" ], [ "-", "8" ], [ "i", "st" ], [ "lo", "ck" ], [ "quen", "ce" ], [ "w", "eight" ], [ "ãģ", "ĺ" ], [ "ãģĵ", "ãģ¨" ], [ "å°", "ı" ], [ "åı", "·" ], [ "æ", "İ" ], [ "ç", "´" ], [ "ĊĊĠĠĠĠ", "def" ], [ "Ġ2", "0" ], [ "\"", ":" ], [ "id", "ual" ], [ "war", "d" ], [ "ãģª", "ãĤĭ" ], [ "ãĥ³", "ãĥ" ], [ "ãĥ¼ãĥ", "ī" ], [ "ãĥĥ", "ãĤ¯" ], [ "æĮ", "ģ" ], [ "Ġ", "I" ], [ "al", "ization" ], [ "at", "tention" ], [ "d", "ropout" ], [ "ed", "ding" ], [ "mb", "edding" ], [ "n", "um" ], [ "ãģ", "ļ" ], [ "ãĥĥãĥ", "ģ" ], [ "æ", "Ĺ" ], [ "è¦", "ģ" ], [ "é", "ļ" ], [ "Ġ", "h" ], [ "Ġn", "ot" ], [ "[", "float" ], [ "_s", "ize" ], [ "et", "_" ], [ "self", "," ], [ "¾", "éħį" ], [ "ãģ§", "ãģ¯" ], [ "ä»", "¶" ], [ "å¤", "§" ], [ "å¤", "ī" ], [ "åı", "ĸ" ], [ "æ", "¨" ], [ "æ", "Ľ" ], [ "æ", "ŀ" ], [ "æŃ£è¦ı", "åĮĸ" ], [ "è¨", "ĺ" ], [ "é", "Ľ" ], [ "Ċ", ")" ], [ "M", "od" ], [ "a", "d_" ], [ "e", "g" ], [ "le", "arn" ], [ "q", "," ], [ "res", "idual" ], [ "un", "t" ], [ "ãģ", "Ľ" ], [ "ãģ¨", "ãģĹãģ¦" ], [ "ãĤ", "±" ], [ "å", "º" ], [ "èĩ", "ª" ], [ "éĩį", "ãģ¿" ], [ "ĊĠĠĠĠĠĠĠĠ", "return" ], [ "att", "n" ], [ "ran", "s" ], [ "t", "rain" ], [ "ãģĿ", "ãģ®" ], [ "ãĥ", "ĸ" ], [ "ãĥ¼", "ãĥ«" ], [ "å", "¹" ], [ "åħ", "¥" ], [ "æľ", "¬" ], [ "çµ", "Ĳ" ], [ "ģ", "´" ], [ "at", "a" ], [ "l", "or" ], [ "o", "ol" ], [ "¡", "ä»¶" ], [ "ãĤ¹", "ãĤ¯" ], [ "å", "¸" ], [ "Ġ", "<" ], [ "Ģ", "§" ], [ "(n", "p." ], [ "Po", "sition" ], [ "T", "rue" ], [ "_n", "orm" ], [ "ct", "ion" ], [ "e", "w" ], [ "er", "at" ], [ "f", "orm" ], [ "g", "u" ], [ "op", "tim" ], [ "s", "``" ], [ "ãģŁãĤģ", "ãĢģ" ], [ "å", "Ł" ], [ "å¤", "±" ], [ "æ", "Ģ§" ], [ "éģ", "©" ], [ "Ġ", "al" ], [ "A", "_" ], [ "A", "ttention" ], [ "ca", "ch" ], [ "d", "array" ], [ "e", "an" ], [ "e", "quence" ], [ "en", "cod" ], [ "gu", "ag" ], [ "p", "lot" ], [ "r", "or" ], [ "t", "orch." ], [ "ãģĹ", "ãĢģ" ], [ "ãĥĳãĥ©ãĥ¡", "ãĥ¼ãĤ¿" ], [ "åĲĮ", "ãģĺ" ], [ "éł", "Ĩ" ], [ "Ġ", "U" ], [ "Ġdi", "ct" ], [ "Ħ", "ı" ], [ "Q", "u" ], [ "Qu", "ery" ], [ "_", "param" ], [ "ar", "r" ], [ "b", "yt" ], [ "c", "es" ], [ "ch", "ar" ], [ "g", "e(" ], [ "k", "ey" ], [ "rans", "form" ], [ "ãĥ", "¥" ], [ "ä", "¹" ], [ "Ġ", "et" ], [ "ħ", "§" ], [ "None", "``" ], [ "]", ":" ], [ "_", "to" ], [ "c", "on" ], [ "ex", "p" ], [ "n", "darray" ], [ "w", "or" ], [ "ãģĹ", "ãģªãģĦ" ], [ "ãĤ¤ãĤ", "º" ], [ "ãĤ¹", "ãĥĨ" ], [ "ãĤ¹ãĥĨ", "ãĥĥãĥĹ" ], [ "ãĥ", "ŀ" ], [ "ãĥĥãĥ", "ī" ], [ "å¿", "ľ" ], [ "æĿ", "¡ä»¶" ], [ "ç", "Ļ" ], [ "çĶ", "Ł" ], [ "è", "»" ], [ "é", "ĥ" ], [ "éĥ", "¨" ], [ "Ġ", "len(" ], [ "_f", "n" ], [ "a", "yer" ], [ "c", "la" ], [ "d", "o" ], [ "e", "ed" ], [ "el", "t" ], [ "learn", "ing" ], [ "y", "label" ], [ "ãģķ", "ãĤĮ" ], [ "ä", "¿" ], [ "ä½", "ĵ" ], [ "åı", "Ĥç" ], [ "åıĤç", "ħ§" ], [ "æı", "Ľ" ], [ "æĹ", "¢" ], [ "æŀ", "ľ" ], [ "ç©", "º" ], [ "ĠA", "x" ], [ "ĠS", "equence" ], [ "Ĥ", "¹" ], [ "(", "log" ], [ "an", "d" ], [ "lo", "p" ], [ "oken", "izer" ], [ "p", "er" ], [ "¢", "æķ°" ], [ "ãĥĩ", "ãĥ«" ], [ "å", "ĳ" ], [ "å¹", "³" ], [ "ĊĠĠĠĠĠĠĠĠ", "if" ], [ "T", "okenizer" ], [ "at", "ch" ], [ "edi", "a" ], [ "elt", "a" ], [ "token", "izer" ], [ "type", "=" ], [ "v", "el" ], [ "war", "m" ], [ "ä»", "¥" ], [ "å·", "®" ], [ "å¾", "Ĺ" ], [ "åĲ", "Ħ" ], [ "åĲĮ", "ä¸Ģ" ], [ "Ġ", "ax" ], [ "Ġ", "l" ], [ "Ġb", "ool" ], [ "Ġn", "n." ], [ "cla", "ss" ], [ "i", "e" ], [ "le", "[" ], [ "r", "ror" ], [ "ãĥ", "Ń" ], [ "å", "į" ], [ "åĩº", "åĬĽ" ], [ "è»", "¸" ], [ "Ġ", "F" ], [ "Ġ", "Î" ], [ "=", "True" ], [ "E", "rror" ], [ "N", "orm" ], [ "d", "ata" ], [ "en", "d(" ], [ "ro", "ss_" ], [ "t", "itle" ], [ "token", "_id" ], [ "ãģª", "ãģ©" ], [ "ãģĪ", "ãĤĭ" ], [ "ãĤ¹", "ãĤ±" ], [ "ãĥ", "İ" ], [ "ãĥ³", "ãĤ¯" ], [ "äº", "ĭ" ], [ "åħ", "Ī" ], [ "æ", "¯Ķ" ], [ "æ", "Ħı" ], [ "Ġnp.", "ndarray" ], [ "ĠĠ", "#" ], [ "B", "," ], [ "T", "ensor" ], [ "la", "yer" ], [ "r", "id" ], [ "ul", "ti" ], [ "¼", "ĥ" ], [ "ãģ«", "ãĤĪ" ], [ "ãĥĥ", "ãĥĪ" ], [ "å®ļ", "åĢ¤" ], [ "å¼", "ı" ], [ "æ", "ī" ], [ "æı", "ı" ], [ "çĶ", "»" ], [ "è", "©" ], [ "è", "¼ĥ" ], [ "éĻ", "Ĳ" ], [ "ĊĠĠĠĠ", ")" ], [ "ĠS_", "k" ], [ "ĠTensor", "," ], [ "Ġal", ".," ], [ "ĺ", "ãĥĥãĥī" ], [ "2", ")" ], [ "D", "e" ], [ "F", "or" ], [ "For", "ward" ], [ "_b", "u" ], [ "ap", "p" ], [ "co", "lor" ], [ "co", "mpute_" ], [ "e", "," ], [ "et", "wor" ], [ "etwor", "k" ], [ "i", "c" ], [ "p", "ut" ], [ "relative_", "position" ], [ "se", "quen" ], [ "u", "le" ], [ "å¾", "Į" ], [ "åł´åĲĪ", "ãģ¯" ], [ "æ¸", "Ľ" ], [ "è¨", "¼" ], [ "éĢ", "£" ], [ "ĊĊ", "from" ], [ "ĊĠĠĠĠĠĠĠĠ", "ax" ], [ "Ġ", "{" ], [ ".s", "et_" ], [ "K", "ey" ], [ "al", "_" ], [ "it", "s" ], [ "learning", "_rat" ], [ "m", "ean" ], [ "po", "w" ], [ "s", "()" ], [ "t", "(" ], [ "å¤", "ĸ" ], [ "å¯¾", "è" ], [ "å¿", "ħ" ], [ "æ", "º" ], [ "æ¨", "Ļ" ], [ "æ³", "¨" ], [ "æľ", "Ł" ], [ "è¿", "Ķ" ], [ "éĸ", "¢æķ°" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "``" ], [ "L", "U" ], [ "``", ")" ], [ "it", "em" ], [ "j", "s" ], [ "js", "on" ], [ "p", "ath" ], [ "s", "ymbol" ], [ "¤", "ľ" ], [ "ãĢ", "Į" ], [ "ãĢ", "į" ], [ "ãģ", "³" ], [ "ãģ", "ł" ], [ "ãģĮ", "ãģĤ" ], [ "ãĥ", "ł" ], [ "åĪ", "Ŀ" ], [ "éĩ", "ı" ], [ "Ċ", "from" ], [ "ĊĠĠĠĠĠĠĠĠ", "b" ], [ "Ġ", "P" ], [ "Ġd", "type=" ], [ "al", "ue" ], [ "byt", "e_" ], [ "c", "ross_" ], [ "ch", "il" ], [ "chil", "la" ], [ "e", "mbedding" ], [ "e", "valu" ], [ "in", "chilla" ], [ "w", "ik" ], [ "¤ľ", "è¨¼" ], [ "ãģ¯", "ãĤģ" ], [ "ãĤ¤", "ãĥ³" ], [ "ãĥ", "¬" ], [ "ä½", "ľ" ], [ "å¯¾", "å¿ľ" ], [ "į", "å¤±" ], [ "Ĳ", "įå¤±" ], [ "..", ".," ], [ "act", "ory" ], [ "er", "g" ], [ "g", "et" ], [ "ip", "edia" ], [ "it", "h" ], [ "ot", "ation" ], [ "ãģ¦", "ãģ¯ãĤģ" ], [ "ãĥ", "ģ" ], [ "ãĥ", "ķ" ], [ "åº", "¦" ], [ "åĪ", "ĩ" ], [ "åŃ", "ĺ" ], [ "æ¸", "¬" ], [ "æıı", "çĶ»" ], [ "æĹ¢", "å®ļåĢ¤" ], [ "ç·", "ļ" ], [ "è¾", "¼" ], [ "éģ©", "çĶ¨" ], [ "Ġ", "ãģ¤" ], [ "Ġ20", "1" ], [ "1", "," ], [ "[", "int" ], [ "[str", "," ], [ "do", "w" ], [ "warm", "up" ], [ "ãģ§", "ãģį" ], [ "ãģ®", "ãģ¿" ], [ "ãģ¾", "ãģ¾" ], [ "ãĤµ", "ãĤ¤ãĤº" ], [ "ä¹", "Ĺ" ], [ "å", "Ŀ" ], [ "åĨ", "į" ], [ "åĭ", "ķ" ], [ "åıĸ", "å¾Ĺ" ], [ "çĶŁ", "æĪĲ" ], [ "è¡", "¨" ], [ "ĊĠĠĠĠ", "``" ], [ "ĊĠĠĠĠĠĠĠĠ", "x" ], [ "Ġ", "æ" ], [ "ĠAx", "es" ], [ "\"", "``" ], [ "(", "\"" ], [ ".", "5" ], [ "A", "L" ], [ "L", "O" ], [ "_byt", "es" ], [ "_in", "it_" ], [ "_init_", "_" ], [ "_s", "pl" ], [ "el", "se" ], [ "en", "ce" ], [ "i", "x" ], [ "lo", "ss_" ], [ "log", "_n" ], [ "on", "ti" ], [ "to", "t" ], [ "ãģĤ", "ãģ¦ãģ¯ãĤģ" ], [ "ãģĻãĤĭ", "(" ], [ "ãĤĬ", "ãĢģ" ], [ "ãĥ£", "ãĥ³ãĤ¯" ], [ "å¿ħ", "è¦ģ" ], [ "åĭ", "¾éħį" ], [ "åĿ", "ĩ" ], [ "æĮĩ", "å®ļ" ], [ "Ġ", "Query" ], [ "Ġ", "max_" ], [ "--------", "----" ], [ "----------------", "----------------" ], [ "al", "l" ], [ "d", "elta" ], [ "g", "rid" ], [ "gradient", "_norm" ], [ "le", "vel" ], [ "m", "ask" ], [ "r", "onti" ], [ "s", "[" ], [ "se", "q" ], [ "ul", "t" ], [ "ver", "g" ], [ "ãģ", "°" ], [ "ãģ®å", "Ģ¤" ], [ "ãģĵ", "ãģ®" ], [ "ãĤŃ", "ãĥ£" ], [ "ãĥ¢", "ãĥĩãĥ«" ], [ "ãĥĨ", "ãĤŃ" ], [ "åĪĿ", "æľŁ" ], [ "æ¸", "¡" ], [ "æĸĩåŃĹ", "åĪĹ" ], [ "çµĲ", "æŀľ" ], [ "è", "ĥ" ], [ "éķ·", "ãģķ" ], [ "ĊĠĠĠĠĠĠĠĠ", ")" ], [ "Ġ", "W" ], [ "Ġ", "å½¢çĬ¶" ], [ "De", "coder" ], [ "_m", "ask" ], [ "app", "end(" ], [ "d", "am" ], [ "di", "r" ], [ "in", "dow" ], [ "ine", "ar" ], [ "iv", "ation" ], [ "la", "n" ], [ "m", "in" ], [ "q", "r" ], [ "§", "ĭ" ], [ "«", "ĸ" ], [ "ãĤĭ", "(" ], [ "ä¸", "Ń" ], [ "åı", "¯" ], [ "åĲ", "«" ], [ "è", "«ĸ" ], [ "èĥ", "½" ], [ "ĊĊĠĠĠĠ", "``" ], [ "Ĳ", "Ĩ" ], [ "()", "." ], [ "=", "0" ], [ "=", "False" ], [ "B", "lock" ], [ "_", "arr" ], [ "``", "ãĥ»" ], [ "b", "a" ], [ "cach", "e_" ], [ "de", "ca" ], [ "e_", "b" ], [ "gra", "d" ], [ "guag", "e" ], [ "i", "mport" ], [ "ie", "ce" ], [ "ol", "d" ], [ "rc", "." ], [ "ur", "v" ], [ "¤", "ĩ" ], [ "ãģ£", "ãģŁ" ], [ "ãģĶãģ¨", "ãģ«" ], [ "ãĥ©", "ãĥĥãĥģ" ], [ "ãĥĨãĤŃ", "ãĤ¹ãĥĪ" ], [ "å¯¾", "æķ°" ], [ "æ¯Ķ", "è¼ĥ" ], [ "æĸ", "°" ], [ "ç", "¾" ], [ "çļĦ", "ãģ«" ], [ "çĽ", "´" ], [ "è", "¤ĩ" ], [ "è¨", "Ģ" ], [ "è¾¼", "ãģ¿" ], [ "éĢ", "ļ" ], [ "Ċ", "#" ], [ "Ġ", "4" ], [ "Ġ", "a" ], [ "Ġ0", ".0" ], [ "ĠA", "ttention" ], [ "ĠI", "D" ], [ "Ġd_", "k" ], [ "ı", "¾" ], [ ")ãĢĤ", "``" ], [ ".", "append(" ], [ "B", "ias" ], [ "E", "n" ], [ "F", "it" ], [ "H", "ead" ], [ "P", "ar" ], [ "]", ")" ], [ "_spl", "it_" ], [ "_to", "_id" ], [ "act", "ivation" ], [ "ath", ":" ], [ "d_model", ":" ], [ "i", "se" ], [ "ize", "=" ], [ "lo", "ad_" ], [ "t", "ex" ], [ "¢", "º" ], [ "±", "¡" ], [ "ãģ®å", "ł´åĲĪ" ], [ "ãĥ©ãĥĥãĥģ", "å®Łè£ħ" ], [ "ä¸", "ĭ" ], [ "åį", "ĺ" ], [ "æ¸Ľ", "è¡" ], [ "æ¸Ľè¡", "°" ], [ "ç", "¢º" ], [ "ç", "¯" ], [ "è¨Ģ", "èªŀ" ], [ "ĊĠĠĠĠ", "n" ], [ "ĊĠĠĠĠĠĠĠĠ", "f" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "x" ], [ "Ġ", "re" ], [ "ĠT", "rue" ], [ "Ġfloat", "(np." ], [ "Ġn", "um" ], [ "Ġt", "ext" ], [ "ĠĠĠĠ", "ĠĠĠĠ" ], [ "M", "ulti" ], [ "Mod", "ule" ], [ "_w", "eight" ], [ "an", "n" ], [ "attn", "_weight" ], [ "ayer", "Norm" ], [ "ck", "et" ], [ "d_", "ff" ], [ "l", "in" ], [ "p", "air" ], [ "pow", "er_" ], [ "r", "et" ], [ "ver", "tex" ], [ "¨", "ĵ" ], [ "ãģ¦", "ãģĦ" ], [ "ãģ«", "ä½¿ãģĨ" ], [ "ãģ¯", "ãĢģ" ], [ "ãģĻ", "ãĢĤ" ], [ "ãĤĪ", "ãģĨ" ], [ "ãĤĮ", "ãĤĭ" ], [ "ãĥģ", "ãĥ£ãĥ³ãĤ¯" ], [ "ä¸", "į" ], [ "ä¸Ģ", "èĩ" ], [ "ä¸Ģèĩ", "´" ], [ "å®Ł", "é" ], [ "å®Łé", "¨ĵ" ], [ "å¯¾è", "±¡" ], [ "å¸", "¸" ], [ "å¾", "©" ], [ "å¿ħè¦ģ", "ãģĮãģĤ" ], [ "åħ¨", "ä½ĵ" ], [ "åī", "ĩ" ], [ "åĲĮä¸Ģ", "ãģ®" ], [ "ç", "ı¾" ], [ "è¿", "½" ], [ "éĻ", "¤" ], [ "éļ", "Ľ" ], [ "Ġ", "\"\"" ], [ "Ġ", "ran" ], [ "Ġ", "ä»¥" ], [ "ĭ", "ãĤģ" ], [ "_", "token" ], [ "a", "ce" ], [ "ar", "t" ], [ "ar", "y" ], [ "c", "e_" ], [ "deca", "y" ], [ "e", "am" ], [ "evalu", "ation" ], [ "f", "\"" ], [ "in", "it" ], [ "lo", "ss" ], [ "o", "f" ], [ "or", "ies" ], [ "out", "put" ], [ "p", "er_" ], [ "s", "ize=" ], [ "¼", "ãģ³" ], [ "ãģ", "¸" ], [ "ãĤ£", "ãĥ³ãĤ°" ], [ "ãĤ¹ãĤ¯", "ãĥ©ãĥĥãĥģå®Łè£ħ" ], [ "ãĥ", "Ĺ" ], [ "ãĥª", "ãĥ" ], [ "ãĥĪ", "ãĥª" ], [ "ä»", "ĺ" ], [ "ä¾", "ĭ" ], [ "å", "ģ´" ], [ "å¤ī", "æıĽ" ], [ "åĨ", "ħ" ], [ "åŁ", "ĭãĤģ" ], [ "åŃ¦ç¿Ĵ", "çİĩ" ], [ "æ³¨", "æĦı" ], [ "ç", "¶" ], [ "çļĦ", "ãģª" ], [ "è¿½", "åĬł" ], [ "ĊĠĠĠĠĠĠĠĠ", "n" ], [ "Ġ", "R" ], [ "Ġ", "k" ], [ "ĠS_", "q," ], [ "ĠTensor", ":" ], [ "ĠV", "alue" ], [ "Ġãĥ", "Ĳ" ], [ "2", "," ], [ "A_", "M" ], [ "F", "-8" ], [ "T", "F-8" ], [ "_s", "cal" ], [ "_split_", "mod" ], [ "_w", "indow" ], [ "a", "k" ], [ "al", "lab" ], [ "array", "(" ], [ "chunk", "_split_mod" ], [ "e", "ed_" ], [ "eed_", "for" ], [ "gra", "m" ], [ "h", "old" ], [ "la", "w" ], [ "m", "ent" ], [ "qu", "ery" ], [ "res", "hold" ], [ "st", "art" ], [ "step", "_" ], [ "t", "er" ], [ "title", ":" ], [ "§", "ãĥ³" ], [ "ãģ", "¹" ], [ "ãģ§ãģĤ", "ãĤĭ" ], [ "ãĤ·", "ãĥ¥" ], [ "ãĥ", "§ãĥ³" ], [ "ãĥ", "Ĭ" ], [ "ãĥ", "Ļ" ], [ "ãĥĥ", "ãĤ·ãĥ¥" ], [ "ä", "¼" ], [ "å", "İ" ], [ "å¹³", "åĿĩ" ], [ "åıĤçħ§", ")ãĢĤ" ], [ "æİ", "¥" ], [ "æľ", "ī" ], [ "æľĢ", "å°ı" ], [ "ç", "Ħ" ], [ "çķ", "ª" ], [ "éĽ", "Ĩ" ], [ "Ġ1", ".0" ], [ "Ġs", "rc." ], [ "Ġå", "Ľ" ], [ "ľ", "ãĥ«" ], [ "\"", "]" ], [ "(", "d_model" ], [ "00", "6" ], [ "1", ")" ], [ "_", "alpha" ], [ "_", "i" ], [ "_", "step" ], [ "_f", "ir" ], [ "_f", "it" ], [ "_fir", "st" ], [ "ab", "o" ], [ "abo", "la" ], [ "c", "l" ], [ "e", "mp" ], [ "er", "L" ], [ "h", "it" ], [ "i", "o" ], [ "ra", "ise" ], [ "st", "at" ], [ "w", "a" ], [ "®", "ĭ" ], [ "ãĤ·", "ãĥ³ãĥ" ], [ "ãĤ·ãĥ³ãĥ", "ľãĥ«" ], [ "ãĥĲ", "ãĤ¤ãĥĪ" ], [ "åŁĭãĤģ", "è¾¼ãģ¿" ], [ "åł´åĲĪ", "ãĢģ" ], [ "æº", "ĸ" ], [ "æĮģ", "ãģ¤" ], [ "æľĢ", "å¤§" ], [ "ç", "ī" ], [ "ç©", "į" ], [ "çķ", "°" ], [ "çĽ", "®" ], [ "çĽ", "¸" ], [ "è¨ĺ", "äºĭ" ], [ "éĥ¨", "åĪĨ" ], [ "éĽ", "¢" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "if" ], [ "Ġ", "3" ], [ "Ġ", "H" ], [ "Ġ", "â" ], [ "ĠF", "eed" ], [ "Ġ``(", "B," ], [ "Ġdict", "[str," ], [ "Ġãģ¨", "åĲĮãģĺ" ], [ "()", "``" ], [ ")", "ãģ«" ], [ "[", "\"" ], [ "_", "embedding" ], [ "ba", "ck" ], [ "c", "lip" ], [ "cor", "e" ], [ "es", "p" ], [ "est", "_p" ], [ "et", "er" ], [ "hit", "esp" ], [ "ig", "ra" ], [ "igra", "m" ], [ "it", "er" ], [ "l", "r" ], [ "max_", "chunk" ], [ "n", "p." ], [ "optim", "izer" ], [ "power_", "law" ], [ "qr", "t(" ], [ "qu", "ar" ], [ "¼ãģ³", "åĩº" ], [ "ãģª", "ãģı" ], [ "ãģĿãģ®", "ãģ¾ãģ¾" ], [ "ãģł", "ãģĳ" ], [ "ãĤ", "¨" ], [ "ãĤ³", "ãĥ¼ãĥ" ], [ "ãĤ¹ãĤ±", "ãĥ¼ãĥ«" ], [ "ãĤŃãĥ£", "ãĥĥãĤ·ãĥ¥" ], [ "ãĥ©ãĥ", "Ļ" ], [ "ãĥ¼", "ãĥĪ" ], [ "ãĥķ", "ãĤ" ], [ "æ", "Ļ" ], [ "æ¸", "Ī" ], [ "æĽ", "¸" ], [ "ç", "Ĥ¹" ], [ "çĦ", "¡" ], [ "è¡Į", "åĪĹ" ], [ "ĊĠĠĠĠ", "(" ], [ "ĊĠĠĠĠax", ".set_" ], [ "ĊĠĠĠĠĠĠĠĠ", "\"" ], [ "ĊĠĠĠĠĠĠĠĠ", "#" ], [ "ĊĠĠĠĠĠĠĠĠ", "s" ], [ "Ġt", "orch" ], [ "Ġt", "up" ], [ "Ġå", "ģ´" ], [ "(", "æĹ¢å®ļåĢ¤" ], [ ")", "ãģ¯" ], [ ".s", "ub" ], [ "=", "-1" ], [ "B", "E" ], [ "_", "len" ], [ "_", "se" ], [ "_c", "on" ], [ "_co", "e" ], [ "``ãĥ»", "``" ], [ "b", "i" ], [ "byte_", "level" ], [ "d_", "k" ], [ "di", "m" ], [ "ent", "ence" ], [ "g", "ment" ], [ "i", "le" ], [ "ist", "ory" ], [ "norm", "_first" ], [ "o", "is" ], [ "s", "/" ], [ "seq", "_len" ], [ "verg", "ed" ], [ "wik", "ipedia" ], [ "¤", "º" ], [ "ãĤ¤", "ãĥ«" ], [ "ãĥ", "ĺãĥĥãĥī" ], [ "ãĥª", "ãĤ¹ãĥĪ" ], [ "ãĥ«", "ãĥł" ], [ "å¸¸", "ãģ«" ], [ "åħ", "±" ], [ "åĬ", "¹" ], [ "åı", "į" ], [ "åĲ", "į" ], [ "æ", "ĺ" ], [ "æī", "±" ], [ "æİ", "¨" ], [ "æĸ¹", "å¼ı" ], [ "ç", "¤º" ], [ "ç¶", "ļ" ], [ "è¨Īç®Ĺ", "éĩı" ], [ "éł", "Ń" ], [ "ĊĊ", "\"\"\"" ], [ "ĊĠĠĠĠ", "f" ], [ "ĊĠĠĠĠĠĠĠĠ", "num" ], [ "Ġ", "False" ], [ "Ġ", "if" ], [ "Ġ", "ãģĭãĤī" ], [ "Ġ=", "=" ], [ "ĠB", "PE" ], [ "ĠSequence", "[float" ], [ "ĠU", "TF-8" ], [ "Ġco", "mpute_" ], [ "Ġran", "ge(" ], [ "ķ", "æķ°" ], [ "Ŀ", "åŃĺ" ], [ "(00", "6" ], [ "A", "dam" ], [ "AL", "P" ], [ "ALP", "H" ], [ "ALPH", "A_" ], [ "ALPHA_", "BE" ], [ "ALPHA_BE", "T" ], [ "ALPHA_BET", "A_M" ], [ "Decoder", "Block" ], [ "Par", "am" ], [ "Position", "al" ], [ "_", "ALPHA_BETA_M" ], [ "_co", "unt" ], [ "_t", "ransform" ], [ "an", "if" ], [ "c", "le" ], [ "ce_", "length" ], [ "e", "ct" ], [ "eg", "ative_" ], [ "h", "inchilla" ], [ "hitesp", "ace" ], [ "in", "f" ], [ "it", "s_" ], [ "it", "y" ], [ "li", "k" ], [ "orm", "alization" ], [ "re", "q" ], [ "residual", "_fn" ], [ "self", "_" ], [ "sequen", "ce_length" ], [ "to", "p" ], [ "tokenizer", "." ], [ "train", "_" ], [ "w", "ith" ], [ "ylabel", ":" ], [ "¦", "åı·" ], [ "¬", "¦åı·" ], [ "ãģĤ", "ãģŁ" ], [ "ãĤ", "»" ], [ "ãĥ", "¦" ], [ "ãĥ", "ĭ" ], [ "ãĥ©ãĥĻ", "ãĥ«ãĢĤ" ], [ "ãĥ³ãĤ", "½" ], [ "ãĥĨ", "ãĥ³ãĤ½" ], [ "æ¸Ī", "ãģ¿" ], [ "ç", "«" ], [ "çĻ", "½" ], [ "è", "§" ], [ "è¾", "ŀ" ], [ "é", "Ŀ" ], [ "éĿ", "ŀ" ], [ "ĊĠĠĠĠ", "ĠĠ" ], [ "ĊĠĠĠĠax", "." ], [ "ĊĠĠĠĠĠĠĠĠ", "for" ], [ "Ġ", "&" ], [ "ĠU", "n" ], [ "Ġdtype=", "float" ], [ "Ġstr", "," ], [ "(", "b" ], [ "(", "self." ], [ "Forward", "N" ], [ "ForwardN", "etwork" ], [ "_", "a" ], [ "_", "{" ], [ "_b", "od" ], [ "_b", "y" ], [ "_bu", "cket" ], [ "_c", "lip" ], [ "_h", "at" ], [ "al", "_transform" ], [ "cache_", "dir" ], [ "cod", "e" ], [ "cross_", "attention" ], [ "de", "cod" ], [ "e", "``" ], [ "e_", "cross_attention" ], [ "er", "o" ], [ "f", "-8" ], [ "m", "erg" ], [ "max", "(" ], [ "per_", "token" ], [ "position", "al_transform" ], [ "s", "rc" ], [ "us", "e_cross_attention" ], [ "ut", "f-8" ], [ "z", "ero" ], [ "¼ãģ³åĩº", "ãģĹ" ], [ "ãģ§", "ãĢģ" ], [ "ãģ¨", "ãģį" ], [ "ãģ®", "ç³»åĪĹ" ], [ "ãģĤãģŁ", "ãĤĬ" ], [ "ãĤ¤ãĤ", "¶" ], [ "ãĥª", "ãĥ³ãĤ°" ], [ "ãĥĪãĥ¼ãĤ¯", "ãĥĬ" ], [ "ãĥĪãĥ¼ãĤ¯ãĥĬ", "ãĤ¤ãĤ¶" ], [ "ãĥİ", "ãĥ«ãĥł" ], [ "äº", "Į" ], [ "å¤§", "ãģį" ], [ "åħ¥", "åĬĽ" ], [ "æ", "§ĭ" ], [ "æ", "Ĳįå¤±" ], [ "æľ", "ª" ], [ "èªŀå½Ļ", "ãĤµãĤ¤ãĤº" ], [ "ĊĠĠĠĠ", "for" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "#" ], [ "Ġ", "en" ], [ "Ġ", "j" ], [ "ĠAxes", "ãĢĤ" ], [ "ĠC", "allab" ], [ "ĠP", "ath" ], [ "ĠĠ", "Ġ:" ], [ "Ĵ", "Į" ], [ "(x", ")" ], [ "))", "," ], [ ".", "3" ], [ "00", "4" ], [ "00", "5" ], [ "E", "_" ], [ "G", "P" ], [ "GP", "T" ], [ "Mod", "el" ], [ "all", "back" ], [ "an", "ge(" ], [ "an", "guage" ], [ "ar", "ange(" ], [ "at", "ing" ], [ "at", "ur" ], [ "attn_weight", "s" ], [ "atur", "ating" ], [ "cor", "p" ], [ "d", "type" ], [ "eed_for", "war" ], [ "item", "s()" ], [ "its_", "per" ], [ "its_per", "_byt" ], [ "lo", "s" ], [ "log", "(" ], [ "on", "g" ], [ "p", "y" ], [ "param", "s" ], [ "sor", "t" ], [ "th", "reshold" ], [ "vertex", "_" ], [ "ãĢĤ", "``" ], [ "ãģ£", "ãģ¦" ], [ "ãģ¾", "ãģŁ" ], [ "ãģĻãĤĭ", "ãģŁãĤģ" ], [ "ãĤ³ãĥ¼ãĥ", "ĳ" ], [ "ãĤ³ãĥ¼ãĥĳ", "ãĤ¹" ], [ "ãĤ¸", "ãĥ¥" ], [ "ãĤ¿", "ãĤ¤ãĥĪ" ], [ "ãĤĤ", "ãģ®" ], [ "å", "ľ" ], [ "åıį", "å¾©" ], [ "æ³", "ķ" ], [ "æĮģ", "ãģŁ" ], [ "æĮĩ", "æķ°" ], [ "æķ°", "åĢ¤" ], [ "ç©º", "çĻ½" ], [ "ç·ļ", "å½¢" ], [ "è", "¶" ], [ "è¨Ģèªŀ", "ãĥ¢ãĥĩãĥ«" ], [ "è»¸", "ãĥ©ãĥĻãĥ«ãĢĤ" ], [ "ĊĊ", "import" ], [ "ĊĊĊ", "class" ], [ "ĊĠĠĠĠ", "b" ], [ "ĊĠĠĠĠĠĠĠĠ", "self," ], [ "Ġ", ">" ], [ "Ġ", "alpha" ], [ "Ġ", "g" ], [ "Ġ", "v" ], [ "Ġ\"\"", "," ], [ "ĠValue", "Error" ], [ "Ġd_model", ")``" ], [ "ĭ", "¾éħį" ], [ "2", "0" ], [ "B", "PE" ], [ "En", "coder" ], [ "[", ":" ], [ "_param", "s" ], [ "_s", "ymbol" ], [ "an", "ce" ], [ "and", "id" ], [ "andid", "at" ], [ "ativ", "e" ], [ "en", "erat" ], [ "f", "actory" ], [ "id", "th" ], [ "ig", "n" ], [ "it", "ext" ], [ "iz", "e(" ], [ "o", "ot" ], [ "otation", ":" ], [ "p", "iece" ], [ "re", "-" ], [ "re", "po" ], [ "ro", "b" ], [ "st", "ra" ], [ "values", "," ], [ "x", "label" ], [ "¬¦åı·", "åĮĸ" ], [ "ãģ", "»" ], [ "ãģ", "Ĭ" ], [ "ãģ«", "ãģª" ], [ "ãģ«", "ãģªãĤĭ" ], [ "ãģķ", "ãģĽ" ], [ "ãģķãĤĮ", "ãģŁ" ], [ "ãĤ¸ãĥ¥", "ãĥ¼ãĥ«" ], [ "ãĤ¿", "ãĥĹ" ], [ "ãĤĮ", "ãģ°" ], [ "ãĥĨ", "ãĤ£" ], [ "ãĥĲ", "ãĥĥãĥģ" ], [ "ä¼", "Ŀ" ], [ "ä¾", "¡" ], [ "ä¾", "ĿåŃĺ" ], [ "å¯¾", "ä½įç½®" ], [ "åĬł", "ç®Ĺ" ], [ "åı", "Ĺ" ], [ "åı¯", "èĥ½" ], [ "åıĹ", "ãģĳ" ], [ "åĽ", "ŀ" ], [ "æİ¨", "å®ļ" ], [ "ç©º", "éĸĵ" ], [ "çµ", "Ĥ" ], [ "çĻ", "º" ], [ "è", "²" ], [ "è¨ĺ", "åı·" ], [ "è¶", "ħ" ], [ "è¾ŀ", "æĽ¸" ], [ "éĢ£", "çµĲ" ], [ "éĽĨ", "åĲĪ" ], [ "ĊĠĠĠĠĠĠĠĠ", "d_model:" ], [ "Ġ", "chunk" ], [ "Ġ", "ãģª" ], [ "Ġ", "åŃ¦ç¿Ĵ" ], [ "Ġ*", "*" ], [ "Ġ+", "=" ], [ "Ġ1", "e" ], [ "Ġ<", "=" ], [ "ĠN", "ormalization" ], [ "ĠN", "otation:" ], [ "Ġb", "ias" ], [ "Ġdtype=float", ")" ], [ "Ġlist", "[str" ], [ "ĠãĥĲ", "ãĤ¤ãĥĪ" ], [ "Ĥç", "Ĥ¹" ], [ "(", "0" ], [ ")", "ãĢģ" ], [ "-", "Forward" ], [ "=", "1" ], [ "E", "mbedding" ], [ "G", "_ALPHA_BETA_M" ], [ "I", "N" ], [ "LO", "G_ALPHA_BETA_M" ], [ "M", "S" ], [ "Multi", "Head" ], [ "Position", "Bias" ], [ "^", "{" ], [ "_", "le" ], [ "_", "op" ], [ "_byt", "es:" ], [ "_c", "y" ], [ "_coe", "f" ], [ "_init__", "(" ], [ "_size", ":" ], [ "a", "w" ], [ "ar", "get" ], [ "at", "e_" ], [ "d", "ition" ], [ "d_", "factory" ], [ "eed_forwar", "d_factory" ], [ "evaluation", "_window" ], [ "exp", "(log" ], [ "iz", "er_" ], [ "lan", "guag" ], [ "n", "am" ], [ "o", "un" ], [ "oun", "d" ], [ "p", "ing" ], [ "par", "abola" ], [ "ronti", "er" ], [ "src", "/" ], [ "t", "in" ], [ "u", "ct" ], [ "ul", "ary" ], [ "ur", "e_" ], [ "values", "_by" ], [ "y_", "length" ], [ "}", "``" ], [ "¸", "¦" ], [ "ãģ«ãĤĪ", "ãĤĭ" ], [ "ãģĹãģ¦", "ãģĦ" ], [ "ãĤ³", "ãĥ¼ãĥī" ], [ "ãĤĪãģĨ", "ãģ«" ], [ "ãĥ³", "ãĥĪ" ], [ "äºĮ", "ä¹Ĺ" ], [ "å", "¢" ], [ "å®", "Į" ], [ "åħĪ", "éłŃ" ], [ "æ", "¤ľè¨¼" ], [ "æ¬¡åħĥ", "ãĢĤ" ], [ "æ¸¡", "ãģĻ" ], [ "ç", "ĲĨ" ], [ "ç«", "ĭ" ], [ "çķ°", "ãģªãĤĭ" ], [ "éĩįãģ¿", "æ¸Ľè¡°" ], [ "éļ", "ł" ], [ "éļł", "ãĤĮ" ], [ "ĠN", "etwork" ], [ "Ġ[", "]" ], [ "Ġ_", "_init__(" ], [ "Ġ``", "None``" ], [ "Ġ``(", "...," ], [ "Ġd_", "ff" ], [ "Ġãģ¤", "ãģ®" ], [ "(", "1" ], [ "(", "p" ], [ "(n", "um" ], [ ")", "ãģĮ" ], [ "MultiHead", "Attention" ], [ "U", "N" ], [ "]", "]" ], [ "_", "2" ], [ "_b", "eta" ], [ "_con", "dition" ], [ "_in", "it" ], [ "at", "r" ], [ "atr", "ix" ], [ "c", "ent" ], [ "e", "ories" ], [ "e", "ur" ], [ "encod", "ing" ], [ "enerat", "or" ], [ "erat", "e(" ], [ "l", "t" ], [ "li", "b" ], [ "norm", "alization" ], [ "re", "ction" ], [ "th", "eories" ], [ "u", "mp" ], [ "ub", "er" ], [ "wa", "y" ], [ "warmup", "_step" ], [ "¡", "ãĤ¤ãĥ«" ], [ "·", "Ŀ" ], [ "º", "å®ļ" ], [ "ãģ®å", "ĭ¾éħį" ], [ "ãģĨ", "ãĢĤ" ], [ "ãģĵ", "ãĤĮ" ], [ "ãĤ", "«" ], [ "ãĤ¹", "ãĤ³" ], [ "ãĤĮ", "ãģªãģĦ" ], [ "ãĥ¦", "ãĥĭ" ], [ "ãĥŀ", "ãĤ¹ãĤ¯" ], [ "äº", "Ī" ], [ "å¯¾å¿ľ", "ãģĻãĤĭ" ], [ "åİ", "Ł" ], [ "åĽ", "²" ], [ "ç¢º", "çİĩ" ], [ "ç¯", "Ħ" ], [ "ç¯Ħ", "åĽ²" ], [ "ç´", "ł" ], [ "çµ", "Ħ" ], [ "çĽ´", "æİ¥" ], [ "éģ", "ķ" ], [ "éł", "ħ" ], [ "Ċ", "\"\"\"" ], [ "ĊĊĠĠĠĠ", "if" ], [ "ĊĠĠĠĠ", "log" ], [ "ĊĠĠĠĠ", "text" ], [ "ĊĠĠĠĠĠĠĠĠ", "(" ], [ "ĊĠĠĠĠĠĠĠĠ", "``" ], [ "Ġ", "ann" ], [ "Ġ", "else" ], [ "Ġ(", "1" ], [ "Ġ0", "." ], [ "Ġ``", "\"" ], [ "Ġb", "eta" ], [ "Ġlist", "[float" ], [ "Ġp", "lt" ], [ "Ġs", "ymbol" ], [ "Ġãģ®", "ãĤ¿ãĥĹ" ], [ "Ġãĥ", "¢" ], [ "Ġãĥ", "ĺãĥĥãĥī" ], [ "(", "L" ], [ "(", "param" ], [ "(æĹ¢å®ļåĢ¤", ")" ], [ "--", "-" ], [ "------------", "---" ], [ "----------------", "---------------" ], [ "--------------------------------", "-------------------------------" ], [ ".", "items()" ], [ "C", "L" ], [ "[", "0" ], [ "_", "grad" ], [ "_", "layer" ], [ "_", "log" ], [ "_", "threshold" ], [ "_b", "y_" ], [ "_c", "urv" ], [ "_f", "ut" ], [ "_fut", "ure_" ], [ "_future_", "_" ], [ "_p", "rob" ], [ "andidat", "es" ], [ "d", "a" ], [ "de", "x" ], [ "device", ")" ], [ "device", "=" ], [ "e", "\"," ], [ "e", "val" ], [ "gment", "ation" ], [ "i", "m" ], [ "in", "(" ], [ "la", "mb" ], [ "lamb", "da" ], [ "layer", "s." ], [ "lop", "es" ], [ "loss_", "step_" ], [ "oot", "stra" ], [ "ootstra", "p" ], [ "otation", "s" ], [ "r", "ed" ], [ "rat", "io" ], [ "re", "f" ], [ "s", "cal" ], [ "se", "gmentation" ], [ "sequen", "ces" ], [ "st", "ance" ], [ "text", "(" ], [ "th", "eta" ], [ "ub", "er_" ], [ "um", "erate(" ], [ "vocab", "ulary" ], [ "y", ":" ], [ "¯", "Ģ" ], [ "³", "å®ļ" ], [ "·Ŀ", "éĽ¢" ], [ "ãģ¦ãģĦ", "ãĤĭ" ], [ "ãģ®", "ãĥªãĤ¹ãĥĪ" ], [ "ãģ®å", "¯¾" ], [ "ãģĶãģ¨", "ãģ®" ], [ "ãģļ", "ãĤĮ" ], [ "ãĤ", "¦" ], [ "ãĤµ", "ãĥĸ" ], [ "ãĤ·", "ãĥ¼ãĥī" ], [ "ãĤ¹ãĤ³", "ãĤ¢" ], [ "ä¸Ń", "éĸĵ" ], [ "ä¼Ŀ", "æ" ], [ "ä¼Ŀæ", "Ĵ" ], [ "ä¼ĿæĴ", "Ń" ], [ "ä½ľ", "æĪĲ" ], [ "å", "ģ" ], [ "å", "ĴĮ" ], [ "å®Į", "åħ¨" ], [ "å¯¾", "ãģĹãģ¦" ], [ "å°", "¤" ], [ "åĪ", "¥" ], [ "åĪĩ", "ãĤĬ" ], [ "åīį", "ãģ®" ], [ "åĮ", "º" ], [ "åĳ", "¼ãģ³åĩºãģĹ" ], [ "æ", "Ķ" ], [ "æĮģãģŁ", "ãģªãģĦ" ], [ "æķ", "£" ], [ "æķ°", "ãĢĤ" ], [ "æĻ", "Ĥ" ], [ "æĽ", "´" ], [ "ç", "¬¦åı·åĮĸ" ], [ "ç¾", "¤" ], [ "è¤ĩ", "æķ°" ], [ "è§", "£" ], [ "è¨", "Ń" ], [ "è¨Īç®Ĺ", "ãģĻãĤĭãĢĤ" ], [ "è¿Ķ", "ãģĻ" ], [ "éģķ", "ãģĦ" ], [ "ĊĠĠĠĠĠĠĠĠ", "raise" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "\"" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "(" ], [ "Ġ", "---------------------------------------------------------------" ], [ "Ġ", "and" ], [ "ĠCallab", "le[" ], [ "ĠFeed", "-Forward" ], [ "ĠL", "ayerNorm" ], [ "ĠN", "eur" ], [ "ĠS_k", ")``" ], [ "ĠValueError", "(" ], [ "Ġ_", "_future__" ], [ "Ġann", "otations" ], [ "Ġm", "ask" ], [ "Ġn", "orm" ], [ "Ġp", "air" ], [ "Ġãģ§", "è¿½åĬł" ], [ "Ġãģ®ãĤ¿ãĥĹ", "ãĥ«ãĢĤ" ], [ "Ġãĥ", "ĳ" ], [ "Ġç", "¯Ģ" ], [ "ĥ", "³å®ļ" ], [ "()", ")" ], [ ".", "b" ], [ ".", "re" ], [ ".5", "," ], [ "0", ".0" ], [ "=True", ")" ], [ "Adam", "W" ], [ "F", "LO" ], [ "FLO", "P" ], [ "Po", "w" ], [ "Pow", "erL" ], [ "PowerL", "aw" ], [ "PowerLaw", "Fit" ], [ "R", "es" ], [ "U", "n" ], [ "[:", "," ], [ "[int", "," ], [ "]", "]," ], [ "_", "_" ], [ "_", "_init__" ], [ "_", "val" ], [ "_f", "actory" ], [ "_in", "f" ], [ "_scal", "e:" ], [ "_t", "rain" ], [ "c", "at" ], [ "char", "act" ], [ "dropout", ":" ], [ "e(", "\"" ], [ "edi", "an" ], [ "em", "ory" ], [ "entence", "piece" ], [ "et", "r" ], [ "etr", "ic" ], [ "ex", "act" ], [ "exp", "on" ], [ "expon", "ent" ], [ "h", "istory" ], [ "i", "a" ], [ "j", "a" ], [ "j", "o" ], [ "m", "_head" ], [ "max_chunk", "_bytes" ], [ "model", "." ], [ "n", "egative_" ], [ "n", "n." ], [ "pl", "it" ], [ "position", "s" ], [ "r", "\"" ], [ "r", "ent" ], [ "s", "o" ], [ "s_", "negative_" ], [ "t", "r" ], [ "tot", "al" ], [ "tot", "al_" ], [ "ump", "y" ], [ "up", "er" ], [ "ur", "rent" ], [ "way", "s_negative_" ], [ "x", "_c" ], [ "´", "»" ], [ "´»", "æĢ§" ], [ "ãģ", "¡" ], [ "ãģ§", "ãģ®" ], [ "ãģ§ãģį", "ãĤĭ" ], [ "ãģ©", "ãģĨ" ], [ "ãģĤãģŁãĤĬ", "ãģ®" ], [ "ãģĦ", "ãģļãĤĮ" ], [ "ãģĮ", "ãĢģ" ], [ "ãĤ¯", "ãĥª" ], [ "ãĤ¸", "ãĥ§ãĥ³" ], [ "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ", "ãĢĤ" ], [ "ãĤĴ", "è¡Į" ], [ "ãĥªãĥ", "ĵ" ], [ "ãĥªãĥĵ", "ãĤ¸ãĥ§ãĥ³" ], [ "ãĥ³ãĥ", "Ĺ" ], [ "ãĥ¼", "ãĤ¸" ], [ "ãĥĩ", "ãĤ£ãĥ³ãĤ°" ], [ "ãĥĩ", "ãĥ¼ãĤ¿" ], [ "ãĥĳ", "ãĥĩãĤ£ãĥ³ãĤ°" ], [ "ãĥķãĤ", "¡ãĤ¤ãĥ«" ], [ "ä½ľæĪĲ", "ãģĻãĤĭãĢĤ" ], [ "ä¿", "Ĥ" ], [ "å·", "±" ], [ "å¸", "ĥ" ], [ "å½", "±" ], [ "å½", "ĵ" ], [ "åĪĿæľŁ", "åĢ¤" ], [ "åĲ", "ĳ" ], [ "åľ", "¨" ], [ "åŁ", "º" ], [ "æ¨Ļ", "æºĸ" ], [ "æĺ", "İ" ], [ "çµĦ", "ãģ¿" ], [ "è¦ģ", "ç´ł" ], [ "è©", "ķ" ], [ "è©ķ", "ä¾¡" ], [ "èĩª", "å·±" ], [ "ĊĊ", "è¨ĺåı·" ], [ "Ġ", "ch" ], [ "Ġ", "ãģªãģ©" ], [ "Ġ", "æĸĩåŃĹ" ], [ "Ġ&", "=" ], [ "ĠTensor", ")" ], [ "Ġ``", "{" ], [ "Ġb", "est" ], [ "Ġd", "ropout" ], [ "Ġdict", "[int," ], [ "Ġen", "umerate(" ], [ "Ġh", "," ], [ "Ġm", "ath" ], [ "Ġsrc.", "layers." ], [ "Ġtup", "le[" ], [ "Ġãģ®", "åł´åĲĪãĢģ" ], [ "Ġç", "¸¦" ], [ "Ģ", "ĭ" ], [ "(", "m" ], [ "().", "__init__" ], [ "().__init__", "()" ], [ "(d_model", "," ], [ ")", "ãģ§" ], [ ")", "ãģ¨" ], [ "))", ")" ], [ ")ãĢĤ``", "None``" ], [ ".", "m" ], [ ".s", "ize(" ], [ "0", "," ], [ "=", "b" ], [ "=\"", "ign" ], [ "=\"ign", "or" ], [ "A", "X" ], [ "F", "eed" ], [ "G", "LU" ], [ "R", "MS" ], [ "S", "_" ], [ "T", "ransform" ], [ "_", "," ], [ "_", "learning_rat" ], [ "_", "lik" ], [ "_head", "s" ], [ "_head", "s:" ], [ "_le", "ad" ], [ "_lead", "ing" ], [ "_leading", "_w" ], [ "_leading_w", "hitespace" ], [ "activation", "_fn" ], [ "alpha", "=0" ], [ "anif", "est_p" ], [ "ath:", "`" ], [ "cent", "er" ], [ "char", "_to_id" ], [ "charact", "er_" ], [ "chunk_split_mod", "e" ], [ "co", "ver" ], [ "d", "en" ], [ "e", ")" ], [ "ed", ":" ], [ "else", ":" ], [ "er", "i" ], [ "er", "ies" ], [ "f", "ig" ], [ "f", "il" ], [ "gradient", "_clip" ], [ "i", "s_" ], [ "is", "h" ], [ "iter", "bi" ], [ "lop", "s_" ], [ "lops_", "per_token" ], [ "loss_step_", "delta" ], [ "m", "ath:`" ], [ "or", "p" ], [ "out", "_leading_whitespace" ], [ "ronti", "er_" ], [ "so", "FLOP" ], [ "theories", "/" ], [ "train", "ing" ], [ "u", "d" ], [ "up", "le" ], [ "uper", "().__init__()" ], [ "ur", "e" ], [ "w", "_" ], [ "w", "ish" ], [ "}", ")" ], [ "}", "_t" ], [ "¡", "ãĥ³ãĥĪ" ], [ "´»æĢ§", "åĮĸ" ], [ "¼", "ķæķ°" ], [ "ãģ§ãģĤ", "ãĤĬãĢģ" ], [ "ãģ«", "ä½¿" ], [ "ãģ®å", "ī" ], [ "ãģ®åĢ¤", "ãĤĴ" ], [ "ãģĻãĤĭ", ")ãĢĤ" ], [ "ãģĻãĤĭ", "ãģĵãģ¨" ], [ "ãĤ", "°" ], [ "ãĤģ", "ãĤĭ" ], [ "ãĥ", "ļ" ], [ "ãĥ»", "00" ], [ "ãĥ¼", "ãĤ¹" ], [ "ãĥĨãĥ³ãĤ½", "ãĥ«ãĢĤ" ], [ "ä¸", "İ" ], [ "å®ļ", "ç¾" ], [ "å®ļç¾", "©" ], [ "å¯¾è±¡", "ãģ®" ], [ "å°ı", "ãģķ" ], [ "å¾", "ĵ" ], [ "åĩºåĬĽ", "å±¤" ], [ "åĽ", "ºå®ļ" ], [ "æĶ", "¹" ], [ "æĸ°", "è¦ı" ], [ "ç", "§" ], [ "ç·", "´" ], [ "è¨", "ĵ" ], [ "è¨ĵ", "ç·´" ], [ "éĥ¨åĪĨ", "èªŀ" ], [ "ĊĊĠĠĠĠ", "@" ], [ "ĊĠĠĠĠ", "def" ], [ "ĊĠĠĠĠĠĠĠĠ", "dropout:" ], [ "ĊĠĠĠĠĠĠĠĠ", "p" ], [ "ĊĠĠĠĠĠĠĠĠs", "uper().__init__()" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "(``" ], [ "Ġ", "device=" ], [ "Ġ", "ãĤ¹ãĥĨãĥĥãĥĹ" ], [ "Ġ[", "]," ], [ "Ġ_", "LOG_ALPHA_BETA_M" ], [ "Ġc", "lip" ], [ "Ġco", "s" ], [ "Ġfloat", ":" ], [ "Ġstr", ")" ], [ "Ġãĥĳ", "ãĥ©ãĥ¡" ], [ "Ġä»¥", "ä¸Ĭ" ], [ "!", "=" ], [ "(", "[" ], [ "(00", "4" ], [ "(n", "n." ], [ "(æĹ¢å®ļåĢ¤)", "ãģ®åł´åĲĪ" ], [ "8", ")" ], [ "C", "hinchilla" ], [ "DecoderBlock", "``" ], [ "E", "S" ], [ "L", "anguage" ], [ "R", "el" ], [ "Rel", "ative" ], [ "Relative", "PositionBias" ], [ "Un", "igram" ], [ "[0", "]" ], [ "_bod", "y_" ], [ "_c", "orp" ], [ "_cy", "cle" ], [ "_log", "_prob" ], [ "_t", "ext" ], [ "a", "," ], [ "al", "g" ], [ "al", "ways_negative_" ], [ "b", "eam" ], [ "ces", "sor" ], [ "ch", "inchilla" ], [ "con", "tin" ], [ "contin", "ue" ], [ "e", "LU" ], [ "e", "ff" ], [ "e_", "model" ], [ "ed", ")" ], [ "emp", "erat" ], [ "g", "t" ], [ "ha", "pe" ], [ "i", "d_" ], [ "izer_", "json" ], [ "languag", "e_model" ], [ "lin", "alg" ], [ "on", "l" ], [ "ong", "Tensor" ], [ "p", "e_" ], [ "p", "t" ], [ "param", "s:" ], [ "position", "s:" ], [ "res", "ult" ], [ "residual", "s" ], [ "self_", "attn" ], [ "text", ":" ], [ "uct", "_" ], [ "w", "idth" ], [ "war", "d(" ], [ "wik", "itext" ], [ "¢", "ãĤ¹" ], [ "¦ç", "ĲĨ" ], [ "ª", "ĵ" ], [ "®", "æŃ£" ], [ "ãĢĤ", "\"\"\"" ], [ "ãģ®", "ãģŁãĤģ" ], [ "ãģ®å", "ĩ" ], [ "ãģĦ", "ãģ¦" ], [ "ãģĿ", "ãĤĮ" ], [ "ãĤ¤ãĤ", "¢ãĤ¹" ], [ "ãĤ¤ãĥ³", "ãĥĩ" ], [ "ãĤĴ", "ãĢģ" ], [ "ãĥ¬", "ãĥĻ" ], [ "ãĥ¼ãĥ", "¡ãĥ³ãĥĪ" ], [ "ãĥ¼ãĥ", "ķ" ], [ "ãĥĲ", "ãĤ¤ãĤ¢ãĤ¹" ], [ "ãĥŀ", "ãĥ¼ãĤ¸" ], [ "äºĭ", "åīį" ], [ "ä¿", "®æŃ£" ], [ "å®Ł", "éļĽ" ], [ "å¾Į", "ãģ®" ], [ "å¿ħè¦ģãģĮãģĤ", "ãĤĭ" ], [ "åĪĨ", "å¸ĥ" ], [ "åĪĹ", "ãĤĴ" ], [ "åİŁ", "è«ĸ" ], [ "åİŁè«ĸ", "æĸĩ" ], [ "åĲ«", "ãģ¾" ], [ "æ", "®ĭ" ], [ "æ", "ħ" ], [ "æ¨", "ª" ], [ "æº", "Ģ" ], [ "æľĢ", "çµĤ" ], [ "æĿ", "¥" ], [ "çī", "¹" ], [ "çķ", "Į" ], [ "çķª", "åı·" ], [ "è©", "¦" ], [ "è²", "ł" ], [ "è¿", "ĳ" ], [ "éģ©çĶ¨", "ãģĻãĤĭ" ], [ "éĩį", "è¤ĩ" ], [ "éłĨ", "ä¼ĿæĴŃ" ], [ "Ċ", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "Args:" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "Returns:" ], [ "ĊĠĠĠĠ", "compute_" ], [ "ĊĠĠĠĠ", "m" ], [ "ĊĠĠĠĠ", "title:" ], [ "ĊĠĠĠĠ\"", "compute_" ], [ "ĊĠĠĠĠ\"\"\"", "``" ], [ "ĊĠĠĠĠĠĠĠĠ", "m" ], [ "ĊĠĠĠĠĠĠĠĠ", "max_" ], [ "ĊĠĠĠĠĠĠĠĠ", "title:" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", ")" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "if" ], [ "Ġ", "!=" ], [ "Ġ", "D" ], [ "Ġ", "Key" ], [ "Ġ", "æ¬¡åħĥ" ], [ "Ġ201", "7" ], [ "ĠB", "lock" ], [ "ĠFalse", "," ], [ "ĠS", "," ], [ "ĠSequence[float", "]," ], [ "Ġdtype=", "torch." ], [ "Ġn", "ew" ], [ "Ġnn.", "Module" ], [ "Ġnp.", "array(" ], [ "Ġnp.", "as" ], [ "Ġself.", "beta" ], [ "Ġstr", ":" ], [ "Ġâ", "Ī" ], [ "Ġãģ¨", "ãģĹãģ¦" ], [ "Ġãģª", "ãĤī" ], [ "Ġå", "±¤" ], [ "ĥ", "çķĮ" ], [ "Ń", "ī" ], [ "(", "r\"" ], [ "(", "title" ], [ "(nn.", "Module" ], [ ")", "}" ], [ "-", "gram" ], [ "-", "in" ], [ ".", "2" ], [ "..", "." ], [ ".s", "et" ], [ ".sub", "plot" ], [ ".subplot", "s(" ], [ "=\"", "utf-8" ], [ "C", "O" ], [ "Feed", "ForwardNetwork" ], [ "Key", "Positional" ], [ "KeyPositional", "Transform" ], [ "L", "inear" ], [ "W", "ik" ], [ "^", "K" ], [ "_", "decay" ], [ "_", "list" ], [ "_curv", "es" ], [ "_op", "t" ], [ "_size", "," ], [ "anifest_p", "ath" ], [ "cor", "e_b" ], [ "device", ":" ], [ "encoding", "=\"utf-8" ], [ "f", "it_" ], [ "g", "ro" ], [ "ha", "w" ], [ "ip", "y" ], [ "k", "e" ], [ "l", "m_head" ], [ "le", "g" ], [ "log", "it" ], [ "m", "ap" ], [ "ois", "y_" ], [ "on", "t" ], [ "ont", "size=" ], [ "po", "s" ], [ "r", "ight" ], [ "ref", "ix" ], [ "ret", "oken" ], [ "size=", "(" ], [ "sort", "ed_" ], [ "text", "{" ], [ "ti", "l" ], [ "token_id", "s:" ], [ "total_", "step" ], [ "un", "ction" ], [ "ur", "l" ], [ "vertex_", "log_n" ], [ "weight", "_decay" ], [ "y_length", ":" ], [ "±", "Ĥ" ], [ "ãģ®", "ãĤ¿ãĤ¤ãĥĪ" ], [ "ãģ®", "éģķãģĦ" ], [ "ãģ®", "éļłãĤĮ" ], [ "ãģ®ãĤ¿ãĤ¤ãĥĪ", "ãĥ«ãĢĤ" ], [ "ãģ®å", "¤" ], [ "ãģ¹", "ãģį" ], [ "ãģ¹ãģį", "ä¹Ĺ" ], [ "ãģ»", "ãģ©" ], [ "ãģĳ", "ãĤĭ" ], [ "ãģķãĤĮ", "ãĤĭ" ], [ "ãģĻãĤĭ", "(00" ], [ "ãģĻãĤĭ", "ãģŁãĤģãĢģ" ], [ "ãģĻãĤĭãĢĤ", "\"\"\"" ], [ "ãĤ", "Ģ" ], [ "ãĤ", "Ħ" ], [ "ãĤ¯", "ãĥ©" ], [ "ãĤ¯ãĥ©", "ãĤ¹" ], [ "ãĤĪ", "ãĤĭ" ], [ "ãĤı", "ãģªãģĦ" ], [ "ãĥ¬ãĥĻ", "ãĥ«" ], [ "ãĥ¼ãĥ¡ãĥ³ãĥĪ", "æİ¨å®ļ" ], [ "ãĥĪãĥª", "ãĥĥãĤ¯" ], [ "ä¸Ńéĸĵ", "å±¤" ], [ "äº", "¤" ], [ "ä¿", "Ŀ" ], [ "ä¿Ĥ", "æķ°" ], [ "å®", "ī" ], [ "å¯¾æķ°", "ç©ºéĸĵ" ], [ "å¼", "µ" ], [ "åħ±", "éĢļ" ], [ "åĪ", "©" ], [ "æ", "´»æĢ§åĮĸ" ], [ "æĸ¹", "åĲĳ" ], [ "æĽ´", "æĸ°" ], [ "ç", "¸" ], [ "ç", "ĭ" ], [ "ç", "Ł" ], [ "ç¯", "ī" ], [ "çĲĨ", "è«ĸ" ], [ "çĶ", "±" ], [ "çŁ", "¥" ], [ "è", "Ń" ], [ "è£", "ľ" ], [ "è¿Ķ", "ãģĻãĢĤ" ], [ "éĢ£", "ç¶ļ" ], [ "éĸĵ", "ãģ§" ], [ "éĻ", "į" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "#" ], [ "ĊĠĠĠĠĠĠĠĠ", "_," ], [ "ĊĠĠĠĠĠĠĠĠnum", "_heads:" ], [ "Ġ", "o" ], [ "Ġ", "param" ], [ "Ġ", "q" ], [ "Ġ", "ãĥĪãĥ¼ãĤ¯ãĥ³" ], [ "Ġ", "åıĤçħ§" ], [ "Ġ", "éĩįãģ¿" ], [ "Ġ0", ":" ], [ "Ġ1", ")" ], [ "ĠL", "ongTensor" ], [ "ĠS_k", ")" ], [ "Ġ\\\\", "\\\\" ], [ "Ġ\\\\", "beta" ], [ "Ġclip", "ping" ], [ "Ġdi", "m" ], [ "Ġi", "," ], [ "Ġlist", "[" ], [ "Ġlog", "its" ], [ "Ġnp.ndarray", "," ], [ "Ġp", "lot" ], [ "Ġplt", ".subplots(" ], [ "Ġs", "et" ], [ "Ġt", "itle" ], [ "Ġãģ¨åĲĮãģĺ", "å½¢çĬ¶" ], [ "Ġãģ®", "åĩºåĬĽ" ], [ "Ġãģ®", "åł´åĲĪãģ¯" ], [ "Ġãĥ¢", "ãĥĩãĥ«" ], [ "Ġå", "Ģĭ" ], [ "ĠåĽ", "³" ], [ "Ġæ", "Ĳįå¤±" ], [ "ĩ", "ãĤ£ãĥ³ãĤ°" ], [ ")ãģ®", "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãĢĤ" ], [ "-", "n" ], [ "-1", "}" ], [ ".0", "," ], [ ".set", "_t" ], [ ".set_t", "itle" ], [ ".set_title", "(title" ], [ "00", "9" ], [ "5", "," ], [ "=-1", ")" ], [ "=True", "," ], [ "B", "Y" ], [ "BY", "T" ], [ "CO", "D" ], [ "E_", "T" ], [ "E_T", "O" ], [ "E_TO", "_" ], [ "GPT", "Language" ], [ "GPTLanguage", "Model" ], [ "I", "COD" ], [ "I", "D" ], [ "L", "ayer" ], [ "Param", "eter" ], [ "Res", "ult" ], [ "S", "core" ], [ "S_", "q," ], [ "Score", "Bias" ], [ "UN", "ICOD" ], [ "[float", "," ], [ "_co", "st" ], [ "_p", "refix" ], [ "_{", "t" ], [ "``", "(" ], [ "``", ")ãĤĴ" ], [ "``", "ãĢģ" ], [ "a", "ch" ], [ "a", "se" ], [ "alpha=0", ".3" ], [ "ar", "ies" ], [ "ar", "quar" ], [ "arquar", "d" ], [ "at", "{" ], [ "b", "erg" ], [ "c", "urrent" ], [ "cod", "ing" ], [ "corp", "u" ], [ "d_ff", ":" ], [ "de", "d_" ], [ "dropout", ")" ], [ "e_", "s" ], [ "ed", "," ], [ "en", "berg" ], [ "en", "d" ], [ "encod", "e_" ], [ "ep", "s" ], [ "et", "h" ], [ "eth", "od" ], [ "f", "allback" ], [ "f", "t" ], [ "fig", "size=(" ], [ "grid", ":" ], [ "h", "at{" ], [ "i", "GLU" ], [ "i", "code" ], [ "ine", "width" ], [ "ize", "``" ], [ "key", ":" ], [ "linalg", "." ], [ "m", "edian" ], [ "model", "_prefix" ], [ "n", "_" ], [ "normalization", "_factory" ], [ "o", "unt" ], [ "of", "t" ], [ "on", "_embedding" ], [ "ound", "aries" ], [ "pe", "ak" ], [ "plot", "_" ], [ "power_law", "_fit" ], [ "qu", "est" ], [ "r", "od" ], [ "ransform", "er" ], [ "repo", "_id" ], [ "ret", "r" ], [ "s", "\"," ], [ "s", "et" ], [ "t", "uple" ], [ "text", ")" ], [ "to", "(" ], [ "uber_", "delta" ], [ "v", "enberg" ], [ "values_by", "_condition" ], [ "w", "iGLU" ], [ "weight", "s" ], [ "ylabel", "(" ], [ "ylabel", ")" ], [ "«", "ĺ" ], [ "³", "ãĥ¼ãĥ" ], [ "º", "«" ], [ "Î", "²" ], [ "ãģ", "¥" ], [ "ãģ«", "ãģ¯" ], [ "ãģ¸", "ãģ®" ], [ "ãģĦ", "ãģĨ" ], [ "ãģĭ", "ãģ©ãģĨ" ], [ "ãģĹãģ¦ãģĦ", "ãĤĭ" ], [ "ãģĹãģªãģĦ", "ãĢĤ" ], [ "ãģĻãĤĭ", "ãģ¨" ], [ "ãĤ¢", "ãĥĥãĥĹ" ], [ "ãĤ¤ãĥ³ãĥĩ", "ãĥĥãĤ¯" ], [ "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯", "ãĤ¹" ], [ "ãĤ¨", "ãĥ³ãĤ" ], [ "ãĤ¨ãĥ³ãĤ", "³ãĥ¼ãĥ" ], [ "ãĤ¯", "ãĥĪ" ], [ "ãĤ¯ãĥĪ", "ãĥ«" ], [ "ãĤ±", "ãĥĥãĥĪ" ], [ "ãĤ·", "ãĥ§ãĥ³" ], [ "ãĤ¹", "ãĤ¿" ], [ "ãĤī", "ãĤĮ" ], [ "ãĤĴ", "åıĹãģĳ" ], [ "ãĥ", "¯" ], [ "ãĥ¦ãĥĭ", "ãĥĥãĥĪ" ], [ "ãĥªãĥ", "Ŀ" ], [ "ãĥªãĥĿ", "ãĤ¸" ], [ "ãĥĨãĤŃãĤ¹ãĥĪ", "åħ¨ä½ĵ" ], [ "ãĥĸ", "ãĥŃ" ], [ "ãĥĻ", "ãĤ¯ãĥĪãĥ«" ], [ "ä¸Ĭ", "éĻĲ" ], [ "äºĪ", "ç®Ĺ" ], [ "ä¾ĭ", ":" ], [ "å®ī", "å®ļ" ], [ "å°ıãģķ", "ãģĦ" ], [ "å¿ħè¦ģãģĮãģĤ", "ãĤĭãĢĤ" ], [ "åħ±", "æľī" ], [ "åħĪ", "ãģ®" ], [ "åĪ", "¶" ], [ "åĪ©", "çĶ¨" ], [ "åĪĨåī²", "çµĲæŀľ" ], [ "åŁº", "ãģ¥" ], [ "æ¸¬", "å®ļ" ], [ "æĮ", "Ļ" ], [ "æı", "ĥ" ], [ "æııçĶ»", "åħĪãģ®" ], [ "æķ", "´" ], [ "æĸ°è¦ı", "ä½ľæĪĲãģĻãĤĭãĢĤ" ], [ "æľĢ", "éģ©" ], [ "æĿ", "Ł" ], [ "ç", "ªĵ" ], [ "ç", "ľ" ], [ "ç³»åĪĹ", "éķ·" ], [ "ç´", "Ħ" ], [ "çµ", "¶" ], [ "çµ", "Į" ], [ "è", "º«" ], [ "è¾ŀæĽ¸", "ãĢĤ" ], [ "é", "«ĺ" ], [ "éĢ£çµĲ", "ãģĹãģŁ" ], [ "éłĨ", "ãģ«" ], [ "ĊĠĠĠĠ", "(``" ], [ "ĊĠĠĠĠax", ":" ], [ "ĊĠĠĠĠĠĠĠĠ", "d_ff:" ], [ "ĊĠĠĠĠĠĠĠĠ", "vocab" ], [ "ĊĠĠĠĠĠĠĠĠax", ":" ], [ "Ġ", "...," ], [ "Ġ", "M" ], [ "Ġ", "id" ], [ "Ġ", "load_" ], [ "Ġ", "step" ], [ "Ġ", "values" ], [ "Ġ", "ãģ¾" ], [ "Ġ", "ãĤ¹ãĤ³ãĤ¢" ], [ "Ġ", "æ¨ª" ], [ "Ġ", "æııçĶ»åħĪãģ®" ], [ "Ġ:", "math:`" ], [ "ĠAx", "es:" ], [ "ĠCallable[", "[" ], [ "ĠS_k", "," ], [ "ĠTrue", "," ], [ "Ġco", "lor" ], [ "Ġd", "elta" ], [ "Ġd_k", ")" ], [ "Ġint", ")" ], [ "Ġnn.", "Linear" ], [ "Ġplt.subplots(", "figsize=(" ], [ "Ġre", ".sub" ], [ "Ġs", "qrt(" ], [ "Ġself.", "dropout" ], [ "ĠãģªãĤī", "æĸ°è¦ıä½ľæĪĲãģĻãĤĭãĢĤ" ], [ "Ġãģ«", "ãĤĪãĤĭ" ], [ "Ġãģ®", "ãģ¨ãģį" ], [ "ĠåĽ³", "ãģ®ãĤ¿ãĤ¤ãĥĪãĥ«ãĢĤ" ], [ "Ĩ", "ħ" ], [ "\"``", ":" ], [ "(", "alpha=0.3" ], [ "(nn.Module", "):" ], [ ")", "^" ], [ ")ãĢĤ", "\"\"\"" ], [ "-", "Head" ], [ "-", "byt" ], [ "-", "per" ], [ "-per", "-byt" ], [ ".", "ad" ], [ ".", "py" ], [ ".s", "hape" ], [ ".set_title(title", ")" ], [ "0", ")" ], [ "00", "3" ], [ "=", "float" ], [ "B", "i" ], [ "C", "H" ], [ "En", "coding" ], [ "Multi", "-Head" ], [ "O", "p" ], [ "Op", "tim" ], [ "RMS", "Norm" ], [ "Tensor", "," ], [ "[", "i" ], [ "\\\\", "hat{" ], [ "_", "an" ], [ "_", "lr" ], [ "_corp", "us" ], [ "_head", "s)" ], [ "_n", "ew" ], [ "_p", "en" ], [ "_pen", "al" ], [ "_penal", "ty" ], [ "_{t", "-1}" ], [ "ab", "c" ], [ "act", "erL" ], [ "acterL", "e" ], [ "acterLe", "vel" ], [ "acterLevel", "Tokenizer" ], [ "ak", "e_" ], [ "ar", "acterLevelTokenizer" ], [ "ar", "ch" ], [ "ar", "ti" ], [ "arti", "cl" ], [ "ath:", ":" ], [ "ce", "pt" ], [ "co", "unt" ], [ "con", "verged" ], [ "core_b", "ias" ], [ "emperat", "ure" ], [ "est_p", "air" ], [ "ff", "er" ], [ "g", "e" ], [ "gradient_clip", "_threshold" ], [ "grid", "(alpha=0.3" ], [ "h", "aracterLevelTokenizer" ], [ "h", "ead" ], [ "h", "o" ], [ "i", "Bi" ], [ "id", "den" ], [ "ipy", "n" ], [ "its", "-per-byt" ], [ "l", "le" ], [ "label", ")" ], [ "label", "=" ], [ "lan", "guage" ], [ "length", "_penalty" ], [ "lle", "ction" ], [ "log", "_c" ], [ "m", "ethod" ], [ "mask", ":" ], [ "model", ":" ], [ "p.", "grad" ], [ "path", "." ], [ "po", "int" ], [ "r", "m" ], [ "ro", "w" ], [ "s", "c" ], [ "s", "um" ], [ "to", "_" ], [ "token", "_embedding" ], [ "train_", "language_model" ], [ "val", "id" ], [ "w", "h" ], [ "©", "Ł" ], [ "¬", "ç«ĭ" ], [ "®", "å°ı" ], [ "¯Ķ", "è¼ĥ" ], [ "ãģª", "ãģĹ" ], [ "ãģ«", "æ" ], [ "ãģ«ãģª", "ãĤĭãĢĤ" ], [ "ãģ«ä½¿", "ãģ£ãģŁ" ], [ "ãģ®åĩ", "º" ], [ "ãģ¹ãģįä¹Ĺ", "åīĩ" ], [ "ãģ¾ãģŁ", "ãģ¯" ], [ "ãģĦãģļãĤĮ", "ãĤĤ" ], [ "ãģĵãģ¨", "ãģĮ" ], [ "ãģĹ", "ãģ¾" ], [ "ãĤµ", "ãĥ³ãĥĹ" ], [ "ãĤĪ", "ãĤĬ" ], [ "ãĤı", "ãģĽ" ], [ "ãĤĴ", "æĮģãģŁãģªãģĦ" ], [ "ãĥªãĥĿãĤ¸", "ãĥĪãĥª" ], [ "ãĥ¼ãĥĪ", "ãĥĸ" ], [ "ãĥ¼ãĥĪãĥĸ", "ãĥĥãĤ¯" ], [ "ãĥĥãĥ", "Ķ" ], [ "ãĥĥãĥĶ", "ãĥ³ãĤ°" ], [ "ãĥĪãĥªãĥĥãĤ¯", "ãģĤãģ¦ãģ¯ãĤģ" ], [ "ãĥİ", "ãĥ¼ãĥĪãĥĸãĥĥãĤ¯" ], [ "ä¸", "»" ], [ "ä¸Ĭ", "ä½į" ], [ "å¤§ãģį", "ãģĦ" ], [ "å®ŁéļĽ", "ãģ«" ], [ "å°", "İ" ], [ "åģ", "ı" ], [ "åģı", "å·®" ], [ "åĨį", "åĪ©çĶ¨" ], [ "åĩ", "¦çĲĨ" ], [ "åĪ", "¤" ], [ "åĪĨ", "æķ£" ], [ "åįĺ", "ä½į" ], [ "åĲĪ", "ãĤıãģĽ" ], [ "åŁºãģ¥", "ãģı" ], [ "åŃĺ", "åľ¨" ], [ "æ", "ĭ" ], [ "æ³¨", "åħ¥" ], [ "æħ", "ĭ" ], [ "æī±", "ãģĨ" ], [ "æĮĻ", "åĭķ" ], [ "æııçĶ»", "ãģ«ä½¿ãģ£ãģŁ" ], [ "æķ", "Ĺ" ], [ "æķ°", "ãģĪ" ], [ "æľĢå¤§", "åĢ¤" ], [ "æľĢå°ı", "äºĮä¹Ĺ" ], [ "æŃ£è¦ıåĮĸ", "å±¤" ], [ "æŃ£è¦ıåĮĸ", "åīį" ], [ "ç", "¨" ], [ "ç´", "°" ], [ "çµ", "±" ], [ "çµ¶", "å¯¾ä½įç½®" ], [ "çĬ¶", "æħĭ" ], [ "çĭ", "¬ç«ĭ" ], [ "çķ", "¥" ], [ "çľ", "ģ" ], [ "è¡¨", "çı¾" ], [ "è¦", "ĸ" ], [ "èªŀå½Ļ", "ãģ«" ], [ "è¶ħ", "ãģĪãĤĭ" ], [ "éģ", "Ķ" ], [ "Ċ", "import" ], [ "Ċ", "ĠĠ" ], [ "ĊĊĊ", "@" ], [ "ĊĊĠĠĠĠ", ".." ], [ "ĊĠĠĠĠ", "h" ], [ "ĊĠĠĠĠax.", "grid(alpha=0.3" ], [ "ĊĠĠĠĠax.", "leg" ], [ "ĊĠĠĠĠax.set_", "ylabel(" ], [ "ĊĠĠĠĠĠĠĠĠ", "beta" ], [ "ĊĠĠĠĠĠĠĠĠ", "æııçĶ»ãģ«ä½¿ãģ£ãģŁ" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "return" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "ĠĠĠĠ" ], [ "Ġ", "J" ], [ "Ġ", "as" ], [ "Ġ", "decay" ], [ "Ġ", "seq_len" ], [ "Ġ", "èĩª" ], [ "Ġ(", "B," ], [ "Ġ0.0", "," ], [ "Ġ1", ":" ], [ "Ġ20", "20" ], [ "Ġ:", "]" ], [ "ĠAxesãĢĤ", "None" ], [ "ĠH", "uber" ], [ "Ġco", "llection" ], [ "Ġd_model", ")" ], [ "Ġdim", "=-1)" ], [ "Ġfloat", ")" ], [ "Ġfor", "ward(" ], [ "Ġl", "r" ], [ "Ġlist[str", "]" ], [ "Ġlog", "_" ], [ "Ġlog", "_alpha" ], [ "Ġm", "ath::" ], [ "Ġm", "emory" ], [ "Ġre.sub", "(r\"" ], [ "Ġtorch.", "arange(" ], [ "Ġx", "," ], [ "Ġ{", "}" ], [ "Ġãģ®", "å®Łé¨ĵ" ], [ "Ġãģ®", "è¾ŀæĽ¸ãĢĤ" ], [ "ĠãĤĴ", "ä½¿ãģĨ" ], [ "Ġç", "³»åĪĹ" ], [ "Ģ", "Ļ" ], [ "Ħ", "å½±" ], [ "(", "se" ], [ "().", "item" ], [ "(00", "5" ], [ "(``", "src/" ], [ "(param", "s)" ], [ "(x", "label)" ], [ ".", "encod" ], [ ".", "jo" ], [ ".", "mean" ], [ ".", "plot" ], [ ".", "t" ], [ "9", "9" ], [ "=-1", "," ], [ "=False", "``" ], [ "AL", "iBi" ], [ "Attention", "ScoreBias" ], [ "BPE", "Tokenizer" ], [ "CL", "R" ], [ "D", "ropout" ], [ "ID", "Tokenizer" ], [ "Po", "int" ], [ "R", "L" ], [ "S", "haw" ], [ "T", "5" ], [ "V", "alue" ], [ "[", "1" ], [ "[int", "]" ], [ "_b", "o" ], [ "_bu", "ffer" ], [ "_bucket", "s" ], [ "_c", "andidates" ], [ "_f", "req" ], [ "_new", "_token" ], [ "_s", "quar" ], [ "_scal", "ing" ], [ "_train", "_val" ], [ "a", "c" ], [ "ake_", "evaluation_window" ], [ "al", "_f" ], [ "al_", "values" ], [ "at", "s" ], [ "att", "emp" ], [ "attemp", "t" ], [ "b", "``" ], [ "b", "ar" ], [ "b", "est" ], [ "b", "its_per_byt" ], [ "c", "re" ], [ "ca", "us" ], [ "caus", "al" ], [ "chinchilla", "_param" ], [ "color", "_cycle" ], [ "corp", "us" ], [ "cover", "ag" ], [ "d_k", ")" ], [ "data", "/" ], [ "data", "class" ], [ "di", "stance" ], [ "e", "\"" ], [ "e", "arn" ], [ "e(", "self," ], [ "e(\"", "utf-8" ], [ "el", "i" ], [ "eli", "ho" ], [ "eliho", "od" ], [ "en", "er" ], [ "ep", "s:" ], [ "er", "til" ], [ "eri", "or" ], [ "ertil", "ity" ], [ "ex", "ist" ], [ "f", "eed_forward_factory" ], [ "f", "ontsize=" ], [ "g", "enerator" ], [ "gro", "up" ], [ "h", "idden" ], [ "ht", "t" ], [ "i", "ct" ], [ "i", "rection" ], [ "id", "irection" ], [ "idirection", "al" ], [ "ik", "ipedia" ], [ "in", "di" ], [ "int", "erior" ], [ "la", "y" ], [ "label", ":" ], [ "le", "ft" ], [ "los", "s:" ], [ "m", "a" ], [ "max", "_new_token" ], [ "max_chunk", "_bytes:" ], [ "merg", "es" ], [ "n", "an" ], [ "n", "ew" ], [ "n", "o" ], [ "norm_first", ":" ], [ "np.", "log(" ], [ "o", "ver" ], [ "oisy_", "l" ], [ "on", "es" ], [ "optim", "." ], [ "optim", "al_f" ], [ "or", "d(" ], [ "p", "retoken" ], [ "pair", "]" ], [ "par", "ent" ], [ "plit", "_train_val" ], [ "pro", "fil" ], [ "qu", "er" ], [ "r", "n" ], [ "ra", "c" ], [ "s", "qrt(" ], [ "s.", "abc" ], [ "symbol", "_to_id" ], [ "t", "arget" ], [ "ti", "ck" ], [ "un", "it_" ], [ "url", "lib" ], [ "valu", "e" ], [ "verg", "ence" ], [ "xlabel", "(xlabel)" ], [ "xlabel", ":" ], [ "z", "ip" ], [ "}", "\")" ], [ "¬", "¬" ], [ "¯", "ç©į" ], [ "ãĢģ", "æľ¬" ], [ "ãģ", "Ń" ], [ "ãģ¤", "ãģĦãģ¦" ], [ "ãģ§", "ãĤĤ" ], [ "ãģ§ãģ¯", "ãģªãģı" ], [ "ãģ¨", "ãģ¿" ], [ "ãģ«", "ãĢģ" ], [ "ãģ«", "ä¾ĿåŃĺ" ], [ "ãģ«", "å¤īæıĽ" ], [ "ãģ«", "éģ©çĶ¨ãģĻãĤĭ" ], [ "ãģ«ãĤĪ", "ãĤĬ" ], [ "ãģ®", "æĮĩæķ°" ], [ "ãģ®", "éķ·ãģķ" ], [ "ãģ®å", "°" ], [ "ãģ®å", "¾" ], [ "ãģ®åĩº", "åĬĽ" ], [ "ãģ®åĭ¾éħį", "ãĥİãĥ«ãĥł" ], [ "ãģ¹", "ãģ¦" ], [ "ãģĦ", "ãĢģ" ], [ "ãģĦ", "ãģŁ" ], [ "ãģĭ", "ãĢĤ" ], [ "ãģĹ", "ãģª" ], [ "ãģĽ", "ãģļ" ], [ "ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥ", "ĩãĤ£ãĥ³ãĤ°" ], [ "ãĤ¯", "ãĤ·ãĥ§ãĥ³" ], [ "ãĤ¯ãĥª", "ãĥĥãĥĶãĥ³ãĤ°" ], [ "ãĤµãĥ³ãĥĹ", "ãĥªãĥ³ãĤ°" ], [ "ãĤ¹ãĤ±", "ãĤ¸ãĥ¥ãĥ¼ãĥ«" ], [ "ãĤ»", "ãĤ¯ãĤ·ãĥ§ãĥ³" ], [ "ãĤĴ", "åıĤçħ§)ãĢĤ" ], [ "ãĤĴ", "åıĸå¾Ĺ" ], [ "ãĤĴ", "æĮģãģ¤" ], [ "ãĥ", "¡" ], [ "ãĥ", "¤" ], [ "ãĥ", "Ķ" ], [ "ãĥ¢", "ãĤ¸ãĥ¥ãĥ¼ãĥ«" ], [ "ãĥ¼ãĥ", "Ĳ" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "ãĤĴ" ], [ "ãĥİ", "ãĤ¤ãĤº" ], [ "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "æķ°" ], [ "äº¤", "å·®" ], [ "äº¤å·®", "æ³¨æĦı" ], [ "ä»ĺ", "ãģį" ], [ "å", "±" ], [ "å¢", "ĥçķĮ" ], [ "å¤", "ļ" ], [ "å®", "¹" ], [ "å®Įåħ¨", "ãģ«" ], [ "å®Ł", "è¡Į" ], [ "å¯¾æķ°", "ãĤ¹ãĤ±ãĥ¼ãĥ«" ], [ "å¯¾è", "·ĿéĽ¢" ], [ "åº", "ı" ], [ "åĩº", "çı¾" ], [ "åĪ¤", "å®ļ" ], [ "åį", "Ĭ" ], [ "åįĺ", "ä¸Ģ" ], [ "åı", "İ" ], [ "åĲĮä¸Ģãģ®", "æĮĻåĭķ" ], [ "åĲį", ":" ], [ "åĳ", "³" ], [ "åĳ¼ãģ³åĩºãģĹ", "åģ´" ], [ "æ", "©Ł" ], [ "æ¬¡åħĥ", "(" ], [ "æ´»æĢ§åĮĸ", "éĸ¢æķ°" ], [ "æııçĶ»", "ãģĻãĤĭãĢĤ" ], [ "æĶ¹", "è¡Į" ], [ "æķ´", "æķ°" ], [ "æĸ", "Ń" ], [ "æľª", "æºĢ" ], [ "æľª", "çŁ¥" ], [ "ç½®", "(" ], [ "çĶ¨", "ãģ®" ], [ "è¨Īç®Ĺéĩı", "äºĪç®Ĺ" ], [ "èĩªå·±", "æ³¨æĦı" ], [ "éĢļ", "ãĤĬ" ], [ "éł", "ĤçĤ¹" ], [ "éłĨ", "åºı" ], [ "Ċ", "``" ], [ "Ċ", "class" ], [ "ĊĊĊ@", "dataclass" ], [ "ĊĊĠĠĠĠ", "ax" ], [ "ĊĠĠĠĠ", "num" ], [ "ĊĠĠĠĠ", "xlabel:" ], [ "ĊĠĠĠĠ", "ylabel:" ], [ "ĊĠĠĠĠax.leg", "end(" ], [ "ĊĠĠĠĠax.set_ylabel(", "ylabel)" ], [ "ĊĠĠĠĠĠĠĠĠ", "ylabel:" ], [ "ĊĠĠĠĠĠĠĠĠ", "ĠĠ" ], [ "ĊĠĠĠĠĠĠĠĠb", "ia" ], [ "ĊĠĠĠĠĠĠĠĠx", "label:" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "raise" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", ")" ], [ "Ġ", "E" ], [ "Ġ", "G" ], [ "Ġ", "relative_" ], [ "Ġ", "se" ], [ "Ġ", "start" ], [ "Ġ", "warmup_step" ], [ "Ġ", "ãĤŃãĥ£ãĥĥãĤ·ãĥ¥" ], [ "Ġ20", "2" ], [ "ĠC", "hinchilla" ], [ "ĠFeed", "ForwardNetwork" ], [ "ĠLayerNorm", "(x" ], [ "Ġb", "yt" ], [ "Ġco", "unt" ], [ "Ġcollection", "s.abc" ], [ "Ġcos", "ine" ], [ "Ġd", "ata" ], [ "Ġf", "req" ], [ "Ġfloat", "," ], [ "Ġlist[float", "]" ], [ "Ġlist[str", "]:" ], [ "Ġlog", "_n" ], [ "Ġnn.", "Dropout" ], [ "Ġnp.as", "array(" ], [ "Ġplt.subplots(figsize=(", "6" ], [ "Ġstr", "ict" ], [ "Ġt", "arget" ], [ "Ġtext", ")" ], [ "Ġtorch.", "zero" ], [ "Ġtup", "le" ], [ "Ġtuple[", "Tensor," ], [ "Ġx", ":" ], [ "Ġãģ®", "ãģ¿" ], [ "Ġãĥ", "ķ" ], [ "Ġãĥ¢ãĥĩãĥ«", "ãģ®éļłãĤĮ" ], [ "ĠåĽ", "ŀ" ], [ "Ġæ¨ª", "è»¸ãĥ©ãĥĻãĥ«ãĢĤ" ], [ "Ġç", "Ľ" ], [ "Ġç³»åĪĹ", "éķ·" ], [ "Ġç¸¦", "è»¸ãĥ©ãĥĻãĥ«ãĢĤ" ], [ "ĠĠĠĠ", "Ġ:" ], [ "ĢĻ", "è£ľ" ], [ "ĸ", "Ħ" ], [ "\"", "))" ], [ "\"]", ".append(" ], [ "(", "2" ], [ "(", "2)" ], [ "(", "Y" ], [ "(", "dropout)" ], [ "(", "g" ], [ "(", "length" ], [ "(x", "," ], [ "-", "2" ], [ "-", "inf" ], [ "-", "onl" ], [ "-in", "dex" ], [ "-onl", "y" ], [ ".", "to(" ], [ ".jo", "in(" ], [ "/", "/" ], [ "00", "2" ], [ "00", "7" ], [ "3", "," ], [ "6", "," ], [ "8", "," ], [ "=", "color" ], [ "=False", ")" ], [ "D", "I" ], [ "I", "P" ], [ "I", "soFLOP" ], [ "IP", "S" ], [ "Optim", "izer" ], [ "[", "key" ], [ "_", "1" ], [ "_", "re" ], [ "_co", "mpute_" ], [ "_head", "s," ], [ "_m", "n" ], [ "_s", "eries" ], [ "a", "_body_" ], [ "anif", "est" ], [ "arquard", "t" ], [ "at", "e_b" ], [ "ate_b", "its_per_byt" ], [ "attention", "_s" ], [ "aturating", "PowerLawFit" ], [ "ax", "es" ], [ "b", "j" ], [ "b", "s" ], [ "bi", "an" ], [ "c", "urv" ], [ "cache_dir", ":" ], [ "ch", "ed" ], [ "character_", "coverag" ], [ "chinchilla_param", "etric" ], [ "clip", "(log" ], [ "co", "bian" ], [ "coder", "-only" ], [ "con", "str" ], [ "constr", "uct_" ], [ "construct_", "optimal_f" ], [ "cor", "rection" ], [ "d", "er" ], [ "de", "coder-only" ], [ "di", "vergence" ], [ "do", "m" ], [ "e", "ad_" ], [ "e\"", "):" ], [ "ect", "or" ], [ "el", "u" ], [ "em", "or" ], [ "emor", "y_" ], [ "er", "r" ], [ "err", "stat" ], [ "evalu", "ate_bits_per_byt" ], [ "evaluation", "_mask" ], [ "ext", "ra" ], [ "f", "rontier_" ], [ "fontsize=", "8)" ], [ "gram", "s" ], [ "history", "[\"" ], [ "ias", ")" ], [ "il", "ter" ], [ "in", "put" ], [ "is_", "interior" ], [ "lay", "er_" ], [ "log", "its" ], [ "m", "ath" ], [ "m", "o" ], [ "max_", "len" ], [ "nam", "e" ], [ "nam", "e," ], [ "o", "ff" ], [ "o", "u" ], [ "oft", "max" ], [ "ot", "ary" ], [ "out_leading_whitespace", "\":" ], [ "pair", "_in" ], [ "param", "eter" ], [ "pro", "ach" ], [ "pro", "cessor" ], [ "ran", "dom" ], [ "rid", "Point" ], [ "ro", "ss" ], [ "s_", "correction" ], [ "sum", "(" ], [ "theories/", "0" ], [ "token", "izer_json" ], [ "tokenizer.", "json" ], [ "torch.", "optim." ], [ "ud", "get" ], [ "ul", "_" ], [ "v", "id" ], [ "w", "hitespace" ], [ "wikipedia", "_corpus" ], [ "x", ":" ], [ "¢", "«" ], [ "¢«", "è¦" ], [ "¢«è¦", "Ĩ" ], [ "¥", "æľ¬" ], [ "±Ĥ", "ãĤģ" ], [ "²", "åĲĪ" ], [ "´", "ãĥª" ], [ "»", "åĭķ" ], [ "ãĢ", "ľ" ], [ "ãģ¨", "ãģĦãģĨ" ], [ "ãģ¨ãģ¿", "ãģª" ], [ "ãģªãģĦ", "ãĢĤ" ], [ "ãģ«", "ãģĻãĤĭ" ], [ "ãģ®", "å®Ł" ], [ "ãģ®å", "·" ], [ "ãģ®å¯¾", "æķ°" ], [ "ãģ®åī", "²åĲĪ" ], [ "ãģ®ç³»åĪĹ", "éķ·" ], [ "ãģĤ", "ãĤĭ" ], [ "ãģĭãģ©ãģĨ", "ãģĭ" ], [ "ãģķãģĽ", "ãĤĭ" ], [ "ãģĿãģ®", "ãĤĤãģ®" ], [ "ãĤ", "´ãĥª" ], [ "ãĤ", "ĵ" ], [ "ãĤ¦", "ãĥ³ãĥ" ], [ "ãĤ«", "ãĥ©" ], [ "ãĤµãĥĸ", "ãĤ¯ãĥ©ãĤ¹" ], [ "ãĤ¹ãĤ±", "ãĥ¼" ], [ "ãĤ¹ãĤ±ãĥ¼", "ãĥªãĥ³ãĤ°" ], [ "ãĤ¿", "ãĤ¤" ], [ "ãĤĴ", "ä½ľ" ], [ "ãĤĴ", "åĲ«" ], [ "ãĤĴ", "è¡¨" ], [ "ãĥ", "į" ], [ "ãĥ¯", "ãĥ¼ãĤ¯" ], [ "ãĥĥãĥĪ", "ãĥ¯ãĥ¼ãĤ¯" ], [ "ãĥĨãĥ³ãĤ½", "ãĥ«" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "æķ°" ], [ "ãĥį", "ãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯" ], [ "ãĥĲ", "ãĤ±ãĥĥãĥĪ" ], [ "ä½¿", "ãĤı" ], [ "å", "ł" ], [ "å¤§ãģį", "ãģı" ], [ "å¤±", "æķĹ" ], [ "å¤ĸ", "éĥ¨" ], [ "å®ļ", "æķ°" ], [ "å°¤", "åº¦" ], [ "å¾Į", "ãģ«" ], [ "å¾ĵ", "æĿ¥" ], [ "åı¯èĥ½", "ãģª" ], [ "åıİ", "æĿŁ" ], [ "åıĸ", "ãĤĬ" ], [ "åł", "±" ], [ "æ", "ĥ³å®ļ" ], [ "æ§ĭ", "ç¯ī" ], [ "æ¯Ķè¼ĥ", "ãģĻãĤĭ" ], [ "æī", "ĭ" ], [ "æĮĩ", "æ¨Ļ" ], [ "æı", "Ĳ" ], [ "æĽ", "¿" ], [ "æľĢ", "å°¤" ], [ "ç§", "»åĭķ" ], [ "ç´", "¯ç©į" ], [ "çµ±", "è¨Ī" ], [ "çµĦãģ¿", "åĲĪãĤıãģĽ" ], [ "ç·", "ı" ], [ "çĲĨè«ĸ", "ãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³" ], [ "çĶŁæĪĲ", "ãģĻãĤĭ" ], [ "çĽ¸", "å¯¾ä½įç½®" ], [ "çĽ¸", "å¯¾è·ĿéĽ¢" ], [ "çľģ", "çķ¥" ], [ "è", "¢«è¦Ĩ" ], [ "è¦", "ĭ" ], [ "è¦ı", "åīĩ" ], [ "è¨Ģèªŀãĥ¢ãĥĩãĥ«", "ãģ®" ], [ "è¨Ń", "å®ļ" ], [ "è¶ħ", "ãģĪ" ], [ "é", "Ł" ], [ "éģ", "İ" ], [ "éĻ¤", "å¤ĸ" ], [ "éŁ", "¿" ], [ "Ċ", "(" ], [ "ĊĊĊ", "#" ], [ "ĊĊĠĠĠĠ", "for" ], [ "ĊĊĠĠĠĠ", "return" ], [ "ĊĠĠĠĠ", "#" ], [ "ĊĠĠĠĠ", "d_model" ], [ "ĊĠĠĠĠax.legend(", "fontsize=8)" ], [ "ĊĠĠĠĠĠĠĠĠ", "alpha" ], [ "ĊĠĠĠĠĠĠĠĠ", "device:" ], [ "ĊĠĠĠĠĠĠĠĠax", ".set_title(title)" ], [ "ĊĠĠĠĠĠĠĠĠself.", "dropout" ], [ "ĊĠĠĠĠĠĠĠĠx", ":" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "for" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "m" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠx", ":" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "``" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "continue" ], [ "Ġ", "%" ], [ "Ġ", "5" ], [ "Ġ", "residual_fn" ], [ "Ġ", "top" ], [ "Ġ", "w" ], [ "Ġ", "ãĥ©" ], [ "Ġ", "æŃ£" ], [ "Ġ", "èªŀå½ĻãĤµãĤ¤ãĤº" ], [ "Ġ", "è»¸" ], [ "Ġ1", "00" ], [ "Ġ201", "6" ], [ "ĠNeur", "IPS" ], [ "ĠSequence", "[str" ], [ "ĠT", "ransformer" ], [ "ĠTensor", "]" ], [ "ĠTensor", "]:" ], [ "ĠValueError(", "f\"" ], [ "Ġ__init__(", "self," ], [ "Ġc", "on" ], [ "Ġforward(", "self," ], [ "Ġh", "istory" ], [ "Ġl", "inewidth" ], [ "Ġm", "atrix" ], [ "Ġmemory", "," ], [ "Ġn", "_" ], [ "Ġn", "n" ], [ "Ġnn.Dropout", "(dropout)" ], [ "Ġnn.Linear", "(d_model," ], [ "Ġnn.Module", "]" ], [ "Ġnp.ndarray", ":" ], [ "Ġplot", "_" ], [ "Ġs", "entencepiece" ], [ "Ġself.", "norm" ], [ "Ġtitle", ":" ], [ "Ġtorch.", "device" ], [ "ĠÎ", "»" ], [ "Ġãģ§", "ãģ¯" ], [ "Ġãģ¨", "åĲĮä¸Ģãģ®æĮĻåĭķ" ], [ "Ġãģ®", "ãĤµãĥĸãĤ¯ãĥ©ãĤ¹" ], [ "Ġãģ®", "çµĲæŀľ" ], [ "ĠãĤĴ", "æĮģãģ¤" ], [ "Ġãĥĳãĥ©ãĥ¡", "ãĥ¼ãĤ¿" ], [ "Ġä»¥", "ä¸ĭ" ], [ "Ġå", "¼ķæķ°" ], [ "ĠĠ", ":" ], [ "ĠĠĠĠ", "ĠĠĠ:" ], [ "ĠĠĠĠĠĠĠĠ", "Ġ:" ], [ "ĥ", "ħ" ], [ "ĥħ", "åł±" ], [ "(", "G" ], [ "(", "R" ], [ "(", "a" ], [ "(", "chunk" ], [ "(", "dim" ], [ "(", "path" ], [ "(", "ä¾ĭ:" ], [ "(n", "ew" ], [ "(np.", "exp(log" ], [ "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪ", "ãģ¯" ], [ "-index", "ed" ], [ "-inf", "``" ], [ ".", "json" ], [ ".", "weight" ], [ ".2", "))" ], [ "0", "." ], [ "2", "9" ], [ "=\"ignor", "e\"):" ], [ "=\"ignor", "e\"," ], [ "=0", ".0" ], [ "=b", "ias)" ], [ "E", "LU" ], [ "Encoder", "Block" ], [ "Error", ":" ], [ "G", "ridPoint" ], [ "IN", "," ], [ "K", "_" ], [ "L", "ayerNorm" ], [ "N", "," ], [ "P", "iece" ], [ "Par", "abola" ], [ "Parabola", "Fit" ], [ "Param", "etric" ], [ "Parametric", "Fit" ], [ "Position", "Embedding" ], [ "Query", "KeyPositionalTransform" ], [ "\\\\", "text{" ], [ "^", "2" ], [ "^K", "_mn" ], [ "_", "stat" ], [ "_", "v" ], [ "__", "all" ], [ "__all", "__" ], [ "_bu", "il" ], [ "_buil", "d_" ], [ "_co", "l" ], [ "_compute_", "b" ], [ "_compute_b", "udget" ], [ "_lik", "e(" ], [ "_lik", "elihood" ], [ "_likelihood", "_n" ], [ "_n", "egative_" ], [ "_negative_", "log" ], [ "_negative_log", "_likelihood_n" ], [ "_rat", "io" ], [ "_se", "gment" ], [ "_w", "ith" ], [ "``", ")ãĢĤ" ], [ "``", "ãĢģ``" ], [ "ad_", "text(" ], [ "ad_text(", "encoding=\"utf-8" ], [ "atch", "_size," ], [ "att", "er" ], [ "b", "pe_" ], [ "beam", "_size" ], [ "byte_", "fallback" ], [ "c", "map" ], [ "c", "str" ], [ "causal", "_mask" ], [ "ched", "ule" ], [ "chunk", "_" ], [ "chunk_split_mod", "e:" ], [ "co", "st" ], [ "color", "=\"" ], [ "construct_optimal_f", "rontier" ], [ "cstr", "ing" ], [ "d_model", "``" ], [ "def", "a" ], [ "dtype", "," ], [ "e(", "over" ], [ "e(over", "=\"ignore\"," ], [ "e_b", "and" ], [ "e_s", "chedule" ], [ "edi", "ct" ], [ "eed_forward_factory", ":" ], [ "ener", "ation" ], [ "entence", "Piece" ], [ "ep", "s)" ], [ "errstat", "e(over=\"ignore\"," ], [ "ex", "cl" ], [ "f", "init" ], [ "f", "rac" ], [ "f", "ul" ], [ "finit", "e(" ], [ "float", "3" ], [ "frac", "{" ], [ "g", "res" ], [ "gt", "_mask" ], [ "h", "," ], [ "htt", "p" ], [ "i", "ent" ], [ "ic", "ient" ], [ "in", "stance" ], [ "indi", "ces" ], [ "ine", "d_" ], [ "init", "i" ], [ "is", "finite(" ], [ "is", "of" ], [ "isof", "lop" ], [ "iterbi", "_segment" ], [ "l", "_inf" ], [ "l", "p" ], [ "label", "s" ], [ "lambda", "_" ], [ "le", "x" ], [ "learn", "_b" ], [ "learning_rat", "e_schedule" ], [ "lex", "ity" ], [ "m", "ake_evaluation_window" ], [ "max_", "position" ], [ "max_", "sequence_length" ], [ "normalization_factory", ":" ], [ "o", "k" ], [ "ois", "e_band" ], [ "on_embedding", "_param" ], [ "ot", "at" ], [ "otary", "PositionEmbedding" ], [ "ount", "er" ], [ "p", "lexity" ], [ "par", "se" ], [ "parabola", "." ], [ "parameter", "s()" ], [ "params", "0" ], [ "pe", "." ], [ "peak", "_learning_rat" ], [ "pretoken", "ize``" ], [ "pro", "p" ], [ "qr", "t" ], [ "qu", "e" ], [ "query", ":" ], [ "r", "edict" ], [ "ra", "ction" ], [ "rain", "ing" ], [ "ran", "k" ], [ "ro", "t" ], [ "s", "entencepiece" ], [ "s", "plit_train_val" ], [ "s[", "i" ], [ "self", ")" ], [ "sequences", "_a" ], [ "st", "d" ], [ "theta", "_i" ], [ "token_id", "s)" ], [ "total", "_negative_log_likelihood_n" ], [ "tr", "y:" ], [ "train_language_model", "``" ], [ "tuple", "[str," ], [ "u", "cket" ], [ "v", "e_" ], [ "v", "ector" ], [ "v", "max" ], [ "valid", "=\"ignore\"):" ], [ "vocab", "_size" ], [ "weight", "s:" ], [ "wh", "ile" ], [ "wor", "d" ], [ "x", "tick" ], [ "y", "s" ], [ "¥æľ¬", "èªŀ" ], [ "§", "ãĥ¼ãĤ¹" ], [ "©", "ãĥ¼ãĥ«" ], [ "©ãĥ¼ãĥ«", "ãĥĲ" ], [ "©ãĥ¼ãĥ«ãĥĲ", "ãĥĥãĤ¯" ], [ "«", "å°" ], [ "«å°", "¾" ], [ "®", "µ" ], [ "±", "æķ°" ], [ "¾", "åĢ¤" ], [ "Ã", "Ĺ" ], [ "Î", "¸" ], [ "Ï", "ĥ" ], [ "ãĢģ", "n" ], [ "ãģ", "Ĳ" ], [ "ãģ§ãģ¯", "ãĢģ" ], [ "ãģ§ãģĤ", "ãĤĭãĢĤ" ], [ "ãģ§ãģĤãĤĭ", "å¿ħè¦ģãģĮãģĤãĤĭ" ], [ "ãģ¨", "ãģ®" ], [ "ãģ¨", "åĲĮãģĺ" ], [ "ãģ¨ãģĹãģ¦", "æī±" ], [ "ãģ©", "ãģ®" ], [ "ãģª", "ãĤĭãĢĤ" ], [ "ãģ«ä½¿ãģĨ", "ãĢĤ" ], [ "ãģ®", "éĽĨåĲĪ" ], [ "ãģ®å", "Ĩħ" ], [ "ãģ®å·", "®" ], [ "ãģĬ", "ãĤĪ" ], [ "ãģĮ", "ä¸Ģ" ], [ "ãģĮ", "å°ıãģķãģĦ" ], [ "ãģĳ", "ãĤĮãģ°" ], [ "ãģĵ", "ãģ®æ" ], [ "ãģĵ", "ãģĵ" ], [ "ãģĵãĤĮ", "ãĤĴ" ], [ "ãģĻ", "ãģ¹ãģ¦" ], [ "ãģŁãĤģ", ")ãĢĤ" ], [ "ãĤ¤ãĥ³", "ãĤ¿" ], [ "ãĤ»", "ãĥ«" ], [ "ãĤ¿ãĤ¤ãĥĪ", "ãĥ«" ], [ "ãĤī", "ãģļ" ], [ "ãĤĴ", "ãģ¾ãģŁ" ], [ "ãĤĴæ", "¯Ķè¼ĥ" ], [ "ãĤĴæ", "ĥ³å®ļ" ], [ "ãĤĴè¡Į", "ãģĨ" ], [ "ãĥ", "ĵ" ], [ "ãĥ¢", "ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ" ], [ "ãĥª", "ãĥ³ãĤ¯" ], [ "ãĥ³", "ãĤ¹" ], [ "ãĥ³", "ãĥĨãĤ£" ], [ "ãĥ»", "Key" ], [ "ãĥ¼", "ãĢĤ" ], [ "ãĥ¼ãĥķ", "ãĤ" ], [ "ãĥĥãĥ", "ĳ" ], [ "ãĥĩ", "ãĤ£" ], [ "ãĥĪ", "ãĥĶ" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "åŁĭãĤģè¾¼ãģ¿" ], [ "ãĥĪãĥĶ", "ãĥĥãĤ¯" ], [ "ãĥŃ", "ãĥ³ãĥĨãĤ£" ], [ "ãĥŃ", "ãĥķãĤ¡ãĤ¤ãĥ«" ], [ "ä¸", "¦" ], [ "ä¸»", "å¼µ" ], [ "ä¹", "±æķ°" ], [ "äºĭåīį", "åĪĨåī²" ], [ "ä»¥", "å¤ĸ" ], [ "ä¼", "¼" ], [ "ä½¿ãģĨ", "ãĢĤ" ], [ "ä½įç½®", "ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°" ], [ "å", "ĢĻè£ľ" ], [ "å¢", "Ĺ" ], [ "å¤ī", "ãĤı" ], [ "å®īå®ļ", "åĮĸ" ], [ "å®Łè¡Į", "åīįãģ®" ], [ "å°", "Ħå½±" ], [ "å·®", "åĪĨ" ], [ "å¼", "ķæķ°" ], [ "å¾ĵæĿ¥", "éĢļãĤĬ" ], [ "åħ¨", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿" ], [ "åħ¨ä½ĵ", "ãģ®" ], [ "åĨįåĪ©çĶ¨", "ãģĻãĤĭ" ], [ "åĪĿæľŁ", "åĮĸ" ], [ "åĬł", "ãģĪãĤĭ" ], [ "åĭ¾éħį", "ãĥİãĥ«ãĥł" ], [ "åĽ", "ł" ], [ "åŁ", "Ł" ], [ "åŃĺåľ¨", "ãģĹãģªãģĦ" ], [ "æ¸Ľ", "ç®Ĺ" ], [ "æī", "Ģ" ], [ "æķ°ãģĪ", "æĸ¹" ], [ "æĸ¹", "éĩ" ], [ "æĸ¹éĩ", "Ŀ" ], [ "æĺİ", "ç¤º" ], [ "æľ", "«å°¾" ], [ "æľªçŁ¥", "èªŀ" ], [ "æľ¬", "æĸĩ" ], [ "æĿ¡ä»¶", "ãģĶãģ¨ãģ«" ], [ "æŃ£è¦ıåĮĸ", "å¾Į" ], [ "ç", "Ńī" ], [ "çµĮ", "çĶ±" ], [ "ç·ı", "åĴĮ" ], [ "çķª", "å®Łè¡Įåīįãģ®" ], [ "è", "·ĿéĽ¢" ], [ "è©", "³" ], [ "è©¦", "è¡Į" ], [ "èª", "į" ], [ "èº«", "ãģ®" ], [ "é«ĺ", "ãģĦ" ], [ "éĢ", "Ĩ" ], [ "éģ©çĶ¨", "åīį" ], [ "éĸ", "¾åĢ¤" ], [ "éĿŀ", "ç·ļå½¢" ], [ "Ċ", "]" ], [ "ĊĊ", "__all__" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "if" ], [ "ĊĠĠĠĠax.grid(alpha=0.3", ")" ], [ "ĊĠĠĠĠĠĠ", "Ġ" ], [ "ĊĠĠĠĠĠĠĠĠ", "-" ], [ "ĊĠĠĠĠĠĠĠĠ", "else:" ], [ "ĊĠĠĠĠĠĠĠĠ", "eps:" ], [ "ĊĠĠĠĠĠĠĠĠ", "log" ], [ "ĊĠĠĠĠĠĠĠĠ", "norm_first:" ], [ "ĊĠĠĠĠĠĠĠĠ", "normalization_factory:" ], [ "ĊĠĠĠĠĠĠĠĠ", "relative_position" ], [ "ĊĠĠĠĠĠĠĠĠ", "text:" ], [ "ĊĠĠĠĠĠĠĠĠf", "eed_forward_factory:" ], [ "Ġ", "6" ], [ "Ġ", "K" ], [ "Ġ", "Model" ], [ "Ġ", "Multi-Head" ], [ "Ġ", "MultiHeadAttention" ], [ "Ġ", "cache_dir" ], [ "Ġ", "key" ], [ "Ġ", "label=" ], [ "Ġ", "max(" ], [ "Ġ", "relative_position" ], [ "Ġ", "token_id" ], [ "Ġ", "un" ], [ "Ġ", "urllib" ], [ "Ġ", "åĪĨåī²" ], [ "Ġ", "æ¬¡" ], [ "Ġ", "æľªæºĢ" ], [ "Ġ", "è¨ĺäºĭ" ], [ "Ġ(", "log" ], [ "Ġ(", "{" ], [ "Ġ0", ".5" ], [ "Ġ1e", "-" ], [ "Ġ201", "9" ], [ "Ġ4", ".2))" ], [ "ĠC", "ounter" ], [ "ĠL", "anguage" ], [ "ĠL", "inear" ], [ "ĠLayerNorm", "alization" ], [ "ĠNeur", "al" ], [ "ĠNone", ")" ], [ "ĠR", "^{" ], [ "ĠR", "eLU" ], [ "ĠUn", "icode" ], [ "Ġ_LOG_ALPHA_BETA_M", "AX" ], [ "Ġ_LOG_ALPHA_BETA_M", "IN," ], [ "Ġ``", "DecoderBlock``" ], [ "Ġal", "." ], [ "Ġb", "est_pair" ], [ "Ġbias", "=bias)" ], [ "Ġcon", "verged" ], [ "Ġd_model", "," ], [ "Ġdtype=torch.", "float3" ], [ "Ġfloat", "]" ], [ "Ġin", "valid=\"ignore\"):" ], [ "Ġmax_", "len" ], [ "Ġnp.", "errstate(over=\"ignore\"," ], [ "Ġnp.", "exp(log" ], [ "Ġnum", "_heads" ], [ "Ġp", "re" ], [ "Ġp", "ro" ], [ "Ġplt.subplots(figsize=(6", ".5," ], [ "Ġtext", ":" ], [ "Ġtuple", "[float," ], [ "Ġx", "_hat" ], [ "ĠÎ", "²" ], [ "ĠÎ", "µ" ], [ "ĠâĪ", "Ī" ], [ "Ġãģ¨åĲĮãģĺ", "éķ·ãģķ" ], [ "Ġãģ®", "ç¯ĦåĽ²" ], [ "Ġãģ®", "çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³" ], [ "Ġãĥ", "Ĺ" ], [ "ĠãĥĹ", "ãĥŃãĥķãĤ¡ãĤ¤ãĥ«" ], [ "Ġãĥĺãĥĥãĥī", "ãģĤãģŁãĤĬãģ®" ], [ "Ġãĥĺãĥĥãĥī", "æķ°" ], [ "Ġæ", "¤ľè¨¼" ], [ "ĠĠĠĠ", ":" ], [ "į", "ĺ" ], [ "\"", ")," ], [ "(", "id" ], [ "(", "re-" ], [ "(", "torch." ], [ "()", "," ], [ "(``", "n" ], [ "(d_model", ")" ], [ "(log", "_n" ], [ "(log", "it" ], [ "(num", "_heads)" ], [ "(x", ":" ], [ "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪ", "ãĢģ" ], [ ")", "." ], [ ")", "]," ], [ ")", "ãģĭãĤī" ], [ "**", "2" ], [ "-", "M" ], [ "-", "start" ], [ "-M", "arquardt" ], [ ".", "co" ], [ ".", "norm" ], [ ".", "p" ], [ ".encod", "e(\"utf-8" ], [ ".items()", "):" ], [ ".mean", "(dim" ], [ ".re", "ad_text(encoding=\"utf-8" ], [ "1", "_f" ], [ "1_f", "ound" ], [ "1_found", "ation" ], [ "1_foundation", "s/" ], [ "2", "]" ], [ "=color", "," ], [ "=float", "(np.exp(log" ], [ "BYT", "E_TO_" ], [ "BYTE_TO_", "UNICOD" ], [ "BYTE_TO_UNICOD", "E" ], [ "G", "ELU" ], [ "I", "K" ], [ "IK", "I" ], [ "R", "E" ], [ "R", "eLU" ], [ "R", "otaryPositionEmbedding" ], [ "RL", "S" ], [ "S", "cal" ], [ "S", "entencePiece" ], [ "Shaw", "RelativePositionBias" ], [ "Unigram", "Tokenizer" ], [ "V", "as" ], [ "Vas", "w" ], [ "Vasw", "an" ], [ "Vaswan", "i" ], [ "W", "IKI" ], [ "[", "d_model" ], [ "[d_model", "]" ], [ "[int", "]," ], [ "[key", "]" ], [ "\\", "n" ], [ "\\", "s" ], [ "\\\\", "_lr" ], [ "\\\\_lr", "}" ], [ "\\\\hat{", "v" ], [ "\\\\hat{v", "}_t" ], [ "_", "ax" ], [ "_", "grid:" ], [ "_", "ok" ], [ "_", "parabola" ], [ "_", "power_law" ], [ "_", "values" ], [ "_an", "al" ], [ "_anal", "ys" ], [ "_b", ":" ], [ "_b", "atch" ], [ "_coef", "_init" ], [ "_con", "verged" ], [ "_f", "ilter" ], [ "_f", "raction" ], [ "_fn", "(d_model)" ], [ "_grad", "()" ], [ "_h", "ub" ], [ "_layer", ":" ], [ "_m", "ulti" ], [ "_op", "t(" ], [ "_opt(", "C" ], [ "_p", "air" ], [ "_p", "la" ], [ "_pla", "in" ], [ "_s", "aturating" ], [ "_s", "core" ], [ "_s", "lopes" ], [ "_saturating", "_power_law" ], [ "_scaling", "_analys" ], [ "_se", "ed" ], [ "_t", "er" ], [ "_t", "gt" ], [ "_t", "r" ], [ "_ter", "m" ], [ "_tr", "ig" ], [ "_trig", "g" ], [ "_trigg", "er" ], [ "_w", "idth" ], [ "_with", "out_leading_whitespace\":" ], [ "``", ")ãģ®" ], [ "``", "ãĢĤ" ], [ "``)", "ãģ§" ], [ "``)", "ãģ¯" ], [ "a", "3" ], [ "ab", "le" ], [ "ab", "s_" ], [ "abs_", "r" ], [ "act", "ion" ], [ "activation_fn", ":" ], [ "am", "ma" ], [ "an", "h" ], [ "ann", "otat" ], [ "app", "s/" ], [ "ar", "_width" ], [ "ar", "ization" ], [ "ar", "k" ], [ "articl", "es" ], [ "atch", "_rat" ], [ "atch", "_size:" ], [ "attention_s", "core_bias" ], [ "b", "ucket" ], [ "bj", "ect" ], [ "bucket", "s" ], [ "byte_level", ":" ], [ "byte_level", "_symbol" ], [ "cal", "_" ], [ "character_", "level" ], [ "cl", "ud" ], [ "clud", "e_" ], [ "cor", "es" ], [ "d", "ot" ], [ "d", "type=" ], [ "d(", "id" ], [ "d_ff", "," ], [ "d_model", "," ], [ "decod", "e_" ], [ "decode_", "byte_level_symbol" ], [ "defa", "ult" ], [ "dtype", ":" ], [ "e", "ep" ], [ "e", "venberg" ], [ "e(\"", "log" ], [ "e(\"log", "\")" ], [ "e_", "embedding" ], [ "eed_for", "ward" ], [ "eep", "dim" ], [ "eg", "ul" ], [ "egul", "arization" ], [ "er", "ror" ], [ "er", "s" ], [ "es", "(" ], [ "es", "pe" ], [ "espe", "ar" ], [ "et_", "random" ], [ "et_random", "_batch" ], [ "evenberg", "-Marquardt" ], [ "ex", "cept" ], [ "exist", "_ok" ], [ "exp", "(" ], [ "ext", "_token" ], [ "f", "in" ], [ "for", "ward" ], [ "frontier_", "log_c" ], [ "g", "amma" ], [ "ha", "k" ], [ "hak", "espear" ], [ "i", ")" ], [ "i", "an" ], [ "ic", "method" ], [ "id", "s" ], [ "ie", "ces" ], [ "isoflop", "_parabola" ], [ "ist", "ic" ], [ "it", "_saturating_power_law" ], [ "its-per-byt", "e" ], [ "ja", "cobian" ], [ "json", "." ], [ "ke", "y_length:" ], [ "l", "at" ], [ "l", "ine" ], [ "l", "inear" ], [ "l", "l" ], [ "la", "ce" ], [ "label", "s:" ], [ "learning", "_curves" ], [ "length", ")" ], [ "lik", "e(" ], [ "lo", "ad" ], [ "log_n", "_opt" ], [ "make_evaluation_window", "s``" ], [ "max_chunk_bytes", "``" ], [ "merg", "e_" ], [ "model.", "parameters()" ], [ "mp", "t" ], [ "mpt", "_id" ], [ "n", "one" ], [ "nn.", "Parameter" ], [ "o", "cstring" ], [ "ootstrap", "_scaling_analys" ], [ "optimizer", "``" ], [ "or", "der" ], [ "ord(", "\"" ], [ "out", "," ], [ "output", "," ], [ "p", "e(" ], [ "p", "lace" ], [ "p", "proach" ], [ "per", "plexity" ], [ "position", "s." ], [ "positional_transform", "``" ], [ "quer", "y_length:" ], [ "ra", "w" ], [ "rn", "g" ], [ "rontier", "Result" ], [ "s", "]" ], [ "s.", "to" ], [ "s.to", "list" ], [ "s:", "//" ], [ "s_", "like(" ], [ "scal", "e(\"log\")" ], [ "se", "ed:" ], [ "segmentation", "_a" ], [ "sequence_length", ":" ], [ "sort", "ed" ], [ "stat", "icmethod" ], [ "step", "_count" ], [ "str", "ip" ], [ "symbol", "s" ], [ "symbol", "s," ], [ "t", "emperature" ], [ "theories/0", "1_foundations/" ], [ "tim", "e" ], [ "top", "-" ], [ "total_step", "s" ], [ "u", "cle" ], [ "ulti", "-start" ], [ "unk", "_token" ], [ "ut", "il" ], [ "v", "is" ], [ "warm", "_fit" ], [ "wiGLU", "FeedForwardNetwork" ], [ "y_", "arr" ], [ "y_", "decode_byte_level_symbol" ], [ "y_length", "," ], [ "z", "order" ], [ "|", "}" ], [ "¥", "µ" ], [ "¨", "æ³" ], [ "¨æ³", "¢æķ°" ], [ "®å°ı", "å®ļæķ°" ], [ "®ĭ", "å·®" ], [ "³", "åĿĩ" ], [ "·", "ļ" ], [ "·ļ", "å½¢" ], [ "¹", "³åĿĩ" ], [ "¼", "Ķ" ], [ "¼Ķ", "ç®Ĺ" ], [ "â", "ĸ" ], [ "âĸ", "ģ" ], [ "ãģ", "¼" ], [ "ãģ§ãģĤ", "ãĤĮãģ°" ], [ "ãģ§ãģį", "ãģªãģĦ" ], [ "ãģ¨", "ãĢģ" ], [ "ãģ¨", "ãĤģ" ], [ "ãģª", "ãĤĬ" ], [ "ãģ«", "ä½įç½®" ], [ "ãģ«", "å¯¾å¿ľ" ], [ "ãģ«", "åĬłç®Ĺ" ], [ "ãģ«", "åŁºãģ¥ãģı" ], [ "ãģ«ãģªãĤĭ", "(" ], [ "ãģ«ãģĻãĤĭ", "ãģĭãĢĤ" ], [ "ãģ®", "ä¸ĬéĻĲ" ], [ "ãģ®", "åıįå¾©" ], [ "ãģ®", "åĲĦ" ], [ "ãģ®", "æķ°" ], [ "ãģ®", "æľĢå¤§åĢ¤" ], [ "ãģ®ãģŁãĤģ", "ãģ®å¾" ], [ "ãģ®ãģŁãĤģãģ®å¾", "®å°ıå®ļæķ°" ], [ "ãģ®å", "¼ķæķ°" ], [ "ãģ®å¯¾æķ°", "å°¤åº¦" ], [ "ãģ®æ", "¤ľè¨¼" ], [ "ãģ®æ", "¨" ], [ "ãģ»", "ãģ¼" ], [ "ãģĨ", "ãĤĭ" ], [ "ãģĬãĤĪ", "ãģ³" ], [ "ãģĭ", "ãģ£ãģŁ" ], [ "ãģĭãĤī", "ãģ®" ], [ "ãģĵãģ¨", "ãĤĴ" ], [ "ãģķ", "ãģªãģĦ" ], [ "ãģķãĤĮ", "ãĤĭãĢĤ" ], [ "ãģĹãģ¦", "ãģĬ" ], [ "ãģĹãģ¦", "ãģĭãĤī" ], [ "ãģĻãĤĭ", "(``" ], [ "ãģĻãĤĭ(00", "7" ], [ "ãĤ", "ª" ], [ "ãĤ¢", "ãĤ¦" ], [ "ãĤ¤", "ãĥ¤" ], [ "ãĤ¤ãĥ¤", "ãĥ¼" ], [ "ãĤ¤ãĥ³ãĤ¿", "ãĥ¼ãĥķãĤ" ], [ "ãĤ¦ãĥ³ãĥ", "ī" ], [ "ãĤ¦ãĥ³ãĥī", "ä¿®æŃ£" ], [ "ãĤ¨", "ãĤ¹ãĥĪ" ], [ "ãĤ«", "ãĥĨ" ], [ "ãĤ«ãĥĨ", "ãĤ´ãĥª" ], [ "ãĤ¯", "ãĤ¨ãĤ¹ãĥĪ" ], [ "ãĤ³", "ãĤ¹ãĥĪ" ], [ "ãĤ¹", "ãĢĤ" ], [ "ãĤ¿ãĤ¤", "ãĥł" ], [ "ãĤ¿ãĤ¤ãĥł", "ãĤ¢ãĤ¦" ], [ "ãĤĬ", "å½ĵ" ], [ "ãĤĬå½ĵ", "ãģ¦" ], [ "ãĤĭ", "ãģĵãģ¨" ], [ "ãĤĮãĤĭ", ")ãĢĤ" ], [ "ãĤĴ", "ãģĿãģ®ãģ¾ãģ¾" ], [ "ãĤĴ", "å¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«" ], [ "ãĤĴ", "æ¸¡" ], [ "ãĤĴ", "è¨Īç®Ĺ" ], [ "ãĤĴ", "è¿ĶãģĻãĢĤ" ], [ "ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«", "ãģ«ãģĻãĤĭãģĭãĢĤ" ], [ "ãĤĴæ", "Ħı" ], [ "ãĤĴæĦı", "åĳ³" ], [ "ãĤŃãĥ£", "ãĤ¹ãĥĪ" ], [ "ãĥ©", "ãĥĥãĥĳ" ], [ "ãĥª", "ãĤ¯ãĤ¨ãĤ¹ãĥĪ" ], [ "ãĥ«", "ãĥ¼ãĥ" ], [ "ãĥ¬", "ãĤ¤ãĥ¤ãĥ¼" ], [ "ãĥ¬", "ãĥ¼ãĥĪ" ], [ "ãĥ³ãĥ", "Ģ" ], [ "ãĥ»", "``" ], [ "ãĥ¼ãĥī", "ãĤŃãĥ£ãĤ¹ãĥĪ" ], [ "ãĥ¼ãĥķ", "ãĥŃ" ], [ "ãĥ¼ãĥķãĥŃ", "ãĥ¼" ], [ "ãĥģãĥ£ãĥ³ãĤ¯", "ãģ®" ], [ "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶", "æĿ¡ä»¶" ], [ "ãĥĲãĥĥãĥģ", "ãĤµãĤ¤ãĤº" ], [ "ãĥĳ", "ãĤ¤" ], [ "ãĥĸãĥŃ", "ãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪ" ], [ "ãĥŃãĥ³ãĥĨãĤ£", "ãĤ¢" ], [ "ä¸Ģ", "åĪĩ" ], [ "äº", "Ĩ" ], [ "äºĪ", "æ¸¬" ], [ "ä»ĺ", "ä¸İ" ], [ "ä½¿ãģĨ", ")ãĢĤ" ], [ "ä½ĵ", "ãģ¯" ], [ "å", "ķ" ], [ "å", "ĸĦ" ], [ "å¤ĸéĥ¨", "ãģĭãĤī" ], [ "å®īå®ļåĮĸ", "ãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°" ], [ "å±¤", "æŃ£è¦ıåĮĸ" ], [ "å·", "¦" ], [ "å¸", "°" ], [ "å¸¸ãģ«", "è²ł" ], [ "å¹³åĿĩ", "æ¸Ľç®Ĺ" ], [ "å¾", "´" ], [ "å¾©", "åħĥ" ], [ "åĢ¤", "ãģ¯" ], [ "åħ¥", "åĩºåĬĽ" ], [ "åħ¥åĬĽ", "ãĢĤ" ], [ "åħ¨", "ãĥĺãĥĥãĥī" ], [ "åħĥ", "ãģ®" ], [ "åĨħ", "éĥ¨" ], [ "åĬ¹", "çİĩ" ], [ "åĭ¾éħį", "ãģĮ" ], [ "åĮº", "éĸĵ" ], [ "åİ", "»" ], [ "åıĸ", "ãĤĭ" ], [ "åĲ«ãģ¾", "ãĤĮãģªãģĦ" ], [ "åĳ", "¨æ³¢æķ°" ], [ "åķ", "ı" ], [ "åĽ", "³" ], [ "åĽŀ", "å¸°" ], [ "åł´åĲĪ", "(" ], [ "åŃ¦ç¿Ĵçİĩ", "ãĤĴ" ], [ "æ§ĭ", "æĪĲ" ], [ "æ¨Ļæºĸ", "åģıå·®" ], [ "æ®ĭ", "å·®" ], [ "æ¸Īãģ¿", "ãģ®" ], [ "æĸĩ", "è" ], [ "æĸĩåŃĹ", "ãģ®" ], [ "æĸĩåŃĹ", "ãģĮ" ], [ "æĸĩè", "Ħ" ], [ "æĸĩèĦ", "Ī" ], [ "æĹ", "¥æľ¬èªŀ" ], [ "æľĢå°¤", "åĪĨåī²" ], [ "æľĢå°ı", "åĢ¤" ], [ "æľĢéģ©", "åĮĸ" ], [ "æľī", "éĻĲ" ], [ "æĿ¡ä»¶", "éĸĵãģ§" ], [ "æŃ£", "ãģĹ" ], [ "ç¢º", "èªį" ], [ "ç¨", "®" ], [ "ç´°", "ãģ¯" ], [ "ç¸", "®" ], [ "ç½®", "æıĽ" ], [ "çĦ¡", "è¦ĸ" ], [ "çī¹", "å¾´" ], [ "çĶ¨", "ãģ«" ], [ "çķªå®Łè¡Įåīįãģ®", "ä¿®æŃ£" ], [ "çĻº", "çĶŁ" ], [ "çľģçķ¥", "åı¯" ], [ "è", "Ĳ" ], [ "è¡Į", "ãĤı" ], [ "è§", "Ĵ" ], [ "è©³", "ç´°ãģ¯" ], [ "èª", "Ń" ], [ "è»¸", "ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ" ], [ "è¿ĳ", "ä¼¼" ], [ "èĩª", "åĪĨ" ], [ "èĩªå·±", "åĽŀå¸°" ], [ "èĲ", "½" ], [ "èŃ", "ĺ" ], [ "èŃĺ", "åĪ¥" ], [ "é", "¡" ], [ "é", "£" ], [ "éĢ", "ı" ], [ "éģ", "¿" ], [ "éģĶ", "ãģĻãĤĭ" ], [ "éĥ¨åĪĨèªŀ", "ãĤ·ãĥ³ãĥľãĥ«" ], [ "éĸ", "¢" ], [ "éĽĨ", "è¨Ī" ], [ "éłĨä¼ĿæĴŃ", "ãĢĤ" ], [ "éłĨä¼ĿæĴŃ", "ãĥįãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯" ], [ "Ċ", "_" ], [ "ĊĊĠĠĠĠ", "#" ], [ "ĊĊĠĠĠĠ", "R" ], [ "ĊĊĠĠĠĠ@", "staticmethod" ], [ "ĊĊĠĠĠĠax", ".set_" ], [ "ĊĠĠĠĠ", "B" ], [ "ĊĠĠĠĠ", "S" ], [ "ĊĠĠĠĠ", "V" ], [ "ĊĠĠĠĠ", "data" ], [ "ĊĠĠĠĠ", "vocab" ], [ "ĊĠĠĠĠ", "with" ], [ "ĊĠĠĠĠ", "x" ], [ "ĊĠĠĠĠ", "Î" ], [ "ĊĠĠĠĠax.set_", "xlabel(xlabel)" ], [ "ĊĠĠĠĠb", "est" ], [ "ĊĠĠĠĠĠĠĠĠ", "activation_fn:" ], [ "ĊĠĠĠĠĠĠĠĠ", "h" ], [ "ĊĠĠĠĠĠĠĠĠax", ".set_" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "b" ], [ "Ġ", "$" ], [ "Ġ", "003" ], [ "Ġ", "004" ], [ "Ġ", "DecoderBlock" ], [ "Ġ", "Encoding" ], [ "Ġ", "ar" ], [ "Ġ", "attn_weights" ], [ "Ġ", "decod" ], [ "Ġ", "e" ], [ "Ġ", "exist_ok" ], [ "Ġ", "item" ], [ "Ġ", "length)" ], [ "Ġ", "p.grad" ], [ "Ġ", "r" ], [ "Ġ", "sequence_length" ], [ "Ġ", "token" ], [ "Ġ", "warmup" ], [ "Ġ", "zip" ], [ "Ġ", "ÃĹ" ], [ "Ġ", "ãĤµãĥ³ãĥĹãĥªãĥ³ãĤ°" ], [ "Ġ", "ä¸»å¼µ" ], [ "Ġ", "åı" ], [ "Ġ", "åıĤçħ§)ãĢĤ" ], [ "Ġ", "åĲĦ" ], [ "Ġ", "è¨ĵç·´" ], [ "Ġ", "éĩįãģ¿æ¸Ľè¡°" ], [ "Ġ(", "S_q," ], [ "Ġ0", ".5," ], [ "Ġ100", "00" ], [ "Ġ1e", "-8" ], [ "Ġ<", "-" ], [ "ĠA", "dam" ], [ "ĠBPE", "Tokenizer" ], [ "ĠI", "CLR" ], [ "ĠI", "soFLOP" ], [ "ĠPath", ")" ], [ "ĠS", ")``" ], [ "ĠS", "_tgt" ], [ "ĠSequence", "[" ], [ "ĠSequence[float", "])" ], [ "ĠUn", "igram" ], [ "ĠW", "^" ], [ "ĠW", "ikipedia" ], [ "Ġ``", "-inf``" ], [ "Ġb", "," ], [ "Ġbest", "_params" ], [ "Ġc", "urv" ], [ "Ġco", "st" ], [ "Ġd", "ocstring" ], [ "Ġd_", "v" ], [ "Ġd_k", ")``" ], [ "Ġdropout", "=" ], [ "Ġdtype=torch.float3", "2)" ], [ "Ġf", "\"" ], [ "Ġf", "it_" ], [ "Ġfloat(np.", "sum(" ], [ "Ġis", "instance" ], [ "Ġk", "eepdim" ], [ "Ġm", "ean" ], [ "Ġm", "in(" ], [ "Ġmax_", "chunk" ], [ "Ġn", "-gram" ], [ "Ġn", "am" ], [ "Ġnorm", "_fn(d_model)" ], [ "Ġq", "_m" ], [ "Ġt", "h" ], [ "Ġtorch.", "Tensor" ], [ "Ġtorch.", "cat" ], [ "Ġtorch.zero", "s_like(" ], [ "Ġx", "." ], [ "Ġâ", "Ĩ" ], [ "ĠâĨ", "Ĵ" ], [ "Ġãģ¨", "å®Įåħ¨ãģ«" ], [ "Ġãģ«", "å¯¾ãģĹãģ¦" ], [ "Ġãģ®", "ä¸Ńéĸĵå±¤" ], [ "Ġãģ®", "å¯¾å¿ľ" ], [ "ĠãĤĴ", "åıĤçħ§)ãĢĤ" ], [ "Ġãĥ©", "ãĤ¦ãĥ³ãĥīä¿®æŃ£" ], [ "ĠæĲįå¤±", "ãģ®" ], [ "Ġç", "¬¬" ], [ "Ġç", "·ļå½¢" ], [ "Ġç", "İĩ" ], [ "Ġç", "Ńī" ], [ "Ġç¯Ģ", ")ãĢĤ" ], [ "Ġçİĩ", "ãĢĤ" ], [ "\"", "]," ], [ "(", "B," ], [ "(", "True" ], [ "(", "c" ], [ "(", "len(" ], [ "(", "parent" ], [ "(", "self," ], [ "(", "vocabulary" ], [ "(", "weight" ], [ "(", "å¯¾æķ°ç©ºéĸĵ" ], [ "().item", "()" ], [ "(00", "7" ], [ "(00", "9" ], [ "(L", "ayerNorm" ], [ "(LayerNorm", "(x" ], [ "(``", "pretokenize``" ], [ "(``n", "umpy" ], [ "(a", ")" ], [ "(b", "eam" ], [ "(num", "_window" ], [ "(num_window", "s," ], [ "(parent", "s" ], [ "(parents", "=True," ], [ "(se", "ed)" ], [ ")", "]" ], [ ")", "ãģ§ãģ¯" ], [ ")", "ãĥ»" ], [ "))", ":" ], [ "**", "(" ], [ "-", "P" ], [ "-", "attention" ], [ ".", "device)" ], [ ".", "get" ], [ ".", "ipyn" ], [ ".", "max" ], [ ".", "values" ], [ ".b", "ias" ], [ ".items()", "}" ], [ ".m", "k" ], [ ".mk", "dir" ], [ ".mkdir", "(parents=True," ], [ ".size(", "-1" ], [ ".t", "x" ], [ ".to(", "dtype=" ], [ ".tx", "t" ], [ "006", "ãĥ»00" ], [ "006ãĥ»00", "8" ], [ "2", ")ãĢĤ" ], [ "4", "," ], [ "5", "6" ], [ "9", "0" ], [ "=", "str" ], [ "=", "{" ], [ "=1", "." ], [ "=1", ".0" ], [ "=True", "``" ], [ "AdamW", "``" ], [ "B", "yt" ], [ "BYT", "E" ], [ "Byt", "e" ], [ "C", "haracterLevelTokenizer" ], [ "CH", "UN" ], [ "CHUN", "K_" ], [ "D", "``" ], [ "E_TO_", "BYTE" ], [ "G", "enerator" ], [ "H", "off" ], [ "Hoff", "m" ], [ "Hoffm", "ann" ], [ "K", "V" ], [ "L", "2" ], [ "L", "earn" ], [ "LayerNorm", "alization" ], [ "None", "," ], [ "O", "S" ], [ "P", "I" ], [ "P", "re-" ], [ "Parameter", "(torch." ], [ "Pre-", "Layer" ], [ "S", "wish" ], [ "Tensor", "]," ], [ "UNICOD", "E_TO_BYTE" ], [ "Value", "Error:" ], [ "Wik", "i" ], [ "[", "None," ], [ "[", "[" ], [ "[", "tuple[str," ], [ "\\\\", "math" ], [ "\\\\math", "rm" ], [ "\\\\mathrm", "{" ], [ "_", "BYTE_TO_UNICODE" ], [ "_", "ap" ], [ "_", "dir" ], [ "_", "exact" ], [ "_", "if" ], [ "_", "l" ], [ "_", "loss_" ], [ "_", "model" ], [ "_", "residual" ], [ "_", "seq" ], [ "_arr", "," ], [ "_b", ")" ], [ "_bod", "y" ], [ "_buffer", "(\"" ], [ "_bytes", ")" ], [ "_coe", "ff" ], [ "_compute_budget", "s:" ], [ "_count", "s[" ], [ "_counts[", "pair]" ], [ "_cy", "c" ], [ "_cyc", "le[" ], [ "_fit", "_n" ], [ "_fit_n", "onl" ], [ "_fit_nonl", "inear" ], [ "_fit_nonlinear", "_h" ], [ "_fit_nonlinear_h", "uber" ], [ "_grad", "()``" ], [ "_h", "uber_" ], [ "_if", "_l" ], [ "_if_l", "ar" ], [ "_in", "ter" ], [ "_loss_", "per_token" ], [ "_m", "anifest" ], [ "_n", "am" ], [ "_n", "um" ], [ "_n", "umpy" ], [ "_opt(C", ")" ], [ "_p", "rod" ], [ "_param", "s," ], [ "_prod", "uct_" ], [ "_product_", "attention" ], [ "_s", "u" ], [ "_se", "arch" ], [ "_t", "y" ], [ "_train", "_text" ], [ "_ty", "pe" ], [ "a", "_bo" ], [ "a", "ck" ], [ "a", "is" ], [ "a", "u" ], [ "action", "=" ], [ "ad", "d_" ], [ "ad", "ded_" ], [ "ad", "ding" ], [ "ag", "e" ], [ "ais", "es:" ], [ "al", "f" ], [ "allab", "le" ], [ "always_negative_", "unit_" ], [ "always_negative_unit_", "ratio" ], [ "ap", "i" ], [ "ark", "er" ], [ "arr", "," ], [ "at", "ed" ], [ "at", "plot" ], [ "ate_", "causal_mask" ], [ "ate_", "h" ], [ "ate_h", "alf" ], [ "atplot", "lib" ], [ "attn_weight", "s)" ], [ "b", "ool" ], [ "bs", "ol" ], [ "bsol", "ut" ], [ "bsolut", "e" ], [ "c", "andidates" ], [ "c", "ation" ], [ "cache_", "path" ], [ "cat", "eg" ], [ "center", "ing" ], [ "chunk_split_mode", "=\"" ], [ "clip", "_threshold" ], [ "co", "eff" ], [ "corpu", "s_" ], [ "cross_", "attn" ], [ "data", "[\"" ], [ "decay", "_fraction" ], [ "ded_", "param" ], [ "ded_param", "_id" ], [ "device", "," ], [ "dex", "[" ], [ "di", "st" ], [ "dot", "_product_attention" ], [ "dropout", "," ], [ "e", "z" ], [ "e(", "æĹ¢å®ļåĢ¤" ], [ "ed_", "dot_product_attention" ], [ "edia", "Wiki" ], [ "elu", "_exact" ], [ "es", "ign" ], [ "est", "y" ], [ "esty", "le" ], [ "estyle", "=\"" ], [ "excl", "u" ], [ "exclu", "ded_param_id" ], [ "ez", "e(" ], [ "f", "ile" ], [ "fit_", "chinchilla_parametric" ], [ "g", "g" ], [ "gg", "ing" ], [ "http", "s://" ], [ "i", "es:" ], [ "id", "al" ], [ "id_", "tokenizer_json" ], [ "ik", "itext" ], [ "in", "n" ], [ "in", "us" ], [ "initi", "al_values" ], [ "inus", "o" ], [ "inuso", "idal" ], [ "ist", "ories" ], [ "it", ")" ], [ "it", "e_" ], [ "it_", "alpha" ], [ "jo", "ined_" ], [ "k", "," ], [ "l", "st" ], [ "layer_", "output" ], [ "learn_b", "pe" ], [ "learning_rat", "e" ], [ "linalg.", "lst" ], [ "linalg.lst", "s" ], [ "linalg.lsts", "q" ], [ "lo", "se" ], [ "load_", "wikipedia_corpus" ], [ "lose", "st" ], [ "lp", "(Y" ], [ "m", "an" ], [ "m", "t" ], [ "m", "}_t" ], [ "max_", "relative_position" ], [ "mean", "_loss_per_token" ], [ "min", "(" ], [ "min", "_learning_rat" ], [ "no", "_grad()" ], [ "np.", "ndarray" ], [ "np.log(", "warm_fit" ], [ "o", "PE" ], [ "o", "m" ], [ "ol", "d_" ], [ "ootstrap", "Result" ], [ "ootstrap_scaling_analys", "is" ], [ "path.", "w" ], [ "path.w", "r" ], [ "path.wr", "ite_" ], [ "position", "s)" ], [ "positional_transform", ":" ], [ "pro", "gres" ], [ "pro", "mpt_id" ], [ "profil", "es" ], [ "que", "eze(" ], [ "r", "_squar" ], [ "r", "u" ], [ "raw", "_norm" ], [ "re", "ak" ], [ "re", "quest" ], [ "redict", "ed_" ], [ "residual", "s:" ], [ "rod", "uct" ], [ "rot", "ate_half" ], [ "ru", "e(æĹ¢å®ļåĢ¤" ], [ "s", "a" ], [ "s", "er" ], [ "s", "h" ], [ "s", "lopes" ], [ "s_", "with" ], [ "s_with", "out_leading_whitespace" ], [ "step", "()" ], [ "symbol", "s)" ], [ "t", ")" ], [ "theories/", "ãĥ»" ], [ "theories/ãĥ»", "apps/" ], [ "theta_i", "^{" ], [ "ti", "e_embedding" ], [ "to", "_numpy" ], [ "token_id", "s," ], [ "train_", "loss" ], [ "u", "gging" ], [ "uber_delta", ":" ], [ "ucle", "us" ], [ "un", "igram" ], [ "unction", "al" ], [ "use_cross_attention", ":" ], [ "v", "ar" ], [ "v", "e" ], [ "v", "iterbi_segment" ], [ "values_by", "_series" ], [ "venberg", "_m" ], [ "vi", "ew" ], [ "wikipedia", "_manifest" ], [ "y", "scale(\"log\")" ], [ "zorder", "=" ], [ "}", "\"," ], [ "}", "," ], [ "¡", "çĶ¨" ], [ "§", "ç¸®" ], [ "§", "Ķ" ], [ "§", "ĺ" ], [ "§ĭ", "æĪĲ" ], [ "ª", "çĽ®" ], [ "²", "ç·ļ" ], [ "·", "·" ], [ "·", "åº¦" ], [ "»", "åº¦" ], [ "»", "æĦı" ], [ "½", "åĴĮ" ], [ "½åĴĮ", "ãģ¹ãģįä¹Ĺåīĩ" ], [ "ãĢģ", "ãģĿãģ®" ], [ "ãĢį", "ãĤĴ" ], [ "ãģ", "Ĵ" ], [ "ãģ§ãģ®", "éĢ£" ], [ "ãģ§ãģ®éĢ£", "çķª" ], [ "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭ", ":" ], [ "ãģ§ãģįãĤĭ", "ãĤĪãģĨãģ«" ], [ "ãģªãģĦ", ")ãĢĤ" ], [ "ãģ«", "ãģ¤ãģĦãģ¦" ], [ "ãģ«", "å¯¾" ], [ "ãģ«", "å¾ĵ" ], [ "ãģ«ãģª", "ãĤĬãĢģ" ], [ "ãģ«ä¾ĿåŃĺ", "ãģĻãĤĭ" ], [ "ãģ®", "ãĤ·ãĥ³ãĥľãĥ«" ], [ "ãģ®", "ãĥ¦ãĥĭãĥĥãĥĪ" ], [ "ãģ®", "åħ¨" ], [ "ãģ®", "ç·ıåĴĮ" ], [ "ãģ®ãĥªãĤ¹ãĥĪ", "ãĢĤ" ], [ "ãģ®å", "¹³åĿĩ" ], [ "ãģ®å", "½" ], [ "ãģ®å", "ĴĮ" ], [ "ãģ®å", "Ľ" ], [ "ãģ®å¤", "§" ], [ "ãģ®å¯¾", "å¿ľ" ], [ "ãģ®å°", "İ" ], [ "ãģ®æ", "®ĭå·®" ], [ "ãģ®æ", "ī" ], [ "ãģ®æ", "Ĺ" ], [ "ãģ®æĮĩæķ°", "ç§»åĭķ" ], [ "ãģ®æĮĩæķ°ç§»åĭķ", "å¹³åĿĩ" ], [ "ãģ¹", "ãģŁ" ], [ "ãģ¾", "ãĤĭ" ], [ "ãģį", "æıĽ" ], [ "ãģįæıĽ", "ãģĪ" ], [ "ãģĵãģ®", "éĸ¢æķ°" ], [ "ãģĹ", "ãĢģ``" ], [ "ãģĹãģ¦", "ãĤĤ" ], [ "ãģĹãģªãģĦ", ")ãĢĤ" ], [ "ãģĻ", ")ãĢĤ" ], [ "ãģĻãĤĭ", "ãģ®" ], [ "ãģĿãģ®ãģ¾ãģ¾", "æ¸¡ãģĻ" ], [ "ãĤ¢ãĥĥãĥĹ", "ãĥŃ" ], [ "ãĤ¢ãĥĥãĥĹãĥŃ", "ãĥ¼ãĥī" ], [ "ãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ", "§ãĥ¼ãĤ¹" ], [ "ãĤª", "ãĥ¼ãĥĲ" ], [ "ãĤªãĥ¼ãĥĲ", "ãĥ¼ãĥķãĥŃãĥ¼" ], [ "ãĤ«ãĥ©", "ãĥ¼" ], [ "ãĤ¯", "ãĥĪãĥª" ], [ "ãĤ°", "ãĥª" ], [ "ãĤ°ãĥª", "ãĥĥãĥī" ], [ "ãĤµãĥĸ", "ãĥ¬ãĤ¤ãĥ¤ãĥ¼" ], [ "ãĤ¹ãĤ¿", "ãĤ¤ãĥ«" ], [ "ãĤ¹ãĤ¿", "ãĥ³ãĤ¹" ], [ "ãĤ¹ãĥĨãĥĥãĥĹ", "æķ°" ], [ "ãĤ¹ãĥĪ", "ãĥ©" ], [ "ãĤ»", "ãĥĥãĥĪ" ], [ "ãĤ¿ãĤ¤ãĥłãĤ¢ãĤ¦", "ãĥĪ" ], [ "ãĤĮãģªãģĦ", "ãĢĤ" ], [ "ãĤĴ", "ä¿Ŀ" ], [ "ãĤĴ", "åħ±æľī" ], [ "ãĤĴæ", "¤ľè¨¼" ], [ "ãĤĴæ", "Ĭ" ], [ "ãĤĴè¡¨", "ãģĻ" ], [ "ãĤĵ", "ãģł" ], [ "ãĥ", "ī" ], [ "ãĥ", "Ł" ], [ "ãĥ¦ãĥĭ", "ãĥ¼ãĤ¯" ], [ "ãĥ©", "ãĥ³ãĥĢ" ], [ "ãĥ«", "ãĥĨãĤ£" ], [ "ãĥ«ãĥ¼ãĥ", "Ĺ" ], [ "ãĥ¬", "ãĤ¯ãĥĪãĥª" ], [ "ãĥ¼ãĤ¯", "ãĤ¢ãĥĥãĥĹ" ], [ "ãĥ¼ãĥ", "ł" ], [ "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵ", "ãĤĴ" ], [ "ãĥĩ", "ãĤ³ãĥ¼ãĥī" ], [ "ãĥĩãĤ£", "ãĥ¬ãĤ¯ãĥĪãĥª" ], [ "ãĥĬ", "ãĥ«ãĥĨãĤ£" ], [ "ãĥĲãĤ¤ãĥĪ", "ãĥ¬ãĥĻãĥ«" ], [ "ãĥĳãĥ©ãĥ¡", "ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ" ], [ "ãĥĳãĥĩãĤ£ãĥ³ãĤ°", "ä½įç½®" ], [ "ãĥķãĤ", "©ãĥ¼ãĥ«ãĥĲãĥĥãĤ¯" ], [ "ãĥĸ", "ãĥ»" ], [ "ãĥĸãĥŃ", "ãĥĥãĤ¯" ], [ "ãĥĺãĥĥãĥī", "æķ°" ], [ "ãĥļ", "ãĤ¢" ], [ "ãĥļ", "ãĥĬãĥ«ãĥĨãĤ£" ], [ "ãĥŀ", "ãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ" ], [ "ãĥŀãĥ¼ãĤ¸", "è¦ıåīĩ" ], [ "ä¸", "¡" ], [ "ä¸¦", "ãģ¹ãģŁ" ], [ "ä¸Ģèĩ´", "ãģĻãĤĭ" ], [ "ä¸Ĭ", "æĺ" ], [ "ä¸Ĭæĺ", "ĩ" ], [ "ä¸ĭ", "éĻĲ" ], [ "ä¸į", "å¤ī" ], [ "ä¸į", "æŃ£" ], [ "ä¸į", "èĥ½" ], [ "ä¸įèĥ½", "ãģª" ], [ "ä¹", "ĸ" ], [ "äº", "Ĵ" ], [ "ä»", "»æĦı" ], [ "ä»ĺ", "ãģĳãĤĭ" ], [ "ä½", "µ" ], [ "ä½¿ãģĨ", "(" ], [ "å", "§Ķ" ], [ "å", "¯" ], [ "å", "Ĥ" ], [ "å®Ł", "æ¸¬" ], [ "å®Łè£ħ", "æĸ¹éĩĿ" ], [ "å¯¾æķ°", "ç¢ºçİĩ" ], [ "å°", "ĳ" ], [ "å¹³", "æĸĩ" ], [ "å½±", "éŁ¿" ], [ "å¾", "ħ" ], [ "å¾Ĺ", "ãģŁ" ], [ "åĢ¤", "ãĤĴ" ], [ "åħ", "¬" ], [ "åħ", "¸" ], [ "åħĪéłŃ", "ãģ«" ], [ "åĨį", "åŃ¦ç¿Ĵ" ], [ "åĪ¶", "ç´Ħ" ], [ "åĪĿæľŁ", "èªŀå½Ļ" ], [ "åīĩ", "åĮĸ" ], [ "åīį", "æıĲ" ], [ "åĮº", "åĪĩ" ], [ "åĮĸ", "ãģĹãģŁ" ], [ "åı¯", "éĢĨ" ], [ "åĲ«", "ãĤģãĤĭ" ], [ "åĳ", "¼" ], [ "åĽł", "æŀľ" ], [ "åľ", "§ç¸®" ], [ "åŁĭãĤģè¾¼ãģ¿", "è¡ĮåĪĹ" ], [ "åł´åĲĪ", "ãĢĤ" ], [ "åŃ¦ç¿Ĵ", "ãĤ³ãĥ¼ãĥĳãĤ¹" ], [ "æ", "¯" ], [ "æ", "µ" ], [ "æ", "¼Ķç®Ĺ" ], [ "æ", "ĥħåł±" ], [ "æ", "ł" ], [ "æ¨Ļ", "æľ¬" ], [ "æ¯Ķ", "ä¾ĭ" ], [ "æ¯Ķè¼ĥ", "å¯¾è±¡" ], [ "æīĭ", "æ³ķ" ], [ "æĭ", "¡" ], [ "æĮĩå®ļ", "ãģķãĤĮãģŁ" ], [ "æİ¥", "ç¶ļ" ], [ "æĶ¹", "åĸĦ" ], [ "æķ°åĢ¤", "å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°" ], [ "æĸĩåŃĹ", "ãĤĴ" ], [ "æĸĩåŃĹ", "ãĥ¬ãĥĻãĥ«" ], [ "æĻ", "ĤçĤ¹" ], [ "æĽ", "²ç·ļ" ], [ "æĽ¿", "ãģĪãĤĭ" ], [ "æľ¬", "ä½ĵ" ], [ "æŀ", "Ĳ" ], [ "æŃ£è¦ıåĮĸåīį", "ç½®(" ], [ "æŃ£è¦ıåĮĸåīįç½®(", "Pre-Layer" ], [ "ç¤º", "ãģĻ" ], [ "ç©į", "ãĤĴ" ], [ "ç³»åĪĹ", "åĲį:" ], [ "ç´¯ç©į", "ç¢ºçİĩ" ], [ "çµ¶å¯¾ä½įç½®", "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹" ], [ "çµĤ", "äºĨ" ], [ "ç½®", "ãģįæıĽãģĪ" ], [ "çĦ¡", "ãģĦ" ], [ "çī", "Ī" ], [ "çĶ¨", ")ãĢĤ" ], [ "çĽ®", "æ¨Ļ" ], [ "çĽ´", "åīį" ], [ "çĽ¸", "å¯¾" ], [ "è", "ī" ], [ "è¡¨", "ç¤º" ], [ "è¢«è¦Ĩ", "çİĩ" ], [ "è§£", "æŀĲ" ], [ "è¨Ń", "è¨Ī" ], [ "èª", "¤" ], [ "èªŃ", "ãģ¿" ], [ "è¶ħãģĪ", "ãģ¦ãģĦãĤĭ" ], [ "èĩª", "åĭķ" ], [ "èŃĺåĪ¥", "ä¸įèĥ½ãģª" ], [ "é", "Į" ], [ "é", "ĺ" ], [ "é£", "½åĴĮãģ¹ãģįä¹Ĺåīĩ" ], [ "éĢ", "Ł" ], [ "éĢı", "éģİ" ], [ "éĢļ", "å¸¸" ], [ "éģ", "¸" ], [ "éĥ¨åĪĨ", "ç©ºéĸĵ" ], [ "éĩį", "ãģŃ" ], [ "éĮ", "²" ], [ "éĺ", "²" ], [ "éĻ¤", "åİ»" ], [ "éĻ¤", "ç®Ĺ" ], [ "éļ", "İ" ], [ "éļłãĤĮ", "çĬ¶æħĭ" ], [ "éĿŀ", "éĩįè¤ĩ" ], [ "éł", "»åº¦" ], [ "éł", "ĺ" ], [ "éłĺ", "åŁŁ" ], [ "ĊĊ", "**" ], [ "ĊĊ\"\"\"", "theories/ãĥ»apps/" ], [ "ĊĊĠĠĠĠ", "Attention" ], [ "ĊĊĠĠĠĠR", "aises:" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "``" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "x" ], [ "ĊĠĠĠĠ", "2" ], [ "ĊĠĠĠĠ", "d_" ], [ "ĊĠĠĠĠ", "p" ], [ "ĊĠĠĠĠ", "plot_" ], [ "ĊĠĠĠĠ", "ãĤĴ" ], [ "ĊĠĠĠĠ\"", "plot_" ], [ "ĊĠĠĠĠ\"\"\"", "åŃ¦ç¿Ĵ" ], [ "ĊĠĠĠĠlog", "_" ], [ "ĊĠĠĠĠlog", "_scale:" ], [ "ĊĠĠĠĠĠĠĠĠ", "1" ], [ "ĊĠĠĠĠĠĠĠĠ", "ValueError:" ], [ "ĊĠĠĠĠĠĠĠĠ", "dtype:" ], [ "ĊĠĠĠĠĠĠĠĠ", "history[\"" ], [ "ĊĠĠĠĠĠĠĠĠ", "key_length:" ], [ "ĊĠĠĠĠĠĠĠĠ", "model:" ], [ "ĊĠĠĠĠĠĠĠĠ", "param" ], [ "ĊĠĠĠĠĠĠĠĠ", "positional_transform:" ], [ "ĊĠĠĠĠĠĠĠĠ", "query_length:" ], [ "ĊĠĠĠĠĠĠĠĠ", "residuals" ], [ "ĊĠĠĠĠĠĠĠĠ", "seq_len" ], [ "ĊĠĠĠĠĠĠĠĠax", ".plot" ], [ "ĊĠĠĠĠĠĠĠĠax.set_", "yscale(\"log\")" ], [ "ĊĠĠĠĠĠĠĠĠb", "idirectional" ], [ "ĊĠĠĠĠĠĠĠĠbia", "s:" ], [ "ĊĠĠĠĠĠĠĠĠlog", "_scale:" ], [ "ĊĠĠĠĠĠĠĠĠself.", "norm" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "False" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "np.log(warm_fit" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "positions:" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "self." ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "f\"" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "self." ], [ "Ġ", "decoder-only" ], [ "Ġ", "eps" ], [ "Ġ", "gradient_clip_threshold" ], [ "Ġ", "loss_" ], [ "Ġ", "or" ], [ "Ġ", "params" ], [ "Ġ", "point" ], [ "Ġ", "positions" ], [ "Ġ", "sorted_" ], [ "Ġ", "to" ], [ "Ġ", "y" ], [ "Ġ", "ãģ¾ãģŁãģ¯" ], [ "Ġ", "ãĤ¹ãĤ±ãĤ¸ãĥ¥ãĥ¼ãĥ«" ], [ "Ġ", "ãĤĪãĤĬ" ], [ "Ġ", "ä¸Ńéĸĵå±¤" ], [ "Ġ", "ä¿Ĥæķ°" ], [ "Ġ", "æ³ķ" ], [ "Ġ", "æķ°åĢ¤å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°" ], [ "Ġ", "æĸ¹å¼ı" ], [ "Ġ", "çĶŁæĪĲ" ], [ "Ġ", "è¨Ģèªŀãĥ¢ãĥĩãĥ«ãģ®" ], [ "Ġ(", "0." ], [ "Ġ(", "2" ], [ "Ġ(", "n" ], [ "Ġ(", "name," ], [ "Ġ(", "np." ], [ "Ġ(", "x" ], [ "Ġ0", "1" ], [ "Ġ1", "))" ], [ "Ġ1e", "-1" ], [ "Ġ2", "," ], [ "Ġ4", ".0" ], [ "ĠA", "PI" ], [ "ĠA", "pproach" ], [ "ĠBPE", "IDTokenizer" ], [ "ĠC", "haracterLevelTokenizer" ], [ "ĠCallab", "le" ], [ "ĠCallable[", "[int]," ], [ "ĠH", "ub" ], [ "ĠI", "ter" ], [ "ĠLongTensor", "(" ], [ "ĠNone", "]" ], [ "ĠP", "air" ], [ "ĠPath", "(" ], [ "ĠQuery", "KeyPositionalTransform" ], [ "ĠS", "aturatingPowerLawFit" ], [ "ĠSequence[str", "]," ], [ "ĠV", ")" ], [ "ĠW", "_2" ], [ "Ġ\\\\", "frac{" ], [ "Ġ\\\\", "text{" ], [ "Ġ\\\\", "theta" ], [ "Ġ_", "to_numpy" ], [ "Ġ``", "0.0" ], [ "Ġ``", "_" ], [ "Ġ``", "beam_size" ], [ "Ġ``", "d_model``" ], [ "Ġ``\"", "loss_step_delta" ], [ "Ġ``(", "S_q," ], [ "Ġ``{", "ç³»åĪĹåĲį:" ], [ "Ġa", "^K_mn" ], [ "Ġax", "is" ], [ "Ġb", "eam" ], [ "Ġbyt", "e_" ], [ "Ġc", "allable" ], [ "Ġc", "ross_" ], [ "Ġexist_ok", "=True)" ], [ "Ġfloat", "(\"" ], [ "Ġfloat(np.", "median" ], [ "Ġh", "uber_delta" ], [ "Ġh", "ãĢĤ" ], [ "Ġk", "_clip" ], [ "Ġkeepdim", "=True)" ], [ "Ġl", "_inf" ], [ "Ġlen(", "color_cycle" ], [ "Ġlist", "(" ], [ "Ġlog", "_scale:" ], [ "Ġm", "anifest_path" ], [ "Ġm", "atplotlib" ], [ "Ġm", "emory_" ], [ "Ġn", "oise_band" ], [ "Ġnn.", "Parameter(torch." ], [ "Ġnp.", "arange(" ], [ "Ġnp.", "isfinite(" ], [ "Ġnp.array(", "[" ], [ "Ġo", "bject" ], [ "Ġparam", "s," ], [ "Ġrange(", "n" ], [ "Ġs", "in" ], [ "Ġself.", "_" ], [ "Ġself.", "_m" ], [ "Ġself.", "eps)" ], [ "Ġseq_len", "," ], [ "Ġsqrt(", "d_k)" ], [ "Ġstrict", "=True" ], [ "Ġsymbol", "_to_id" ], [ "Ġt", "gt_mask" ], [ "Ġth", "reshold" ], [ "Ġtorch.", "dtype," ], [ "Ġtorch.", "s" ], [ "Ġtorch.", "t" ], [ "ĠÎ", "¸" ], [ "Ġãģ¤ãģ®", "æĸĩåŃĹåĪĹ" ], [ "Ġãģ§è¿½åĬł", ")" ], [ "Ġãģ«", "å¯¾å¿ľãģĻãĤĭ" ], [ "Ġãģ«", "ç½®ãģįæıĽãģĪ" ], [ "Ġãģ®", "ãĥĺãĥĥãĥīæķ°" ], [ "Ġãģ®", "å®ļç¾©" ], [ "Ġãģ®", "åħ¥åĬĽãĢĤ" ], [ "Ġãģ¾", "ãģ§ãģ®éĢ£çķª" ], [ "ĠãģĭãĤī", "åĨįåĪ©çĶ¨ãģĻãĤĭ" ], [ "ĠãĤ¹ãĥĨãĥĥãĥĹ", "ãģĶãģ¨" ], [ "ĠãĤĴ", "åĬłç®Ĺ" ], [ "Ġãĥ", "ģ" ], [ "Ġãĥģ", "ãĥ£ãĥ³ãĤ¯" ], [ "ĠãĥĲ", "ãĥĥãĥģ" ], [ "ĠãĥĲãĤ¤ãĥĪ", "æķ°" ], [ "ĠãĥĲãĤ¤ãĥĪ", "éķ·" ], [ "Ġãĥĳãĥ©ãĥ¡", "ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ" ], [ "Ġä»¥", "éĻį" ], [ "Ġå", "±" ], [ "Ġåģ´", "ãģ®" ], [ "Ġæ", "§ĭæĪĲ" ], [ "ĠæŃ£", "åīĩåĮĸ" ], [ "Ġç", "ķ" ], [ "Ġç¸¦", "è»¸ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ" ], [ "Ġçķ", "ªçĽ®" ], [ "ĠĠĠĠĠĠĠĠ", "ĠĠĠ:" ], [ "ĠĠĠĠĠĠĠĠ", "ĠĠĠĠĠ:" ], [ "Ĥ", "åĬł" ], [ "ĳ", "¼ãģ³åĩºãģĹ" ], [ "Ļ", "æºĸ" ], [ "Ļ", "ĤçĤ¹" ], [ "ŀ", "æĢ§" ], [ "!", "r" ], [ "\"", "axes" ], [ "\"", "en" ], [ "\"]", ".b" ], [ "\"].b", "y_" ], [ "\"].by_", "key" ], [ "\"].by_key", "()" ], [ "\"].by_key()", "[\"" ], [ "\"].by_key()[\"", "color" ], [ "\"].by_key()[\"color", "\"]" ], [ "\"axes", "." ], [ "\"axes.", "prop" ], [ "\"axes.prop", "_cycle" ], [ "\"axes.prop_cycle", "\"].by_key()[\"color\"]" ], [ "(", "Byte" ], [ "(", "C" ], [ "(", "N," ], [ "(", "T" ], [ "(", "W" ], [ "(", "de" ], [ "(", "hidden" ], [ "(", "residual_fn" ], [ "(", "ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°" ], [ "(", "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶" ], [ "(", "å®Łé¨ĵ" ], [ "(", "åĭ¾éħį" ], [ "(", "çľģçķ¥åı¯" ], [ "().item", "())" ], [ "(00", "1" ], [ "(00", "3" ], [ "(006", "ãĢģ" ], [ "(1", "-indexed" ], [ "(2", "i" ], [ "(L", "ayer" ], [ "(N,", "D" ], [ "(N,D", ")" ], [ "(``numpy", "." ], [ "(``src/", "data/" ], [ "(b", "atch" ], [ "(g", "," ], [ "(m", ":" ], [ "(p", "re-" ], [ "(path", ")" ], [ "(re-", "centering" ], [ "(se", "quence" ], [ "(self.", "beta" ], [ "(self.", "f" ], [ "(x", ")))" ], [ "(æĹ¢å®ļåĢ¤", ")ãģ®" ], [ ")", "\"," ], [ ")ãĢĤ``None``", "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯" ], [ ")ãģ«", "å¯¾ãģĹãģ¦" ], [ "*", "?" ], [ "**", "t" ], [ "-", "Layer" ], [ "-1", "," ], [ "-P", "roduct" ], [ "-gram", "s" ], [ "-in", "f\"" ], [ "-n", "_val" ], [ ".", "*?" ], [ ".", "dim" ], [ ".", "g" ], [ ".", "grad" ], [ ".", "min" ], [ ".", "power_law_fit" ], [ ".", "rc" ], [ ".dim", "()" ], [ ".items()", ":" ], [ ".m", "ul_" ], [ ".mean(dim", "=-1," ], [ ".mul_", "(self.beta" ], [ ".rc", "Param" ], [ ".rcParam", "s[" ], [ ".rcParams[", "\"axes.prop_cycle\"].by_key()[\"color\"]" ], [ ".re", "quest" ], [ ".s", "c" ], [ ".s", "qrt(" ], [ ".size(", "1)" ], [ ".txt", "\"" ], [ "/", "D" ], [ "/", "N" ], [ "/D", "^" ], [ "/D^", "beta" ], [ "/N", "^" ], [ "/N^", "alpha" ], [ "00", "1" ], [ "004", "_norm" ], [ "1", "(" ], [ "1", ")ãĢĤ" ], [ "1", "2," ], [ "1", "6" ], [ "3", ")" ], [ "99", "9" ], [ "<", "language" ], [ "" ], [ "=", "1)" ], [ "=", "repo_id" ], [ "=1", "e" ], [ "=repo_id", "," ], [ "=str", "(" ], [ ">", "\"," ], [ "A", "C" ], [ "A", "CL" ], [ "A", "N" ], [ "A", "bsolute" ], [ "AC", "E_" ], [ "ALiBi", "PositionBias" ], [ "AN", "I" ], [ "ANI", "F" ], [ "ANIF", "ES" ], [ "ANIFES", "T" ], [ "ANIFEST", "_" ], [ "ANIFEST_", "DI" ], [ "ANIFEST_DI", "R" ], [ "A_M", "ANIFEST_DIR" ], [ "Absolute", "Positional" ], [ "AbsolutePositional", "Embedding" ], [ "BPE", "IDTokenizer" ], [ "C", "A" ], [ "Chinchilla", "ParametricFit" ], [ "DI", "A_MANIFEST_DIR" ], [ "ES", "P" ], [ "ESP", "ACE_" ], [ "EncoderBlock", "``" ], [ "F", "F" ], [ "F", "rontierResult" ], [ "GPT", "-2" ], [ "GPTLanguageModel", "``" ], [ "I", "RLS" ], [ "I", "mp" ], [ "L", "o" ], [ "L2", "R" ], [ "L2R", "egularization" ], [ "M", "ediaWiki" ], [ "MultiHeadAttention", "``" ], [ "O", "N" ], [ "Optimizer", "``" ], [ "PE", "DIA_MANIFEST_DIR" ], [ "Po", "st" ], [ "Positional", "Encoding" ], [ "Post", "-Layer" ], [ "Q", "," ], [ "Query", "ãĥ»Key" ], [ "S", "ON" ], [ "S", "plit" ], [ "Split", "Mod" ], [ "T5", "RelativePositionBias" ], [ "Tensor", "]" ], [ "WIKI", "PEDIA_MANIFEST_DIR" ], [ "X", "iv" ], [ "Xiv", ":" ], [ "Y", "ou" ], [ "[", "j" ], [ "[", "start" ], [ "[1", "]" ], [ "[i", "]" ], [ "[str", "]" ], [ "\\\\", "alpha" ], [ "\\\\hat{", "m}_t" ], [ "\\s", "*" ], [ "^", "(" ], [ "^", "t)" ], [ "_", "den" ], [ "_", "exponent" ], [ "_", "grid" ], [ "_", "length_penalty" ], [ "_", "max" ], [ "_", "position" ], [ "_", "pow" ], [ "_", "ro" ], [ "_", "sequences" ], [ "_", "ylabel:" ], [ "_2", "_slopes" ], [ "_an", "d_" ], [ "_and_", "activation" ], [ "_b", "lock" ], [ "_b", "ootstrap" ], [ "_by_", "unit_" ], [ "_by_unit_", "group" ], [ "_count", ":" ], [ "_curv", "es:" ], [ "_cycle[", "i" ], [ "_den", "om" ], [ "_h", "e" ], [ "_he", "at" ], [ "_head", "s(" ], [ "_heat", "map" ], [ "_i", ")" ], [ "_i", "^" ], [ "_list", ".append(" ], [ "_log_prob", "s," ], [ "_m", "atch_rat" ], [ "_m", "edian" ], [ "_m", "in" ], [ "_n", "grams" ], [ "_n", "on_embedding_param" ], [ "_nam", "es" ], [ "_non_embedding_param", "eter" ], [ "_p", "red" ], [ "_p", "ret" ], [ "_param", "s)" ], [ "_plain", "text(" ], [ "_pret", "raining" ], [ "_ratio", ":" ], [ "_s", "hakespear" ], [ "_se", "ed_" ], [ "_seed_", "sc" ], [ "_seed_sc", "atter" ], [ "_spl", "it" ], [ "_split", "_heads(" ], [ "_squar", "e" ], [ "_to", "t" ], [ "_token", "izer_json" ], [ "_tot", "al" ], [ "_trigger", "ed" ], [ "_type", "=\"" ], [ "_v", "[key]" ], [ "_w", "ikitext" ], [ "``", "ãĤĴ" ], [ "``)", "ãģ®åł´åĲĪ" ], [ "``)ãģ®åł´åĲĪ", ":" ], [ "a", ":" ], [ "a", "sel" ], [ "a_body_", "list" ], [ "ab", ":" ], [ "ab", "s(" ], [ "ac", "c" ], [ "acc", "ar" ], [ "accar", "d" ], [ "ack", "([" ], [ "activation_fn", "=" ], [ "ad", "for" ], [ "ad_", "am" ], [ "ad_am", "ount" ], [ "adfor", "d" ], [ "al", "_norm" ], [ "al", "lo" ], [ "al_", "allo" ], [ "al_", "encoding" ], [ "al_allo", "cation" ], [ "al_allocation", "_exponent" ], [ "allback", "_" ], [ "alpha", ":" ], [ "am", "pl" ], [ "an", "g" ], [ "annotat", "e:" ], [ "articl", "e_" ], [ "at", "eg" ], [ "atch", "es" ], [ "ate_", "f" ], [ "ate_f", "lops_per_token" ], [ "attention", "_heatmap" ], [ "attn", "_" ], [ "au", "ss" ], [ "b", "lock" ], [ "b", "ootstrap_scaling_analysis" ], [ "ba", "se" ], [ "byte_", "to_" ], [ "c", "es:" ], [ "c", "i" ], [ "cach", "e" ], [ "cach", "e(" ], [ "cach", "ed" ], [ "categ", "ories" ], [ "char", "s" ], [ "character_level", "_tokenizer_json" ], [ "chinchilla", "_residual" ], [ "chunk_split_mod", "e," ], [ "chunk_split_mode=\"", "whitespace" ], [ "chunk_split_mode=\"whitespace", "\"``" ], [ "clip(log", "_alpha" ], [ "cmap", ":" ], [ "co", "mput" ], [ "co", "s" ], [ "color", "bar" ], [ "con", "d" ], [ "count", "_pair" ], [ "d", "el" ], [ "d", "esign" ], [ "d_model", ")" ], [ "data", "." ], [ "de", "p" ], [ "dep", "th" ], [ "dex[", "pair]" ], [ "di", "ct" ], [ "dist", "in" ], [ "distin", "ct" ], [ "dtype", ")" ], [ "e", "()``" ], [ "e", "=" ], [ "e_", "corpus" ], [ "e_", "init" ], [ "e_", "m" ], [ "e_", "per_" ], [ "e_", "std" ], [ "e_", "x" ], [ "e_m", "n" ], [ "e_per_", "head" ], [ "ed", "=" ], [ "ed", "AbsolutePositionalEmbedding" ], [ "eed", "\"," ], [ "el", "f" ], [ "elf", "Attention" ], [ "em", "ent" ], [ "en", "_" ], [ "encod", "e(" ], [ "encode_", "chunk" ], [ "encode_", "corpus" ], [ "encoding=\"utf-8", "\")" ], [ "eneration", "\"," ], [ "ent", "=False)" ], [ "ent", "ro" ], [ "entro", "py" ], [ "er", "f" ], [ "eries", "_names" ], [ "ers", "ist" ], [ "ersist", "ent=False)" ], [ "est", "im" ], [ "estim", "ate_flops_per_token" ], [ "estyle=\"", "--" ], [ "et_", "learning_rat" ], [ "eval", "_step" ], [ "evaluate_bits_per_byt", "e``" ], [ "evaluation_mask", ":" ], [ "evaluation_window", "s:" ], [ "exact", "_match_rat" ], [ "f", "ertility" ], [ "f", "i" ], [ "feed_forward_factory", "``" ], [ "fit_", "power_law" ], [ "fit_chinchilla_parametric", "``" ], [ "ful", "l" ], [ "g", "/" ], [ "g", "ist" ], [ "g", "n" ], [ "gist", "er" ], [ "gn", "mt" ], [ "gnmt", "_length_penalty" ], [ "gra", "y" ], [ "gradient_norm", "_by_unit_group" ], [ "gradient_norm", "s:" ], [ "gram", "s)" ], [ "gray", "\"," ], [ "h", "unk" ], [ "ha", "pe(" ], [ "hunk", "SplitMod" ], [ "ia", "s_correction" ], [ "id", "ent" ], [ "id_", "to_" ], [ "id_to_", "char" ], [ "in", "clude_" ], [ "in", "e_" ], [ "in", "it_alpha" ], [ "in", "v" ], [ "init", "_beta" ], [ "inusoidal", "PositionalEncoding" ], [ "inv", "_freq" ], [ "ipyn", "b``" ], [ "is", "m" ], [ "it_", "se" ], [ "it_se", "cond" ], [ "ith", "L2Regularization" ], [ "iv", "e" ], [ "iz", "e_per_head" ], [ "json.", "load" ], [ "json.load", "s(" ], [ "l", "ong" ], [ "label", "=\"" ], [ "languag", "e:" ], [ "language", "}" ], [ "learning_rat", "e(" ], [ "len", "am" ], [ "lin", "es" ], [ "linalg.", "norm" ], [ "lo", "w" ], [ "logit", "s," ], [ "logit", "s``" ], [ "lop", "s" ], [ "los", "s)" ], [ "los", "s," ], [ "losest", "_pow" ], [ "loss_step_delta", "s:" ], [ "lr", ":" ], [ "m", "ul_" ], [ "mask", ")" ], [ "mask", "]" ], [ "max_len", "," ], [ "max_new_token", "s)" ], [ "max_new_token", "s:" ], [ "mean", "_square" ], [ "mo", "ve_" ], [ "mod", "ule" ], [ "n_", "query" ], [ "norm", "s" ], [ "num", "_heads" ], [ "of", "_2_slopes" ], [ "ois", "e_std" ], [ "old", "id" ], [ "on", "-" ], [ "on_embedding", "_params" ], [ "ones", "_like(" ], [ "optim", "al_allocation_exponent" ], [ "optimizer", "." ], [ "or", "g/" ], [ "ot", "-Product" ], [ "ou", "s_" ], [ "oundaries", "_a" ], [ "out", ")" ], [ "p", ")" ], [ "p", "90" ], [ "p", "i" ], [ "p", "ie" ], [ "p", "m" ], [ "p", "red" ], [ "p.", "add_" ], [ "path.write_", "text(" ], [ "po", "s," ], [ "pow", "(2)" ], [ "power_", "of_2_slopes" ], [ "pro", "b" ], [ "profil", "es:" ], [ "r", "(" ], [ "r", "int" ], [ "ran", "ge(" ], [ "ratio", "s" ], [ "re", "construct_optimal_frontier" ], [ "re", "gister" ], [ "re", "move_" ], [ "re", "vid" ], [ "relative_position", "_bucket" ], [ "repo_id", "=repo_id," ], [ "res", "hape(" ], [ "res", "s" ], [ "residual", "s)" ], [ "residual_fn", ":" ], [ "residuals", "**2" ], [ "retr", "y_" ], [ "retry_", "wa" ], [ "retry_wa", "it_second" ], [ "rom", "ot" ], [ "rontier_", "log_n_opt" ], [ "row", "s." ], [ "s", "\"" ], [ "s", "\":" ], [ "s", "=" ], [ "s", "core" ], [ "s", "eries_names" ], [ "s", "ize_per_head" ], [ "s", "queeze(" ], [ "s", "rc." ], [ "s.", "items()" ], [ "s.", "reshape(" ], [ "s.tolist", "()" ], [ "s_", "train" ], [ "s_", "training" ], [ "self.", "w_" ], [ "self_attn", "``" ], [ "seq_len", "," ], [ "src/", "data/" ], [ "st", "ack([" ], [ "t", "ab:" ], [ "target", "_compute_budgets:" ], [ "text", "," ], [ "th", "er_" ], [ "theories/01_foundations/", "004_norm" ], [ "tin", "y" ], [ "tiny", "_shakespear" ], [ "tokenizer", ":" ], [ "tokenizer.", "char_to_id" ], [ "torch.", "nn.Parameter" ], [ "torch.", "no_grad()" ], [ "torch.optim.", "Optimizer``" ], [ "total", "_bytes:" ], [ "total_negative_log_likelihood_n", "ats" ], [ "total_step", "s``" ], [ "training", "." ], [ "training", "/" ], [ "u", "s)" ], [ "ual", "_seed" ], [ "um", "_log_prob" ], [ "un", "squeeze(" ], [ "unigram", "_model" ], [ "unk_token", ":" ], [ "use_cross_attention", "=False" ], [ "v", "_hat" ], [ "valu", "e:" ], [ "valu", "e_" ], [ "values_by_condition", ":" ], [ "vertex_log_n", "," ], [ "vi", "ous_" ], [ "vocab", ")" ], [ "vocabulary", "_size:" ], [ "wa", "s_training" ], [ "warmup_step", "s" ], [ "weight", "_grad" ], [ "weight_decay", ":" ], [ "wikipedia_manifest", "s/" ], [ "x_c", "," ], [ "y", "(" ], [ "y", "," ], [ "}", "\"" ], [ "¢", "ç´" ], [ "¢", "ĥçķĮ" ], [ "¢ç´", "¢" ], [ "§ĭ", "ç¯ī" ], [ "¬", "æĺİ" ], [ "®", "æ¨Ļ" ], [ "®µ", "éļİ" ], [ "±", "º" ], [ "²", "æ³ķ" ], [ "³", "å¯" ], [ "³å¯", "Ĩ" ], [ "¶", "æķ°" ], [ "¾", "¤" ], [ "ãĢģ", "006" ], [ "ãĢģ", "009" ], [ "ãĢģ", "ãģĿãĤĮ" ], [ "ãĢģ", "åĲĮãģĺ" ], [ "ãĢĤ", "åŃĺåľ¨ãģĹãģªãģĦ" ], [ "ãĢĤåŃĺåľ¨ãģĹãģªãģĦ", "åł´åĲĪãģ¯" ], [ "ãĢĮ", "æľ¬ä½ĵ" ], [ "ãĢľ", "004" ], [ "ãģ", "¶" ], [ "ãģ¡", "ãĢģ" ], [ "ãģ£", "ãģ¦ãģĦ" ], [ "ãģ¤", "ãģ¤" ], [ "ãģ¤", "ãģį" ], [ "ãģ¦", "ãĤĤ" ], [ "ãģ§", "ãģªãģĦ" ], [ "ãģ§", "åº" ], [ "ãģ§ãģĤãĤĭ", "å¿ħè¦ģãģĮãģĤãĤĭãĢĤ" ], [ "ãģ§åº", "ĥ" ], [ "ãģ¨", "åĩºåĬĽå±¤" ], [ "ãģ¨ãĤģ", "ãģ¦" ], [ "ãģª", "ãģ®" ], [ "ãģª", "ãĤĬãĢģ" ], [ "ãģªãģ©", ")" ], [ "ãģªãģ©", ")ãĤĴ" ], [ "ãģªãģĦ", "(" ], [ "ãģ«", "ãģ¯ãĢģ" ], [ "ãģ«", "ãģĹãģŁ" ], [ "ãģ«", "å¯¾å¿ľãģĻãĤĭ" ], [ "ãģ«", "åĩºçı¾" ], [ "ãģ«ãĤĪ", "ãĤīãģļ" ], [ "ãģ«ãĤĪ", "ãĤĬãĢģ" ], [ "ãģ«å¯¾å¿ľ", "ãģĻãĤĭãĢĤ" ], [ "ãģ«åŁºãģ¥ãģı", "ãĢĤ" ], [ "ãģ®", "äºĮä¹Ĺ" ], [ "ãģ®", "ä½įç½®" ], [ "ãģ®", "å®ļç¾©" ], [ "ãģ®", "åıĸå¾Ĺ" ], [ "ãģ®", "è¨ĺäºĭ" ], [ "ãģ®", "è¾ŀæĽ¸" ], [ "ãģ®", "éłĨåºı" ], [ "ãģ®ãĥªãĤ¹ãĥĪ", "}``" ], [ "ãģ®å", "Ĭ" ], [ "ãģ®å", "įĺ" ], [ "ãģ®å", "ĳ¼ãģ³åĩºãģĹ" ], [ "ãģ®å", "Ł" ], [ "ãģ®å¤", "ī" ], [ "ãģ®å¤§", "ãģį" ], [ "ãģ®å°İ", "åĩº" ], [ "ãģ®åĭ¾éħį", "ãĤĴ" ], [ "ãģ®åł´åĲĪ", "ãģ¯" ], [ "ãģ®æ", "Ŀ" ], [ "ãģ®æ¨", "Ļæºĸ" ], [ "ãģ®æ¨Ļæºĸ", "åģıå·®" ], [ "ãģ®æī", "ĭ" ], [ "ãģ®æĹ", "¢" ], [ "ãģ®æĿ", "¡ä»¶" ], [ "ãģ¯", "çĦ¡è¦ĸ" ], [ "ãģ°", "ãĤī" ], [ "ãģ°ãĤī", "ãģ¤ãģį" ], [ "ãģ¹ãģįä¹Ĺåīĩ", "ãģĤãģ¦ãģ¯ãĤģ" ], [ "ãģ¿", "ãĤĴ" ], [ "ãģĤãģ¦ãģ¯ãĤģ", "(" ], [ "ãģĦ", ")ãĢĤ" ], [ "ãģĪ", "ãĢģ" ], [ "ãģĪ", "ãģ°" ], [ "ãģĪ", "ãģŁ" ], [ "ãģĭ", "ãĤĴ" ], [ "ãģĭãĤī", "éĻ¤å¤ĸ" ], [ "ãģĮ", "ãĢĮ" ], [ "ãģĮ", "ä¸Ģèĩ´" ], [ "ãģĮ", "å¤§ãģįãģĦ" ], [ "ãģĮ", "é«ĺãģĦ" ], [ "ãģĮãģĤ", "ãĤĭãĢĤ" ], [ "ãģı", ")ãĢĤ" ], [ "ãģĵ", "ãģ®å¼ķæķ°" ], [ "ãģĵ", "ãģ®åĢ¤ãĤĴ" ], [ "ãģĵãģ¨", "ãĤĴæĦıåĳ³" ], [ "ãģĵãĤĮãĤĴ", "è¶ħãģĪãĤĭ" ], [ "ãģĹ", "ãģĨãĤĭ" ], [ "ãģĹãģ¦", "ãĤŃãĥ£ãĥĥãĤ·ãĥ¥" ], [ "ãģĹãģŁ", "è¨ĺäºĭ" ], [ "ãģĻ", "ãĤĮãģ°" ], [ "ãģĻãĤĭ", "(006" ], [ "ãģĻãĤĭ", "éļĽ" ], [ "ãģļ", "ãĢģ" ], [ "ãģĽ", "ãĤĭ" ], [ "ãģŁãĤģ", "ãĢģ``" ], [ "ãģłãģĳ", "ãģ§" ], [ "ãĤ£", "ãĥª" ], [ "ãĤ£", "ãĥ«" ], [ "ãĤ£ãĥª", "ãĥĨãĤ£" ], [ "ãĤ£ãĥ«", "ãĤ¿" ], [ "ãĤ¤", "ãĤ¹ãĢĤ" ], [ "ãĤ¤ãĥ³", "ãĤ¹ãĤ¿ãĥ³ãĤ¹" ], [ "ãĤ¨", "ãĥ©" ], [ "ãĤ¯ãĥª", "ãĥĥãĥĹ" ], [ "ãĤ°", "ãĥ©ãĥ" ], [ "ãĤ°ãĥ©ãĥ", "ķ" ], [ "ãĤ¹ãĥĨãĥĥãĥĹ", "ãģ§" ], [ "ãĤ¹ãĥĨãĥĥãĥĹ", "çķªåı·" ], [ "ãĤ¹ãĥĪãĥ©", "ãĥĥãĥĹ" ], [ "ãĤĦ", "ãģĻ" ], [ "ãĤī", "ãĤĮãĤĭ" ], [ "ãĤīãĤĮ", "ãģŁ" ], [ "ãĤīãĤĮ", "ãĤĭãĢĤ" ], [ "ãĤĭãĢĤ", "ãģĵãģ®" ], [ "ãĤĴ", "ä½¿" ], [ "ãĤĴ", "ä½¿ãģĨ" ], [ "ãĤĴ", "æĮģãģŁ" ], [ "ãĤĴ", "è¨Īç®ĹãģĻãĤĭãĢĤ" ], [ "ãĤĴ", "è©¦" ], [ "ãĤĴ", "è¿ĶãģĻ" ], [ "ãĤĴ", "éĢ£çµĲãģĹãģŁ" ], [ "ãĤĴãģ¾ãģŁ", "ãģĦ" ], [ "ãĤĴåĲ«", "ãĤĢ" ], [ "ãĤĴæ", "®ĭ" ], [ "ãĤĴæ", "İ" ], [ "ãĤĴæĥ³å®ļ", ")ãĢĤ" ], [ "ãĤĴè¨Īç®Ĺ", "ãģĻãĤĭ(007" ], [ "ãĤŃ", "ãĥ¼" ], [ "ãĥ", "Ģ" ], [ "ãĥ¡", "ãĤ¤ãĥ³" ], [ "ãĥ¡", "ãĥ¢" ], [ "ãĥ¦", "ãĥ¼ãĥ" ], [ "ãĥ¦ãĥ¼ãĥ", "Ĩ" ], [ "ãĥ¦ãĥ¼ãĥĨ", "ãĤ£ãĥªãĥĨãĤ£" ], [ "ãĥ©", "ãĤ¤ãĥ³" ], [ "ãĥ©ãĥ³ãĥĢ", "ãĥł" ], [ "ãĥ©ãĥĥãĥĳ", "ãĥ¼ãĢĤ" ], [ "ãĥ«", "ãĥ»" ], [ "ãĥ³ãĥĹ", "ãĥ¬ãĥ¼ãĥĪ" ], [ "ãĥ»", "ALiBi" ], [ "ãĥ»", "D" ], [ "ãĥ»", "å®Łé¨ĵ" ], [ "ãĥ»00", "2" ], [ "ãĥ¼ãĥ", "ŀ" ], [ "ãĥ¼ãĥĪ", "ãĤ¹ãĥĪãĥ©ãĥĥãĥĹ" ], [ "ãĥĨ", "ãĥ³ãĥĹãĥ¬ãĥ¼ãĥĪ" ], [ "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵ", "ãģ®" ], [ "ãĥĩ", "ãĥĲ" ], [ "ãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥª", "ãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯" ], [ "ãĥĩãĥĲ", "ãĤ¤ãĤ¹ãĢĤ" ], [ "ãĥī", "ãĥ¡ãĤ¤ãĥ³" ], [ "ãĥĲãĤ¤ãĥĪ", "åĪĹ" ], [ "ãĥĳãĤ¤", "ãĥĹ" ], [ "ãĥĳãĤ¤ãĥĹ", "ãĥ©ãĤ¤ãĥ³" ], [ "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "ç¾¤" ], [ "ãĥĳãĥĩãĤ£ãĥ³ãĤ°", "ãĥŀãĤ¹ãĤ¯" ], [ "ãĥķ", "ãĥŃãĥ³ãĥĨãĤ£ãĤ¢" ], [ "ãĥķãĤ¡ãĤ¤ãĥ«", "åĲį" ], [ "ä¸", "¸" ], [ "ä¸¡", "æĸ¹" ], [ "ä¸Ĭ", "ãģĴ" ], [ "ä¸įæŃ£", "ãģª" ], [ "ä¹±æķ°", "ãĤ·ãĥ¼ãĥī" ], [ "ä¹ĸ", "éĽ¢" ], [ "ä»", "£" ], [ "ä»¥", "ä¸ĭ" ], [ "ä»¥å¤ĸ", "ãģ®" ], [ "ä½", "İ" ], [ "ä½", "ķ" ], [ "ä½įç½®", "å¤īæıĽ" ], [ "ä½ķ", "ãĤĤ" ], [ "ä¾ĭ", "ãģĪãģ°" ], [ "ä¾ĭ", "å¤ĸ" ], [ "ä¿", "ĿåŃĺ" ], [ "å", "ĥ" ], [ "å", "Ħ" ], [ "å", "ŀ" ], [ "å®Įåħ¨", "ä¸Ģèĩ´" ], [ "å±", "ķ" ], [ "å±¤æŃ£è¦ıåĮĸ", "(Layer" ], [ "å±ķ", "éĸ" ], [ "å±ķéĸ", "ĭ" ], [ "å¸¸ãģ«è²ł", "ãģ®ãĥ¦ãĥĭãĥĥãĥĪ" ], [ "å¹", "ħ" ], [ "å¹³åĿĩæ¸Ľç®Ĺ", "(re-centering" ], [ "å¹³æĸĩ", "ãģ«å¤īæıĽ" ], [ "å¼", "·åº¦" ], [ "å½¢", "å¼ı" ], [ "åģ", "¶æķ°" ], [ "åģ´", "ãģ§" ], [ "åħ¥", "ãĤĮ" ], [ "åħ¥åĬĽ", "åŁĭãĤģè¾¼ãģ¿" ], [ "åħ¨ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "ãģ®åĭ¾éħįãĤĴ" ], [ "åħ¬", "å¹³" ], [ "åħĪ", "ãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯" ], [ "åĨħ", "ãģ®" ], [ "åĨħ", "å®¹" ], [ "åĨį", "çı¾" ], [ "åĨį", "è©¦è¡Į" ], [ "åī²", "ãĤĬå½ĵãģ¦" ], [ "åī²", "åĲĪ" ], [ "åįĺä½į", "ãģ®" ], [ "åı¯éĢĨ", "æĢ§" ], [ "åıĸãĤĬ", "åĩº" ], [ "åĲ«ãģ¾", "ãĤĮãĤĭ" ], [ "åĲĦ", "ä½įç½®" ], [ "åĳ¼ãģ³åĩºãģĹåģ´", "ãģĮ" ], [ "åķı", "é¡" ], [ "åķıé¡", "Į" ], [ "åľ§ç¸®", "åĬ¹çİĩ" ], [ "åŀ", "ĭ" ], [ "åŁº", "æºĸ" ], [ "åł´åĲĪ", "ãĢģ``" ], [ "åł´åĲĪ", "ãģ«" ], [ "åŃ", "Ĳ" ], [ "åŃ¦ç¿Ĵ", "ãĥ«ãĥ¼ãĥĹ" ], [ "æ", "¥µ" ], [ "æ", "§ĺ" ], [ "æ", "±ĤãĤģ" ], [ "æ", "ģ" ], [ "æ", "į" ], [ "æ©Ł", "æ§ĭ" ], [ "æ¬", "²æ³ķ" ], [ "æ¬", "ł" ], [ "æ®ĭ", "ãĤĬ" ], [ "æ³¨æĦı", "æ©Łæ§ĭ" ], [ "æ¸¡ãģĻ", "æĥ³å®ļ" ], [ "æ¸Ľè¡°", "çİĩ" ], [ "æºĢ", "ãģŁ" ], [ "æĢ§", "(" ], [ "æĪ", "»" ], [ "æĮ", "¿" ], [ "æİ¨", "è«ĸ" ], [ "æıĥ", "ãģĪãĤĭ" ], [ "æķ°åĢ¤", "çļĦãģ«" ], [ "æĸĩåŃĹ", "æķ°" ], [ "æĸŃ", "çī" ], [ "æĸŃçī", "ĩ" ], [ "æĹ¢", "ãģ«" ], [ "æĹ¢", "å®ļ" ], [ "æĺİç¤º", "çļĦãģ«" ], [ "æĽ¸", "ãģį" ], [ "æľĢ", "åĪĿ" ], [ "æľĢçµĤ", "æ¬¡åħĥ" ], [ "æľī", "åĬ¹" ], [ "æľī", "çĦ¡" ], [ "æŃ", "¢" ], [ "æŃ£è¦ı", "åĪĨå¸ĥ" ], [ "æŃ£è¦ıåĮĸå¾Į", "ç½®(" ], [ "æŃ£è¦ıåĮĸå¾Įç½®(", "Post-Layer" ], [ "ç", "²" ], [ "ç¨", "ĭ" ], [ "ç©ºçĻ½", "ãģ®" ], [ "çµ±è¨Ī", "éĩı" ], [ "çµĲæŀľ", "ãģĮ" ], [ "ç¸", "®å°ı" ], [ "çĦ¡", "åĬ¹" ], [ "çĶ¨", "ãģĦãģŁ" ], [ "çĻº", "æķ£" ], [ "è", "·" ], [ "è", "Ī" ], [ "è¡Į", "ãģĨ" ], [ "è¡Į", "ãĤıãģªãģĦ" ], [ "è¤ĩæķ°", "ãģ®" ], [ "è¦ģç´ł", "ãģ¯" ], [ "è¦ı", "ç´Ħ" ], [ "è§Ĵ", "åĳ¨æ³¢æķ°" ], [ "è¨", "±" ], [ "è¨Ģèªŀãĥ¢ãĥĩãĥ«", "(" ], [ "è¨Īç®Ĺéĩı", "æľĢéģ©" ], [ "è¨ĵç·´", "æĲįå¤±" ], [ "è©ķä¾¡", "å¯¾è±¡ãģ®" ], [ "èª", "¬æĺİ" ], [ "è²", "ª" ], [ "è»", "¢" ], [ "è¾¼", "ãĤĢ" ], [ "èĩª", "ä½ĵãģ¯" ], [ "èĩª", "ä½ľ" ], [ "èĩªåĭķ", "çļĦãģ«" ], [ "éĢ", "Ĳ" ], [ "éĢ", "Ķ" ], [ "éĢĲ", "æ¬¡" ], [ "éģ©çĶ¨", "ãģĻãĤĭãĢĤ" ], [ "éĩįãģ¿", "ä»ĺãģį" ], [ "éĩįãģŃ", "ãģ¦" ], [ "éĺ²", "ãģĲ" ], [ "éĻį", "éłĨãģ«" ], [ "éĻĲ", "ãģªãģı" ], [ "éļĽ", "éĻĲãģªãģı" ], [ "ĊĊ", "#" ], [ "ĊĊ", "``" ], [ "ĊĊĠĠĠĠ", "color_cycle" ], [ "ĊĊĠĠĠĠ", "è¨ĺåı·" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "self." ], [ "ĊĊĠĠĠĠĠĠĠĠĠĠĠĠ", "m" ], [ "ĊĠĠĠĠ", "W" ], [ "ĊĠĠĠĠ", "a" ], [ "ĊĠĠĠĠ", "converged" ], [ "ĊĠĠĠĠ", "d_ff" ], [ "ĊĠĠĠĠ", "else:" ], [ "ĊĠĠĠĠ", "load_" ], [ "ĊĠĠĠĠ", "max_" ], [ "ĊĠĠĠĠ", "plot" ], [ "ĊĠĠĠĠ", "right" ], [ "ĊĠĠĠĠ", "while" ], [ "ĊĠĠĠĠ", "}" ], [ "ĊĠĠĠĠ", "ãģĿãģ®ãģ¾ãģ¾" ], [ "ĊĠĠĠĠ\"", "plot" ], [ "ĊĠĠĠĠax.set_", "xtick" ], [ "ĊĠĠĠĠlog", "_alpha" ], [ "ĊĠĠĠĠĠĠĠĠ", "a" ], [ "ĊĠĠĠĠĠĠĠĠ", "cache_dir:" ], [ "ĊĠĠĠĠĠĠĠĠ", "chunk" ], [ "ĊĠĠĠĠĠĠĠĠ", "d_k" ], [ "ĊĠĠĠĠĠĠĠĠ", "k" ], [ "ĊĠĠĠĠĠĠĠĠ", "lr:" ], [ "ĊĠĠĠĠĠĠĠĠ", "max_chunk_bytes:" ], [ "ĊĠĠĠĠĠĠĠĠ", "params:" ], [ "ĊĠĠĠĠĠĠĠĠ", "path" ], [ "ĊĠĠĠĠĠĠĠĠ", "t" ], [ "ĊĠĠĠĠĠĠĠĠ", "token_ids:" ], [ "ĊĠĠĠĠĠĠĠĠ", "v" ], [ "ĊĠĠĠĠĠĠĠĠ", "values" ], [ "ĊĠĠĠĠĠĠĠĠ", "weight_decay:" ], [ "ĊĠĠĠĠĠĠĠĠbeta", "s:" ], [ "ĊĠĠĠĠĠĠĠĠbia", "s_correction" ], [ "ĊĠĠĠĠĠĠĠĠmax_", "distance" ], [ "ĊĠĠĠĠĠĠĠĠparam", "s_" ], [ "ĊĠĠĠĠĠĠĠĠresiduals", "=" ], [ "ĊĠĠĠĠĠĠĠĠself.", "beta" ], [ "ĊĠĠĠĠĠĠĠĠself.", "eps" ], [ "ĊĠĠĠĠĠĠĠĠself.", "lr" ], [ "ĊĠĠĠĠĠĠĠĠself.", "register" ], [ "ĊĠĠĠĠĠĠĠĠself.", "step_count" ], [ "ĊĠĠĠĠĠĠĠĠself.", "w_" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "[" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "ax" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "continue" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "current" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "f" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "i" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "n" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "relative_position" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "å½¢çĬ¶" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "åł´åĲĪãģ¯" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠnp.log(warm_fit", "." ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "(" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "param" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "best" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "j" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "np.log(" ], [ "Ġ", "005" ], [ "Ġ", "@" ], [ "Ġ", "Decoder" ], [ "Ġ", "Encoder" ], [ "Ġ", "GPTLanguageModel" ], [ "Ġ", "PowerLawFit" ], [ "Ġ", "You" ], [ "Ġ", "_build_" ], [ "Ġ", "attempt" ], [ "Ġ", "char" ], [ "Ġ", "e_mn" ], [ "Ġ", "encoding=\"utf-8\")" ], [ "Ġ", "exp" ], [ "Ġ", "gradient" ], [ "Ġ", "gradient_norm" ], [ "Ġ", "grid" ], [ "Ġ", "json.loads(" ], [ "Ġ", "parabola." ], [ "Ġ", "pos" ], [ "Ġ", "positions:" ], [ "Ġ", "query" ], [ "Ġ", "self_" ], [ "Ġ", "sorted" ], [ "Ġ", "tokenizer.json" ], [ "Ġ", "use_cross_attention" ], [ "Ġ", "val" ], [ "Ġ", "vertex_log_n" ], [ "Ġ", "vocab" ], [ "Ġ", "vocabulary" ], [ "Ġ", "ãģ¸" ], [ "Ġ", "ãģĶãģ¨ãģ«" ], [ "Ġ", "ãĥªãĥĿãĤ¸ãĥĪãĥª" ], [ "Ġ", "åĢ¤" ], [ "Ġ", "åħ¥åĩºåĬĽ" ], [ "Ġ", "åĪĨåī²çµĲæŀľ" ], [ "Ġ", "åıĸå¾Ĺ" ], [ "Ġ", "åŃ¦ç¿Ĵçİĩ" ], [ "Ġ", "æĸ¹åĲĳ" ], [ "Ġ", "æľĢå¤§" ], [ "Ġ", "æŃ£è¦ıåĮĸ" ], [ "Ġ", "è¡ĮåĪĹ" ], [ "Ġ\"", "Attention" ], [ "Ġ\"", "T" ], [ "Ġ\"\"", ".join(" ], [ "Ġ-", "inf" ], [ "Ġ/", "/" ], [ "Ġ0", "," ], [ "Ġ0.", "999" ], [ "Ġ0.0", ":" ], [ "Ġ1", "," ], [ "Ġ1", "0" ], [ "Ġ2", "1)ãĢĤ" ], [ "Ġ4", "29" ], [ "ĠA", "ll" ], [ "ĠB", "ãĢĤ" ], [ "ĠBlock", "(" ], [ "ĠC", "^" ], [ "ĠC", "on" ], [ "ĠCallable[[", "Tensor]," ], [ "ĠCallable[[", "]," ], [ "ĠD", "ot-Product" ], [ "ĠI", "s" ], [ "ĠIsoFLOP", "ParabolaFit" ], [ "ĠJ", "accard" ], [ "ĠN", "eed\"," ], [ "ĠR", "MS" ], [ "ĠR^{", "d_model" ], [ "ĠS", "elfAttention" ], [ "ĠS", "quar" ], [ "ĠS", "wish" ], [ "ĠS", "ãĢĤ" ], [ "ĠS_", "0" ], [ "ĠS_tgt", "," ], [ "ĠSequence", "[int]" ], [ "ĠTensor", "):" ], [ "ĠV", ")``" ], [ "ĠV", "iterbi" ], [ "Ġ\\\\beta", "_" ], [ "Ġ\\\\beta", "_1" ], [ "Ġ\\\\beta", "_2" ], [ "Ġ\\\\theta", "_{t-1}" ], [ "Ġ_", "," ], [ "Ġ_", "count_pair" ], [ "Ġ_", "merge_" ], [ "Ġ_", "power_of_2_slopes" ], [ "Ġ_LOG_ALPHA_BETA_MAX", ")" ], [ "Ġ``", "0" ], [ "Ġ``", "C" ], [ "Ġ``", "log" ], [ "Ġ``", "n" ], [ "Ġ``", "nn." ], [ "Ġ``", "torch." ], [ "Ġ``", "âĸģ" ], [ "Ġ``\"loss_step_delta", "\"``" ], [ "Ġar", "Xiv:" ], [ "Ġb", "ase" ], [ "Ġb", "atch" ], [ "Ġb", "ias_correction" ], [ "Ġb", "its-per-byte" ], [ "Ġb", "pe." ], [ "Ġbest", "_cost" ], [ "Ġbias", "(" ], [ "Ġc", "re" ], [ "Ġc", "urrent" ], [ "Ġcache_dir", "," ], [ "Ġchunk", "s" ], [ "Ġco", "eff" ], [ "Ġcolor", "=color," ], [ "Ġd_ff", "'" ], [ "Ġd_ff", "," ], [ "Ġd_model", "``" ], [ "Ġd_model)``", "ãĢĤ" ], [ "Ġdevice=", "device)" ], [ "Ġdtype=torch.", "bool" ], [ "Ġf", "ertility" ], [ "Ġf", "lops_per_token" ], [ "Ġfloat", "(" ], [ "Ġfloat", "]:" ], [ "Ġfloat(\"", "-inf\"" ], [ "Ġfloat(np.", "exp(log" ], [ "Ġfreq", ")" ], [ "Ġj", "o" ], [ "Ġl", "p" ], [ "Ġlinewidth", "=1." ], [ "Ġlist[float", "]:" ], [ "Ġlog", "it" ], [ "Ġm", "in" ], [ "Ġmanifest_path", "=" ], [ "Ġmask", ")" ], [ "Ġmath", "." ], [ "Ġmath", ".sqrt(" ], [ "Ġmax_", "exact" ], [ "Ġn", ":" ], [ "Ġn", "_layer" ], [ "Ġn", "eg" ], [ "Ġn", "umpy" ], [ "Ġnn.", "Embedding" ], [ "Ġnorm", "ed," ], [ "Ġnp.", "inf" ], [ "Ġnp.", "log" ], [ "Ġnp.", "ones_like(" ], [ "Ġnp.as", "array" ], [ "Ġnp.ndarray", "]," ], [ "Ġnum", "_buckets" ], [ "Ġnum", "_heads," ], [ "Ġp", "ersistent=False)" ], [ "Ġplt", ".rcParams[\"axes.prop_cycle\"].by_key()[\"color\"]" ], [ "Ġq", "." ], [ "Ġre", "vid" ], [ "Ġre.sub(r\"", "<" ], [ "Ġrelative_", "vector" ], [ "Ġs", "cal" ], [ "Ġs", "ize" ], [ "Ġs", "um" ], [ "Ġs", "wish" ], [ "Ġself.", "params:" ], [ "Ġself.", "self_attn" ], [ "Ġself.", "w_" ], [ "Ġself.", "weight_decay" ], [ "Ġself.beta", "2" ], [ "Ġself.dropout", "1(" ], [ "Ġself.dropout", "2" ], [ "Ġself.norm", "1" ], [ "Ġsrc.", "data." ], [ "Ġstr", "]]" ], [ "Ġt", "emperature" ], [ "Ġtorch.", "sqrt(" ], [ "Ġtorch.cat", "([" ], [ "Ġtorch.t", "ensor" ], [ "Ġtorch.zeros_like(", "p)" ], [ "Ġurllib", "." ], [ "Ġvalues", ")" ], [ "Ġx", ")" ], [ "ĠÎ", "³" ], [ "ĠÎ", "·" ], [ "ĠÎ", "¼" ], [ "ĠÎµ", "ãĢĤ" ], [ "ĠâĪ", "Ŀ" ], [ "Ġãģ¤", "ãĤĤ" ], [ "Ġãģ¤ãģ®", "åĪĨåī²" ], [ "Ġãģ¨", "ãģĻãĤĭãĢĤ" ], [ "Ġãģ¨å®Įåħ¨ãģ«", "åĲĮä¸Ģãģ®" ], [ "Ġãģ¨åĲĮãģĺå½¢çĬ¶", "ãģ®" ], [ "Ġãģ¨åĲĮãģĺéķ·ãģķ", ")ãĢĤ" ], [ "Ġãģ«", "ãģªãĤĭ" ], [ "Ġãģ«", "ãģªãĤĭãĢĤ" ], [ "Ġãģ«", "ãģĻãĤĭãĢĤ" ], [ "Ġãģ«", "ãģĿãģ®ãģ¾ãģ¾æ¸¡ãģĻ" ], [ "Ġãģ«", "ä½¿ãģĨ" ], [ "Ġãģ®", "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ" ], [ "Ġãģ®", "ãĥªãĤ¹ãĥĪ" ], [ "Ġãģ®", "ãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ" ], [ "Ġãģ®", "å®Łè£ħæĸ¹éĩĿ" ], [ "Ġãģ®", "åĪĿæľŁåĢ¤" ], [ "Ġãģ®ãģ¨ãģį", "ãģ¯" ], [ "Ġãģ®ãĤ¿ãĥĹãĥ«ãĢĤ", "ãģĦãģļãĤĮãĤĤ" ], [ "Ġãģ®åĩºåĬĽ", ")ãĢĤ" ], [ "Ġãģ®çµĲæŀľ", "ãĢĤ" ], [ "Ġãģ¾ãģ§ãģ®éĢ£çķª", "ãģ¨ãģĹãģ¦æī±" ], [ "ĠãĤ¹ãĤ³ãĤ¢", "ãģ«" ], [ "ĠãĤĴ", "è¶ħãģĪãģ¦ãģĦãĤĭ" ], [ "ĠãĤĴ", "è¶ħãģĪãĤĭ" ], [ "ĠãĤĴæĮģãģ¤", ")ãĢĤ" ], [ "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥", "åħĪãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯" ], [ "Ġãĥ", "ŀ" ], [ "Ġãĥ¢", "ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ" ], [ "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮ", "æ¬¡åħĥ" ], [ "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮ", "æ¬¡åħĥãĢĤ" ], [ "ĠãĥĲãĥĥãĥģ", "ãĤµãĤ¤ãĤº" ], [ "Ġãĥķ", "ãĤ" ], [ "Ġå", "Ĩħ" ], [ "Ġå±", "ŀæĢ§" ], [ "ĠåĢĭ", "ãģ®" ], [ "Ġåģ´", "ãĢģn" ], [ "Ġåģ´", "ãģ®ç³»åĪĹéķ·" ], [ "ĠåĪĨåī²", "å¯¾è±¡ãģ®" ], [ "Ġæ¬¡", "ãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ" ], [ "Ġæ¬¡åħĥ", "ãģ®" ], [ "ĠæľªæºĢ", "ãģ§" ], [ "Ġç", "¾¤" ], [ "ĠçĽ", "®æ¨Ļ" ], [ "Ġèªŀå½ĻãĤµãĤ¤ãĤº", "(vocabulary" ], [ "Ġèĩª", "èº«ãģ®" ], [ "Ģ", "æīĢ" ], [ "Ĥç", "³»åĪĹ" ], [ "\"``", "ãĥ»``" ], [ "\"``ãĥ»``", "\"en" ], [ "\"``ãĥ»``\"en", "\"``" ], [ "'", "'" ], [ "(", "5" ], [ "(", "?" ], [ "(", "AttentionScoreBias" ], [ "(", "De" ], [ "(", "S" ], [ "(", "V" ], [ "(", "con" ], [ "(", "i" ], [ "(", "pos," ], [ "(", "positions." ], [ "(", "query" ], [ "(", "symbol" ], [ "(", "token" ], [ "(", "ãģ¾ãģŁãģ¯" ], [ "(", "ãĤ¹" ], [ "(", "å½¢çĬ¶" ], [ "(", "æ¯Ķè¼ĥå¯¾è±¡" ], [ "(", "èªŀå½Ļ" ], [ "(", "èªŀå½ĻãĤµãĤ¤ãĤº" ], [ "(0", ")}" ], [ "(001", "ãĥ»002" ], [ "(001ãĥ»002", "ãĥ»00" ], [ "(001ãĥ»002ãĥ»00", "3" ], [ "(AttentionScoreBias", "):" ], [ "(L", "2" ], [ "(LayerNorm(x", ")," ], [ "(``", "\"" ], [ "(``", "compute_" ], [ "(``", "for" ], [ "(``", "list" ], [ "(``", "train_language_model``" ], [ "(``for", "ward(" ], [ "(``forward(", "token_ids)" ], [ "(``numpy.", "linalg.lstsq" ], [ "(c", "ross" ], [ "(chunk", ")" ], [ "(id", "s," ], [ "(length", "s_without_leading_whitespace" ], [ "(n", "," ], [ "(n", "ext" ], [ "(n", "orm" ], [ "(new", "_symbol" ], [ "(np.", "arange(" ], [ "(p", ")" ], [ "(p", ")}" ], [ "(positions.", "max" ], [ "(positions.max", "().item())" ], [ "(pre-", "token" ], [ "(pre-token", "ization" ], [ "(x", "))" ], [ "(x", "_arr" ], [ "(çľģçķ¥åı¯", ")ãĢĤ" ], [ ")", "ãģ§ãĢģ" ], [ ")", "ãģ¯ãĢģ" ], [ ")", "ãĤ¹ãĤ¿ãĤ¤ãĥ«" ], [ ")^", "2" ], [ ")``", "ãĢģ" ], [ ")``", "ãĢĤ" ], [ ")ãĢĤ``None``", "(æĹ¢å®ļåĢ¤)" ], [ ")ãģ«", "ãģª" ], [ ")ãĤĴ", "ä½¿ãģĨãĢĤ" ], [ ")ãĤĴ", "å" ], [ ")ãĤĴ", "è¨Īç®ĹãģĻãĤĭãĢĤ" ], [ "**", ":" ], [ "**", "alpha" ], [ ",", ")" ], [ "-", "3" ], [ "-", "A" ], [ "-", "S" ], [ "-", "rn" ], [ "-", "theories" ], [ "-1", "]" ], [ "-A", "g" ], [ "-Ag", "ent" ], [ "-rn", "n" ], [ ".", "SentencePiece" ], [ ".", "ax" ], [ ".", "decod" ], [ ".", "key" ], [ ".", "n" ], [ ".", "org/" ], [ ".", "wikipedia" ], [ ".5", "))" ], [ ".ad", "d" ], [ ".ad", "d_" ], [ ".append(", "float" ], [ ".encode(\"utf-8", "\"))" ], [ ".json", "\"" ], [ ".key", "s()" ], [ ".n", "dim" ], [ ".norm", "al" ], [ ".plot", "(" ], [ ".py", "``" ], [ ".re", "place" ], [ ".read_text(encoding=\"utf-8", "\"))" ], [ ".replace", "(\"" ], [ ".request", "." ], [ ".s", "cal" ], [ ".s", "le" ], [ ".sc", "atter" ], [ ".sle", "ep" ], [ ".t", "rans" ], [ ".to(dtype=", "dtype)" ], [ ".trans", "po" ], [ ".transpo", "se" ], [ ".transpose", "(" ], [ ".wikipedia", ".org/" ], [ "/", "sqrt(" ], [ "0.0", "," ], [ "1", "." ], [ "1", ".0" ], [ "1", ":" ], [ "2", ".0" ], [ "2", "5" ], [ "4", "29" ], [ "4", "4" ], [ "7", "," ], [ "7", "1" ], [ ":", "2]" ], [ ":", "math:`" ], [ "<", "unk" ], [ "" ], [ "=", "-" ], [ "=", "None" ], [ "=", "data[\"" ], [ "=", "lambda" ], [ "=\"", "ident" ], [ "=\"", "y" ], [ "=\"ident", "ity" ], [ "=\"y", "\")" ], [ "=0", "." ], [ ">", "]" ], [ ">", "``" ], [ "A", "ACL" ], [ "AX", "_" ], [ "AX_", "CHUNK_" ], [ "AX_CHUNK_", "CA" ], [ "AX_CHUNK_CA", "CH" ], [ "AX_CHUNK_CACH", "E_" ], [ "AX_CHUNK_CACHE_", "E" ], [ "AX_CHUNK_CACHE_E", "N" ], [ "AX_CHUNK_CACHE_EN", "T" ], [ "AX_CHUNK_CACHE_ENT", "R" ], [ "AX_CHUNK_CACHE_ENTR", "I" ], [ "AX_CHUNK_CACHE_ENTRI", "ES" ], [ "AdamW", "ithL2Regularization" ], [ "B", "C" ], [ "Error", "," ], [ "F", "ile" ], [ "FeedForwardNetwork", "``" ], [ "H", "ol" ], [ "Hol", "t" ], [ "Holt", "z" ], [ "Holtz", "man" ], [ "I", "CLR" ], [ "Imp", "le" ], [ "Imple", "ment" ], [ "Implement", "ed" ], [ "Implemented", "Error" ], [ "IsoFLOP", "ParabolaFit" ], [ "K", "ap" ], [ "Kap", "lan" ], [ "L", "evenberg-Marquardt" ], [ "L", "ik" ], [ "LayerNormalization", "``" ], [ "Learn", "edAbsolutePositionalEmbedding" ], [ "Lo", "ss" ], [ "Loss", "\"," ], [ "M", "AX_CHUNK_CACHE_ENTRIES" ], [ "Module", "``" ], [ "N", "ot" ], [ "Not", "e:" ], [ "O", "ut" ], [ "Optimizer", "Lik" ], [ "P", "ath" ], [ "P", "ress" ], [ "P", "ro" ], [ "P", "y" ], [ "PE", "(pos," ], [ "Pro", "cessor" ], [ "QueryKeyPositionalTransform", "``" ], [ "R", "oPE" ], [ "S", "+" ], [ "S", "," ], [ "S", "equence" ], [ "S", "wiGLUFeedForwardNetwork" ], [ "Scal", "ed" ], [ "Scal", "ing" ], [ "U", "RL" ], [ "U", "TF-8" ], [ "V", "^{" ], [ "V^{", "(p)}" ], [ "Value", "\"," ], [ "Wik", "im" ], [ "Wik", "ipedia" ], [ "Wikim", "edia" ], [ "Y", "_" ], [ "[", "^" ], [ "[", "b" ], [ "[", "ch" ], [ "[", "end" ], [ "[^", ">]" ], [ "[b", "]" ], [ "[ch", "]" ], [ "[int", "]:" ], [ "\\", "S+" ], [ "\\", "{" ], [ "\\\\", "eta" ], [ "\\\\", "theta_i^{" ], [ "\\\\text{", "-grams" ], [ "^", "(2i" ], [ "^", "T" ], [ "_", "ag" ], [ "_", "attempt" ], [ "_", "code" ], [ "_", "k" ], [ "_", "labels" ], [ "_", "language_model" ], [ "_", "lin" ], [ "_", "or" ], [ "_", "r" ], [ "_", "rm" ], [ "_", "time" ], [ "_", "values," ], [ "_", "x" ], [ "_ag", "re" ], [ "_agre", "ement" ], [ "_agreement", "_rat" ], [ "_and_activation", ".ipyn" ], [ "_ap", "pro" ], [ "_appro", "x" ], [ "_approx", "im" ], [ "_approxim", "ation" ], [ "_ax", "is_" ], [ "_axis_", "curv" ], [ "_b", "," ], [ "_b", "ar" ], [ "_b", "its_per_byt" ], [ "_b", "oundaries" ], [ "_bar", "_by_" ], [ "_bar_by_", "layer" ], [ "_beta", "3" ], [ "_bits_per_byt", "e" ], [ "_bod", "y:" ], [ "_body_", "output" ], [ "_boundaries", "(" ], [ "_bucket", "s:" ], [ "_by", "_condition" ], [ "_by_", "d_model" ], [ "_by_", "d_model:" ], [ "_by_", "max" ], [ "_by_max", "_bytes" ], [ "_co", "ver" ], [ "_col", "s," ], [ "_condition", "s" ], [ "_f", "et" ], [ "_f", "ile" ], [ "_f", "its" ], [ "_f", "unction" ], [ "_fet", "ch" ], [ "_fetch", "_w" ], [ "_fetch_w", "ikipedia" ], [ "_fetch_wikipedia", "_re" ], [ "_fetch_wikipedia_re", "vis" ], [ "_fetch_wikipedia_revis", "ion" ], [ "_function", "_curves" ], [ "_h", "ead_" ], [ "_head_", "attention" ], [ "_huber_", "weight" ], [ "_if_lar", "ge" ], [ "_inter", "val" ], [ "_k", "_filter" ], [ "_lin", "k" ], [ "_link", "s(" ], [ "_list", "," ], [ "_log_prob", "s" ], [ "_min", "," ], [ "_multi", "_head_attention" ], [ "_p", "_filter" ], [ "_p", "er_" ], [ "_per_", "layer" ], [ "_r", "ul" ], [ "_rm", "s_" ], [ "_rms_", "ratio" ], [ "_ro", "ot" ], [ "_rul", "e_" ], [ "_rule_", "name" ], [ "_rule_name", "=\"identity" ], [ "_s", "et" ], [ "_s", "ing" ], [ "_s", "qrt" ], [ "_seq", "_term" ], [ "_seq_term", ":" ], [ "_sing", "l" ], [ "_singl", "e_" ], [ "_single_", "step_" ], [ "_size", ")" ], [ "_split_", "chunk" ], [ "_split_chunk", "_by_max_bytes" ], [ "_split_heads(", "self.w_" ], [ "_stat", "es:" ], [ "_stat", "istic" ], [ "_su", "c" ], [ "_suc", "ces" ], [ "_succes", "s" ], [ "_success", "ful" ], [ "_t", "anh" ], [ "_tanh", "_approximation" ], [ "_text", "," ], [ "_to", "_rms_ratio" ], [ "_type=\"", "model" ], [ "_type=\"model", "\"," ], [ "_w", "ord(" ], [ "_wikitext", "_links(" ], [ "``ãĥ»``", "zero" ], [ "a", "_coef_init" ], [ "a", "st" ], [ "a", "vi" ], [ "ab", "stra" ], [ "able", "," ], [ "abstra", "ct" ], [ "abstract", "method" ], [ "ac", "cept" ], [ "action=", "parse" ], [ "activation", "s:" ], [ "ad", "d(id" ], [ "add(id", "(" ], [ "adding", "_mask" ], [ "allback", "_train_text" ], [ "ampl", "ing" ], [ "an", "es" ], [ "an", "on" ], [ "andidat", "es:" ], [ "anifest_path", ")" ], [ "anon", "i" ], [ "anoni", "cal_" ], [ "anonical_", "tokenizer" ], [ "anonical_tokenizer", "s." ], [ "ap", "anes" ], [ "ap", "pl" ], [ "api", "." ], [ "appl", "y(" ], [ "ar", "path" ], [ "arker", "=\"" ], [ "arpath", "y" ], [ "arquard", "t(" ], [ "articles", "_dir" ], [ "at", "e" ], [ "ate_", "p" ], [ "ate_p", "adding_mask" ], [ "ateg", "ory" ], [ "ats", ")ãĢĤ" ], [ "att", "r(" ], [ "attention", "_seq_term:" ], [ "attn_weight", "s," ], [ "aturating", "_fit" ], [ "aturating", "_fits" ], [ "auss", "ian" ], [ "b", "oundaries_a" ], [ "beam", "_search" ], [ "best", "_param" ], [ "bits_per_byt", "e" ], [ "c", "anonical_tokenizers." ], [ "c", "to" ], [ "c", "ur" ], [ "cach", "ed_" ], [ "cached_", "pro" ], [ "cached_pro", "per" ], [ "cached_proper", "ty" ], [ "cal_", "jacobian" ], [ "cent", "ile" ], [ "center", ":" ], [ "ch", "()." ], [ "ch", "an" ], [ "char", "s)" ], [ "chunk_", "length" ], [ "chunk_length", "_statistic" ], [ "cle", "an" ], [ "clip(log", "_beta" ], [ "clip(log_alpha", "," ], [ "clip_threshold", ":" ], [ "co", "uple" ], [ "coeff", "icient" ], [ "color=\"", "tab:" ], [ "con", "text" ], [ "construct_optimal_frontier", "(" ], [ "corpu", "s," ], [ "cos", "ine" ], [ "count", "_non_embedding_parameter" ], [ "couple", "d" ], [ "cre", "ase" ], [ "cross_", "attn_weights" ], [ "ct", "x" ], [ "ct", "x:" ], [ "cto", "ol" ], [ "cur", "sor" ], [ "d", "," ], [ "d", "eta" ], [ "d", "ual" ], [ "d(id", "x" ], [ "d(idx", ")" ], [ "d_", "bar" ], [ "d_", "i" ], [ "d_", "values" ], [ "de", "{" ], [ "deca", "y_" ], [ "deta", "ch()." ], [ "de{", "g" ], [ "de{g", "}_t" ], [ "dow", "n" ], [ "dual", "_axis_curv" ], [ "e_", "canonical_tokenizers." ], [ "e_", "decay_" ], [ "e_b", "pe_" ], [ "e_bpe_", "id_tokenizer_json" ], [ "ect", "iv" ], [ "ectiv", "e_decay_" ], [ "ed", "." ], [ "eff", "ective_decay_" ], [ "eg", "ative" ], [ "el", "()" ], [ "el", "if" ], [ "elu", "_tanh_approximation" ], [ "emperat", "ur" ], [ "en", "n" ], [ "encod", "e(self," ], [ "encode(", "text)" ], [ "enn", "r" ], [ "ennr", "i" ], [ "ennri", "ch" ], [ "ent", "er" ], [ "er", "e(" ], [ "er", "s\"," ], [ "er_", "clean" ], [ "eri", "cal_jacobian" ], [ "eval", "()" ], [ "eval", "_bits_per_byte" ], [ "eval", "_interval" ], [ "excluded_param_id", "s." ], [ "excluded_param_ids.", "add(id(" ], [ "exist", "s()" ], [ "exists()", ":" ], [ "ext", "_code" ], [ "ext_token", "_log" ], [ "f", "it_saturating_power_law" ], [ "f\"", "{" ], [ "fi", "le[" ], [ "fil", "ter" ], [ "file", "(" ], [ "file[", "\"" ], [ "fit_", "isoflop_parabola" ], [ "forward", "``" ], [ "full", "_like(" ], [ "g", "elu_exact" ], [ "g", "et_random_batch" ], [ "g", "pt" ], [ "gradient_clip", "_triggered" ], [ "gradient_norm", "_per_layer" ], [ "gradient_norm", "s)" ], [ "gro", "u" ], [ "grou", "pe" ], [ "groupe", "d_bar" ], [ "hunkSplitMod", "e" ], [ "i", "-theories" ], [ "i", ":" ], [ "i", "LU" ], [ "i", "ance" ], [ "i", "cal" ], [ "i", "lenam" ], [ "id", "=-1," ], [ "id", "x" ], [ "id_", "to" ], [ "idx", "]" ], [ "ieces", ".append(" ], [ "if", "orm" ], [ "ig", "size_per_head" ], [ "im", "e" ], [ "in", "al_values" ], [ "in", "crease" ], [ "init", "." ], [ "inn", "er_clean" ], [ "input", "s," ], [ "ip", "t" ], [ "ipt", "s/" ], [ "ipts/", "p" ], [ "ipts/p", "romot" ], [ "ipts/promot", "e_canonical_tokenizers." ], [ "is_interior", ":" ], [ "ism", "atches" ], [ "ist", "or" ], [ "istor", "ies:" ], [ "istories", "_by_condition" ], [ "it", "." ], [ "iter", ":" ], [ "its-per-byt", "e(" ], [ "j", "apanes" ], [ "ja", "\"``ãĥ»``\"en\"``" ], [ "joined_", "a" ], [ "ke", "y_" ], [ "ke", "y_length," ], [ "key_", "embedding" ], [ "l", "_values" ], [ "l", "ation" ], [ "l", "inewidth" ], [ "label", "s(" ], [ "lat", "_mask" ], [ "le", "venberg_m" ], [ "learning_curves", "_multi" ], [ "levenberg_m", "arquardt(" ], [ "lin", "e(" ], [ "linear", "1" ], [ "lm_head", "``)ãģ®" ], [ "lo", "or" ], [ "log", "_x" ], [ "log_n", "_arr" ], [ "lop", "en" ], [ "los", "s." ], [ "lr", ")" ], [ "m", "," ], [ "m", "_h" ], [ "m", "atrix" ], [ "max_", "distance" ], [ "max_chunk_bytes", "," ], [ "max_position", "," ], [ "mean", "_to_rms_ratio" ], [ "merg", "es:" ], [ "min", "," ], [ "min", "\\\\_lr}" ], [ "min", "us" ], [ "model", "," ], [ "model_prefix", "." ], [ "n", "_rat" ], [ "n_", "key" ], [ "name", "}" ], [ "no", "w" ], [ "norm", "_fn" ], [ "norm", "ed" ], [ "norm_first", "=True" ], [ "normalization", "_rule_name=\"identity" ], [ "now", "n_rat" ], [ "np.ndarray", "]," ], [ "num", "_buckets" ], [ "num", "_heads," ], [ "num", "el()" ], [ "o", "q" ], [ "oft", "max(" ], [ "oisy_", "grid" ], [ "on_embedding_param", "s:" ], [ "optimizer", ":" ], [ "oq", "a:" ], [ "ord(\"", "Â" ], [ "ormalization", "_factory" ], [ "ot", "ImplementedError" ], [ "out", "))" ], [ "output", "_layer:" ], [ "p", "``" ], [ "p", "redicted_" ], [ "pair", "_counts[pair]" ], [ "pair_in", "_word(" ], [ "pair_in", "dex[pair]" ], [ "parabola", "_coeff" ], [ "params0", ":" ], [ "parent", ".mkdir(parents=True," ], [ "path.", "exists():" ], [ "pe", "[:," ], [ "peak", "\\\\_lr}" ], [ "peak_learning_rat", "e" ], [ "per", "centile" ], [ "pm", ".SentencePiece" ], [ "point", "er" ], [ "position", ":" ], [ "position", "_embedding" ], [ "position", "al_encoding" ], [ "positions", "]" ], [ "pro", "bs" ], [ "qu", "ir" ], [ "quer", "y_length," ], [ "query", ".device)" ], [ "r", "ipts/promote_canonical_tokenizers." ], [ "r", "y_decode_byte_level_symbol" ], [ "r", "|" ], [ "ran", "g" ], [ "ransformer", "_block" ], [ "re", "p" ], [ "re", "quir" ], [ "relative_", "buckets" ], [ "relative_", "key_embedding" ], [ "relative_", "vector" ], [ "relative_position", ")" ], [ "relative_position_bucket", "(" ], [ "repo", "_root" ], [ "repo_id", ":" ], [ "request", "_time" ], [ "residual", "s," ], [ "residuals**2", "))" ], [ "result", "_bo" ], [ "result", "_body" ], [ "retoken", "ize(" ], [ "retr", "ies" ], [ "ross", "Attention" ], [ "row", "s" ], [ "s", "))," ], [ "s", "])" ], [ "s", "aturating_fit" ], [ "s", "ion" ], [ "s", "oftmax" ], [ "s[", "idx]" ], [ "s_", "q," ], [ "sc", "ripts/promote_canonical_tokenizers." ], [ "scal", "ing" ], [ "segmentation", "_agreement_rat" ], [ "ser", "-Agent" ], [ "split_train_val", "``" ], [ "src.", "training." ], [ "src.training.", "optimizer``" ], [ "src/data/", "wikipedia_manifests/" ], [ "stat", "istic" ], [ "step", ")" ], [ "step", ":" ], [ "step()", "``ãĥ»``zero" ], [ "step_", "accept" ], [ "strip", "_wikitext_links(" ], [ "t", "j" ], [ "t", "ransformer_block" ], [ "temperature", "=0.0" ], [ "ter", "," ], [ "th", "on" ], [ "theories/01_foundations/004_norm", "alization" ], [ "til", "de{g}_t" ], [ "time", ".sleep" ], [ "token", "s:" ], [ "token", "s_train" ], [ "token_embedding", ".weight" ], [ "token_id", "s" ], [ "token_id", "s." ], [ "token_id", "s``" ], [ "tokenizer_json", "_p" ], [ "top", "k" ], [ "top-", "k" ], [ "total", "_bytes)" ], [ "total_", "eval" ], [ "total_negative_log_likelihood_n", "at" ], [ "tr", "y_decode_byte_level_symbol" ], [ "train", "(" ], [ "train", "er" ], [ "train_", "unigram_model" ], [ "ual_seed", "(seed)" ], [ "ul", "t(" ], [ "un", "ctool" ], [ "un", "iform" ], [ "unk", "nown_rat" ], [ "up", "load_" ], [ "ur", "lopen" ], [ "use_cross_attention", "=False``" ], [ "util", "s." ], [ "v", "_term" ], [ "v", "e_bpe_id_tokenizer_json" ], [ "valid", "ation" ], [ "ve_", "character_level_tokenizer_json" ], [ "ver", "se" ], [ "vertex_", "loss" ], [ "vious_", "loss" ], [ "vmax", ":" ], [ "vocab", "``" ], [ "w", "indow" ], [ "wh", "ere(" ], [ "x", "``" ], [ "y_", "labels" ], [ "zero", "s(" ], [ "{", "\"" ], [ "{", "|" ], [ "|}", "{|" ], [ "}", "{" ], [ "}", "}``" ], [ "¡", "ä¾ĭ" ], [ "¡çĶ¨", "ãģĻãĤĭ" ], [ "£", "Ĵ" ], [ "¦", "(x)" ], [ "¦", "æ³" ], [ "¦æ³", "¢" ], [ "¨", "ãģ¦" ], [ "©", "ç·ļ" ], [ "ª", "è»¸" ], [ "®", "åĪĨ" ], [ "®", "åĭķ" ], [ "®", "Ĭ" ], [ "®µéļİ", "ãģ§" ], [ "®åĭķ", "å°ı" ], [ "®åĭķå°ı", "æķ°" ], [ "®åĭķå°ıæķ°", "çĤ¹" ], [ "¯", "å±¤" ], [ "°", "´" ], [ "°", "ãĤĬ" ], [ "±", "ãģĦ" ], [ "±", "èªŀ" ], [ "±", "éŁ¿" ], [ "±éŁ¿", "ãĤĴåıĹãģĳ" ], [ "µ", "·" ], [ "·", "ãģķ" ], [ "¹", "°ãĤĬ" ], [ "¾", "ãģį" ], [ "¾", "çī" ], [ "¾çī", "©ç·ļ" ], [ "Î²", "1," ], [ "Ïĥ", "(" ], [ "Ïĥ", "^2" ], [ "ãĢģ", "007" ], [ "ãĢģ", "False" ], [ "ãĢģ", "ãģĬãĤĪãģ³" ], [ "ãĢģ", "ãģĻ" ], [ "ãĢģ", "ä¾ĭ:" ], [ "ãĢģ", "å½¢çĬ¶" ], [ "ãĢģ", "å¾Į" ], [ "ãĢģ", "éłĤçĤ¹" ], [ "ãĢģæľ¬", "çķªå®Łè¡Įåīįãģ®ä¿®æŃ£" ], [ "ãĢĤ", "æĹ¢å®ļåĢ¤" ], [ "ãĢĤ``", "None``" ], [ "ãĢį", "ãģ¨ãģĦãģĨ" ], [ "ãĢį", "ãģ®" ], [ "ãģ", "ľ" ], [ "ãģ¡", "ãĤī" ], [ "ãģ£", "ãģ¦ãģĦãĤĭ" ], [ "ãģ£ãģ¦", "ãģĹãģ¾" ], [ "ãģ£ãģ¦", "ãĤĤ" ], [ "ãģ£ãģŁ", "ãģ¾ãģ¾" ], [ "ãģ¤ãģĦãģ¦", "ãĢģ" ], [ "ãģ§", "ãĤĪ" ], [ "ãģ§", "å¾ĹãģŁ" ], [ "ãģ§", "æ¯Ķè¼ĥãģĻãĤĭ" ], [ "ãģ§", "æııçĶ»ãģĻãĤĭãĢĤ" ], [ "ãģ§", "è¡ĮãģĨ" ], [ "ãģ§ãģ¯ãģªãģı", "ãĢģ" ], [ "ãģ§ãģį", "ãĢģ" ], [ "ãģ¨", "ãģªãĤĭ" ], [ "ãģ¨", "ãģ«" ], [ "ãģ¨", "ãģĹ" ], [ "ãģ¨", "ãģĹãĢģ" ], [ "ãģ¨", "åĲĮä¸Ģ" ], [ "ãģ¨ãģ®", "ä½µ" ], [ "ãģ¨ãģį", "ãģ®" ], [ "ãģ¨ãģį", "ãģ¯" ], [ "ãģ©", "ãģ¡ãĤī" ], [ "ãģ©", "ãĤĮ" ], [ "ãģ©ãģĨ", "ãģĹ" ], [ "ãģ©ãĤĮ", "ãģłãģĳ" ], [ "ãģª", "ãģĳãĤĮãģ°" ], [ "ãģª", "ãĤı" ], [ "ãģªãģ©", ")ãĢĤ" ], [ "ãģªãģ©", "ãĤĴ" ], [ "ãģªãģĦ", "ãģŁãĤģãĢģ" ], [ "ãģªãģı", "ãģ¦ãĤĤ" ], [ "ãģªãģĹ", "ãģ¨" ], [ "ãģªãģĹ", "ãģ®" ], [ "ãģªãĤı", "ãģ¡" ], [ "ãģ«", "ãģĻãĤĭãĢĤ" ], [ "ãģ«", "ãģĿãģ®ãģ¾ãģ¾" ], [ "ãģ«", "ãĤĤ" ], [ "ãģ«", "ãĤı" ], [ "ãģ«", "å¤±æķĹ" ], [ "ãģ«", "å¯¾ãģĹãģ¦" ], [ "ãģ«", "å¿ľ" ], [ "ãģ«", "åĪĨåī²" ], [ "ãģ«", "æ³¨æĦı" ], [ "ãģ«", "æ¸¡ãģĻ" ], [ "ãģ«", "éģĶãģĻãĤĭ" ], [ "ãģ«ãģĹãģŁ", "æĿ¡ä»¶" ], [ "ãģ«ãĤĪ", "ãģ£ãģ¦" ], [ "ãģ«ãĤı", "ãģŁ" ], [ "ãģ«ä½¿ãģĨ", "(005" ], [ "ãģ«ä½¿ãģĨ", ")ãĢĤ" ], [ "ãģ«å¤īæıĽ", "ãģĻãĤĭãĢĤ\"\"\"" ], [ "ãģ«å¯¾", "ãģĻãĤĭ" ], [ "ãģ«å¿ľ", "ãģĺ" ], [ "ãģ«å¿ľãģĺ", "ãģ¦" ], [ "ãģ«åĩºçı¾", "ãģĻãĤĭ" ], [ "ãģ«åĬłç®Ĺ", "ãģĻãĤĭ" ], [ "ãģ«æ¸¡ãģĻ", "ãĥŀãĤ¹ãĤ¯" ], [ "ãģ®", "ãģ¾ãģ¾" ], [ "ãģ®", "ãģĨ" ], [ "ãģ®", "ãģŁãĤģãĢģ" ], [ "ãģ®", "ãĥªãĥ³ãĤ¯" ], [ "ãģ®", "å®Łè£ħ" ], [ "ãģ®", "åħ¥åĬĽ" ], [ "ãģ®", "åĪĿæľŁåĢ¤" ], [ "ãģ®", "åŃ¦ç¿Ĵ" ], [ "ãģ®", "æ¬¡åħĥãĢĤ" ], [ "ãģ®", "æ¸Ľè¡°çİĩ" ], [ "ãģ®", "çµĦãģ¿åĲĪãĤıãģĽ" ], [ "ãģ®", "ç·ļå½¢" ], [ "ãģ®", "çĲĨè«ĸ" ], [ "ãģ®", "è§£æŀĲ" ], [ "ãģ®ãģ¿", "ãģ«ä¾ĿåŃĺãģĻãĤĭ" ], [ "ãģ®å", "¢ĥçķĮ" ], [ "ãģ®å", "¸" ], [ "ãģ®å", "º" ], [ "ãģ®å", "¼" ], [ "ãģ®å¤§ãģį", "ãģķ" ], [ "ãģ®å®Ł", "åĬ¹" ], [ "ãģ®å®Ł", "æķ°åĢ¤" ], [ "ãģ®å®Łæķ°åĢ¤", "ãģ®ç³»åĪĹ" ], [ "ãģ®å°", "Ħå½±" ], [ "ãģ®å½", "±éŁ¿ãĤĴåıĹãģĳ" ], [ "ãģ®å¾", "Į" ], [ "ãģ®åħ¥åĬĽ", "ãĥĨãĥ³ãĤ½ãĥ«ãĢĤ" ], [ "ãģ®åĨħ", "ç©įãĤĴ" ], [ "ãģ®åĩ", "¦çĲĨ" ], [ "ãģ®åĩºåĬĽ", "ãĥĨãĥ³ãĤ½ãĥ«ãĢĤ" ], [ "ãģ®åī", "¯å±¤" ], [ "ãģ®åī", "į" ], [ "ãģ®åĬ", "¹" ], [ "ãģ®åıįå¾©", "çļĦãģª" ], [ "ãģ®åıįå¾©çļĦãģª", "ç¸®å°ı" ], [ "ãģ®åĴĮ", "ãģĮ" ], [ "ãģ®åŁ", "º" ], [ "ãģ®åł´åĲĪ", "ãĢģ" ], [ "ãģ®æ", "§ĭç¯ī" ], [ "ãģ®æ", "¯Ķ" ], [ "ãģ®æ", "¯Ķè¼ĥ" ], [ "ãģ®æ", "Ĳįå¤±" ], [ "ãģ®æĹ¢", "å®ļ" ], [ "ãģ®ç³»åĪĹéķ·", "ä¾ĿåŃĺ" ], [ "ãģ®ç³»åĪĹéķ·ä¾ĿåŃĺ", "éłħ" ], [ "ãģ®è§£æŀĲ", "ãĥĳãĤ¤ãĥĹãĥ©ãĤ¤ãĥ³" ], [ "ãģ¾", "ãģ¨ãĤģãģ¦" ], [ "ãģ¾ãĤĭ", "ãģĶãģ¨" ], [ "ãģĤãģ¦ãģ¯ãĤģ", "ãĢģéłĤçĤ¹" ], [ "ãģĤãģŁãĤĬãģ®", "è¨Īç®Ĺéĩı" ], [ "ãģĦ", "ãĢĤ" ], [ "ãģĦ", "ãĥ©ãĥĥãĥĳãĥ¼ãĢĤ" ], [ "ãģĦ", "åł´åĲĪãģ¯" ], [ "ãģĪãĤĭ", "ãĤĪãģĨãģ«" ], [ "ãģĭ", "(" ], [ "ãģĭ", "ãģ¤" ], [ "ãģĭãĤī", "éłĨãģ«" ], [ "ãģĮ", "ãĤŃãĥ£ãĥĥãĤ·ãĥ¥" ], [ "ãģĮ", "å¢Ĺ" ], [ "ãģĮ", "çķ°ãģªãĤĭ" ], [ "ãģĮ", "éļĽéĻĲãģªãģı" ], [ "ãģĮä¸Ģèĩ´", "ãģĹãģ¾" ], [ "ãģĮä¸Ģèĩ´ãģĹãģ¾", "ãģĽ" ], [ "ãģĮä¸Ģèĩ´ãģĹãģ¾ãģĽ", "ãĤĵ" ], [ "ãģĮå°ıãģķãģĦ", "ãģ»ãģ©" ], [ "ãģı", "ãĢģ" ], [ "ãģı", "ä½¿ãĤı" ], [ "ãģıä½¿ãĤı", "ãĤĮãĤĭ" ], [ "ãģĵ", "ãģ®åĢ¤" ], [ "ãģĵãģ¨", "ãģ§" ], [ "ãģĵãģ¨", "ãģ¯" ], [ "ãģĵãģ¨ãģĮ", "æľŁ" ], [ "ãģĵãģ¨ãģĮæľŁ", "å¾ħ" ], [ "ãģĵãģ¨ãģĮæľŁå¾ħ", "ãģķãĤĮãĤĭãĢĤ" ], [ "ãģĵãģ®æ", "¯Ķ" ], [ "ãģĵãģĵ", "ãģ«" ], [ "ãģĵãĤĮ", "ãģĮ" ], [ "ãģķãģĽ", "ãĤĭ(" ], [ "ãģķãģĽ", "ãĤĭãĢĤ" ], [ "ãģķãĤĮãģŁ", "ãĥĪãĥ¼ãĤ¯ãĥ³" ], [ "ãģĹ", "åģ´" ], [ "ãģĹãĢģ", "æ®ĭãĤĬ" ], [ "ãģĹãģ¦ãģĦãĤĭ", "ãģŁãĤģãĢģ" ], [ "ãģĹãģ¦ãģĬ", "ãģı" ], [ "ãģĹãģŁ", ")ãĢĤ" ], [ "ãģĹãģŁ", "ãģĦ" ], [ "ãģĹãģŁ", "åıįå¾©" ], [ "ãģĹãģŁè¨ĺäºĭ", "æľ¬æĸĩ" ], [ "ãģĹãģŁè¨ĺäºĭæľ¬æĸĩ", "ãĤĴéĢ£çµĲãģĹãģŁ" ], [ "ãģĻ", "ãģĵãģ¨" ], [ "ãģĻãĤĭ", ")" ], [ "ãģĻãĤĭ", "åŃ¦ç¿Ĵçİĩ" ], [ "ãģĻãĤĭ(00", "9" ], [ "ãģĻãĤĭãģĵãģ¨", "ãģ§" ], [ "ãģĻãĤĭãģĵãģ¨", "ãģ§ãĢģ" ], [ "ãģĻãĤĭãģŁãĤģ", "ãģ«ä½¿ãģĨãĢĤ" ], [ "ãģĻãĤĭãģŁãĤģ", "ãģ®" ], [ "ãģļ", "ãģ«" ], [ "ãģł", "ãģĮãĢģ" ], [ "ãģłãģĳ", "ãģĮ" ], [ "ãĤ", "º" ], [ "ãĤ", "½" ], [ "ãĤ¢", "ãĥ³" ], [ "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹", "(m:" ], [ "ãĤ¨", "ãĤ¤" ], [ "ãĤ¨ãĥ©", "ãĥ¼" ], [ "ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°", "éģ©çĶ¨åīį" ], [ "ãĤ³", "ãĥĵ" ], [ "ãĤ³ãĤ¹ãĥĪ", "ãģĮ" ], [ "ãĤ³ãĥ¼ãĥ", "ĩãĤ£ãĥ³ãĤ°" ], [ "ãĤ³ãĥ¼ãĥĳãĤ¹", "ãĤĴåıĸå¾Ĺ" ], [ "ãĤ³ãĥĵ", "ãĤ¢ãĥ³" ], [ "ãĤµ", "ãĥ¼ãĥ" ], [ "ãĤµãĥ¼ãĥ", "ģ" ], [ "ãĤµãĥĸ", "ãĥ¯" ], [ "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼", "åĩºåĬĽ" ], [ "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽ", "ãģ«éģ©çĶ¨ãģĻãĤĭ" ], [ "ãĤµãĥĸãĥ¯", "ãĥ¼ãĥī" ], [ "ãĤµãĥĸãĥ¯ãĥ¼ãĥī", "åĪĨåī²" ], [ "ãĤ·ãĥ³ãĥľãĥ«", "ãģ®" ], [ "ãĤ·ãĥ¼ãĥī", "ãģĶãģ¨ãģ®" ], [ "ãĤ·ãĥ¼ãĥī", "éĸĵ" ], [ "ãĤ¹", "ãĤŃ" ], [ "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ", "ãģĻãĤĭãĢĤ" ], [ "ãĤ¹ãĤ±ãĥ¼ãĥªãĥ³ãĤ°", "åīĩ" ], [ "ãĤ¹ãĥĨãĥĥãĥĹ", "ãģ¨ãģ®" ], [ "ãĤ¹ãĥĨãĥĥãĥĹ", "æĲįå¤±" ], [ "ãĤ¿", "ãĥĸãĥ»" ], [ "ãĤ¿ãĥĸãĥ»", "æĶ¹è¡Į" ], [ "ãĤĦãģĻ", "ãģĦ" ], [ "ãĤĪãģĨ", "ãģª" ], [ "ãĤĪãĤĬ", "ä¸Ĭä½į" ], [ "ãĤĪãĤĬä¸Ĭä½į", "ãģ®" ], [ "ãĤī", "ãģªãģĦ)ãĢĤ" ], [ "ãĤĭ", "(``" ], [ "ãĤĭ(", "Hoffmann" ], [ "ãĤĭãģĵãģ¨", "ãģ§ãĢģ" ], [ "ãĤĮ", "åĢ¤" ], [ "ãĤı", "ãĤĮ" ], [ "ãĤĴ", "ä¸ĢåĪĩ" ], [ "ãĤĴ", "ä¸İ" ], [ "ãĤĴ", "åĬłãģĪãĤĭ" ], [ "ãĤĴ", "åĽºå®ļ" ], [ "ãĤĴ", "çĶŁæĪĲ" ], [ "ãĤĴ", "è¦ĭ" ], [ "ãĤĴ", "éĺ²ãģĲ" ], [ "ãĤĴ", "éĻ¤" ], [ "ãĤĴãģ¾ãģŁ", "ãģĲ" ], [ "ãĤĴä¸İ", "ãģĪãĤĭ" ], [ "ãĤĴä¿Ŀ", "ãģ£ãģŁãģ¾ãģ¾" ], [ "ãĤĴåĲ«", "ãĤģ" ], [ "ãĤĴæ", "±ĤãĤģ" ], [ "ãĤĴæ", "Ľ" ], [ "ãĤĴæ¸¡", "ãģĻãĢĤ" ], [ "ãĤĴæĬ", "ĳ" ], [ "ãĤĴè¡Į", "ãģĨãĢĤ" ], [ "ãĤĴè¡Į", "ãĤıãģªãģĦ" ], [ "ãĤĴè¡ĮãģĨ", "çī¹å¾´" ], [ "ãĥ", "ı" ], [ "ãĥ¢ãĤ¸ãĥ¥ãĥ¼ãĥ«", "ãģ¯" ], [ "ãĥ¢ãĥĩãĥ«", "ãģ®" ], [ "ãĥ¤", "ãĤ³ãĥĵãĤ¢ãĥ³" ], [ "ãĥª", "ãĤ¢" ], [ "ãĥ«", "ãĥģ" ], [ "ãĥ»", "2" ], [ "ãĥ»", "åĲĦ" ], [ "ãĥ»", "æ¤ľè¨¼" ], [ "ãĥ»åĲĦ", "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ" ], [ "ãĥ¼ãĤ¹", "ãģ®åıįå¾©çļĦãģªç¸®å°ı" ], [ "ãĥ¼ãĥĲ", "ãĥ¼" ], [ "ãĥ¼ãĥł", "ãĤµãĥ¼ãĥģ" ], [ "ãĥģãĥ£ãĥ³ãĤ¯", "éķ·" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "ãģ®" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "å¢ĥçķĮ" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "åĮĸ" ], [ "ãĥĪãĥ¼ãĤ¯ãĥ³", "æķ°ãĢĤ" ], [ "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶", "ãĢĤ" ], [ "ãĥİãĤ¤ãĤº", "åº" ], [ "ãĥİãĤ¤ãĤºåº", "Ĭ" ], [ "ãĥĲãĤ¤ãĤ¢ãĤ¹", "ãĤĴåĬłãģĪãĤĭ" ], [ "ãĥĲãĤ¤ãĤ¢ãĤ¹", "éłħ" ], [ "ãĥĲãĤ±ãĥĥãĥĪ", "åĮĸ" ], [ "ãĥĳ", "ãĤ¹" ], [ "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "ãĢģå½¢çĬ¶" ], [ "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "ãĤĴæĮģãģŁãģªãģĦ" ], [ "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "æĽ´æĸ°" ], [ "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æĽ´æĸ°", "å¼ı" ], [ "ãĥĵ", "ãĥ¼ãĥłãĤµãĥ¼ãĥģ" ], [ "ãĥķ", "ãĤ£ãĥ«ãĤ¿" ], [ "ãĥĸ", "ãĥ¼ãĥĪãĤ¹ãĥĪãĥ©ãĥĥãĥĹ" ], [ "ãĥĸãĥŃãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪ", "åı¯èĥ½" ], [ "ãĥŀ", "ãĥ«ãĥģ" ], [ "ãĥŀ", "ãĥĥãĥĹ" ], [ "ä¸¸", "ãĤģ" ], [ "ä¸Ģ", "èĪ" ], [ "ä¸ĢèĪ", "¬" ], [ "ä¸Ĭ", "ãģ®" ], [ "ä¸Ĭæĺĩ", "å¹ħ" ], [ "ä¸į", "ä¸Ģèĩ´" ], [ "ä¸į", "è¦ģ" ], [ "ä¸İ", "ãģĪ" ], [ "ä¸Ń", "ãģ®" ], [ "ä¸Ń", "å¿" ], [ "ä¸Ńå¿", "ĥ" ], [ "ä¸Ńå¿ĥ", "å·®åĪĨ" ], [ "ä¸Ńå¿ĥå·®åĪĨ", "ãģ«ãĤĪãĤĭ" ], [ "ä¸Ńéĸĵ", "æ¬¡åħĥ" ], [ "äºĭåīįåĪĨåī²", "(pre-tokenization" ], [ "äºĮä¹Ĺ", "å¹³åĿĩ" ], [ "äºĴ", "æıĽ" ], [ "äºĴæıĽ", "æĢ§" ], [ "ä»¥", "éĻį" ], [ "ä»»æĦı", "å½¢çĬ¶" ], [ "ä»»æĦıå½¢çĬ¶", "ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ" ], [ "ä¼Ŀ", "ãĤı" ], [ "ä½¿", "ãĤıãģªãģĦ" ], [ "ä½įç½®", "ãģ§" ], [ "ä½įç½®", "ãģ®ãģ¿ãģ«ä¾ĿåŃĺãģĻãĤĭ" ], [ "ä½įç½®", "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:" ], [ "ä¿Ŀ", "æĮģ" ], [ "å¤§ãģį", "ãģª" ], [ "å¤§ãģįãģı", "ãģªãĤĬ" ], [ "å¤ī", "æĽ´" ], [ "å¤īæıĽ", "ãģĻãĤĭ" ], [ "å¤ĸ", "ãĤĮåĢ¤" ], [ "å¤ļ", "ãģĦ" ], [ "å§Ķ", "ãģŃ" ], [ "å®Įåħ¨", "ãģª" ], [ "å®ļåĢ¤", "ãģ®ç³»åĪĹ" ], [ "å®Ł", "ãĥĪãĥ¼ãĤ¯ãĥ³" ], [ "å®Łè£ħ", "ãģ§ãģ¯" ], [ "å®Łè£ħ", "ãģĹãģªãģĦ" ], [ "å®Łè£ħ", "ãģĻãĤĭãĢĤ" ], [ "å¯¾å¿ľãģĻãĤĭ", "æĲįå¤±" ], [ "å¯¾æķ°ç©ºéĸĵ", "ãģ§" ], [ "å°İ", "åħ¥" ], [ "å°İ", "åĩº" ], [ "å°İ", "éĸ¢æķ°" ], [ "å±", "ĢæīĢ" ], [ "å±¤", "ãģ®" ], [ "å±¤", "ãģĶãģ¨" ], [ "å±¤", "æķ°" ], [ "å·®", "ãģĹ" ], [ "å¹³", "æĸ¹" ], [ "å¹³", "è¡Į" ], [ "å¹³è¡Į", "ç§»åĭķ" ], [ "åº¦", "ãĤĤ" ], [ "å¼", "¦æ³¢" ], [ "å¼·åº¦", "ãģĮ" ], [ "å½¢çĬ¶", "ãģ¯" ], [ "å½ĵ", "ãģĻãĤĭ" ], [ "å¾", "®åĪĨ" ], [ "å¾©åħĥ", "ãģ§ãģįãĤĭ" ], [ "å¾Į", "åįĬ" ], [ "å¿ħ", "ãģļ" ], [ "å¿ħè¦ģãģĮãģĤ", "ãĤĭ(" ], [ "åĢ¤", "ãģ¨ãģĹãģ¦" ], [ "åĢĻè£ľ", "èªŀå½Ļ" ], [ "åĢĻè£ľèªŀå½Ļ", "ãģĭãĤīãģ®" ], [ "åĤ", "¾ãģį" ], [ "åĥ", "ı" ], [ "åħ¨ä½ĵ", "ãĤĴ" ], [ "åħ¬å¹³", "ãģ«" ], [ "åħ±éĢļ", "ãģ®" ], [ "åħĪéłŃãģ«", "ä»ĺ" ], [ "åĨį", "æ§ĭæĪĲ" ], [ "åĩ", "¡ä¾ĭ" ], [ "åĩºåĬĽ", "å°Ħå½±" ], [ "åĩºçı¾", "ãģĹãģŁ" ], [ "åĪ¶", "éĻĲ" ], [ "åĪĨ", "ãĤĴ" ], [ "åĪĨ", "æ¯" ], [ "åĪĨæ¯", "į" ], [ "åĪĩãĤĬ", "åĩº" ], [ "åĪĩãĤĬ", "æį" ], [ "åĪĩãĤĬæį", "¨ãģ¦" ], [ "åĪĹ", ")ãĢĤ" ], [ "åī", "Ĭ" ], [ "åī²ãĤĬå½ĵãģ¦", "ãĤĭ" ], [ "åīį", "ãģ«" ], [ "åīį", "åįĬ" ], [ "åīįæıĲ", "æĿ¡ä»¶" ], [ "åĬ", "Ł" ], [ "åĬ¹", "æŀľ" ], [ "åįĺ", "èªŀ" ], [ "åįĺä¸Ģ", "ãĤ¹ãĥĨãĥĥãĥĹæĲįå¤±" ], [ "åįĺä¸Ģ", "ãĥĪãĥ¼ãĤ¯ãĥ³" ], [ "åİ", "³å¯Ĩ" ], [ "åİŁè«ĸæĸĩ", "ãģ®" ], [ "åİŁè«ĸæĸĩ", "ãģ¯" ], [ "åı", "³" ], [ "åıĸå¾Ĺ", "ãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ" ], [ "åĲĦ", "è¦ģç´łãģ¯" ], [ "åĲĪ", "ãģ¯" ], [ "åĲį", "åīį" ], [ "åĳ", "¼ãģ³" ], [ "åĳ", "¼ãģ³åĩº" ], [ "åĽºå®ļ", "åŃ¦ç¿Ĵçİĩ" ], [ "åĽŀ", "æķ°" ], [ "åĽŀ", "æķ°ãĢĤ" ], [ "åĽłæŀľ", "ãĥŀãĤ¹ãĤ¯" ], [ "åľ¨", "ãģ®" ], [ "åŁĭãĤģè¾¼ãģ¿", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°" ], [ "åŃ¦ç¿Ĵ", "ãģĮ" ], [ "åŃ¦ç¿Ĵ", "ãģĹãģŁ" ], [ "åŃ¦ç¿Ĵ", "çĶ¨" ], [ "åŃ¦ç¿ĴçĶ¨", "ãĥ»æ¤ľè¨¼" ], [ "æ", "®Ĭ" ], [ "æ", "±Ĥ" ], [ "æ", "··" ], [ "æ¤ľè¨¼", "ãĥĩãĥ¼ãĤ¿" ], [ "æ¨Ļæºĸ", "ãģ®" ], [ "æ¨Ļæºĸ", "æŃ£è¦ıåĪĨå¸ĥ" ], [ "æ¨Ļæºĸ", "çļĦãģª" ], [ "æ¯Ķ", "çİĩ" ], [ "æ¯Ķä¾ĭ", "ãģĹãģ¦" ], [ "æµ", "®åĭķå°ıæķ°çĤ¹" ], [ "æ¸¡", "ãģĻãĢĤ" ], [ "æ¸¬", "ãĤĭ" ], [ "æ¸¬", "å®ļåĢ¤ãģ®ç³»åĪĹ" ], [ "æĢ§", "èĥ½" ], [ "æģ", "Ĵ" ], [ "æĦı", "åĳ³" ], [ "æħ", "®" ], [ "æĪĲ", "åĬŁ" ], [ "æī±", "ãģĨãĢĤ" ], [ "æī±ãģĨ", ")ãĢĤ" ], [ "æĭ¡", "å¼µ" ], [ "æĮĩå®ļ", "ãģĹãģŁ" ], [ "æĮĩå®ļ", "ãģĻãĤĭãģ¨" ], [ "æĮĩå®ļãģķãĤĮãģŁ", "åł´åĲĪãĢģ" ], [ "æĮĩæķ°", "æİ¨å®ļ" ], [ "æİ", "¢ç´¢" ], [ "æķ°ãģĪæĸ¹", "ãĢĮæľ¬ä½ĵ" ], [ "æĸ°", "çīĪ" ], [ "æĸ¹", "äºĴæıĽæĢ§" ], [ "æĸĩåŃĹ", "åįĺä½įãģ®" ], [ "æĸĩåŃĹåĪĹ", "ãĤĴ" ], [ "æĺİç¤º", "çļĦãģª" ], [ "æĻĤ", "ãģ®" ], [ "æĻĤçĤ¹", "ãģ§" ], [ "æĽ¸ãģį", "åĩº" ], [ "æĽ¿", "ãģĪ" ], [ "æľ¬", "çķªå®Łè¡Įåīįãģ®ä¿®æŃ£" ], [ "æľĢ", "æĸ°çīĪ" ], [ "æľĢ", "èī" ], [ "æľĢå°ı", "éĻĲ" ], [ "æľĢçµĤ", "æŃ£è¦ıåĮĸå±¤" ], [ "æľĢèī", "¯" ], [ "æľīéĻĲ", "åĢ¤" ], [ "æĿ¡ä»¶", "åĲį:" ], [ "æł", "¹" ], [ "æŃ", "»" ], [ "æŃ£", "å¼¦æ³¢" ], [ "æŃ£ãģĹ", "ãģķ" ], [ "æŃ£è¦ıåĮĸ", "æ¸Īãģ¿" ], [ "æŃ£è¦ıåĮĸå¾Į", "ç½®" ], [ "æŃ£è¦ıåĮĸåīį", "ç½®" ], [ "ç", "¬¦åı·" ], [ "ç", "¬¬" ], [ "ç", "¸¦" ], [ "ç", "¹°ãĤĬ" ], [ "ç", "Ń" ], [ "ç¢º", "ç«ĭ" ], [ "ç¨ĭ", "åº¦" ], [ "ç©ºçĻ½", "ãģĮ" ], [ "ç©ºçĻ½", "ãĤĴ" ], [ "ç«", "¯" ], [ "ç³»åĪĹ", "ãģ®" ], [ "ç´", "Ķ" ], [ "çµ¶å¯¾ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹", "(å½¢çĬ¶" ], [ "çµĮ", "è·" ], [ "çµĮçĶ±", "ãģ§" ], [ "çµĮè·", "¯" ], [ "çī¹", "æ®Ĭ" ], [ "çĲĨ", "çĶ±" ], [ "çĶ¨", "ãģĦãģ¦" ], [ "çĶ¨ãģĦãģŁ", "éĢĲæ¬¡" ], [ "çĶ¨ãģĦãģŁéĢĲæ¬¡", "æİ¨è«ĸ" ], [ "çĶ»", "åĥı" ], [ "çĶŁ", "ãģĺ" ], [ "çĶŁæĪĲ", "ãģ®åĲĦ" ], [ "çĻ", "¾" ], [ "çĻº", "åĭķ" ], [ "çĽ®", "è¦ĸ" ], [ "çĽ®è¦ĸ", "ãģ§" ], [ "çĽ´åīį", "ãģ§" ], [ "çŃī", "ä¾¡" ], [ "è", "µ·" ], [ "è", "Ģ" ], [ "è", "ĭ" ], [ "è", "ĸĦ" ], [ "è¡Į", "ãģĮ" ], [ "è¡ĮãĤı", "ãģļãĢģ" ], [ "è¤ĩæķ°", "ãĤ·ãĥ¼ãĥī" ], [ "è¨±", "å®¹" ], [ "è¨Ģèªŀãĥ¢ãĥĩãĥ«", "(``forward(token_ids)" ], [ "è¨Īç®Ĺ", "ãģĻãĤĭ(" ], [ "è¨ĺ", "éĮ²" ], [ "è¨Ń", "ãģĳ" ], [ "è¨Ńå®ļ", "ãģ§" ], [ "è©", "²" ], [ "èª", "¿" ], [ "èª¤", "ãĤĬ" ], [ "èªŀå½Ļ", "åŃ¦ç¿Ĵ" ], [ "èªŀå½Ļãģ«", "åĲ«ãģ¾ãĤĮãģªãģĦ" ], [ "èªŀå½Ļãģ«", "åĲ«ãģ¾ãĤĮãĤĭ" ], [ "è²ª", "æ¬²æ³ķ" ], [ "è²ł", "ãģ®å¯¾æķ°å°¤åº¦" ], [ "è²ł", "éłĺåŁŁ" ], [ "è¿½åĬł", "ãģĹãģŁ" ], [ "èĢ", "ĥ" ], [ "èĢĥ", "æħ®" ], [ "èĩªä½ľ", "ãģ®" ], [ "èĭ", "±èªŀ" ], [ "èĸĦ", "ãģĦãĥ©ãĥĥãĥĳãĥ¼ãĢĤ" ], [ "èŃĺåĪ¥ä¸įèĥ½ãģª", "ãĥĩãĥ¼ãĤ¿" ], [ "éĢ", "Ģ" ], [ "éĢ", "ł" ], [ "éĢıéģİ", "ãģĻãĤĭ" ], [ "éĩįãģ¿", "ãĤĴåħ±æľī" ], [ "éĩįãģ¿æ¸Ľè¡°", "ãģ®å®ŁåĬ¹" ], [ "éĩįãģŃãģ¦", "æııçĶ»ãģĻãĤĭãĢĤ" ], [ "éķ·", "ãģĦ" ], [ "éķ·ãģķ", "ãĥļãĥĬãĥ«ãĥĨãĤ£" ], [ "éĸ¢", "ä¿Ĥ" ], [ "éĸ¢æķ°", "ãģ®" ], [ "éĸ¾åĢ¤", "ãĢĤ" ], [ "éĽĨè¨Ī", "ãģĻãĤĭãĢĤ" ], [ "éĿŀ", "åŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°" ], [ "éĿŀç·ļå½¢", "æľĢå°ıäºĮä¹Ĺ" ], [ "éĿŀéĩįè¤ĩ", "çªĵ" ], [ "Ċ", "(``" ], [ "Ċ", "-" ], [ "Ċ", "ICLR" ], [ "Ċ", "åŃ¦ç¿Ĵ" ], [ "ĊĊ", "Vaswani" ], [ "ĊĊĠĠĠĠ", "1." ], [ "ĊĊĠĠĠĠ", "A" ], [ "ĊĊĠĠĠĠ", "åĲĦ" ], [ "ĊĊĠĠĠĠ@", "torch.no_grad()" ], [ "ĊĊĠĠĠĠax.set_", "xlabel(xlabel)" ], [ "ĊĊĠĠĠĠax.set_", "xtick" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "\\\\mathrm{" ], [ "ĊĊĠĠĠĠĠĠĠĠ", "for" ], [ "ĊĠĠ", "``" ], [ "ĊĠĠĠĠ", "$" ], [ "ĊĠĠĠĠ", "3" ], [ "ĊĠĠĠĠ", "L" ], [ "ĊĠĠĠĠ", "]" ], [ "ĊĠĠĠĠ", "alpha:" ], [ "ĊĠĠĠĠ", "always_negative_" ], [ "ĊĠĠĠĠ", "api." ], [ "ĊĠĠĠĠ", "chunk_split_mode:" ], [ "ĊĠĠĠĠ", "corpus_" ], [ "ĊĠĠĠĠ", "d_k" ], [ "ĊĠĠĠĠ", "device:" ], [ "ĊĠĠĠĠ", "generator" ], [ "ĊĠĠĠĠ", "i" ], [ "ĊĠĠĠĠ", "max_chunk_bytes:" ], [ "ĊĠĠĠĠ", "model:" ], [ "ĊĠĠĠĠ", "params" ], [ "ĊĠĠĠĠ", "path" ], [ "ĊĠĠĠĠ", "profiles:" ], [ "ĊĠĠĠĠ", "residual_fn:" ], [ "ĊĠĠĠĠ", "residuals:" ], [ "ĊĠĠĠĠ", "result" ], [ "ĊĠĠĠĠ", "sorted_" ], [ "ĊĠĠĠĠ", "try:" ], [ "ĊĠĠĠĠ", "v" ], [ "ĊĠĠĠĠ", "weights:" ], [ "ĊĠĠĠĠ", "ãĢĮ" ], [ "ĊĠĠĠĠ", "ãģĮ" ], [ "ĊĠĠĠĠ", "æĲįå¤±" ], [ "ĊĠĠĠĠ\"", "load_" ], [ "ĊĠĠĠĠ2", "." ], [ "ĊĠĠĠĠax.grid(alpha=0.3", "," ], [ "ĊĠĠĠĠb", "atch_size:" ], [ "ĊĠĠĠĠf", "it_" ], [ "ĊĠĠĠĠf", "lops_per_token" ], [ "ĊĠĠĠĠh", "uber_delta:" ], [ "ĊĠĠĠĠn", "_" ], [ "ĊĠĠĠĠtext", ":" ], [ "ĊĠĠĠĠvocab", ":" ], [ "ĊĠĠĠĠĠĠĠĠ", "chunk_split_mode:" ], [ "ĊĠĠĠĠĠĠĠĠ", "color" ], [ "ĊĠĠĠĠĠĠĠĠ", "curv" ], [ "ĊĠĠĠĠĠĠĠĠ", "e" ], [ "ĊĠĠĠĠĠĠĠĠ", "except" ], [ "ĊĠĠĠĠĠĠĠĠ", "log_n" ], [ "ĊĠĠĠĠĠĠĠĠ", "merges" ], [ "ĊĠĠĠĠĠĠĠĠ", "norm_fn" ], [ "ĊĠĠĠĠĠĠĠĠ", "ratio" ], [ "ĊĠĠĠĠĠĠĠĠ", "seed:" ], [ "ĊĠĠĠĠĠĠĠĠ", "try:" ], [ "ĊĠĠĠĠĠĠĠĠ", "vocabulary_size:" ], [ "ĊĠĠĠĠĠĠĠĠ", "weights:" ], [ "ĊĠĠĠĠĠĠĠĠ", "åıĸå¾ĹãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ" ], [ "ĊĠĠĠĠĠĠĠĠ\"", "median" ], [ "ĊĠĠĠĠĠĠĠĠ\"", "p90" ], [ "ĊĠĠĠĠĠĠĠĠ\"\"\"", "éłĨä¼ĿæĴŃãĢĤ" ], [ "ĊĠĠĠĠĠĠĠĠb", "atch_size:" ], [ "ĊĠĠĠĠĠĠĠĠb", "eam" ], [ "ĊĠĠĠĠĠĠĠĠbeta", ":" ], [ "ĊĠĠĠĠĠĠĠĠbidirectional", ":" ], [ "ĊĠĠĠĠĠĠĠĠf", "lops_per_token" ], [ "ĊĠĠĠĠĠĠĠĠm", "," ], [ "ĊĠĠĠĠĠĠĠĠmax_distance", ":" ], [ "ĊĠĠĠĠĠĠĠĠn", ":" ], [ "ĊĠĠĠĠĠĠĠĠn", "ew" ], [ "ĊĠĠĠĠĠĠĠĠn", "n." ], [ "ĊĠĠĠĠĠĠĠĠnum", "_buckets:" ], [ "ĊĠĠĠĠĠĠĠĠself.", "d_k" ], [ "ĊĠĠĠĠĠĠĠĠself.", "num_heads" ], [ "ĊĠĠĠĠĠĠĠĠself.register", "_buffer(\"" ], [ "ĊĠĠĠĠĠĠĠĠvocab", ":" ], [ "ĊĠĠĠĠĠĠĠĠx", "_c" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "(004" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "color=\"tab:" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "f\"" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "feed_forward_factory" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "history" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "label=" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "lin" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "log" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "new" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "output" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "p.grad" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "self_" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "ãģ®" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(", "output," ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠb", "reak" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠx", "_hat" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "b" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "for" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "s" ], [ "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠparam", "s," ], [ "Ġ", "..." ], [ "Ġ", "Embedding" ], [ "Ġ", "Error" ], [ "Ġ", "OS" ], [ "Ġ", "_boundaries(" ], [ "Ġ", "_fit_nonlinear_huber" ], [ "Ġ", "attn_weights)" ], [ "Ġ", "cache_path" ], [ "Ġ", "default" ], [ "Ġ", "encode(self," ], [ "Ġ", "evaluate_bits_per_byt" ], [ "Ġ", "evaluation_window" ], [ "Ġ", "extra" ], [ "Ġ", "len" ], [ "Ġ", "lo" ], [ "Ġ", "model" ], [ "Ġ", "model." ], [ "Ġ", "model.parameters()" ], [ "Ġ", "of" ], [ "Ġ", "optimizer" ], [ "Ġ", "positional_transform" ], [ "Ġ", "raw_norm" ], [ "Ġ", "s_" ], [ "Ġ", "total_bytes)" ], [ "Ġ", "value" ], [ "Ġ", "values," ], [ "Ġ", "vocab_size" ], [ "Ġ", "weights" ], [ "Ġ", "wikitext" ], [ "Ġ", "ãģĵãģ®" ], [ "Ġ", "ãĤ«ãĥ©" ], [ "Ġ", "ãĥĨãĥ³ãĤ½ãĥ«" ], [ "Ġ", "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶" ], [ "Ġ", "ä»»æĦıå½¢çĬ¶ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ" ], [ "Ġ", "ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:" ], [ "Ġ", "å¯¾" ], [ "Ġ", "å¯¾å¿ľãģĻãĤĭæĲįå¤±" ], [ "Ġ", "å½¢å¼ı" ], [ "Ġ", "æķ´æķ°" ], [ "Ġ", "çµĤäºĨ" ], [ "Ġ", "çĶŁæĪĲãģĻãĤĭ" ], [ "Ġ", "è¡Į" ], [ "Ġ", "èªŀå½Ļ" ], [ "Ġ", "èĩªå·±æ³¨æĦı" ], [ "Ġ", "éĿŀåŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°" ], [ "Ġ\"", "G" ], [ "Ġ\"", "Loss\"," ], [ "Ġ\"", "S" ], [ "Ġ\"", "Value\"," ], [ "Ġ\"", "log" ], [ "Ġ(", "-" ], [ "Ġ(", "...," ], [ "Ġ(", "009" ], [ "Ġ(", "20" ], [ "Ġ(", "b" ], [ "Ġ(", "h," ], [ "Ġ(", "i" ], [ "Ġ(", "x_c" ], [ "Ġ(0.", "1," ], [ "Ġ-", "(" ], [ "Ġ-", "=" ], [ "Ġ0", ".3" ], [ "Ġ0.", "7" ], [ "Ġ0.", "8" ], [ "Ġ01", "0)" ], [ "Ġ1", "):" ], [ "Ġ1", "]" ], [ "Ġ1.0", ")" ], [ "Ġ1.0", "," ], [ "Ġ10000", "^(2i" ], [ "Ġ1e-1", "2)" ], [ "Ġ2", ")" ], [ "Ġ2", "-3" ], [ "Ġ2", ".0" ], [ "Ġ2", ".5" ], [ "Ġ2", ":" ], [ "Ġ2", "]" ], [ "Ġ201", "8" ], [ "Ġ2017", ")ãĢĤ" ], [ "Ġ202", "4" ], [ "Ġ3", ":" ], [ "Ġ4", ".5))" ], [ "Ġ:", "," ], [ "Ġ<-", ">" ], [ "ĠA", "," ], [ "ĠA", "/N^alpha" ], [ "ĠA", "BC" ], [ "ĠAdam", "W" ], [ "ĠB", "/D^beta" ], [ "ĠC", "rossAttention" ], [ "ĠCallable[[", "np.ndarray]," ], [ "ĠCharacterLevelTokenizer", ":" ], [ "ĠChinchilla", "ParametricFit" ], [ "ĠCounter", "[tuple[str," ], [ "ĠE", "M" ], [ "ĠF", "ace" ], [ "ĠF", "ig" ], [ "ĠF", "unction" ], [ "ĠJ", "SON" ], [ "ĠL", "2" ], [ "ĠL", "aw" ], [ "ĠL", "earn" ], [ "ĠL", "evenberg-Marquardt" ], [ "ĠLayerNorm(x", ")," ], [ "ĠLongTensor", "(``" ], [ "ĠLongTensor", "ãĢĤ" ], [ "ĠM", "ean" ], [ "ĠMultiHeadAttention", "(d_model," ], [ "ĠN", "," ], [ "ĠN", "_opt" ], [ "ĠN", "otImplementedError" ], [ "ĠNormalization", "\"," ], [ "ĠPath", "," ], [ "ĠQuery", "ãĥ»Key" ], [ "ĠRMS", "Norm" ], [ "ĠR^{", "d_ff" ], [ "ĠS", "cal" ], [ "ĠS", "ennrich" ], [ "ĠS", "haw" ], [ "ĠS", "iLU" ], [ "ĠS", "t" ], [ "ĠS", "wiGLU" ], [ "ĠSequence[str", "])" ], [ "ĠSquar", "e" ], [ "ĠT", "ext" ], [ "ĠT", "rue(æĹ¢å®ļåĢ¤" ], [ "ĠUn", "it" ], [ "ĠUnigram", "Tokenizer" ], [ "ĠValueError", "(\"" ], [ "ĠW^", "O" ], [ "Ġ[", "p." ], [ "Ġ\\\\", "lambda" ], [ "Ġ_", "WIKIPEDIA_MANIFEST_DIR" ], [ "Ġ_", "chinchilla_residual" ], [ "Ġ_", "gnmt_length_penalty" ], [ "Ġ_", "levenberg_marquardt(" ], [ "Ġ_", "strip_wikitext_links(" ], [ "Ġ_count_pair", "s(" ], [ "Ġ_merge_", "pair_in_word(" ], [ "Ġ``", "(num_windows," ], [ "Ġ``", "6" ], [ "Ġ``", "L" ], [ "Ġ``", "MultiHeadAttention``" ], [ "Ġ``", "lp(Y" ], [ "Ġ``", "p``" ], [ "Ġ``\"", "gradient_norm" ], [ "Ġ``(", "1," ], [ "Ġ``(", "N," ], [ "Ġ``(", "S," ], [ "Ġ``C", "``" ], [ "Ġ``beam_size", "``" ], [ "Ġ``n", "``" ], [ "Ġ``{", "æĿ¡ä»¶åĲį:" ], [ "Ġ``âĸģ", "``" ], [ "Ġa", "_coef" ], [ "Ġalpha", "=-" ], [ "Ġaxis", "=\"y\")" ], [ "Ġb", "_coef" ], [ "Ġb", "lock" ], [ "Ġb", "oundaries" ], [ "Ġbest", "_score" ], [ "Ġbeta", ":" ], [ "Ġbool", "]:" ], [ "Ġc", "enter" ], [ "Ġc", "um_log_prob" ], [ "Ġchunk", ":" ], [ "Ġchunk", "_freq" ], [ "Ġcolor", "=\"" ], [ "Ġcolor", "_cycle[i" ], [ "Ġcre", "ate_causal_mask" ], [ "Ġcross_", "attn_weights" ], [ "Ġd_", "values," ], [ "Ġd_k", "/" ], [ "Ġd_model", "))" ], [ "Ġd_model", ":" ], [ "Ġd_model)``", "ãĢģ" ], [ "Ġdata", "[\"" ], [ "Ġdata", "class" ], [ "Ġdecod", "e(self," ], [ "Ġdevice=", "query.device)" ], [ "Ġdict", "(" ], [ "Ġdict", "[float," ], [ "Ġdropout=", "dropout)" ], [ "Ġdropout=", "dropout," ], [ "Ġen", "d" ], [ "Ġf", "ontsize=" ], [ "Ġf", "unctional" ], [ "Ġf\"", "{" ], [ "Ġfloat", "]," ], [ "Ġfloat(\"-inf\"", "))" ], [ "Ġfloat(np.", "mean" ], [ "Ġg", "_t" ], [ "Ġg", "enerator" ], [ "Ġg", "et_random_batch" ], [ "Ġgrid", "]," ], [ "Ġh", "ead" ], [ "Ġhuber_delta", ")" ], [ "Ġid", "(p)" ], [ "Ġin", "put" ], [ "Ġint", "]" ], [ "Ġitem", ":" ], [ "Ġjson.loads(", "Path" ], [ "Ġk", "," ], [ "Ġl", "in" ], [ "Ġl", "ine" ], [ "Ġl", "m_head" ], [ "Ġlabel=", "name," ], [ "Ġlen(color_cycle", ")]" ], [ "Ġlist[", "Tensor]" ], [ "Ġlist[float", "]]," ], [ "Ġload_", "bpe_" ], [ "Ġload_", "wikipedia_corpus" ], [ "Ġlog_alpha", "_max" ], [ "Ġlog_n", "_opt" ], [ "Ġlogit", "s``" ], [ "Ġlogits", "ãĢĤ" ], [ "Ġlr", ":" ], [ "Ġm", "," ], [ "Ġm", "ulti-start" ], [ "Ġmask", ".dim()" ], [ "Ġmatrix", ".ndim" ], [ "Ġmax_", "relative_position" ], [ "Ġmax_chunk", "_bytes:" ], [ "Ġmax_exact", ")" ], [ "Ġmemory_", "mask)" ], [ "Ġn", "oqa:" ], [ "Ġn", "ormalization_factory" ], [ "Ġn", "p" ], [ "Ġn_", "ctx" ], [ "Ġnam", "e" ], [ "Ġnew", "_symbol" ], [ "Ġnp.", "log(" ], [ "Ġnp.array([", "p." ], [ "Ġnp.ndarray", ")" ], [ "Ġnum", "_total" ], [ "Ġo", "ther_" ], [ "Ġp", "ath" ], [ "Ġp", "iece" ], [ "Ġp", "retokenize(" ], [ "Ġpair", "," ], [ "Ġpair", "_in" ], [ "Ġparam", "s)" ], [ "Ġpoint", "s]" ], [ "Ġrange(", "num" ], [ "Ġre", "construct_optimal_frontier(" ], [ "Ġre.sub(r\"", "(?" ], [ "Ġrelative_", "buckets" ], [ "Ġresidual_fn", "(param" ], [ "Ġresidual_fn", "(params)" ], [ "Ġs", "ampling" ], [ "Ġs", "core" ], [ "Ġs", "oftmax" ], [ "Ġs", "s_" ], [ "Ġse", "quence" ], [ "Ġself.", "_split_heads(self.w_" ], [ "Ġself.", "attention_score_bias" ], [ "Ġself.", "gamma" ], [ "Ġself.norm", "2" ], [ "Ġsequence_length", ")" ], [ "Ġset", "()" ], [ "Ġset", "[str]" ], [ "Ġsize", ")" ], [ "Ġsrc.", "utils." ], [ "Ġsrc.layers.", "attention" ], [ "Ġsrc.layers.", "normalization" ], [ "Ġsrc.layers.", "positional_encoding" ], [ "Ġstr", "]" ], [ "Ġstr", "]," ], [ "Ġstrict", "=True)" ], [ "Ġsymbol", "s" ], [ "Ġsymbol", "s[i" ], [ "Ġt", "rain" ], [ "Ġt", "ry_decode_byte_level_symbol" ], [ "Ġtarget", "s" ], [ "Ġtoken_id", "s" ], [ "Ġtorch.", "Generator" ], [ "Ġtorch.", "m" ], [ "Ġtorch.arange(", "key_length," ], [ "Ġtorch.arange(", "query_length," ], [ "Ġtorch.zero", "s(" ], [ "Ġun", "it)" ], [ "Ġv", "max" ], [ "Ġvalues", "_arr" ], [ "Ġwarmup_step", "s" ], [ "Ġx", "^" ], [ "Ġzip", "(" ], [ "Ġ{", "\"" ], [ "Ġ{", "ch" ], [ "ĠÎµ", ")" ], [ "Ġâ", "ī" ], [ "Ġâī", "Ī" ], [ "Ġãģ¤ãģ®æĸĩåŃĹåĪĹ", "ãĢĤ" ], [ "Ġãģ§", "ãĢģ" ], [ "Ġãģ§", "ãģĤãĤĭ" ], [ "Ġãģ§è¿½åĬł", ")ãĢĤ" ], [ "Ġãģ¨", "ãģ®åĨħç©įãĤĴ" ], [ "Ġãģ¨åĲĮãģĺå½¢çĬ¶", "ãģ®åĩºåĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ" ], [ "Ġãģ¨åĲĮä¸Ģãģ®æĮĻåĭķ", ")ãĢĤ" ], [ "Ġãģªãģ©", ")ãĢĤ" ], [ "Ġãģ«", "åŁºãģ¥ãģı" ], [ "Ġãģ«", "éģĶ" ], [ "Ġãģ®", "ãĥŀãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ" ], [ "Ġãģ®", "åĨħå®¹" ], [ "Ġãģ®", "æ¸¬å®ļåĢ¤ãģ®ç³»åĪĹ" ], [ "Ġãģ®", "æĹ¢å®ļ" ], [ "Ġãģ®", "ç³»åĪĹ" ], [ "Ġãģ®", "è¨Ģèªŀ" ], [ "Ġãģ®", "éķ·ãģķ" ], [ "Ġãģ®", "éłĨä¼ĿæĴŃãĢĤ" ], [ "Ġãģ®ãĤµãĥĸãĤ¯ãĥ©ãĤ¹", "ãĢģä¾ĭ:" ], [ "Ġãģ®ä¸Ńéĸĵå±¤", "æ¬¡åħĥ" ], [ "Ġãģ®ä¸Ńéĸĵå±¤", "æ¬¡åħĥãĢĤ" ], [ "Ġãģ®å®Łé¨ĵ", "5" ], [ "Ġãģ®çµĲæŀľãĢĤ", "L" ], [ "Ġãģ®çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³", "åıĤçħ§)ãĢĤ" ], [ "Ġãģ¯", "å½¢çĬ¶" ], [ "Ġãģ¾ãģ§ãģ®éĢ£çķªãģ¨ãģĹãģ¦æī±", "ãģĨãĢĤ" ], [ "ĠãģĭãĤī", "ãģ®åıĸå¾Ĺ" ], [ "ĠãģĮ", "ãĥĳãĥĩãĤ£ãĥ³ãĤ°" ], [ "ĠãģĮ", "å®ŁãĥĪãĥ¼ãĤ¯ãĥ³" ], [ "ĠãĤ¹ãĥĨãĥĥãĥĹ", "çķªåı·" ], [ "ĠãĤ¹ãĥĨãĥĥãĥĹãģĶãģ¨", "ãģ®åĭ¾éħįãĥİãĥ«ãĥł" ], [ "ĠãĤĴ", "ãĢģ" ], [ "ĠãĤĴ", "è¨Īç®ĹãģĻãĤĭãĢĤ" ], [ "ĠãĤĴ", "è¿ĶãģĻ" ], [ "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥", "ãĤĴ" ], [ "Ġãĥ", "ĩ" ], [ "Ġãĥ", "İ" ], [ "Ġãĥ", "Ļ" ], [ "Ġãĥ©ãĤ¦ãĥ³ãĥīä¿®æŃ£", "ãĢģ" ], [ "ĠãĥĪãĥ¼ãĤ¯ãĥ³", "ãģĤãģŁãĤĬ" ], [ "ĠãĥĪãĥ¼ãĤ¯ãĥ³", "ãģĤãģŁãĤĬãģ®è¨Īç®Ĺéĩı" ], [ "ĠãĥĲ", "ãĤ±ãĥĥãĥĪ" ], [ "ĠãĥĲãĤ¤ãĥĪéķ·", "ãģ®ä¸ĬéĻĲ" ], [ "ĠãĥĻ", "ãĥ¼ãĤ¹ãģ®åıįå¾©çļĦãģªç¸®å°ı" ], [ "Ġä»¥ä¸Ĭ", ")ãĢĤ" ], [ "Ġä»¥ä¸ĭ", ")ãĢĤ" ], [ "Ġå", "ĭ¾éħį" ], [ "Ġå", "įĺ" ], [ "Ġå±¤", "çĽ®" ], [ "Ġåģ´", ")" ], [ "Ġåģ´ãĢģn", ":" ], [ "Ġåı", "ĤåĬł" ], [ "ĠåıĤçħ§", "ãĢģ006" ], [ "Ġæ", "ĻĤçĤ¹" ], [ "ĠæĸĩåŃĹ", "ãĤĴ" ], [ "ĠæŃ£", "ãģ®å®Łæķ°åĢ¤ãģ®ç³»åĪĹ" ], [ "ĠæŃ£è¦ıåĮĸ", "ãĤĴè¡ĮãģĨçī¹å¾´" ], [ "Ġç", "¬¦åı·åĮĸ" ], [ "Ġç³»åĪĹéķ·", "(sequence" ], [ "Ġç¾¤", "ãģĶãģ¨ãģ®" ], [ "ĠçĽ", "¸" ], [ "Ġèĩªèº«ãģ®", "åĪĨåī²çµĲæŀľ" ], [ "Ġéĩįãģ¿", "ãĥ»åĲĦãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ" ], [ "Ġéĩįãģ¿æ¸Ľè¡°", "ä¿Ĥæķ°" ], [ "İ", "åĽŀ" ], [ "İ", "ç¤º" ], [ "ĺ", "İç¤º" ], [ "!r", "}" ], [ "\"", ".join(" ], [ "\"\"", ".join(" ], [ "\"\".join(", "current" ], [ "\"\".join(current", "))" ], [ "\")", "]" ], [ "\"].append(", "step)" ], [ "&", "oldid" ], [ "&", "prop" ], [ "&oldid", "=" ], [ "&oldid=", "<" ], [ "&oldid=<", "ãĥªãĥĵãĤ¸ãĥ§ãĥ³" ], [ "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³", "ID" ], [ "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³ID", ">``" ], [ "&prop", "=" ], [ "&prop=", "extra" ], [ "&prop=extra", "ct" ], [ "&prop=extract", "s``" ], [ "(", "...," ], [ "(", "1)" ], [ "(", "429" ], [ "(", "=" ], [ "(", "A" ], [ "(", "Attention" ], [ "(", "D" ], [ "(", "Encoder" ], [ "(", "Key" ], [ "(", "MediaWiki" ], [ "(", "N" ], [ "(", "Out" ], [ "(", "Q," ], [ "(", "Query" ], [ "(", "Res" ], [ "(", "Scaling" ], [ "(", "Vaswani" ], [ "(", "Wikipedia" ], [ "(", "`" ], [ "(", "alpha" ], [ "(", "cache_dir:" ], [ "(", "cl" ], [ "(", "clip" ], [ "(", "cross_" ], [ "(", "d_" ], [ "(", "d_ff," ], [ "(", "data[\"" ], [ "(", "design" ], [ "(", "exponent" ], [ "(", "htt" ], [ "(", "iter" ], [ "(", "key" ], [ "(", "l_inf" ], [ "(", "model.parameters()" ], [ "(", "module" ], [ "(", "repo_id=repo_id," ], [ "(", "retry_wait_second" ], [ "(", "s)" ], [ "(", "self)" ], [ "(", "seq_len," ], [ "(", "set" ], [ "(", "text" ], [ "(", "text:" ], [ "(", "token_ids," ], [ "(", "values_by_condition" ], [ "(", "{" ], [ "(", "ãģĦãģļãĤĮãĤĤ" ], [ "(", "ãĤªãĥ¼ãĥĲãĥ¼ãĥķãĥŃãĥ¼" ], [ "(", "ãĤ¹ãĥĨãĥĥãĥĹ" ], [ "(", "ãĥĲãĥĥãĥģ" ], [ "(", "äºĭåīįåĪĨåī²" ], [ "(", "å¤ļ" ], [ "(", "å¯¾æķ°" ], [ "(", "å¹³åĿĩ" ], [ "(", "å¹³åĿĩæ¸Ľç®Ĺ" ], [ "(", "åĢĻè£ľèªŀå½ĻãģĭãĤīãģ®" ], [ "(", "åħ¨ãĥĺãĥĥãĥī" ], [ "(", "åįĺ" ], [ "(", "åŃ¦ç¿Ĵ" ], [ "(", "æ¬¡" ], [ "(", "æµ®åĭķå°ıæķ°çĤ¹" ], [ "(", "æĸĩ" ], [ "(", "æľĢ" ], [ "(", "ç©ºçĻ½ãģ®" ], [ "(", "çĽ´åīį" ], [ "(", "è©³ç´°ãģ¯" ], [ "(", "éĩįãģ¿" ], [ "(\"", "." ], [ "(\"", "\\n" ], [ "(\"", "en" ], [ "(\"", "ja" ], [ "(\"", "nan" ], [ "(\"en", "\"," ], [ "(\"ja", "\"," ], [ "(\"nan", "\")," ], [ "(00", "2" ], [ "(006", ")ãĢĤ" ], [ "(1", ".0" ], [ "(1-indexed", ")ãĢĤ" ], [ "(C", "D" ], [ "(CD", "F" ], [ "(G", "N" ], [ "(G", "ated" ], [ "(G", "aussian" ], [ "(GN", "M" ], [ "(GNM", "T" ], [ "(GNMT", ")ãĤ¹ãĤ¿ãĤ¤ãĥ«" ], [ "(GNMT)ãĤ¹ãĤ¿ãĤ¤ãĥ«", "ãģ®éķ·ãģķ" ], [ "(LayerNorm(x", "))" ], [ "(Out", "-" ], [ "(Out-", "of" ], [ "(Out-of", "-" ], [ "(Out-of-", "V" ], [ "(Out-of-V", "ocab" ], [ "(Out-of-Vocab", "ulary" ], [ "(R", "adford" ], [ "(Res", "idual" ], [ "(V", ")" ], [ "(Wikipedia", ")" ], [ "(Wikipedia)", "ãĢ" ], [ "(Wikipedia)ãĢ", "ı" ], [ "(Y", ")" ], [ "([", "]," ], [ "(``", "1" ], [ "(``", "DecoderBlock``" ], [ "(``", "GPTLanguageModel``" ], [ "(``", "None``" ], [ "(``", "QueryKeyPositionalTransform``" ], [ "(``", "_le" ], [ "(``", "encod" ], [ "(``", "learn_bpe" ], [ "(``", "lm_head``)ãģ®" ], [ "(``", "make_evaluation_windows``" ], [ "(``", "nn.Parameter" ], [ "(``\"", "ja\"``ãĥ»``\"en\"``" ], [ "(``_le", "venberg_m" ], [ "(``_levenberg_m", "arquardt" ], [ "(``list", "[int]" ], [ "(``list[int]", "``" ], [ "(``nn.Parameter", "``" ], [ "(``numpy.linalg.lstsq", "``)" ], [ "(``src/", "training/" ], [ "(``src/data/", "text" ], [ "(alpha", "+" ], [ "(alpha+", "beta" ], [ "(b", "atch_size," ], [ "(b", "est" ], [ "(b", "pe" ], [ "(b", "pe." ], [ "(beam", "_sequences" ], [ "(beam_sequences", "," ], [ "(best", "_params)" ], [ "(best_params)", "," ], [ "(bpe", "," ], [ "(con", "text)" ], [ "(d_model", "))" ], [ "(de", "ad" ], [ "(design", "," ], [ "(g", "re" ], [ "(gre", "ed" ], [ "(greed", "y" ], [ "(hidden", ")" ], [ "(id", "s[i" ], [ "(length", "s))," ], [ "(lengths_without_leading_whitespace", "))," ], [ "(log", "its" ], [ "(log_n", "," ], [ "(logit", "s," ], [ "(logit", "s:" ], [ "(m", "_hat" ], [ "(m", "atrix" ], [ "(m", "emory" ], [ "(model.parameters()", "," ], [ "(n", "-" ], [ "(n", "\\\\text{-grams" ], [ "(n", "ats)ãĢĤ" ], [ "(n", "grams" ], [ "(n", "grams)" ], [ "(n", "oisy_l" ], [ "(n", "on-" ], [ "(n\\\\text{-grams", "})" ], [ "(n\\\\text{-grams})", "|}{|" ], [ "(n\\\\text{-grams})|}{|", "n" ], [ "(new_symbol", ")" ], [ "(next", "(iter" ], [ "(ngrams", "))" ], [ "(nn.Module", "," ], [ "(norm", "ed," ], [ "(np.", "as" ], [ "(np.", "exp(" ], [ "(np.as", "array(" ], [ "(param", "s," ], [ "(path", ":" ], [ "(path)", ".read_text(encoding=\"utf-8\"))" ], [ "(re-", "scaling" ], [ "(residual_fn", "," ], [ "(retry_wait_second", "s" ], [ "(self.", "dropout" ], [ "(self.f", "eed_forwar" ], [ "(self.f", "eed_forward" ], [ "(self.feed_forwar", "d(" ], [ "(self.feed_forward", "(x)))" ], [ "(self.feed_forward(", "self." ], [ "(self.feed_forward(self.", "norm" ], [ "(weight", "s)" ], [ "(x", "_n" ], [ "(ãĤ¹", "ãĥļ" ], [ "(ãĤ¹ãĥļ", "ãĥ¼ãĤ¹" ], [ "(ãĤ¹ãĥļãĥ¼ãĤ¹", "ãĥ»" ], [ "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶", "ãģ®" ], [ "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®", "ç²" ], [ "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²", "Ĵ" ], [ "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²Ĵ", "åº¦" ], [ "(äºĭåīįåĪĨåī²", "ãģ§å¾ĹãģŁ" ], [ "(å¯¾æķ°", "å°¤åº¦" ], [ "(åħ¨ãĥĺãĥĥãĥī", "åĪĨãĤĴ" ], [ "(æµ®åĭķå°ıæķ°çĤ¹", "æ¼Ķç®Ĺ" ], [ "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®Ĺ", "åĽŀæķ°" ], [ "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®ĹåĽŀæķ°", ")" ], [ "(æĹ¢å®ļåĢ¤", "ãĢģ" ], [ "(çĽ´åīį", "ãĤ¹ãĥĨãĥĥãĥĹãģ¨ãģ®" ], [ "(èªŀå½Ļ", "ãģ¸" ], [ "(èªŀå½Ļãģ¸", "ãģ®å°Ħå½±" ], [ ")", "\"" ], [ ")", "\")" ], [ ")", "ãģ¨ãģĹãģ¦" ], [ ")", "ãģªãģ©" ], [ ")", "ãģ«ãĤĪãĤĬ" ], [ ")", "ãģ®åł´åĲĪãĢģ" ], [ ")", "ãģ¸" ], [ ")", "ãģĵãģ¨ãģĮæľŁå¾ħãģķãĤĮãĤĭãĢĤ" ], [ ")", "ãģĵãģ¨ãĤĴæĦıåĳ³" ], [ ")", "ãģŁãĤģãĢģ" ], [ ")", "ãĤĦ" ], [ ")", "åł´åĲĪãģ¯" ], [ ")^", "T" ], [ ")^", "alpha" ], [ ")``", ")ãĢĤ" ], [ ")ãĢģ", "False" ], [ ")ãĢģ", "æİ¥ç¶ļ" ], [ ")ãĢģæİ¥ç¶ļ", "ç¢ºç«ĭ" ], [ ")ãĢĤ", "**" ], [ ")ãĢĤ", "False" ], [ ")ãĢĤ``None``", "(æĹ¢å®ļåĢ¤)ãģ®" ], [ ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯", "å¾ĵæĿ¥éĢļãĤĬ" ], [ ")ãģ«", "ä¾ĿåŃĺ" ], [ ")ãģ®", "ãģ¿" ], [ ")ãģ®", "ãĤĪãģĨãģ«" ], [ ")ãģ®", "ãĥ©ãĥĻãĥ«ãĢĤ" ], [ ")ãģ®", "åħ±éĢļ" ], [ ")ãģ®ãĤĪãģĨãģ«", "ãĢģ" ], [ ")ãģ®åħ±éĢļ", "ãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹" ], [ ")ãģ®åħ±éĢļãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹", "ãĢĤ" ], [ ")ãģĵãģ¨ãĤĴæĦıåĳ³", "ãģĻãĤĭãĢĤ" ], [ ")ãĤĴ", "ä½ľ" ], [ ")ãĤĴ", "å¤±" ], [ ")ãĤĴ", "æĮģãģ¤" ], [ ")ãĤĴ", "è¨Īç®Ĺ" ], [ ")ãĤĴ", "éģ¿" ], [ "**", "8" ], [ "**", "beta" ], [ "+", "\\s*" ], [ "+", "åĩºåĬĽå±¤" ], [ "+åĩºåĬĽå±¤", "ãĢįãģ®" ], [ "-", "\"," ], [ "-", "2)" ], [ "-", "Decoder" ], [ "-", "N" ], [ "-", "level" ], [ "-", "max" ], [ "-", "w" ], [ "-", "x" ], [ "-", "}" ], [ "-", "ç¶" ], [ "-1", "})" ], [ "-8", "h" ], [ "-8h", "/" ], [ "-8h/", "H" ], [ "-8h/H", ")" ], [ "-S", "A" ], [ "-attention", ")" ], [ "-n", "egative" ], [ "-n_val", ":" ], [ "-n_val", "]," ], [ "-n_val:", "]" ], [ "-w", "ise" ], [ "-x", "2," ], [ "-}", "n" ], [ "-ç¶", "Ļ" ], [ "-ç¶Ļ", "æī" ], [ "-ç¶Ļæī", "¿" ], [ ".", "4" ], [ ".", "6" ], [ ".", "D" ], [ ".", "T" ], [ ".", "cross_" ], [ ".", "default" ], [ ".", "detach()." ], [ ".", "dtype," ], [ ".", "exponent" ], [ ".", "fig" ], [ ".", "forward``" ], [ ".", "l" ], [ ".", "law" ], [ ".", "log(" ], [ ".", "pow(2)" ], [ ".", "view" ], [ ".", "vocab``" ], [ "...", ")``" ], [ "...", "]," ], [ "...", "]]" ] ], "vocab": [ "!", "!=", "!r", "!r}", "\"", "\"\"", "\"\"\"", "\"\".join(", "\"\".join(current", "\"\".join(current))", "\")", "\"))", "\"),", "\")]", "\",", "\".join(", "\":", "\"]", "\"],", "\"].append(", "\"].append(step)", "\"].b", "\"].by_", "\"].by_key", "\"].by_key()", "\"].by_key()[\"", "\"].by_key()[\"color", "\"].by_key()[\"color\"]", "\"``", "\"``:", "\"``ãĥ»``", "\"``ãĥ»``\"en", "\"``ãĥ»``\"en\"``", "\"axes", "\"axes.", "\"axes.prop", "\"axes.prop_cycle", "\"axes.prop_cycle\"].by_key()[\"color\"]", "\"en", "#", "$", "%", "&", "&oldid", "&oldid=", "&oldid=<", "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³", "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³ID", "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³ID>``", "&prop", "&prop=", "&prop=extra", "&prop=extract", "&prop=extracts``", "'", "''", "(", "(\"", "(\".", "(\"\\n", "(\"en", "(\"en\",", "(\"ja", "(\"ja\",", "(\"nan", "(\"nan\"),", "()", "())", "(),", "().", "().__init__", "().__init__()", "().item", "().item()", "().item())", "()``", "(...,", "(0", "(0)}", "(00", "(001", "(001ãĥ»002", "(001ãĥ»002ãĥ»00", "(001ãĥ»002ãĥ»003", "(002", "(003", "(004", "(005", "(006", "(006)ãĢĤ", "(006ãĢģ", "(007", "(009", "(1", "(1)", "(1-indexed", "(1-indexed)ãĢĤ", "(1.0", "(2", "(2)", "(2i", "(429", "(5", "(=", "(?", "(A", "(Attention", "(AttentionScoreBias", "(AttentionScoreBias):", "(B,", "(Byte", "(C", "(CD", "(CDF", "(D", "(De", "(Encoder", "(G", "(GN", "(GNM", "(GNMT", "(GNMT)ãĤ¹ãĤ¿ãĤ¤ãĥ«", "(GNMT)ãĤ¹ãĤ¿ãĤ¤ãĥ«ãģ®éķ·ãģķ", "(Gated", "(Gaussian", "(Key", "(L", "(L2", "(Layer", "(LayerNorm", "(LayerNorm(x", "(LayerNorm(x))", "(LayerNorm(x),", "(MediaWiki", "(N", "(N,", "(N,D", "(N,D)", "(Out", "(Out-", "(Out-of", "(Out-of-", "(Out-of-V", "(Out-of-Vocab", "(Out-of-Vocabulary", "(Q,", "(Query", "(R", "(Radford", "(Res", "(Residual", "(S", "(Scaling", "(T", "(True", "(V", "(V)", "(Vaswani", "(W", "(Wikipedia", "(Wikipedia)", "(Wikipedia)ãĢ", "(Wikipedia)ãĢı", "(Y", "(Y)", "([", "([],", "(`", "(``", "(``\"", "(``\"ja\"``ãĥ»``\"en\"``", "(``1", "(``DecoderBlock``", "(``GPTLanguageModel``", "(``None``", "(``QueryKeyPositionalTransform``", "(``_le", "(``_levenberg_m", "(``_levenberg_marquardt", "(``compute_", "(``encod", "(``for", "(``forward(", "(``forward(token_ids)", "(``learn_bpe", "(``list", "(``list[int]", "(``list[int]``", "(``lm_head``)ãģ®", "(``make_evaluation_windows``", "(``n", "(``nn.Parameter", "(``nn.Parameter``", "(``numpy", "(``numpy.", "(``numpy.linalg.lstsq", "(``numpy.linalg.lstsq``)", "(``pretokenize``", "(``src/", "(``src/data/", "(``src/data/text", "(``src/training/", "(``train_language_model``", "(a", "(a)", "(alpha", "(alpha+", "(alpha+beta", "(alpha=0.3", "(b", "(batch", "(batch_size,", "(beam", "(beam_sequences", "(beam_sequences,", "(best", "(best_params)", "(best_params),", "(bpe", "(bpe,", "(bpe.", "(c", "(cache_dir:", "(chunk", "(chunk)", "(cl", "(clip", "(con", "(context)", "(cross", "(cross_", "(d_", "(d_ff,", "(d_model", "(d_model)", "(d_model))", "(d_model,", "(data[\"", "(de", "(dead", "(design", "(design,", "(dim", "(dropout)", "(exponent", "(g", "(g,", "(gre", "(greed", "(greedy", "(hidden", "(hidden)", "(htt", "(i", "(id", "(ids,", "(ids[i", "(iter", "(key", "(l_inf", "(len(", "(length", "(lengths)),", "(lengths_without_leading_whitespace", "(lengths_without_leading_whitespace)),", "(log", "(log_n", "(log_n,", "(logit", "(logits", "(logits,", "(logits:", "(m", "(m:", "(m_hat", "(matrix", "(memory", "(model.parameters()", "(model.parameters(),", "(module", "(n", "(n,", "(n-", "(n\\\\text{-grams", "(n\\\\text{-grams})", "(n\\\\text{-grams})|}{|", "(n\\\\text{-grams})|}{|n", "(nats)ãĢĤ", "(new", "(new_symbol", "(new_symbol)", "(next", "(next(iter", "(ngrams", "(ngrams)", "(ngrams))", "(nn.", "(nn.Module", "(nn.Module):", "(nn.Module,", "(noisy_l", "(non-", "(norm", "(normed,", "(np.", "(np.arange(", "(np.as", "(np.asarray(", "(np.exp(", "(np.exp(log", "(num", "(num_heads)", "(num_window", "(num_windows,", "(p", "(p)", "(p)}", "(param", "(params)", "(params,", "(parent", "(parents", "(parents=True,", "(path", "(path)", "(path).read_text(encoding=\"utf-8\"))", "(path:", "(pos,", "(positions.", "(positions.max", "(positions.max().item())", "(pre-", "(pre-token", "(pre-tokenization", "(query", "(r\"", "(re-", "(re-centering", "(re-scaling", "(repo_id=repo_id,", "(residual_fn", "(residual_fn,", "(retry_wait_second", "(retry_wait_seconds", "(s)", "(se", "(seed)", "(self)", "(self,", "(self.", "(self.beta", "(self.dropout", "(self.f", "(self.feed_forwar", "(self.feed_forward", "(self.feed_forward(", "(self.feed_forward(self.", "(self.feed_forward(self.norm", "(self.feed_forward(x)))", "(seq_len,", "(sequence", "(set", "(symbol", "(text", "(text:", "(title", "(token", "(token_ids,", "(torch.", "(values_by_condition", "(vocabulary", "(weight", "(weights)", "(x", "(x)", "(x))", "(x)))", "(x,", "(x:", "(x_arr", "(x_n", "(xlabel)", "({", "(ãģ¾ãģŁãģ¯", "(ãģĦãģļãĤĮãĤĤ", "(ãĤªãĥ¼ãĥĲãĥ¼ãĥķãĥŃãĥ¼", "(ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°", "(ãĤ¹", "(ãĤ¹ãĥĨãĥĥãĥĹ", "(ãĤ¹ãĥļ", "(ãĤ¹ãĥļãĥ¼ãĤ¹", "(ãĤ¹ãĥļãĥ¼ãĤ¹ãĥ»", "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶", "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®", "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²", "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²Ĵ", "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²Ĵåº¦", "(ãĥĲãĥĥãĥģ", "(äºĭåīįåĪĨåī²", "(äºĭåīįåĪĨåī²ãģ§å¾ĹãģŁ", "(ä¾ĭ:", "(å¤ļ", "(å®Łé¨ĵ", "(å¯¾æķ°", "(å¯¾æķ°å°¤åº¦", "(å¯¾æķ°ç©ºéĸĵ", "(å¹³åĿĩ", "(å¹³åĿĩæ¸Ľç®Ĺ", "(å½¢çĬ¶", "(åĢĻè£ľèªŀå½ĻãģĭãĤīãģ®", "(åħ¨ãĥĺãĥĥãĥī", "(åħ¨ãĥĺãĥĥãĥīåĪĨãĤĴ", "(åĭ¾éħį", "(åįĺ", "(åŃ¦ç¿Ĵ", "(æ¬¡", "(æ¯Ķè¼ĥå¯¾è±¡", "(æµ®åĭķå°ıæķ°çĤ¹", "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®Ĺ", "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®ĹåĽŀæķ°", "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®ĹåĽŀæķ°)", "(æĸĩ", "(æĹ¢å®ļåĢ¤", "(æĹ¢å®ļåĢ¤)", "(æĹ¢å®ļåĢ¤)ãģ®", "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪ", "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãĢģ", "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯", "(æĹ¢å®ļåĢ¤ãĢģ", "(æľĢ", "(ç©ºçĻ½ãģ®", "(çĽ´åīį", "(çĽ´åīįãĤ¹ãĥĨãĥĥãĥĹãģ¨ãģ®", "(çľģçķ¥åı¯", "(çľģçķ¥åı¯)ãĢĤ", "(è©³ç´°ãģ¯", "(èªŀå½Ļ", "(èªŀå½Ļãģ¸", "(èªŀå½Ļãģ¸ãģ®å°Ħå½±", "(èªŀå½ĻãĤµãĤ¤ãĤº", "(éĩįãģ¿", ")", ")\"", ")\")", ")\",", "))", ")))", ")),", ")):", "),", ").", "):", ")]", ")],", ")^", ")^2", ")^T", ")^alpha", ")``", ")``)ãĢĤ", ")``ãĢģ", ")``ãĢĤ", ")}", ")ãĢģ", ")ãĢģFalse", ")ãĢģæİ¥ç¶ļ", ")ãĢģæİ¥ç¶ļç¢ºç«ĭ", ")ãĢĤ", ")ãĢĤ\"\"\"", ")ãĢĤ**", ")ãĢĤFalse", ")ãĢĤ``", ")ãĢĤ``None``", ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)", ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®", ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯", ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯å¾ĵæĿ¥éĢļãĤĬ", ")ãģ§", ")ãģ§ãĢģ", ")ãģ§ãģ¯", ")ãģ¨", ")ãģ¨ãģĹãģ¦", ")ãģªãģ©", ")ãģ«", ")ãģ«ãģª", ")ãģ«ãĤĪãĤĬ", ")ãģ«ä¾ĿåŃĺ", ")ãģ«å¯¾ãģĹãģ¦", ")ãģ®", ")ãģ®ãģ¿", ")ãģ®ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãĢĤ", ")ãģ®ãĤĪãģĨãģ«", ")ãģ®ãĤĪãģĨãģ«ãĢģ", ")ãģ®ãĥ©ãĥĻãĥ«ãĢĤ", ")ãģ®åħ±éĢļ", ")ãģ®åħ±éĢļãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹", ")ãģ®åħ±éĢļãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹ãĢĤ", ")ãģ®åł´åĲĪãĢģ", ")ãģ¯", ")ãģ¯ãĢģ", ")ãģ¸", ")ãģĭãĤī", ")ãģĮ", ")ãģĵãģ¨ãģĮæľŁå¾ħãģķãĤĮãĤĭãĢĤ", ")ãģĵãģ¨ãĤĴæĦıåĳ³", ")ãģĵãģ¨ãĤĴæĦıåĳ³ãģĻãĤĭãĢĤ", ")ãģŁãĤģãĢģ", ")ãĤ¹ãĤ¿ãĤ¤ãĥ«", ")ãĤĦ", ")ãĤĴ", ")ãĤĴä½¿ãģĨãĢĤ", ")ãĤĴä½ľ", ")ãĤĴå", ")ãĤĴå¤±", ")ãĤĴæĮģãģ¤", ")ãĤĴè¨Īç®Ĺ", ")ãĤĴè¨Īç®ĹãģĻãĤĭãĢĤ", ")ãĤĴéģ¿", ")ãĥ»", ")åł´åĲĪãģ¯", "*", "**", "**(", "**2", "**8", "**:", "**alpha", "**beta", "**t", "*?", "+", "+\\s*", "+åĩºåĬĽå±¤", "+åĩºåĬĽå±¤ãĢįãģ®", ",", ",)", "-", "-\",", "--", "---", "----", "--------", "------------", "---------------", "----------------", "-------------------------------", "--------------------------------", "---------------------------------------------------------------", "-1", "-1,", "-1]", "-1}", "-1})", "-2", "-2)", "-3", "-8", "-8h", "-8h/", "-8h/H", "-8h/H)", "-A", "-Ag", "-Agent", "-Decoder", "-Forward", "-Head", "-Layer", "-M", "-Marquardt", "-N", "-P", "-Product", "-S", "-SA", "-attention", "-attention)", "-byt", "-gram", "-grams", "-in", "-index", "-indexed", "-inf", "-inf\"", "-inf``", "-level", "-max", "-n", "-n_val", "-n_val:", "-n_val:]", "-n_val],", "-negative", "-onl", "-only", "-per", "-per-byt", "-rn", "-rnn", "-start", "-theories", "-w", "-wise", "-x", "-x2,", "-}", "-}n", "-ç¶", "-ç¶Ļ", "-ç¶Ļæī", "-ç¶Ļæī¿", ".", ".*?", ".,", "..", "...", "...)``", "...,", "...],", "...]]", ".0", ".0,", ".2", ".2))", ".3", ".4", ".5", ".5))", ".5,", ".6", ".D", ".SentencePiece", ".T", ".ad", ".add", ".add_", ".append(", ".append(float", ".ax", ".b", ".bias", ".co", ".cross_", ".decod", ".default", ".detach().", ".device)", ".dim", ".dim()", ".dtype,", ".encod", ".encode(\"utf-8", ".encode(\"utf-8\"))", ".exponent", ".fig", ".forward``", ".g", ".get", ".grad", ".ipyn", ".items()", ".items()):", ".items():", ".items()}", ".jo", ".join(", ".json", ".json\"", ".key", ".keys()", ".l", ".law", ".log(", ".m", ".max", ".mean", ".mean(dim", ".mean(dim=-1,", ".min", ".mk", ".mkdir", ".mkdir(parents=True,", ".mul_", ".mul_(self.beta", ".n", ".ndim", ".norm", ".normal", ".org/", ".p", ".plot", ".plot(", ".pow(2)", ".power_law_fit", ".py", ".py``", ".rc", ".rcParam", ".rcParams[", ".rcParams[\"axes.prop_cycle\"].by_key()[\"color\"]", ".re", ".read_text(encoding=\"utf-8", ".read_text(encoding=\"utf-8\"))", ".replace", ".replace(\"", ".request", ".request.", ".s", ".sc", ".scal", ".scatter", ".set", ".set_", ".set_t", ".set_title", ".set_title(title", ".set_title(title)", ".shape", ".size(", ".size(-1", ".size(1)", ".sle", ".sleep", ".sqrt(", ".sub", ".subplot", ".subplots(", ".t", ".to(", ".to(dtype=", ".to(dtype=dtype)", ".trans", ".transpo", ".transpose", ".transpose(", ".tx", ".txt", ".txt\"", ".values", ".view", ".vocab``", ".weight", ".wikipedia", ".wikipedia.org/", "/", "//", "/D", "/D^", "/D^beta", "/N", "/N^", "/N^alpha", "/sqrt(", "0", "0)", "0,", "0.", "0.0", "0.0,", "00", "001", "002", "003", "004", "004_norm", "005", "006", "006ãĥ»00", "006ãĥ»008", "007", "009", "1", "1(", "1)", "1)ãĢĤ", "1,", "1.", "1.0", "12,", "16", "1:", "1_f", "1_found", "1_foundation", "1_foundations/", "2", "2)", "2)ãĢĤ", "2,", "2.0", "20", "25", "29", "2]", "3", "3)", "3,", "4", "4,", "429", "44", "5", "5,", "56", "6", "6,", "7", "7,", "71", "8", "8)", "8,", "9", "90", "99", "999", ":", ":2]", ":math:`", ";", "<", "", "", "=", "=\"", "=\"ident", "=\"identity", "=\"ign", "=\"ignor", "=\"ignore\"):", "=\"ignore\",", "=\"utf-8", "=\"y", "=\"y\")", "=-", "=-1", "=-1)", "=-1,", "=0", "=0.", "=0.0", "=1", "=1)", "=1.", "=1.0", "=1e", "=False", "=False)", "=False``", "=None", "=True", "=True)", "=True,", "=True``", "=b", "=bias)", "=color", "=color,", "=data[\"", "=float", "=float(np.exp(log", "=lambda", "=repo_id", "=repo_id,", "=str", "=str(", "={", ">", ">\",", ">]", ">``", "?", "@", "A", "AACL", "AC", "ACE_", "ACL", "AL", "ALP", "ALPH", "ALPHA_", "ALPHA_BE", "ALPHA_BET", "ALPHA_BETA_M", "ALiBi", "ALiBiPositionBias", "AN", "ANI", "ANIF", "ANIFES", "ANIFEST", "ANIFEST_", "ANIFEST_DI", "ANIFEST_DIR", "AX", "AX_", "AX_CHUNK_", "AX_CHUNK_CA", "AX_CHUNK_CACH", "AX_CHUNK_CACHE_", "AX_CHUNK_CACHE_E", "AX_CHUNK_CACHE_EN", "AX_CHUNK_CACHE_ENT", "AX_CHUNK_CACHE_ENTR", "AX_CHUNK_CACHE_ENTRI", "AX_CHUNK_CACHE_ENTRIES", "A_", "A_M", "A_MANIFEST_DIR", "Absolute", "AbsolutePositional", "AbsolutePositionalEmbedding", "Adam", "AdamW", "AdamW``", "AdamWithL2Regularization", "Arg", "Args:", "Attention", "AttentionScoreBias", "B", "B,", "BC", "BE", "BPE", "BPEIDTokenizer", "BPETokenizer", "BY", "BYT", "BYTE", "BYTE_TO_", "BYTE_TO_UNICOD", "BYTE_TO_UNICODE", "Bi", "Bias", "Block", "Byt", "Byte", "C", "CA", "CH", "CHUN", "CHUNK_", "CL", "CLR", "CO", "COD", "CharacterLevelTokenizer", "Chinchilla", "ChinchillaParametricFit", "D", "DI", "DIA_MANIFEST_DIR", "D``", "De", "Decoder", "DecoderBlock", "DecoderBlock``", "Dropout", "E", "ELU", "ES", "ESP", "ESPACE_", "E_", "E_T", "E_TO", "E_TO_", "E_TO_BYTE", "Embedding", "En", "Encoder", "EncoderBlock", "EncoderBlock``", "Encoding", "Error", "Error,", "Error:", "F", "F-8", "FF", "FLO", "FLOP", "Fal", "False", "Feed", "FeedForwardNetwork", "FeedForwardNetwork``", "File", "Fit", "For", "Forward", "ForwardN", "ForwardNetwork", "FrontierResult", "G", "GELU", "GLU", "GP", "GPT", "GPT-2", "GPTLanguage", "GPTLanguageModel", "GPTLanguageModel``", "G_ALPHA_BETA_M", "Generator", "GridPoint", "H", "Head", "Hoff", "Hoffm", "Hoffmann", "Hol", "Holt", "Holtz", "Holtzman", "I", "ICLR", "ICOD", "ID", "IDTokenizer", "IK", "IKI", "IN", "IN,", "IP", "IPS", "IRLS", "Imp", "Imple", "Implement", "Implemented", "ImplementedError", "IsoFLOP", "IsoFLOPParabolaFit", "J", "K", "KV", "K_", "Kap", "Kaplan", "Key", "KeyPositional", "KeyPositionalTransform", "L", "L2", "L2R", "L2Regularization", "LO", "LOG_ALPHA_BETA_M", "LU", "Language", "Layer", "LayerNorm", "LayerNormalization", "LayerNormalization``", "Learn", "LearnedAbsolutePositionalEmbedding", "Levenberg-Marquardt", "Lik", "Linear", "Lo", "Loss", "Loss\",", "M", "MAX_CHUNK_CACHE_ENTRIES", "MS", "MediaWiki", "Mod", "Model", "Module", "Module``", "Multi", "Multi-Head", "MultiHead", "MultiHeadAttention", "MultiHeadAttention``", "N", "N,", "None", "None,", "None``", "Norm", "Not", "Note:", "O", "ON", "OS", "Op", "Optim", "Optimizer", "OptimizerLik", "Optimizer``", "Out", "P", "PE", "PE(pos,", "PEDIA_MANIFEST_DIR", "PI", "Par", "Parabola", "ParabolaFit", "Param", "Parameter", "Parameter(torch.", "Parametric", "ParametricFit", "Path", "Piece", "Po", "Point", "Position", "PositionBias", "PositionEmbedding", "Positional", "PositionalEncoding", "Post", "Post-Layer", "Pow", "PowerL", "PowerLaw", "PowerLawFit", "Pre-", "Pre-Layer", "Press", "Pro", "Processor", "Py", "Q", "Q,", "Qu", "Query", "QueryKeyPositionalTransform", "QueryKeyPositionalTransform``", "Queryãĥ»Key", "R", "RE", "RL", "RLS", "RMS", "RMSNorm", "ReLU", "Rel", "Relative", "RelativePositionBias", "Res", "Result", "Return", "Returns:", "RoPE", "RotaryPositionEmbedding", "S", "S+", "S,", "SON", "S_", "S_q,", "Scal", "Scaled", "Scaling", "Score", "ScoreBias", "SentencePiece", "Sequence", "Shaw", "ShawRelativePositionBias", "Split", "SplitMod", "SwiGLUFeedForwardNetwork", "Swish", "T", "T5", "T5RelativePositionBias", "TF-8", "Tensor", "Tensor,", "Tensor]", "Tensor],", "Tokenizer", "Transform", "True", "U", "UN", "UNICOD", "UNICODE_TO_BYTE", "URL", "UTF-8", "Un", "Unigram", "UnigramTokenizer", "V", "V^{", "V^{(p)}", "Value", "Value\",", "ValueError:", "Vas", "Vasw", "Vaswan", "Vaswani", "W", "WIKI", "WIKIPEDIA_MANIFEST_DIR", "Wik", "Wiki", "Wikim", "Wikimedia", "Wikipedia", "X", "Xiv", "Xiv:", "Y", "Y_", "You", "Z", "[", "[\"", "[0", "[0]", "[1", "[1]", "[:", "[:,", "[None,", "[[", "[^", "[^>]", "[b", "[b]", "[ch", "[ch]", "[d_model", "[d_model]", "[end", "[float", "[float,", "[i", "[i]", "[int", "[int,", "[int]", "[int],", "[int]:", "[j", "[key", "[key]", "[start", "[str", "[str,", "[str]", "[tuple[str,", "\\", "\\S+", "\\\\", "\\\\_lr", "\\\\_lr}", "\\\\alpha", "\\\\eta", "\\\\hat{", "\\\\hat{m}_t", "\\\\hat{v", "\\\\hat{v}_t", "\\\\math", "\\\\mathrm", "\\\\mathrm{", "\\\\text{", "\\\\text{-grams", "\\\\theta_i^{", "\\n", "\\s", "\\s*", "\\{", "]", "])", "],", "]:", "]]", "]],", "^", "^(", "^(2i", "^2", "^K", "^K_mn", "^T", "^t)", "^{", "_", "_,", "_1", "_2", "_2_slopes", "_ALPHA_BETA_M", "_BYTE_TO_UNICODE", "__", "__all", "__all__", "__init__", "_a", "_ag", "_agre", "_agreement", "_agreement_rat", "_alpha", "_an", "_anal", "_analys", "_and_", "_and_activation", "_and_activation.ipyn", "_ap", "_appro", "_approx", "_approxim", "_approximation", "_arr", "_arr,", "_attempt", "_ax", "_axis_", "_axis_curv", "_b", "_b)", "_b,", "_b:", "_bar", "_bar_by_", "_bar_by_layer", "_batch", "_beta", "_beta3", "_bits_per_byt", "_bits_per_byte", "_block", "_bo", "_bod", "_body", "_body:", "_body_", "_body_output", "_bootstrap", "_boundaries", "_boundaries(", "_bu", "_bucket", "_buckets", "_buckets:", "_buffer", "_buffer(\"", "_buil", "_build_", "_by", "_by_", "_by_condition", "_by_d_model", "_by_d_model:", "_by_max", "_by_max_bytes", "_by_unit_", "_by_unit_group", "_byt", "_bytes", "_bytes)", "_bytes:", "_c", "_candidates", "_clip", "_co", "_code", "_coe", "_coef", "_coef_init", "_coeff", "_col", "_cols,", "_compute_", "_compute_b", "_compute_budget", "_compute_budgets:", "_con", "_condition", "_conditions", "_converged", "_corp", "_corpus", "_cost", "_count", "_count:", "_counts[", "_counts[pair]", "_cover", "_curv", "_curves", "_curves:", "_cy", "_cyc", "_cycle", "_cycle[", "_cycle[i", "_decay", "_den", "_denom", "_dir", "_embedding", "_exact", "_exponent", "_f", "_factory", "_fet", "_fetch", "_fetch_w", "_fetch_wikipedia", "_fetch_wikipedia_re", "_fetch_wikipedia_revis", "_fetch_wikipedia_revision", "_file", "_filter", "_fir", "_first", "_fit", "_fit_n", "_fit_nonl", "_fit_nonlinear", "_fit_nonlinear_h", "_fit_nonlinear_huber", "_fits", "_fn", "_fn(d_model)", "_fraction", "_freq", "_function", "_function_curves", "_fut", "_future_", "_future__", "_grad", "_grad()", "_grad()``", "_grid", "_grid:", "_h", "_hat", "_he", "_head", "_head_", "_head_attention", "_heads", "_heads(", "_heads)", "_heads,", "_heads:", "_heat", "_heatmap", "_hub", "_huber_", "_huber_weight", "_i", "_i)", "_i^", "_id", "_if", "_if_l", "_if_lar", "_if_large", "_in", "_inf", "_init", "_init_", "_init__", "_init__(", "_inter", "_interval", "_k", "_k_filter", "_l", "_labels", "_language_model", "_layer", "_layer:", "_le", "_lead", "_leading", "_leading_w", "_leading_whitespace", "_learning_rat", "_len", "_length_penalty", "_lik", "_like(", "_likelihood", "_likelihood_n", "_lin", "_link", "_links(", "_list", "_list,", "_list.append(", "_log", "_log_prob", "_log_probs", "_log_probs,", "_loss_", "_loss_per_token", "_lr", "_m", "_manifest", "_mask", "_match_rat", "_max", "_median", "_min", "_min,", "_mn", "_model", "_multi", "_multi_head_attention", "_n", "_nam", "_names", "_negative_", "_negative_log", "_negative_log_likelihood_n", "_new", "_new_token", "_ngrams", "_non_embedding_param", "_non_embedding_parameter", "_norm", "_num", "_numpy", "_ok", "_op", "_opt", "_opt(", "_opt(C", "_opt(C)", "_or", "_p", "_p_filter", "_pair", "_parabola", "_param", "_params", "_params)", "_params,", "_pen", "_penal", "_penalty", "_per_", "_per_layer", "_pla", "_plain", "_plaintext(", "_position", "_pow", "_power_law", "_pred", "_prefix", "_pret", "_pretraining", "_prob", "_prod", "_product_", "_product_attention", "_r", "_rat", "_ratio", "_ratio:", "_re", "_residual", "_rm", "_rms_", "_rms_ratio", "_ro", "_root", "_rul", "_rule_", "_rule_name", "_rule_name=\"identity", "_s", "_saturating", "_saturating_power_law", "_scal", "_scale:", "_scaling", "_scaling_analys", "_score", "_se", "_search", "_seed", "_seed_", "_seed_sc", "_seed_scatter", "_segment", "_seq", "_seq_term", "_seq_term:", "_sequences", "_series", "_set", "_shakespear", "_sing", "_singl", "_single_", "_single_step_", "_size", "_size)", "_size,", "_size:", "_slopes", "_spl", "_split", "_split_", "_split_chunk", "_split_chunk_by_max_bytes", "_split_heads(", "_split_heads(self.w_", "_split_mod", "_sqrt", "_squar", "_square", "_stat", "_states:", "_statistic", "_step", "_su", "_suc", "_succes", "_success", "_successful", "_symbol", "_t", "_tanh", "_tanh_approximation", "_ter", "_term", "_text", "_text,", "_tgt", "_threshold", "_time", "_to", "_to_id", "_to_rms_ratio", "_token", "_tokenizer_json", "_tot", "_total", "_tr", "_train", "_train_text", "_train_val", "_transform", "_trig", "_trigg", "_trigger", "_triggered", "_ty", "_type", "_type=\"", "_type=\"model", "_type=\"model\",", "_v", "_v[key]", "_val", "_values", "_values,", "_w", "_weight", "_width", "_wikitext", "_wikitext_links(", "_window", "_with", "_without_leading_whitespace\":", "_word(", "_x", "_ylabel:", "_{", "_{t", "_{t-1}", "`", "``", "``(", "``)", "``)ãĢĤ", "``)ãģ§", "``)ãģ®", "``)ãģ®åł´åĲĪ", "``)ãģ®åł´åĲĪ:", "``)ãģ¯", "``)ãĤĴ", "``ãĢģ", "``ãĢģ``", "``ãĢĤ", "``ãĤĴ", "``ãĥ»", "``ãĥ»``", "``ãĥ»``zero", "a", "a,", "a3", "a:", "a_bo", "a_body_", "a_body_list", "a_coef_init", "ab", "ab:", "abc", "able", "able,", "abo", "abola", "abs(", "abs_", "abs_r", "abstra", "abstract", "abstractmethod", "ac", "acc", "accar", "accard", "accept", "ace", "ach", "ack", "ack([", "act", "acterL", "acterLe", "acterLevel", "acterLevelTokenizer", "action", "action=", "action=parse", "activation", "activation_fn", "activation_fn:", "activation_fn=", "activations:", "actory", "ad", "ad_", "ad_am", "ad_amount", "ad_text(", "ad_text(encoding=\"utf-8", "add(id", "add(id(", "add_", "added_", "adding", "adding_mask", "adfor", "adford", "ag", "age", "air", "ais", "aises:", "ak", "ake_", "ake_evaluation_window", "al", "al_", "al_allo", "al_allocation", "al_allocation_exponent", "al_encoding", "al_f", "al_norm", "al_transform", "al_values", "alf", "alg", "alization", "all", "allab", "allable", "allback", "allback_", "allback_train_text", "allo", "alp", "alpha", "alpha:", "alpha=0", "alpha=0.3", "alue", "always_negative_", "always_negative_unit_", "always_negative_unit_ratio", "am", "amma", "ampl", "ampling", "an", "ance", "and", "andid", "andidat", "andidates", "andidates:", "anes", "ang", "ange(", "anguage", "anh", "anif", "anifest", "anifest_p", "anifest_path", "anifest_path)", "ann", "annotat", "annotate:", "anon", "anoni", "anonical_", "anonical_tokenizer", "anonical_tokenizers.", "ap", "apanes", "api", "api.", "app", "append(", "appl", "apply(", "apps/", "ar", "ar_width", "aracterLevelTokenizer", "arange(", "arch", "arget", "aries", "arization", "ark", "arker", "arker=\"", "arn", "arpath", "arpathy", "arquar", "arquard", "arquardt", "arquardt(", "arr", "arr,", "arra", "array", "array(", "art", "arti", "articl", "article_", "articles", "articles_dir", "ary", "as", "ase", "asel", "ask", "ast", "at", "ata", "atch", "atch_rat", "atch_size,", "atch_size:", "atches", "ate", "ate_", "ate_b", "ate_bits_per_byt", "ate_causal_mask", "ate_f", "ate_flops_per_token", "ate_h", "ate_half", "ate_p", "ate_padding_mask", "ated", "ateg", "ategory", "ath", "ath:", "ath::", "ath:`", "ating", "ation", "ativ", "ative", "ative_", "atplot", "atplotlib", "atr", "atrix", "ats", "ats)ãĢĤ", "att", "attemp", "attempt", "attention", "attention_heatmap", "attention_s", "attention_score_bias", "attention_seq_term:", "atter", "attn", "attn_", "attn_weight", "attn_weights", "attn_weights)", "attn_weights,", "attr(", "atur", "aturating", "aturatingPowerLawFit", "aturating_fit", "aturating_fits", "at{", "au", "auss", "aussian", "avi", "aw", "ax", "axes", "ayer", "ayerNorm", "b", "b``", "ba", "back", "bar", "base", "beam", "beam_search", "beam_size", "berg", "best", "best_param", "beta", "bi", "bian", "bits_per_byt", "bits_per_byte", "bj", "bject", "block", "bool", "bootstrap_scaling_analysis", "boundaries_a", "bpe_", "bs", "bsol", "bsolut", "bsolute", "bucket", "buckets", "byt", "byte_", "byte_fallback", "byte_level", "byte_level:", "byte_level_symbol", "byte_to_", "c", "ca", "cab", "cach", "cache", "cache(", "cache_", "cache_dir", "cache_dir:", "cache_path", "cached", "cached_", "cached_pro", "cached_proper", "cached_property", "cal", "cal_", "cal_jacobian", "candidates", "canonical_tokenizers.", "cat", "categ", "categories", "cation", "caus", "causal", "causal_mask", "ce", "ce_", "ce_length", "cent", "center", "center:", "centering", "centile", "cept", "ces", "ces:", "cessor", "ch", "ch().", "chan", "char", "char_to_id", "charact", "character_", "character_coverag", "character_level", "character_level_tokenizer_json", "chars", "chars)", "ched", "chedule", "chil", "chilla", "chinchilla", "chinchilla_param", "chinchilla_parametric", "chinchilla_residual", "chunk", "chunk_", "chunk_length", "chunk_length_statistic", "chunk_split_mod", "chunk_split_mode", "chunk_split_mode,", "chunk_split_mode:", "chunk_split_mode=\"", "chunk_split_mode=\"whitespace", "chunk_split_mode=\"whitespace\"``", "ci", "ck", "cket", "cl", "cla", "class", "cle", "clean", "clip", "clip(log", "clip(log_alpha", "clip(log_alpha,", "clip(log_beta", "clip_threshold", "clip_threshold:", "clud", "clude_", "cmap", "cmap:", "co", "cobian", "cod", "code", "coder", "coder-only", "coding", "coeff", "coefficient", "color", "color=\"", "color=\"tab:", "color_cycle", "colorbar", "comput", "compute_", "con", "cond", "constr", "construct_", "construct_optimal_f", "construct_optimal_frontier", "construct_optimal_frontier(", "context", "contin", "continue", "converged", "cor", "core", "core_b", "core_bias", "cores", "corp", "corpu", "corpus", "corpus,", "corpus_", "correction", "cos", "cosine", "cost", "count", "count_non_embedding_parameter", "count_pair", "couple", "coupled", "cover", "coverag", "cre", "crease", "cross_", "cross_attention", "cross_attn", "cross_attn_weights", "cstr", "cstring", "ct", "ction", "cto", "ctool", "ctx", "ctx:", "cur", "current", "cursor", "curv", "d", "d(", "d(id", "d(idx", "d(idx)", "d,", "d_", "d_bar", "d_factory", "d_ff", "d_ff,", "d_ff:", "d_i", "d_k", "d_k)", "d_model", "d_model)", "d_model,", "d_model:", "d_model``", "d_values", "da", "dam", "darray", "data", "data.", "data/", "data[\"", "dataclass", "de", "deca", "decay", "decay_", "decay_fraction", "decod", "decode_", "decode_byte_level_symbol", "decoder-only", "ded_", "ded_param", "ded_param_id", "def", "defa", "default", "del", "delta", "den", "dep", "depth", "der", "design", "deta", "detach().", "devi", "device", "device)", "device,", "device:", "device=", "dex", "dex[", "dex[pair]", "de{", "de{g", "de{g}_t", "di", "dict", "dient", "dim", "ding", "dir", "dist", "distance", "distin", "distinct", "dition", "divergence", "do", "dom", "dot", "dot_product_attention", "dow", "down", "dropout", "dropout)", "dropout,", "dropout:", "dtype", "dtype)", "dtype,", "dtype:", "dtype=", "dual", "dual_axis_curv", "e", "e\"", "e\"):", "e\",", "e(", "e(\"", "e(\"log", "e(\"log\")", "e(\"utf-8", "e()``", "e(over", "e(over=\"ignore\",", "e(self,", "e(æĹ¢å®ļåĢ¤", "e)", "e,", "e:", "e=", "eLU", "e_", "e_b", "e_band", "e_bpe_", "e_bpe_id_tokenizer_json", "e_canonical_tokenizers.", "e_corpus", "e_cross_attention", "e_decay_", "e_embedding", "e_init", "e_m", "e_mn", "e_model", "e_per_", "e_per_head", "e_s", "e_schedule", "e_std", "e_x", "e``", "ead", "ead_", "eam", "ean", "earn", "ect", "ectiv", "ective_decay_", "ector", "ed", "ed)", "ed,", "ed.", "ed:", "ed=", "edAbsolutePositionalEmbedding", "ed_", "ed_dot_product_attention", "edding", "edi", "edia", "ediaWiki", "edian", "edict", "eed", "eed\",", "eed_", "eed_for", "eed_forwar", "eed_forward", "eed_forward_factory", "eed_forward_factory:", "eep", "eepdim", "eff", "effective_decay_", "eg", "egative", "egative_", "egul", "egularization", "eight", "el", "el()", "elative_", "elf", "elfAttention", "eli", "elif", "eliho", "elihood", "else", "else:", "elt", "elta", "elu", "elu_exact", "elu_tanh_approximation", "em", "embedding", "ement", "emor", "emory", "emory_", "emp", "emperat", "emperatur", "emperature", "en", "en_", "enberg", "ence", "encod", "encode(", "encode(self,", "encode(text)", "encode_", "encode_chunk", "encode_corpus", "encoding", "encoding=\"utf-8", "encoding=\"utf-8\")", "end", "end(", "ener", "enerat", "eneration", "eneration\",", "enerator", "enn", "ennr", "ennri", "ennrich", "ensor", "ent", "ent=False)", "entence", "entencePiece", "entencepiece", "enter", "entro", "entropy", "eories", "ep", "eps", "eps)", "eps:", "equence", "er", "erL", "er_", "er_clean", "erat", "erate(", "ere(", "erf", "erg", "eri", "erical_jacobian", "eries", "eries_names", "erior", "ero", "err", "error", "errstat", "errstate(over=\"ignore\",", "ers", "ers\",", "ersist", "ersistent=False)", "ertil", "ertility", "ery", "es", "es(", "es:", "esign", "esp", "espe", "espear", "est", "est_p", "est_pair", "estim", "estimate_flops_per_token", "esty", "estyle", "estyle=\"", "estyle=\"--", "et", "et_", "et_learning_rat", "et_random", "et_random_batch", "eta", "eter", "eth", "ethod", "etr", "etric", "etur", "eturn", "etwor", "etwork", "eur", "eval", "eval()", "eval_bits_per_byte", "eval_interval", "eval_step", "evalu", "evaluate_bits_per_byt", "evaluate_bits_per_byte``", "evaluation", "evaluation_mask", "evaluation_mask:", "evaluation_window", "evaluation_windows:", "evenberg", "evenberg-Marquardt", "ew", "ex", "exact", "exact_match_rat", "except", "excl", "exclu", "excluded_param_id", "excluded_param_ids.", "excluded_param_ids.add(id(", "exist", "exist_ok", "exists()", "exists():", "exp", "exp(", "exp(log", "expon", "exponent", "ext", "ext_code", "ext_token", "ext_token_log", "extra", "ey", "ez", "eze(", "f", "f\"", "f\"{", "f-8", "factory", "fallback", "feed_forward_factory", "feed_forward_factory``", "fertility", "ff", "ffer", "fi", "fig", "figsize=(", "fil", "file", "file(", "file[", "file[\"", "filter", "fin", "finit", "finite(", "fit_", "fit_chinchilla_parametric", "fit_chinchilla_parametric``", "fit_isoflop_parabola", "fit_power_law", "fit_saturating_power_law", "float", "float3", "fontsize=", "fontsize=8)", "for", "form", "forward", "forward``", "frac", "frac{", "from", "frontier_", "frontier_log_c", "ft", "ful", "full", "full_like(", "g", "g/", "gamma", "ge", "ge(", "gelu_exact", "generator", "get", "get_random_batch", "gg", "gging", "gist", "gister", "gment", "gmentation", "gn", "gnmt", "gnmt_length_penalty", "gpt", "gra", "grad", "gradient", "gradient_clip", "gradient_clip_threshold", "gradient_clip_triggered", "gradient_norm", "gradient_norm_by_unit_group", "gradient_norm_per_layer", "gradient_norms)", "gradient_norms:", "gram", "grams", "grams)", "gray", "gray\",", "gres", "grid", "grid(alpha=0.3", "grid:", "gro", "grou", "group", "groupe", "grouped_bar", "gt", "gt_mask", "gth", "gu", "guag", "guage", "h", "h,", "ha", "hak", "hakespear", "hape", "hape(", "haracterLevelTokenizer", "hat{", "haw", "head", "hidden", "hinchilla", "history", "history[\"", "hit", "hitesp", "hitespace", "ho", "hold", "ht", "htt", "http", "https://", "hunk", "hunkSplitMod", "hunkSplitMode", "i", "i)", "i-theories", "i:", "iBi", "iGLU", "iLU", "ia", "ian", "iance", "ias", "ias)", "ias_correction", "ic", "ical", "icient", "icmethod", "icode", "ict", "id", "id=-1,", "id_", "id_to", "id_to_", "id_to_char", "id_tokenizer_json", "idal", "idden", "ident", "idirection", "idirectional", "ids", "idth", "idual", "idx", "idx]", "ie", "iece", "ieces", "ieces.append(", "ient", "ies", "ies:", "if", "iform", "ig", "ight", "ign", "igra", "igram", "igsize_per_head", "ik", "ikipedia", "ikitext", "il", "ile", "ilenam", "ilter", "im", "ime", "import", "in", "in(", "inal_values", "inchilla", "include_", "increase", "indi", "indices", "indow", "ine", "ine_", "inear", "ined_", "inewidth", "inf", "ing", "init", "init.", "init_alpha", "init_beta", "initi", "initial_values", "inn", "inner_clean", "input", "inputs,", "instance", "int", "interior", "inus", "inuso", "inusoidal", "inusoidalPositionalEncoding", "inv", "inv_freq", "io", "ion", "ip", "ipedia", "ipt", "ipts/", "ipts/p", "ipts/promot", "ipts/promote_canonical_tokenizers.", "ipy", "ipyn", "ipynb``", "ir", "irection", "is", "is_", "is_interior", "is_interior:", "ise", "isfinite(", "ish", "ism", "ismatches", "isof", "isoflop", "isoflop_parabola", "ist", "istic", "istor", "istories", "istories:", "istories_by_condition", "istory", "it", "it)", "it.", "it_", "it_alpha", "it_saturating_power_law", "it_se", "it_second", "ite_", "item", "items()", "iter", "iter:", "iterbi", "iterbi_segment", "itext", "ith", "ithL2Regularization", "ition", "itle", "its", "its-per-byt", "its-per-byte", "its-per-byte(", "its_", "its_per", "its_per_byt", "ity", "iv", "ivation", "ive", "ix", "iz", "ization", "ize", "ize(", "ize=", "ize_per_head", "ize``", "izer", "izer_", "izer_json", "j", "ja", "ja\"``ãĥ»``\"en\"``", "jacobian", "japanes", "jo", "joined_", "joined_a", "js", "json", "json.", "json.load", "json.loads(", "k", "k,", "ke", "ken", "key", "key:", "key_", "key_embedding", "key_length,", "key_length:", "l", "l_inf", "l_values", "la", "lab", "label", "label)", "label:", "label=", "label=\"", "labels", "labels(", "labels:", "lace", "lamb", "lambda", "lambda_", "lan", "languag", "language", "language:", "language_model", "language}", "lat", "lat_mask", "lation", "law", "lay", "layer", "layer_", "layer_output", "layers.", "le", "le[", "learn", "learn_b", "learn_bpe", "learning", "learning_curves", "learning_curves_multi", "learning_rat", "learning_rate", "learning_rate(", "learning_rate_schedule", "left", "leg", "len", "len(", "lenam", "length", "length)", "length_penalty", "level", "levenberg_m", "levenberg_marquardt(", "lex", "lexity", "li", "lib", "lik", "like(", "lin", "linalg", "linalg.", "linalg.lst", "linalg.lsts", "linalg.lstsq", "linalg.norm", "line", "line(", "linear", "linear1", "lines", "linewidth", "lip", "list", "ll", "lle", "llection", "lm_head", "lm_head``)ãģ®", "lo", "load", "load_", "load_wikipedia_corpus", "loat", "lock", "log", "log(", "log_c", "log_n", "log_n_arr", "log_n_opt", "log_x", "logit", "logits", "logits,", "logits``", "long", "loor", "lop", "lopen", "lopes", "lops", "lops_", "lops_per_token", "lor", "los", "lose", "losest", "losest_pow", "loss", "loss)", "loss,", "loss.", "loss:", "loss_", "loss_step_", "loss_step_delta", "loss_step_deltas:", "lot", "low", "lp", "lp(Y", "lr", "lr)", "lr:", "lst", "lt", "m", "m,", "m_h", "m_head", "ma", "make_evaluation_window", "make_evaluation_windows``", "man", "map", "mask", "mask)", "mask:", "mask]", "math", "math:`", "matrix", "max", "max(", "max_", "max_chunk", "max_chunk_bytes", "max_chunk_bytes,", "max_chunk_bytes:", "max_chunk_bytes``", "max_distance", "max_len", "max_len,", "max_new_token", "max_new_tokens)", "max_new_tokens:", "max_position", "max_position,", "max_relative_position", "max_sequence_length", "mb", "mbedding", "mbol", "mean", "mean_loss_per_token", "mean_square", "mean_to_rms_ratio", "median", "ment", "merg", "merge_", "merges", "merges:", "method", "min", "min(", "min,", "min\\\\_lr}", "min_learning_rat", "minus", "mo", "mod", "model", "model,", "model.", "model.parameters()", "model:", "model_prefix", "model_prefix.", "module", "move_", "mp", "mport", "mpt", "mpt_id", "mput", "mpute_", "mt", "mul_", "m}_t", "n", "n.", "n_", "n_key", "n_query", "n_rat", "nam", "name", "name,", "name}", "nan", "ndarray", "negative_", "new", "nn.", "nn.Parameter", "no", "no_grad()", "none", "norm", "norm_first", "norm_first:", "norm_first=True", "norm_fn", "normalization", "normalization_factory", "normalization_factory:", "normalization_rule_name=\"identity", "normed", "norms", "now", "nown_rat", "np.", "np.log(", "np.log(warm_fit", "np.ndarray", "np.ndarray],", "num", "num_buckets", "num_heads", "num_heads,", "numel()", "o", "oPE", "ocab", "ocstring", "od", "of", "of_2_slopes", "off", "oft", "oftmax", "oftmax(", "ois", "oise_band", "oise_std", "oisy_", "oisy_grid", "oisy_l", "ok", "oken", "okenizer", "ol", "old", "old_", "oldid", "om", "on", "on-", "on_embedding", "on_embedding_param", "on_embedding_params", "on_embedding_params:", "one", "ones", "ones_like(", "ong", "ongTensor", "onl", "ont", "onti", "ontsize=", "ool", "oot", "ootstra", "ootstrap", "ootstrapResult", "ootstrap_scaling_analys", "ootstrap_scaling_analysis", "op", "optim", "optim.", "optimal_allocation_exponent", "optimal_f", "optimizer", "optimizer.", "optimizer:", "optimizer``", "oq", "oqa:", "or", "orch", "orch.", "ord(", "ord(\"", "ord(\"Â", "order", "org/", "ories", "orm", "ormalization", "ormalization_factory", "orp", "ort", "ory", "ot", "ot-Product", "otImplementedError", "otary", "otaryPositionEmbedding", "otat", "otation", "otation:", "otations", "ou", "oun", "ound", "oundaries", "oundaries_a", "ount", "ounter", "ous_", "out", "out)", "out))", "out,", "out_leading_whitespace", "out_leading_whitespace\":", "output", "output,", "output_layer:", "over", "p", "p)", "p.", "p.add_", "p.grad", "p90", "p``", "pair", "pair]", "pair_counts[pair]", "pair_in", "pair_in_word(", "pair_index[pair]", "par", "parabola", "parabola.", "parabola_coeff", "param", "parameter", "parameters()", "params", "params0", "params0:", "params:", "parent", "parent.mkdir(parents=True,", "parse", "path", "path.", "path.exists():", "path.w", "path.wr", "path.write_", "path.write_text(", "pe", "pe(", "pe.", "pe[:,", "pe_", "peak", "peak\\\\_lr}", "peak_learning_rat", "peak_learning_rate", "per", "per_", "per_token", "percentile", "perplexity", "pi", "pie", "piece", "ping", "pl", "place", "plexity", "plit", "plit_train_val", "plot", "plot_", "pm", "pm.SentencePiece", "po", "point", "pointer", "pos", "pos,", "position", "position:", "position_embedding", "positional_encoding", "positional_transform", "positional_transform:", "positional_transform``", "positions", "positions)", "positions.", "positions:", "positions]", "pout", "pow", "pow(2)", "power_", "power_law", "power_law_fit", "power_of_2_slopes", "pproach", "pred", "predicted_", "pretoken", "pretokenize``", "pro", "proach", "prob", "probs", "processor", "profil", "profiles", "profiles:", "progres", "prompt_id", "prop", "pt", "put", "py", "q", "q,", "qr", "qrt", "qrt(", "qu", "quar", "que", "queeze(", "quen", "quence", "quer", "query", "query.device)", "query:", "query_length,", "query_length:", "quest", "quir", "r", "r\"", "r(", "r_squar", "ra", "rac", "raction", "rain", "raining", "raise", "ran", "random", "rang", "range(", "rank", "rans", "ransform", "ransformer", "ransformer_block", "rat", "ratio", "ratios", "raw", "raw_norm", "rc", "rc.", "re", "re-", "reak", "reconstruct_optimal_frontier", "rection", "red", "redict", "redicted_", "ref", "refix", "register", "relative_", "relative_buckets", "relative_key_embedding", "relative_position", "relative_position)", "relative_position_bucket", "relative_position_bucket(", "relative_vector", "remove_", "rent", "rep", "repo", "repo_id", "repo_id:", "repo_id=repo_id,", "repo_root", "req", "request", "request_time", "requir", "res", "reshape(", "reshold", "residual", "residual_fn", "residual_fn:", "residuals", "residuals)", "residuals**2", "residuals**2))", "residuals,", "residuals:", "ress", "result", "result_bo", "result_body", "ret", "retoken", "retokenize(", "retr", "retries", "retry_", "retry_wa", "retry_wait_second", "return", "revid", "rg", "rid", "ridPoint", "right", "rint", "ripts/promote_canonical_tokenizers.", "rm", "rn", "rng", "ro", "rob", "rod", "roduct", "rom", "romot", "ronti", "rontier", "rontierResult", "rontier_", "rontier_log_n_opt", "ropout", "ror", "ross", "rossAttention", "ross_", "rot", "rotate_half", "row", "rows", "rows.", "rror", "ru", "rue", "rue(æĹ¢å®ļåĢ¤", "ry_decode_byte_level_symbol", "r|", "s", "s\"", "s\",", "s\":", "s(", "s()", "s)", "s)),", "s,", "s.", "s.abc", "s.items()", "s.reshape(", "s.to", "s.tolist", "s.tolist()", "s/", "s:", "s://", "s=", "s[", "s[i", "s[idx]", "s]", "s])", "s_", "s_correction", "s_like(", "s_negative_", "s_q,", "s_train", "s_training", "s_with", "s_without_leading_whitespace", "s``", "sa", "saturating_fit", "sc", "scal", "scale(\"log\")", "scaling", "score", "scripts/promote_canonical_tokenizers.", "se", "seed:", "segmentation", "segmentation_a", "segmentation_agreement_rat", "sel", "self", "self)", "self,", "self.", "self.w_", "self_", "self_attn", "self_attn``", "sentencepiece", "seq", "seq_len", "seq_len,", "sequen", "sequence_length", "sequence_length:", "sequences", "sequences_a", "ser", "ser-Agent", "series_names", "set", "sh", "sion", "sition", "size=", "size=(", "size_per_head", "slopes", "so", "soFLOP", "softmax", "sor", "sort", "sorted", "sorted_", "split_train_val", "split_train_val``", "sqrt(", "squeeze(", "src", "src.", "src.training.", "src.training.optimizer``", "src/", "src/data/", "src/data/wikipedia_manifests/", "ss", "ss_", "st", "stack([", "stance", "start", "stat", "staticmethod", "statistic", "std", "step", "step()", "step()``ãĥ»``zero", "step)", "step:", "step_", "step_accept", "step_count", "str", "stra", "strip", "strip_wikitext_links(", "sum", "sum(", "symbol", "symbol_to_id", "symbols", "symbols)", "symbols,", "t", "t(", "t)", "tab:", "target", "target_compute_budgets:", "temperature", "temperature=0.0", "tent", "tention", "ter", "ter,", "tex", "text", "text(", "text)", "text,", "text:", "text{", "th", "theories", "theories/", "theories/0", "theories/01_foundations/", "theories/01_foundations/004_norm", "theories/01_foundations/004_normalization", "theories/ãĥ»", "theories/ãĥ»apps/", "ther_", "theta", "theta_i", "theta_i^{", "thon", "threshold", "ti", "tick", "tie_embedding", "til", "tilde{g}_t", "tim", "time", "time.sleep", "tin", "tiny", "tiny_shakespear", "title", "title:", "tj", "to", "to(", "to_", "to_numpy", "token", "token_embedding", "token_embedding.weight", "token_id", "token_ids", "token_ids)", "token_ids,", "token_ids.", "token_ids:", "token_ids``", "tokenizer", "tokenizer.", "tokenizer.char_to_id", "tokenizer.json", "tokenizer:", "tokenizer_json", "tokenizer_json_p", "tokens:", "tokens_train", "top", "top-", "top-k", "topk", "torch.", "torch.nn.Parameter", "torch.no_grad()", "torch.optim.", "torch.optim.Optimizer``", "tot", "total", "total_", "total_bytes)", "total_bytes:", "total_eval", "total_negative_log_likelihood_n", "total_negative_log_likelihood_nat", "total_negative_log_likelihood_nats", "total_step", "total_steps", "total_steps``", "tr", "train", "train(", "train_", "train_language_model", "train_language_model``", "train_loss", "train_unigram_model", "trainer", "training", "training.", "training/", "transformer_block", "try:", "try_decode_byte_level_symbol", "ttention", "tuple", "tuple[str,", "ty", "type", "type=", "u", "ual", "ual_seed", "ual_seed(seed)", "ub", "uber", "uber_", "uber_delta", "uber_delta:", "ucket", "ucle", "ucleus", "uct", "uct_", "ud", "udget", "ue", "ugging", "ul", "ul_", "ulary", "ule", "ult", "ult(", "ulti", "ulti-start", "um", "um_log_prob", "umerate(", "ump", "umpy", "un", "unction", "unctional", "unctool", "uniform", "unigram", "unigram_model", "unit_", "unk", "unk_token", "unk_token:", "unknown_rat", "unsqueeze(", "unt", "up", "uper", "uper().__init__()", "uple", "upload_", "ur", "ure", "ure_", "url", "urllib", "urlopen", "urrent", "urv", "us", "us)", "use_cross_attention", "use_cross_attention:", "use_cross_attention=False", "use_cross_attention=False``", "ut", "utf-8", "util", "utils.", "v", "v_hat", "v_term", "val", "valid", "valid=\"ignore\"):", "validation", "valu", "value", "value:", "value_", "values", "values,", "values_by", "values_by_condition", "values_by_condition:", "values_by_series", "var", "ve", "ve_", "ve_bpe_id_tokenizer_json", "ve_character_level_tokenizer_json", "vector", "vel", "venberg", "venberg_m", "ver", "verg", "verged", "vergence", "verse", "vertex", "vertex_", "vertex_log_n", "vertex_log_n,", "vertex_loss", "vi", "vid", "view", "vious_", "vious_loss", "vis", "viterbi_segment", "vmax", "vmax:", "vocab", "vocab)", "vocab_size", "vocab``", "vocabulary", "vocabulary_size:", "w", "w_", "wa", "war", "ward", "ward(", "warm", "warm_fit", "warmup", "warmup_step", "warmup_steps", "was_training", "way", "ways_negative_", "weight", "weight_decay", "weight_decay:", "weight_grad", "weights", "weights:", "wh", "where(", "while", "whitespace", "wiGLU", "wiGLUFeedForwardNetwork", "width", "wik", "wikipedia", "wikipedia_corpus", "wikipedia_manifest", "wikipedia_manifests/", "wikitext", "window", "wish", "with", "wor", "word", "x", "x:", "x_c", "x_c,", "x``", "xlabel", "xlabel(xlabel)", "xlabel:", "xtick", "y", "y(", "y,", "y:", "y_", "y_arr", "y_decode_byte_level_symbol", "y_labels", "y_length", "y_length,", "y_length:", "yer", "ylabel", "ylabel(", "ylabel)", "ylabel:", "ymbol", "ys", "yscale(\"log\")", "yt", "z", "zero", "zeros(", "zip", "zorder", "zorder=", "{", "{\"", "{|", "|", "|}", "|}{|", "}", "}\"", "}\")", "}\",", "})", "},", "}_t", "}``", "}{", "}}``", "~", "¡", "¡ãĤ¤ãĥ«", "¡ãĥ³ãĥĪ", "¡ä»¶", "¡ä¾ĭ", "¡çĶ¨", "¡çĶ¨ãģĻãĤĭ", "¢", "¢«", "¢«è¦", "¢«è¦Ĩ", "¢º", "¢ãĤ¹", "¢æķ°", "¢ç´", "¢ç´¢", "¢ĥçķĮ", "£", "£Ĵ", "¤", "¤º", "¤ĩ", "¤ľ", "¤ľè¨¼", "¥", "¥µ", "¥æľ¬", "¥æľ¬èªŀ", "¦", "¦(x)", "¦åı·", "¦æ³", "¦æ³¢", "¦ç", "¦çĲĨ", "§", "§ãĥ³", "§ãĥ¼ãĤ¹", "§ç¸®", "§ĭ", "§ĭæĪĲ", "§ĭç¯ī", "§Ķ", "§ĺ", "¨", "¨ãģ¦", "¨æ³", "¨æ³¢æķ°", "¨ĵ", "©", "©ãĥ¼ãĥ«", "©ãĥ¼ãĥ«ãĥĲ", "©ãĥ¼ãĥ«ãĥĲãĥĥãĤ¯", "©ç·ļ", "©Ł", "ª", "ªçĽ®", "ªè»¸", "ªĵ", "«", "«å°", "«å°¾", "«ĸ", "«ĺ", "¬", "¬¦åı·", "¬¦åı·åĮĸ", "¬¬", "¬æĺİ", "¬ç«ĭ", "®", "®µ", "®µéļİ", "®µéļİãģ§", "®å°ı", "®å°ıå®ļæķ°", "®åĪĨ", "®åĭķ", "®åĭķå°ı", "®åĭķå°ıæķ°", "®åĭķå°ıæķ°çĤ¹", "®æ¨Ļ", "®æŃ£", "®Ĭ", "®ĭ", "®ĭå·®", "¯", "¯¾", "¯å±¤", "¯ç©į", "¯Ģ", "¯Ķ", "¯Ķè¼ĥ", "°", "°´", "°ãĤĬ", "±", "±¡", "±¤", "±º", "±ãģĦ", "±æķ°", "±èªŀ", "±éŁ¿", "±éŁ¿ãĤĴåıĹãģĳ", "±Ĥ", "±ĤãĤģ", "²", "²åĲĪ", "²æ³ķ", "²ç·ļ", "³", "³»", "³»åĪĹ", "³ãĥ¼ãĥ", "³å®ļ", "³å¯", "³å¯Ĩ", "³åĿĩ", "´", "´»", "´»æĢ§", "´»æĢ§åĮĸ", "´ãĥª", "µ", "µ·", "¶", "¶æķ°", "·", "··", "·ãģķ", "·åº¦", "·ļ", "·ļå½¢", "·Ŀ", "·ĿéĽ¢", "¸", "¸¦", "¹", "¹°ãĤĬ", "¹³åĿĩ", "º", "º«", "ºå®ļ", "»", "»åº¦", "»åĭķ", "»æĦı", "¼", "¼ãģ³", "¼ãģ³åĩº", "¼ãģ³åĩºãģĹ", "¼ĥ", "¼Ķ", "¼Ķç®Ĺ", "¼ķæķ°", "½", "½®", "½åĴĮ", "½åĴĮãģ¹ãģįä¹Ĺåīĩ", "¾", "¾¤", "¾ãģį", "¾åĢ¤", "¾çī", "¾çī©ç·ļ", "¾éħį", "¿", "¿Ĵ", "À", "Á", "Â", "Ã", "ÃĹ", "Ä", "Å", "Æ", "Ç", "È", "É", "Ê", "Ë", "Ì", "Í", "Î", "Î²", "Î²1,", "Î¸", "Ï", "Ïĥ", "Ïĥ(", "Ïĥ^2", "Ð", "Ñ", "Ò", "Ó", "Ô", "Õ", "Ö", "×", "Ø", "Ù", "Ú", "Û", "Ü", "Ý", "Þ", "ß", "à", "á", "â", "âĸ", "âĸģ", "ã", "ãĢ", "ãĢģ", "ãĢģ006", "ãĢģ007", "ãĢģ009", "ãĢģFalse", "ãĢģ``", "ãĢģn", "ãĢģãģĬãĤĪãģ³", "ãĢģãģĻ", "ãĢģãģĿãģ®", "ãĢģãģĿãĤĮ", "ãĢģä¾ĭ:", "ãĢģå½¢çĬ¶", "ãĢģå¾Į", "ãĢģåĲĮãģĺ", "ãĢģæľ¬", "ãĢģæľ¬çķªå®Łè¡Įåīįãģ®ä¿®æŃ£", "ãĢģéłĤçĤ¹", "ãĢĤ", "ãĢĤ\"\"\"", "ãĢĤ``", "ãĢĤ``None``", "ãĢĤåŃĺåľ¨ãģĹãģªãģĦ", "ãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯", "ãĢĤæĹ¢å®ļåĢ¤", "ãĢĮ", "ãĢĮæľ¬ä½ĵ", "ãĢį", "ãĢįãģ¨ãģĦãģĨ", "ãĢįãģ®", "ãĢįãĤĴ", "ãĢľ", "ãĢľ004", "ãģ", "ãģ¡", "ãģ¡ãĢģ", "ãģ¡ãĤī", "ãģ£", "ãģ£ãģ¦", "ãģ£ãģ¦ãģĦ", "ãģ£ãģ¦ãģĦãĤĭ", "ãģ£ãģ¦ãģĹãģ¾", "ãģ£ãģ¦ãĤĤ", "ãģ£ãģŁ", "ãģ£ãģŁãģ¾ãģ¾", "ãģ¤", "ãģ¤ãģ¤", "ãģ¤ãģĦãģ¦", "ãģ¤ãģĦãģ¦ãĢģ", "ãģ¤ãģį", "ãģ¥", "ãģ¦", "ãģ¦ãģ¯ãĤģ", "ãģ¦ãģĦ", "ãģ¦ãģĦãĤĭ", "ãģ¦ãĤĤ", "ãģ§", "ãģ§ãĢģ", "ãģ§ãģªãģĦ", "ãģ§ãģ®", "ãģ§ãģ®éĢ£", "ãģ§ãģ®éĢ£çķª", "ãģ§ãģ¯", "ãģ§ãģ¯ãĢģ", "ãģ§ãģ¯ãģªãģı", "ãģ§ãģ¯ãģªãģıãĢģ", "ãģ§ãģĤ", "ãģ§ãģĤãĤĬãĢģ", "ãģ§ãģĤãĤĭ", "ãģ§ãģĤãĤĭãĢĤ", "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭ", "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭ:", "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭãĢĤ", "ãģ§ãģĤãĤĮãģ°", "ãģ§ãģį", "ãģ§ãģįãĢģ", "ãģ§ãģįãģªãģĦ", "ãģ§ãģįãĤĭ", "ãģ§ãģįãĤĭãĤĪãģĨãģ«", "ãģ§ãĤĤ", "ãģ§ãĤĪ", "ãģ§åº", "ãģ§åºĥ", "ãģ§å¾ĹãģŁ", "ãģ§æ¯Ķè¼ĥãģĻãĤĭ", "ãģ§æııçĶ»ãģĻãĤĭãĢĤ", "ãģ§è¡ĮãģĨ", "ãģ¨", "ãģ¨ãĢģ", "ãģ¨ãģªãĤĭ", "ãģ¨ãģ«", "ãģ¨ãģ®", "ãģ¨ãģ®ä½µ", "ãģ¨ãģ¿", "ãģ¨ãģ¿ãģª", "ãģ¨ãģĦãģĨ", "ãģ¨ãģį", "ãģ¨ãģįãģ®", "ãģ¨ãģįãģ¯", "ãģ¨ãģĹ", "ãģ¨ãģĹãĢģ", "ãģ¨ãģĹãģ¦", "ãģ¨ãģĹãģ¦æī±", "ãģ¨ãĤģ", "ãģ¨ãĤģãģ¦", "ãģ¨åĩºåĬĽå±¤", "ãģ¨åĲĮãģĺ", "ãģ¨åĲĮä¸Ģ", "ãģ©", "ãģ©ãģ¡ãĤī", "ãģ©ãģ®", "ãģ©ãģĨ", "ãģ©ãģĨãģĹ", "ãģ©ãĤĮ", "ãģ©ãĤĮãģłãģĳ", "ãģª", "ãģªãģ©", "ãģªãģ©)", "ãģªãģ©)ãĢĤ", "ãģªãģ©)ãĤĴ", "ãģªãģ©ãĤĴ", "ãģªãģ®", "ãģªãģĦ", "ãģªãģĦ(", "ãģªãģĦ)ãĢĤ", "ãģªãģĦãĢĤ", "ãģªãģĦãģŁãĤģãĢģ", "ãģªãģı", "ãģªãģıãģ¦ãĤĤ", "ãģªãģĳãĤĮãģ°", "ãģªãģĹ", "ãģªãģĹãģ¨", "ãģªãģĹãģ®", "ãģªãĤĬ", "ãģªãĤĬãĢģ", "ãģªãĤĭ", "ãģªãĤĭãĢĤ", "ãģªãĤı", "ãģªãĤıãģ¡", "ãģ«", "ãģ«ãĢģ", "ãģ«ãģ¤ãģĦãģ¦", "ãģ«ãģª", "ãģ«ãģªãĤĬãĢģ", "ãģ«ãģªãĤĭ", "ãģ«ãģªãĤĭ(", "ãģ«ãģªãĤĭãĢĤ", "ãģ«ãģ¯", "ãģ«ãģ¯ãĢģ", "ãģ«ãģĹãģŁ", "ãģ«ãģĹãģŁæĿ¡ä»¶", "ãģ«ãģĻãĤĭ", "ãģ«ãģĻãĤĭãĢĤ", "ãģ«ãģĻãĤĭãģĭãĢĤ", "ãģ«ãģĿãģ®ãģ¾ãģ¾", "ãģ«ãĤĤ", "ãģ«ãĤĪ", "ãģ«ãĤĪãģ£ãģ¦", "ãģ«ãĤĪãĤīãģļ", "ãģ«ãĤĪãĤĬ", "ãģ«ãĤĪãĤĬãĢģ", "ãģ«ãĤĪãĤĭ", "ãģ«ãĤı", "ãģ«ãĤıãģŁ", "ãģ«ä½¿", "ãģ«ä½¿ãģ£ãģŁ", "ãģ«ä½¿ãģĨ", "ãģ«ä½¿ãģĨ(005", "ãģ«ä½¿ãģĨ)ãĢĤ", "ãģ«ä½¿ãģĨãĢĤ", "ãģ«ä½įç½®", "ãģ«ä¾ĿåŃĺ", "ãģ«ä¾ĿåŃĺãģĻãĤĭ", "ãģ«å¤±æķĹ", "ãģ«å¤īæıĽ", "ãģ«å¤īæıĽãģĻãĤĭãĢĤ\"\"\"", "ãģ«å¯¾", "ãģ«å¯¾ãģĹãģ¦", "ãģ«å¯¾ãģĻãĤĭ", "ãģ«å¯¾å¿ľ", "ãģ«å¯¾å¿ľãģĻãĤĭ", "ãģ«å¯¾å¿ľãģĻãĤĭãĢĤ", "ãģ«å¾ĵ", "ãģ«å¿ľ", "ãģ«å¿ľãģĺ", "ãģ«å¿ľãģĺãģ¦", "ãģ«åĩºçı¾", "ãģ«åĩºçı¾ãģĻãĤĭ", "ãģ«åĪĨåī²", "ãģ«åĬłç®Ĺ", "ãģ«åĬłç®ĹãģĻãĤĭ", "ãģ«åŁºãģ¥ãģı", "ãģ«åŁºãģ¥ãģıãĢĤ", "ãģ«æ", "ãģ«æ³¨æĦı", "ãģ«æ¸¡ãģĻ", "ãģ«æ¸¡ãģĻãĥŀãĤ¹ãĤ¯", "ãģ«éģ©çĶ¨ãģĻãĤĭ", "ãģ«éģĶãģĻãĤĭ", "ãģ®", "ãģ®ãģ¾ãģ¾", "ãģ®ãģ¿", "ãģ®ãģ¿ãģ«ä¾ĿåŃĺãģĻãĤĭ", "ãģ®ãģĨ", "ãģ®ãģŁãĤģ", "ãģ®ãģŁãĤģãĢģ", "ãģ®ãģŁãĤģãģ®å¾", "ãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°", "ãģ®ãĤ·ãĥ³ãĥľãĥ«", "ãģ®ãĤ¿ãĤ¤ãĥĪ", "ãģ®ãĤ¿ãĤ¤ãĥĪãĥ«ãĢĤ", "ãģ®ãĥ¦ãĥĭãĥĥãĥĪ", "ãģ®ãĥªãĤ¹ãĥĪ", "ãģ®ãĥªãĤ¹ãĥĪ}``", "ãģ®ãĥªãĤ¹ãĥĪãĢĤ", "ãģ®ãĥªãĥ³ãĤ¯", "ãģ®ä¸ĬéĻĲ", "ãģ®äºĮä¹Ĺ", "ãģ®ä½įç½®", "ãģ®å", "ãģ®å¢ĥçķĮ", "ãģ®å¤", "ãģ®å¤§", "ãģ®å¤§ãģį", "ãģ®å¤§ãģįãģķ", "ãģ®å¤ī", "ãģ®å®ļç¾©", "ãģ®å®Ł", "ãģ®å®ŁåĬ¹", "ãģ®å®Łæķ°åĢ¤", "ãģ®å®Łæķ°åĢ¤ãģ®ç³»åĪĹ", "ãģ®å®Łè£ħ", "ãģ®å¯¾", "ãģ®å¯¾å¿ľ", "ãģ®å¯¾æķ°", "ãģ®å¯¾æķ°å°¤åº¦", "ãģ®å°", "ãģ®å°Ħå½±", "ãģ®å°İ", "ãģ®å°İåĩº", "ãģ®å·", "ãģ®å·®", "ãģ®å¸", "ãģ®å¹³åĿĩ", "ãģ®åº", "ãģ®å¼", "ãģ®å¼ķæķ°", "ãģ®å½", "ãģ®å½±éŁ¿ãĤĴåıĹãģĳ", "ãģ®å¾", "ãģ®å¾Į", "ãģ®åĢ¤", "ãģ®åĢ¤ãĤĴ", "ãģ®åħ¥åĬĽ", "ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ", "ãģ®åħ¨", "ãģ®åĨħ", "ãģ®åĨħç©įãĤĴ", "ãģ®åĩ", "ãģ®åĩ¦çĲĨ", "ãģ®åĩº", "ãģ®åĩºåĬĽ", "ãģ®åĩºåĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ", "ãģ®åĪĿæľŁåĢ¤", "ãģ®åī", "ãģ®åī¯å±¤", "ãģ®åī²åĲĪ", "ãģ®åīį", "ãģ®åĬ", "ãģ®åĬ¹", "ãģ®åĭ¾éħį", "ãģ®åĭ¾éħįãĤĴ", "ãģ®åĭ¾éħįãĥİãĥ«ãĥł", "ãģ®åįĺ", "ãģ®åıįå¾©", "ãģ®åıįå¾©çļĦãģª", "ãģ®åıįå¾©çļĦãģªç¸®å°ı", "ãģ®åıĸå¾Ĺ", "ãģ®åĲĦ", "ãģ®åĳ¼ãģ³åĩºãģĹ", "ãģ®åĴĮ", "ãģ®åĴĮãģĮ", "ãģ®åĽ", "ãģ®åŁ", "ãģ®åŁº", "ãģ®åł´åĲĪ", "ãģ®åł´åĲĪãĢģ", "ãģ®åł´åĲĪãģ¯", "ãģ®åŃ¦ç¿Ĵ", "ãģ®æ", "ãģ®æ¤ľè¨¼", "ãģ®æ§ĭç¯ī", "ãģ®æ¨", "ãģ®æ¨Ļæºĸ", "ãģ®æ¨Ļæºĸåģıå·®", "ãģ®æ¬¡åħĥãĢĤ", "ãģ®æ®ĭå·®", "ãģ®æ¯Ķ", "ãģ®æ¯Ķè¼ĥ", "ãģ®æ¸Ľè¡°çİĩ", "ãģ®æī", "ãģ®æīĭ", "ãģ®æĮĩæķ°", "ãģ®æĮĩæķ°ç§»åĭķ", "ãģ®æĮĩæķ°ç§»åĭķå¹³åĿĩ", "ãģ®æĲįå¤±", "ãģ®æķ°", "ãģ®æĹ", "ãģ®æĹ¢", "ãģ®æĹ¢å®ļ", "ãģ®æľĢå¤§åĢ¤", "ãģ®æĿ", "ãģ®æĿ¡ä»¶", "ãģ®ç³»åĪĹ", "ãģ®ç³»åĪĹéķ·", "ãģ®ç³»åĪĹéķ·ä¾ĿåŃĺ", "ãģ®ç³»åĪĹéķ·ä¾ĿåŃĺéłħ", "ãģ®çµĦãģ¿åĲĪãĤıãģĽ", "ãģ®ç·ıåĴĮ", "ãģ®ç·ļå½¢", "ãģ®çĲĨè«ĸ", "ãģ®è§£æŀĲ", "ãģ®è§£æŀĲãĥĳãĤ¤ãĥĹãĥ©ãĤ¤ãĥ³", "ãģ®è¨ĺäºĭ", "ãģ®è¾ŀæĽ¸", "ãģ®éģķãģĦ", "ãģ®éķ·ãģķ", "ãģ®éļłãĤĮ", "ãģ®éĽĨåĲĪ", "ãģ®éłĨåºı", "ãģ¯", "ãģ¯ãĢģ", "ãģ¯ãĤģ", "ãģ¯çĦ¡è¦ĸ", "ãģ°", "ãģ°ãĤī", "ãģ°ãĤīãģ¤ãģį", "ãģ³", "ãģ¶", "ãģ¸", "ãģ¸ãģ®", "ãģ¹", "ãģ¹ãģ¦", "ãģ¹ãģį", "ãģ¹ãģįä¹Ĺ", "ãģ¹ãģįä¹Ĺåīĩ", "ãģ¹ãģįä¹ĹåīĩãģĤãģ¦ãģ¯ãĤģ", "ãģ¹ãģŁ", "ãģ»", "ãģ»ãģ©", "ãģ»ãģ¼", "ãģ¼", "ãģ¾", "ãģ¾ãģ¨ãĤģãģ¦", "ãģ¾ãģ¾", "ãģ¾ãģŁ", "ãģ¾ãģŁãģ¯", "ãģ¾ãĤĭ", "ãģ¾ãĤĭãģĶãģ¨", "ãģ¿", "ãģ¿ãĤĴ", "ãģĤ", "ãģĤãģ¦ãģ¯ãĤģ", "ãģĤãģ¦ãģ¯ãĤģ(", "ãģĤãģ¦ãģ¯ãĤģãĢģéłĤçĤ¹", "ãģĤãģŁ", "ãģĤãģŁãĤĬ", "ãģĤãģŁãĤĬãģ®", "ãģĤãģŁãĤĬãģ®è¨Īç®Ĺéĩı", "ãģĤãĤĭ", "ãģĦ", "ãģĦ)ãĢĤ", "ãģĦãĢģ", "ãģĦãĢĤ", "ãģĦãģ¦", "ãģĦãģĨ", "ãģĦãģļãĤĮ", "ãģĦãģļãĤĮãĤĤ", "ãģĦãģŁ", "ãģĦãĥ©ãĥĥãĥĳãĥ¼ãĢĤ", "ãģĦåł´åĲĪãģ¯", "ãģĨ", "ãģĨãĢĤ", "ãģĨãĤĭ", "ãģĪ", "ãģĪãĢģ", "ãģĪãģ°", "ãģĪãģŁ", "ãģĪãĤĭ", "ãģĪãĤĭãĤĪãģĨãģ«", "ãģĬ", "ãģĬãĤĪ", "ãģĬãĤĪãģ³", "ãģĭ", "ãģĭ(", "ãģĭãĢĤ", "ãģĭãģ£ãģŁ", "ãģĭãģ¤", "ãģĭãģ©ãģĨ", "ãģĭãģ©ãģĨãģĭ", "ãģĭãĤī", "ãģĭãĤīãģ®", "ãģĭãĤīéĻ¤å¤ĸ", "ãģĭãĤīéłĨãģ«", "ãģĭãĤĴ", "ãģĮ", "ãģĮãĢģ", "ãģĮãĢĮ", "ãģĮãģĤ", "ãģĮãģĤãĤĭãĢĤ", "ãģĮãĤŃãĥ£ãĥĥãĤ·ãĥ¥", "ãģĮä¸Ģ", "ãģĮä¸Ģèĩ´", "ãģĮä¸Ģèĩ´ãģĹãģ¾", "ãģĮä¸Ģèĩ´ãģĹãģ¾ãģĽ", "ãģĮä¸Ģèĩ´ãģĹãģ¾ãģĽãĤĵ", "ãģĮå¢Ĺ", "ãģĮå¤§ãģįãģĦ", "ãģĮå°ıãģķãģĦ", "ãģĮå°ıãģķãģĦãģ»ãģ©", "ãģĮçķ°ãģªãĤĭ", "ãģĮé«ĺãģĦ", "ãģĮéļĽéĻĲãģªãģı", "ãģį", "ãģįæıĽ", "ãģįæıĽãģĪ", "ãģı", "ãģı)ãĢĤ", "ãģıãĢģ", "ãģıä½¿ãĤı", "ãģıä½¿ãĤıãĤĮãĤĭ", "ãģĲ", "ãģĳ", "ãģĳãĤĭ", "ãģĳãĤĮãģ°", "ãģĴ", "ãģĵ", "ãģĵãģ¨", "ãģĵãģ¨ãģ§", "ãģĵãģ¨ãģ¯", "ãģĵãģ¨ãģĮ", "ãģĵãģ¨ãģĮæľŁ", "ãģĵãģ¨ãģĮæľŁå¾ħ", "ãģĵãģ¨ãģĮæľŁå¾ħãģķãĤĮãĤĭãĢĤ", "ãģĵãģ¨ãĤĴ", "ãģĵãģ¨ãĤĴæĦıåĳ³", "ãģĵãģ®", "ãģĵãģ®å¼ķæķ°", "ãģĵãģ®åĢ¤", "ãģĵãģ®åĢ¤ãĤĴ", "ãģĵãģ®æ", "ãģĵãģ®æ¯Ķ", "ãģĵãģ®éĸ¢æķ°", "ãģĵãģĵ", "ãģĵãģĵãģ«", "ãģĵãĤĮ", "ãģĵãĤĮãģĮ", "ãģĵãĤĮãĤĴ", "ãģĵãĤĮãĤĴè¶ħãģĪãĤĭ", "ãģĶ", "ãģĶãģ¨", "ãģĶãģ¨ãģ«", "ãģĶãģ¨ãģ®", "ãģķ", "ãģķãģªãģĦ", "ãģķãģĽ", "ãģķãģĽãĤĭ", "ãģķãģĽãĤĭ(", "ãģķãģĽãĤĭãĢĤ", "ãģķãĤĮ", "ãģķãĤĮãģŁ", "ãģķãĤĮãģŁãĥĪãĥ¼ãĤ¯ãĥ³", "ãģķãĤĮãĤĭ", "ãģķãĤĮãĤĭãĢĤ", "ãģĹ", "ãģĹãĢģ", "ãģĹãĢģ``", "ãģĹãĢģæ®ĭãĤĬ", "ãģĹãģ¦", "ãģĹãģ¦ãģĦ", "ãģĹãģ¦ãģĦãĤĭ", "ãģĹãģ¦ãģĦãĤĭãģŁãĤģãĢģ", "ãģĹãģ¦ãģĬ", "ãģĹãģ¦ãģĬãģı", "ãģĹãģ¦ãģĭãĤī", "ãģĹãģ¦ãĤĤ", "ãģĹãģ¦ãĤŃãĥ£ãĥĥãĤ·ãĥ¥", "ãģĹãģª", "ãģĹãģªãģĦ", "ãģĹãģªãģĦ)ãĢĤ", "ãģĹãģªãģĦãĢĤ", "ãģĹãģ¾", "ãģĹãģĨãĤĭ", "ãģĹãģŁ", "ãģĹãģŁ)ãĢĤ", "ãģĹãģŁãģĦ", "ãģĹãģŁåıįå¾©", "ãģĹãģŁè¨ĺäºĭ", "ãģĹãģŁè¨ĺäºĭæľ¬æĸĩ", "ãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ", "ãģĹåģ´", "ãģĺ", "ãģĻ", "ãģĻ)ãĢĤ", "ãģĻãĢĤ", "ãģĻãģ¹ãģ¦", "ãģĻãģĵãģ¨", "ãģĻãĤĭ", "ãģĻãĤĭ(", "ãģĻãĤĭ(00", "ãģĻãĤĭ(006", "ãģĻãĤĭ(007", "ãģĻãĤĭ(009", "ãģĻãĤĭ(``", "ãģĻãĤĭ)", "ãģĻãĤĭ)ãĢĤ", "ãģĻãĤĭãĢĤ", "ãģĻãĤĭãĢĤ\"\"\"", "ãģĻãĤĭãģ¨", "ãģĻãĤĭãģ®", "ãģĻãĤĭãģĵãģ¨", "ãģĻãĤĭãģĵãģ¨ãģ§", "ãģĻãĤĭãģĵãģ¨ãģ§ãĢģ", "ãģĻãĤĭãģŁãĤģ", "ãģĻãĤĭãģŁãĤģãĢģ", "ãģĻãĤĭãģŁãĤģãģ«ä½¿ãģĨãĢĤ", "ãģĻãĤĭãģŁãĤģãģ®", "ãģĻãĤĭåŃ¦ç¿Ĵçİĩ", "ãģĻãĤĭéļĽ", "ãģĻãĤĮãģ°", "ãģļ", "ãģļãĢģ", "ãģļãģ«", "ãģļãĤĮ", "ãģĽ", "ãģĽãģļ", "ãģĽãĤĭ", "ãģľ", "ãģĿ", "ãģĿãģ®", "ãģĿãģ®ãģ¾ãģ¾", "ãģĿãģ®ãģ¾ãģ¾æ¸¡ãģĻ", "ãģĿãģ®ãĤĤãģ®", "ãģĿãĤĮ", "ãģŁ", "ãģŁãĤģ", "ãģŁãĤģ)ãĢĤ", "ãģŁãĤģãĢģ", "ãģŁãĤģãĢģ``", "ãģł", "ãģłãģĮãĢģ", "ãģłãģĳ", "ãģłãģĳãģ§", "ãģłãģĳãģĮ", "ãģŃ", "ãĤ", "ãĤ¢", "ãĤ¢ãĤ¦", "ãĤ¢ãĥ³", "ãĤ¢ãĥĥãĥĹ", "ãĤ¢ãĥĥãĥĹãĥŃ", "ãĤ¢ãĥĥãĥĹãĥŃãĥ¼ãĥī", "ãĤ£", "ãĤ£ãĥª", "ãĤ£ãĥªãĥĨãĤ£", "ãĤ£ãĥ«", "ãĤ£ãĥ«ãĤ¿", "ãĤ£ãĥ³ãĤ°", "ãĤ¤", "ãĤ¤ãĤ", "ãĤ¤ãĤ¢ãĤ¹", "ãĤ¤ãĤ¶", "ãĤ¤ãĤ¹ãĢĤ", "ãĤ¤ãĤº", "ãĤ¤ãĥ¤", "ãĤ¤ãĥ¤ãĥ¼", "ãĤ¤ãĥ«", "ãĤ¤ãĥ³", "ãĤ¤ãĥ³ãĤ¹ãĤ¿ãĥ³ãĤ¹", "ãĤ¤ãĥ³ãĤ¿", "ãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ", "ãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹", "ãĤ¤ãĥ³ãĥĩ", "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯", "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹", "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:", "ãĤ¤ãĥĪ", "ãĤ¦", "ãĤ¦ãĥ³ãĥ", "ãĤ¦ãĥ³ãĥī", "ãĤ¦ãĥ³ãĥīä¿®æŃ£", "ãĤ¨", "ãĤ¨ãĤ¤", "ãĤ¨ãĤ¹ãĥĪ", "ãĤ¨ãĥ©", "ãĤ¨ãĥ©ãĥ¼", "ãĤ¨ãĥ³ãĤ", "ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥ", "ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°", "ãĤª", "ãĤªãĥ¼ãĥĲ", "ãĤªãĥ¼ãĥĲãĥ¼ãĥķãĥŃãĥ¼", "ãĤ«", "ãĤ«ãĥ©", "ãĤ«ãĥ©ãĥ¼", "ãĤ«ãĥĨ", "ãĤ«ãĥĨãĤ´ãĥª", "ãĤ¯", "ãĤ¯ãĤ¨ãĤ¹ãĥĪ", "ãĤ¯ãĤ·ãĥ§ãĥ³", "ãĤ¯ãĥ©", "ãĤ¯ãĥ©ãĤ¹", "ãĤ¯ãĥª", "ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°", "ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°éģ©çĶ¨åīį", "ãĤ¯ãĥªãĥĥãĥĹ", "ãĤ¯ãĥĪ", "ãĤ¯ãĥĪãĥª", "ãĤ¯ãĥĪãĥ«", "ãĤ°", "ãĤ°ãĥ©ãĥ", "ãĤ°ãĥ©ãĥķ", "ãĤ°ãĥª", "ãĤ°ãĥªãĥĥãĥī", "ãĤ±", "ãĤ±ãĥĥãĥĪ", "ãĤ³", "ãĤ³ãĤ¹ãĥĪ", "ãĤ³ãĤ¹ãĥĪãģĮ", "ãĤ³ãĥ¼ãĥ", "ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°", "ãĤ³ãĥ¼ãĥī", "ãĤ³ãĥ¼ãĥĳ", "ãĤ³ãĥ¼ãĥĳãĤ¹", "ãĤ³ãĥ¼ãĥĳãĤ¹ãĤĴåıĸå¾Ĺ", "ãĤ³ãĥĵ", "ãĤ³ãĥĵãĤ¢ãĥ³", "ãĤ´ãĥª", "ãĤµ", "ãĤµãĤ¤ãĤº", "ãĤµãĥ³ãĥĹ", "ãĤµãĥ³ãĥĹãĥªãĥ³ãĤ°", "ãĤµãĥ¼ãĥ", "ãĤµãĥ¼ãĥģ", "ãĤµãĥĸ", "ãĤµãĥĸãĤ¯ãĥ©ãĤ¹", "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼", "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽ", "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ", "ãĤµãĥĸãĥ¯", "ãĤµãĥĸãĥ¯ãĥ¼ãĥī", "ãĤµãĥĸãĥ¯ãĥ¼ãĥīåĪĨåī²", "ãĤ·", "ãĤ·ãĥ¥", "ãĤ·ãĥ§ãĥ³", "ãĤ·ãĥ³ãĥ", "ãĤ·ãĥ³ãĥľãĥ«", "ãĤ·ãĥ³ãĥľãĥ«ãģ®", "ãĤ·ãĥ¼ãĥī", "ãĤ·ãĥ¼ãĥīãģĶãģ¨ãģ®", "ãĤ·ãĥ¼ãĥīéĸĵ", "ãĤ¸", "ãĤ¸ãĥ¥", "ãĤ¸ãĥ¥ãĥ¼ãĥ«", "ãĤ¸ãĥ§ãĥ³", "ãĤ¹", "ãĤ¹ãĢĤ", "ãĤ¹ãĤ¯", "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ", "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãĢĤ", "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãģĻãĤĭãĢĤ", "ãĤ¹ãĤ±", "ãĤ¹ãĤ±ãĤ¸ãĥ¥ãĥ¼ãĥ«", "ãĤ¹ãĤ±ãĥ¼", "ãĤ¹ãĤ±ãĥ¼ãĥªãĥ³ãĤ°", "ãĤ¹ãĤ±ãĥ¼ãĥªãĥ³ãĤ°åīĩ", "ãĤ¹ãĤ±ãĥ¼ãĥ«", "ãĤ¹ãĤ³", "ãĤ¹ãĤ³ãĤ¢", "ãĤ¹ãĤ¿", "ãĤ¹ãĤ¿ãĤ¤ãĥ«", "ãĤ¹ãĤ¿ãĥ³ãĤ¹", "ãĤ¹ãĤŃ", "ãĤ¹ãĥĨ", "ãĤ¹ãĥĨãĥĥãĥĹ", "ãĤ¹ãĥĨãĥĥãĥĹãģ§", "ãĤ¹ãĥĨãĥĥãĥĹãģ¨ãģ®", "ãĤ¹ãĥĨãĥĥãĥĹæĲįå¤±", "ãĤ¹ãĥĨãĥĥãĥĹæķ°", "ãĤ¹ãĥĨãĥĥãĥĹçķªåı·", "ãĤ¹ãĥĪ", "ãĤ¹ãĥĪãĥ©", "ãĤ¹ãĥĪãĥ©ãĥĥãĥĹ", "ãĤº", "ãĤ»", "ãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³", "ãĤ»ãĥ«", "ãĤ»ãĥĥãĥĪ", "ãĤ½", "ãĤ¿", "ãĤ¿ãĤ¤", "ãĤ¿ãĤ¤ãĥĪ", "ãĤ¿ãĤ¤ãĥĪãĥ«", "ãĤ¿ãĤ¤ãĥł", "ãĤ¿ãĤ¤ãĥłãĤ¢ãĤ¦", "ãĤ¿ãĤ¤ãĥłãĤ¢ãĤ¦ãĥĪ", "ãĤ¿ãĥĸãĥ»", "ãĤ¿ãĥĸãĥ»æĶ¹è¡Į", "ãĤ¿ãĥĹ", "ãĤĢ", "ãĤģ", "ãĤģãĤĭ", "ãĤĤ", "ãĤĤãģ®", "ãĤĦ", "ãĤĦãģĻ", "ãĤĦãģĻãģĦ", "ãĤĪ", "ãĤĪãģĨ", "ãĤĪãģĨãģª", "ãĤĪãģĨãģ«", "ãĤĪãĤĬ", "ãĤĪãĤĬä¸Ĭä½į", "ãĤĪãĤĬä¸Ĭä½įãģ®", "ãĤĪãĤĭ", "ãĤī", "ãĤīãģªãģĦ)ãĢĤ", "ãĤīãģļ", "ãĤīãĤĮ", "ãĤīãĤĮãģŁ", "ãĤīãĤĮãĤĭ", "ãĤīãĤĮãĤĭãĢĤ", "ãĤĬ", "ãĤĬãĢģ", "ãĤĬå½ĵ", "ãĤĬå½ĵãģ¦", "ãĤĭ", "ãĤĭ(", "ãĤĭ(Hoffmann", "ãĤĭ(``", "ãĤĭãĢĤ", "ãĤĭãĢĤãģĵãģ®", "ãĤĭãģĵãģ¨", "ãĤĭãģĵãģ¨ãģ§ãĢģ", "ãĤĮ", "ãĤĮãģªãģĦ", "ãĤĮãģªãģĦãĢĤ", "ãĤĮãģ°", "ãĤĮãĤĭ", "ãĤĮãĤĭ)ãĢĤ", "ãĤĮåĢ¤", "ãĤı", "ãĤıãģªãģĦ", "ãĤıãģĽ", "ãĤıãĤĮ", "ãĤĴ", "ãĤĴãĢģ", "ãĤĴãģ¾ãģŁ", "ãĤĴãģ¾ãģŁãģĦ", "ãĤĴãģ¾ãģŁãģĲ", "ãĤĴãģĿãģ®ãģ¾ãģ¾", "ãĤĴä¸ĢåĪĩ", "ãĤĴä¸İ", "ãĤĴä¸İãģĪãĤĭ", "ãĤĴä½¿", "ãĤĴä½¿ãģĨ", "ãĤĴä½ľ", "ãĤĴä¿Ŀ", "ãĤĴä¿Ŀãģ£ãģŁãģ¾ãģ¾", "ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«", "ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ", "ãĤĴåħ±æľī", "ãĤĴåĬłãģĪãĤĭ", "ãĤĴåıĤçħ§)ãĢĤ", "ãĤĴåıĸå¾Ĺ", "ãĤĴåıĹãģĳ", "ãĤĴåĲ«", "ãĤĴåĲ«ãĤĢ", "ãĤĴåĲ«ãĤģ", "ãĤĴåĽºå®ļ", "ãĤĴæ", "ãĤĴæ¤ľè¨¼", "ãĤĴæ®ĭ", "ãĤĴæ¯Ķè¼ĥ", "ãĤĴæ±ĤãĤģ", "ãĤĴæ¸¡", "ãĤĴæ¸¡ãģĻãĢĤ", "ãĤĴæĥ³å®ļ", "ãĤĴæĥ³å®ļ)ãĢĤ", "ãĤĴæĦı", "ãĤĴæĦıåĳ³", "ãĤĴæĬ", "ãĤĴæĬĳ", "ãĤĴæĮģãģ¤", "ãĤĴæĮģãģŁ", "ãĤĴæĮģãģŁãģªãģĦ", "ãĤĴæİ", "ãĤĴæĽ", "ãĤĴçĶŁæĪĲ", "ãĤĴè¡¨", "ãĤĴè¡¨ãģĻ", "ãĤĴè¡Į", "ãĤĴè¡ĮãģĨ", "ãĤĴè¡ĮãģĨãĢĤ", "ãĤĴè¡ĮãģĨçī¹å¾´", "ãĤĴè¡ĮãĤıãģªãģĦ", "ãĤĴè¦ĭ", "ãĤĴè¨Īç®Ĺ", "ãĤĴè¨Īç®ĹãģĻãĤĭ(007", "ãĤĴè¨Īç®ĹãģĻãĤĭãĢĤ", "ãĤĴè©¦", "ãĤĴè¿ĶãģĻ", "ãĤĴè¿ĶãģĻãĢĤ", "ãĤĴéĢ£çµĲãģĹãģŁ", "ãĤĴéĺ²ãģĲ", "ãĤĴéĻ¤", "ãĤĵ", "ãĤĵãģł", "ãĤŃ", "ãĤŃãĥ£", "ãĤŃãĥ£ãĤ¹ãĥĪ", "ãĤŃãĥ£ãĥĥãĤ·ãĥ¥", "ãĤŃãĥ¼", "ãĥ", "ãĥ¡", "ãĥ¡ãĤ¤ãĥ³", "ãĥ¡ãĥ¢", "ãĥ¢", "ãĥ¢ãĤ¸ãĥ¥ãĥ¼ãĥ«", "ãĥ¢ãĤ¸ãĥ¥ãĥ¼ãĥ«ãģ¯", "ãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ", "ãĥ¢ãĥĩãĥ«", "ãĥ¢ãĥĩãĥ«ãģ®", "ãĥ£", "ãĥ£ãĥ³ãĤ¯", "ãĥ¤", "ãĥ¤ãĤ³ãĥĵãĤ¢ãĥ³", "ãĥ¥", "ãĥ¦", "ãĥ¦ãĥ¼ãĥ", "ãĥ¦ãĥ¼ãĥĨ", "ãĥ¦ãĥ¼ãĥĨãĤ£ãĥªãĥĨãĤ£", "ãĥ¦ãĥĭ", "ãĥ¦ãĥĭãĥ¼ãĤ¯", "ãĥ¦ãĥĭãĥĥãĥĪ", "ãĥ§ãĥ³", "ãĥ©", "ãĥ©ãĤ¤ãĥ³", "ãĥ©ãĥ", "ãĥ©ãĥ¡", "ãĥ©ãĥ³ãĥĢ", "ãĥ©ãĥ³ãĥĢãĥł", "ãĥ©ãĥĥãĥģ", "ãĥ©ãĥĥãĥģå®Łè£ħ", "ãĥ©ãĥĥãĥĳ", "ãĥ©ãĥĥãĥĳãĥ¼ãĢĤ", "ãĥ©ãĥĻ", "ãĥ©ãĥĻãĥ«ãĢĤ", "ãĥª", "ãĥªãĤ¢", "ãĥªãĤ¯ãĤ¨ãĤ¹ãĥĪ", "ãĥªãĤ¹ãĥĪ", "ãĥªãĥ", "ãĥªãĥ³ãĤ¯", "ãĥªãĥ³ãĤ°", "ãĥªãĥĵ", "ãĥªãĥĵãĤ¸ãĥ§ãĥ³", "ãĥªãĥĿ", "ãĥªãĥĿãĤ¸", "ãĥªãĥĿãĤ¸ãĥĪãĥª", "ãĥ«", "ãĥ«ãĢĤ", "ãĥ«ãĥ»", "ãĥ«ãĥ¼ãĥ", "ãĥ«ãĥ¼ãĥĹ", "ãĥ«ãĥģ", "ãĥ«ãĥĨãĤ£", "ãĥ«ãĥł", "ãĥ¬", "ãĥ¬ãĤ¤ãĥ¤ãĥ¼", "ãĥ¬ãĤ¯ãĥĪãĥª", "ãĥ¬ãĥ¼ãĥĪ", "ãĥ¬ãĥĻ", "ãĥ¬ãĥĻãĥ«", "ãĥ¯", "ãĥ¯ãĥ¼ãĤ¯", "ãĥ³", "ãĥ³ãĤ", "ãĥ³ãĤ¯", "ãĥ³ãĤ°", "ãĥ³ãĤ¹", "ãĥ³ãĤ½", "ãĥ³ãĥ", "ãĥ³ãĥĢ", "ãĥ³ãĥĨãĤ£", "ãĥ³ãĥĪ", "ãĥ³ãĥĹ", "ãĥ³ãĥĹãĥ¬ãĥ¼ãĥĪ", "ãĥ»", "ãĥ»00", "ãĥ»002", "ãĥ»2", "ãĥ»ALiBi", "ãĥ»D", "ãĥ»Key", "ãĥ»``", "ãĥ»å®Łé¨ĵ", "ãĥ»åĲĦ", "ãĥ»åĲĦãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ", "ãĥ»æ¤ľè¨¼", "ãĥ¼", "ãĥ¼ãĢĤ", "ãĥ¼ãĤ¯", "ãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ", "ãĥ¼ãĤ¸", "ãĥ¼ãĤ¹", "ãĥ¼ãĤ¹ãģ®åıįå¾©çļĦãģªç¸®å°ı", "ãĥ¼ãĤ¿", "ãĥ¼ãĥ", "ãĥ¼ãĥ¡ãĥ³ãĥĪ", "ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ", "ãĥ¼ãĥ«", "ãĥ¼ãĥĪ", "ãĥ¼ãĥĪãĤ¹ãĥĪãĥ©ãĥĥãĥĹ", "ãĥ¼ãĥĪãĥĸ", "ãĥ¼ãĥĪãĥĸãĥĥãĤ¯", "ãĥ¼ãĥī", "ãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪ", "ãĥ¼ãĥĲ", "ãĥ¼ãĥĲãĥ¼", "ãĥ¼ãĥķ", "ãĥ¼ãĥķãĤ", "ãĥ¼ãĥķãĥŃ", "ãĥ¼ãĥķãĥŃãĥ¼", "ãĥ¼ãĥŀ", "ãĥ¼ãĥł", "ãĥ¼ãĥłãĤµãĥ¼ãĥģ", "ãĥĢ", "ãĥģ", "ãĥģãĥ£ãĥ³ãĤ¯", "ãĥģãĥ£ãĥ³ãĤ¯ãģ®", "ãĥģãĥ£ãĥ³ãĤ¯éķ·", "ãĥĥ", "ãĥĥãĤ¯", "ãĥĥãĤ·ãĥ¥", "ãĥĥãĥ", "ãĥĥãĥģ", "ãĥĥãĥĪ", "ãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯", "ãĥĥãĥī", "ãĥĥãĥĳ", "ãĥĥãĥĶ", "ãĥĥãĥĶãĥ³ãĤ°", "ãĥĥãĥĹ", "ãĥĨ", "ãĥĨãĤ£", "ãĥĨãĤŃ", "ãĥĨãĤŃãĤ¹ãĥĪ", "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵ", "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵãģ®", "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵãĤĴ", "ãĥĨãĥ³ãĤ½", "ãĥĨãĥ³ãĤ½ãĥ«", "ãĥĨãĥ³ãĤ½ãĥ«ãĢĤ", "ãĥĨãĥ³ãĥĹãĥ¬ãĥ¼ãĥĪ", "ãĥĩ", "ãĥĩãĤ£", "ãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥª", "ãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯", "ãĥĩãĤ£ãĥ³ãĤ°", "ãĥĩãĤ³ãĥ¼ãĥī", "ãĥĩãĥ«", "ãĥĩãĥ¼ãĤ¿", "ãĥĩãĥĲ", "ãĥĩãĥĲãĤ¤ãĤ¹ãĢĤ", "ãĥĪ", "ãĥĪãĥª", "ãĥĪãĥªãĥĥãĤ¯", "ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ", "ãĥĪãĥ¼ãĤ¯", "ãĥĪãĥ¼ãĤ¯ãĥ³", "ãĥĪãĥ¼ãĤ¯ãĥ³ãģ®", "ãĥĪãĥ¼ãĤ¯ãĥ³ãĤĴ", "ãĥĪãĥ¼ãĤ¯ãĥ³å¢ĥçķĮ", "ãĥĪãĥ¼ãĤ¯ãĥ³åĮĸ", "ãĥĪãĥ¼ãĤ¯ãĥ³åŁĭãĤģè¾¼ãģ¿", "ãĥĪãĥ¼ãĤ¯ãĥ³æķ°", "ãĥĪãĥ¼ãĤ¯ãĥ³æķ°ãĢĤ", "ãĥĪãĥ¼ãĤ¯ãĥĬ", "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶", "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãĢĤ", "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶æĿ¡ä»¶", "ãĥĪãĥĶ", "ãĥĪãĥĶãĥĥãĤ¯", "ãĥī", "ãĥīãĥ¡ãĤ¤ãĥ³", "ãĥĬ", "ãĥĬãĥ«ãĥĨãĤ£", "ãĥĭ", "ãĥį", "ãĥįãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯", "ãĥİ", "ãĥİãĤ¤ãĤº", "ãĥİãĤ¤ãĤºåº", "ãĥİãĤ¤ãĤºåºĬ", "ãĥİãĥ«ãĥł", "ãĥİãĥ¼ãĥĪãĥĸãĥĥãĤ¯", "ãĥı", "ãĥĲ", "ãĥĲãĤ¤ãĤ¢ãĤ¹", "ãĥĲãĤ¤ãĤ¢ãĤ¹ãĤĴåĬłãģĪãĤĭ", "ãĥĲãĤ¤ãĤ¢ãĤ¹éłħ", "ãĥĲãĤ¤ãĥĪ", "ãĥĲãĤ¤ãĥĪãĥ¬ãĥĻãĥ«", "ãĥĲãĤ¤ãĥĪåĪĹ", "ãĥĲãĤ±ãĥĥãĥĪ", "ãĥĲãĤ±ãĥĥãĥĪåĮĸ", "ãĥĲãĥĥãĥģ", "ãĥĲãĥĥãĥģãĤµãĤ¤ãĤº", "ãĥĳ", "ãĥĳãĤ¤", "ãĥĳãĤ¤ãĥĹ", "ãĥĳãĤ¤ãĥĹãĥ©ãĤ¤ãĥ³", "ãĥĳãĤ¹", "ãĥĳãĥ©ãĥ¡", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ãĢģå½¢çĬ¶", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ãĤĴæĮģãģŁãģªãģĦ", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æĽ´æĸ°", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æĽ´æĸ°å¼ı", "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ç¾¤", "ãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ", "ãĥĳãĥĩãĤ£ãĥ³ãĤ°", "ãĥĳãĥĩãĤ£ãĥ³ãĤ°ãĥŀãĤ¹ãĤ¯", "ãĥĳãĥĩãĤ£ãĥ³ãĤ°ä½įç½®", "ãĥĵ", "ãĥĵãĥ¼ãĥłãĤµãĥ¼ãĥģ", "ãĥĶ", "ãĥķ", "ãĥķãĤ", "ãĥķãĤ¡ãĤ¤ãĥ«", "ãĥķãĤ¡ãĤ¤ãĥ«åĲį", "ãĥķãĤ£ãĥ«ãĤ¿", "ãĥķãĤ©ãĥ¼ãĥ«ãĥĲãĥĥãĤ¯", "ãĥķãĥŃãĥ³ãĥĨãĤ£ãĤ¢", "ãĥĸ", "ãĥĸãĥ»", "ãĥĸãĥ¼ãĥĪãĤ¹ãĥĪãĥ©ãĥĥãĥĹ", "ãĥĸãĥŃ", "ãĥĸãĥŃãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪ", "ãĥĸãĥŃãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪåı¯èĥ½", "ãĥĸãĥŃãĥĥãĤ¯", "ãĥĹ", "ãĥĺãĥĥãĥī", "ãĥĺãĥĥãĥīæķ°", "ãĥĻ", "ãĥĻãĤ¯ãĥĪãĥ«", "ãĥļ", "ãĥļãĤ¢", "ãĥļãĥĬãĥ«ãĥĨãĤ£", "ãĥŀ", "ãĥŀãĤ¹ãĤ¯", "ãĥŀãĥ«ãĥģ", "ãĥŀãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ", "ãĥŀãĥ¼ãĤ¸", "ãĥŀãĥ¼ãĤ¸è¦ıåīĩ", "ãĥŀãĥĥãĥĹ", "ãĥŁ", "ãĥł", "ãĥŃ", "ãĥŃãĥ³ãĥĨãĤ£", "ãĥŃãĥ³ãĥĨãĤ£ãĤ¢", "ãĥŃãĥķãĤ¡ãĤ¤ãĥ«", "ä", "ä¸", "ä¸¡", "ä¸¡æĸ¹", "ä¸¦", "ä¸¦ãģ¹ãģŁ", "ä¸¸", "ä¸¸ãĤģ", "ä¸»", "ä¸»å¼µ", "ä¸Ģ", "ä¸ĢåĪĩ", "ä¸Ģèĩ", "ä¸Ģèĩ´", "ä¸Ģèĩ´ãģĻãĤĭ", "ä¸ĢèĪ", "ä¸ĢèĪ¬", "ä¸Ĭ", "ä¸Ĭãģ®", "ä¸ĬãģĴ", "ä¸Ĭä½į", "ä¸Ĭæĺ", "ä¸Ĭæĺĩ", "ä¸Ĭæĺĩå¹ħ", "ä¸ĬéĻĲ", "ä¸ĭ", "ä¸ĭéĻĲ", "ä¸į", "ä¸įä¸Ģèĩ´", "ä¸įå¤ī", "ä¸įæŃ£", "ä¸įæŃ£ãģª", "ä¸įè¦ģ", "ä¸įèĥ½", "ä¸įèĥ½ãģª", "ä¸İ", "ä¸İãģĪ", "ä¸Ń", "ä¸Ńãģ®", "ä¸Ńå¿", "ä¸Ńå¿ĥ", "ä¸Ńå¿ĥå·®åĪĨ", "ä¸Ńå¿ĥå·®åĪĨãģ«ãĤĪãĤĭ", "ä¸Ńéĸĵ", "ä¸Ńéĸĵå±¤", "ä¸Ńéĸĵæ¬¡åħĥ", "ä¹", "ä¹±æķ°", "ä¹±æķ°ãĤ·ãĥ¼ãĥī", "ä¹ĸ", "ä¹ĸéĽ¢", "ä¹Ĺ", "äº", "äº¤", "äº¤å·®", "äº¤å·®æ³¨æĦı", "äºĨ", "äºĪ", "äºĪæ¸¬", "äºĪç®Ĺ", "äºĭ", "äºĭåīį", "äºĭåīįåĪĨåī²", "äºĭåīįåĪĨåī²(pre-tokenization", "äºĮ", "äºĮä¹Ĺ", "äºĮä¹Ĺå¹³åĿĩ", "äºĴ", "äºĴæıĽ", "äºĴæıĽæĢ§", "ä»", "ä»£", "ä»¥", "ä»¥ä¸ĭ", "ä»¥å¤ĸ", "ä»¥å¤ĸãģ®", "ä»¥éĻį", "ä»¶", "ä»»æĦı", "ä»»æĦıå½¢çĬ¶", "ä»»æĦıå½¢çĬ¶ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ", "ä»ĺ", "ä»ĺãģį", "ä»ĺãģĳãĤĭ", "ä»ĺä¸İ", "ä¼", "ä¼¼", "ä¼Ŀ", "ä¼ĿãĤı", "ä¼Ŀæ", "ä¼ĿæĴ", "ä¼ĿæĴŃ", "ä½", "ä½µ", "ä½¿", "ä½¿ãģĨ", "ä½¿ãģĨ(", "ä½¿ãģĨ)ãĢĤ", "ä½¿ãģĨãĢĤ", "ä½¿ãĤı", "ä½¿ãĤıãģªãģĦ", "ä½į", "ä½įç½®", "ä½įç½®ãģ§", "ä½įç½®ãģ®ãģ¿ãģ«ä¾ĿåŃĺãģĻãĤĭ", "ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:", "ä½įç½®ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°", "ä½įç½®å¤īæıĽ", "ä½İ", "ä½ĵ", "ä½ĵãģ¯", "ä½ķ", "ä½ķãĤĤ", "ä½ľ", "ä½ľæĪĲ", "ä½ľæĪĲãģĻãĤĭãĢĤ", "ä¾", "ä¾¡", "ä¾ĭ", "ä¾ĭ:", "ä¾ĭãģĪãģ°", "ä¾ĭå¤ĸ", "ä¾ĿåŃĺ", "ä¿", "ä¿®æŃ£", "ä¿Ĥ", "ä¿Ĥæķ°", "ä¿Ŀ", "ä¿ĿåŃĺ", "ä¿ĿæĮģ", "å", "å¢", "å¢ĥçķĮ", "å¢Ĺ", "å¤", "å¤§", "å¤§ãģį", "å¤§ãģįãģª", "å¤§ãģįãģĦ", "å¤§ãģįãģı", "å¤§ãģįãģıãģªãĤĬ", "å¤±", "å¤±æķĹ", "å¤ī", "å¤īãĤı", "å¤īæıĽ", "å¤īæıĽãģĻãĤĭ", "å¤īæĽ´", "å¤ĸ", "å¤ĸãĤĮåĢ¤", "å¤ĸéĥ¨", "å¤ĸéĥ¨ãģĭãĤī", "å¤ļ", "å¤ļãģĦ", "å§Ķ", "å§ĶãģŃ", "å®", "å®¹", "å®ī", "å®īå®ļ", "å®īå®ļåĮĸ", "å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°", "å®Į", "å®Įåħ¨", "å®Įåħ¨ãģª", "å®Įåħ¨ãģ«", "å®Įåħ¨ä¸Ģèĩ´", "å®ļ", "å®ļåĢ¤", "å®ļåĢ¤ãģ®ç³»åĪĹ", "å®ļæķ°", "å®ļç¾", "å®ļç¾©", "å®Ł", "å®ŁãĥĪãĥ¼ãĤ¯ãĥ³", "å®Łæ¸¬", "å®Łè¡Į", "å®Łè¡Įåīįãģ®", "å®Łè£", "å®Łè£ħ", "å®Łè£ħãģ§ãģ¯", "å®Łè£ħãģĹãģªãģĦ", "å®Łè£ħãģĻãĤĭãĢĤ", "å®Łè£ħæĸ¹éĩĿ", "å®Łé", "å®Łé¨ĵ", "å®ŁéļĽ", "å®ŁéļĽãģ«", "å¯", "å¯¾", "å¯¾ãģĹãģ¦", "å¯¾ä½įç½®", "å¯¾å¿ľ", "å¯¾å¿ľãģĻãĤĭ", "å¯¾å¿ľãģĻãĤĭæĲįå¤±", "å¯¾æķ°", "å¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«", "å¯¾æķ°ç¢ºçİĩ", "å¯¾æķ°ç©ºéĸĵ", "å¯¾æķ°ç©ºéĸĵãģ§", "å¯¾è", "å¯¾è±¡", "å¯¾è±¡ãģ®", "å¯¾è·ĿéĽ¢", "å°", "å°¤", "å°¤åº¦", "å°Ħå½±", "å°İ", "å°İåħ¥", "å°İåĩº", "å°İéĸ¢æķ°", "å°ı", "å°ıãģķ", "å°ıãģķãģĦ", "å°ĳ", "å±", "å±¤", "å±¤ãģ®", "å±¤ãģĶãģ¨", "å±¤æķ°", "å±¤æŃ£è¦ıåĮĸ", "å±¤æŃ£è¦ıåĮĸ(Layer", "å±ĢæīĢ", "å±ķ", "å±ķéĸ", "å±ķéĸĭ", "å·", "å·¦", "å·®", "å·®ãģĹ", "å·®åĪĨ", "å·±", "å¸", "å¸°", "å¸¸", "å¸¸ãģ«", "å¸¸ãģ«è²ł", "å¸¸ãģ«è²łãģ®ãĥ¦ãĥĭãĥĥãĥĪ", "å¸ĥ", "å¹", "å¹³", "å¹³åĿĩ", "å¹³åĿĩæ¸Ľç®Ĺ", "å¹³åĿĩæ¸Ľç®Ĺ(re-centering", "å¹³æĸ¹", "å¹³æĸĩ", "å¹³æĸĩãģ«å¤īæıĽ", "å¹³è¡Į", "å¹³è¡Įç§»åĭķ", "å¹ħ", "åº", "åº¦", "åº¦ãĤĤ", "åºı", "å¼", "å¼¦æ³¢", "å¼µ", "å¼·åº¦", "å¼·åº¦ãģĮ", "å¼ı", "å¼ķæķ°", "å½", "å½¢", "å½¢å¼ı", "å½¢çĬ¶", "å½¢çĬ¶ãģ¯", "å½±", "å½±éŁ¿", "å½ĵ", "å½ĵãģĻãĤĭ", "å½Ļ", "å¾", "å¾©", "å¾©åħĥ", "å¾©åħĥãģ§ãģįãĤĭ", "å¾®åĪĨ", "å¾´", "å¾ħ", "å¾Į", "å¾Įãģ«", "å¾Įãģ®", "å¾ĮåįĬ", "å¾ĵ", "å¾ĵæĿ¥", "å¾ĵæĿ¥éĢļãĤĬ", "å¾Ĺ", "å¾ĹãģŁ", "å¿", "å¿ħ", "å¿ħãģļ", "å¿ħè¦ģ", "å¿ħè¦ģãģĮãģĤ", "å¿ħè¦ģãģĮãģĤãĤĭ", "å¿ħè¦ģãģĮãģĤãĤĭ(", "å¿ħè¦ģãģĮãģĤãĤĭãĢĤ", "å¿ľ", "åĢ¤", "åĢ¤ãģ¨ãģĹãģ¦", "åĢ¤ãģ¯", "åĢ¤ãĤĴ", "åĢĻè£ľ", "åĢĻè£ľèªŀå½Ļ", "åĢĻè£ľèªŀå½ĻãģĭãĤīãģ®", "åģ", "åģ´", "åģ´ãģ§", "åģ¶æķ°", "åģı", "åģıå·®", "åĤ", "åĤ¾ãģį", "åĥ", "åĥı", "åĦ", "åħ", "åħ¥", "åħ¥ãĤĮ", "åħ¥åĩºåĬĽ", "åħ¥åĬĽ", "åħ¥åĬĽãĢĤ", "åħ¥åĬĽåŁĭãĤģè¾¼ãģ¿", "åħ¨", "åħ¨ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "åħ¨ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ãģ®åĭ¾éħįãĤĴ", "åħ¨ãĥĺãĥĥãĥī", "åħ¨ä½ĵ", "åħ¨ä½ĵãģ®", "åħ¨ä½ĵãĤĴ", "åħ¬", "åħ¬å¹³", "åħ¬å¹³ãģ«", "åħ±", "åħ±æľī", "åħ±éĢļ", "åħ±éĢļãģ®", "åħ¸", "åħĥ", "åħĥãģ®", "åħĪ", "åħĪãģ®", "åħĪãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯", "åħĪéłŃ", "åħĪéłŃãģ«", "åħĪéłŃãģ«ä»ĺ", "åĨ", "åĨħ", "åĨħãģ®", "åĨħå®¹", "åĨħéĥ¨", "åĨį", "åĨįåĪ©çĶ¨", "åĨįåĪ©çĶ¨ãģĻãĤĭ", "åĨįåŃ¦ç¿Ĵ", "åĨįæ§ĭæĪĲ", "åĨįçı¾", "åĨįè©¦è¡Į", "åĩ", "åĩ¡ä¾ĭ", "åĩ¦çĲĨ", "åĩº", "åĩºåĬĽ", "åĩºåĬĽå°Ħå½±", "åĩºåĬĽå±¤", "åĩºçı¾", "åĩºçı¾ãģĹãģŁ", "åĪ", "åĪ¤", "åĪ¤å®ļ", "åĪ¥", "åĪ©", "åĪ©çĶ¨", "åĪ¶", "åĪ¶ç´Ħ", "åĪ¶éĻĲ", "åĪĨ", "åĪĨãĤĴ", "åĪĨå¸ĥ", "åĪĨåī²", "åĪĨåī²çµĲæŀľ", "åĪĨæ¯", "åĪĨæ¯į", "åĪĨæķ£", "åĪĩ", "åĪĩãĤĬ", "åĪĩãĤĬåĩº", "åĪĩãĤĬæį", "åĪĩãĤĬæį¨ãģ¦", "åĪĹ", "åĪĹ)ãĢĤ", "åĪĹãĤĴ", "åĪĿ", "åĪĿæľŁ", "åĪĿæľŁåĢ¤", "åĪĿæľŁåĮĸ", "åĪĿæľŁèªŀå½Ļ", "åī", "åī²", "åī²ãĤĬå½ĵãģ¦", "åī²ãĤĬå½ĵãģ¦ãĤĭ", "åī²åĲĪ", "åīĩ", "åīĩåĮĸ", "åīĬ", "åīį", "åīįãģ«", "åīįãģ®", "åīįåįĬ", "åīįæıĲ", "åīįæıĲæĿ¡ä»¶", "åĬ", "åĬ¹", "åĬ¹æŀľ", "åĬ¹çİĩ", "åĬĽ", "åĬŁ", "åĬł", "åĬłãģĪãĤĭ", "åĬłç®Ĺ", "åĭ", "åĭ¾éħį", "åĭ¾éħįãģĮ", "åĭ¾éħįãĥİãĥ«ãĥł", "åĭķ", "åĮ", "åĮº", "åĮºåĪĩ", "åĮºéĸĵ", "åĮĸ", "åĮĸãģĹãģŁ", "åį", "åįĬ", "åįĺ", "åįĺä¸Ģ", "åįĺä¸ĢãĤ¹ãĥĨãĥĥãĥĹæĲįå¤±", "åįĺä¸ĢãĥĪãĥ¼ãĤ¯ãĥ³", "åįĺä½į", "åįĺä½įãģ®", "åįĺèªŀ", "åİ", "åİ³å¯Ĩ", "åİ»", "åİŁ", "åİŁè«ĸ", "åİŁè«ĸæĸĩ", "åİŁè«ĸæĸĩãģ®", "åİŁè«ĸæĸĩãģ¯", "åı", "åı¯", "åı¯èĥ½", "åı¯èĥ½ãģª", "åı¯éĢĨ", "åı¯éĢĨæĢ§", "åı³", "åı·", "åıĤç", "åıĤçħ§", "åıĤçħ§)ãĢĤ", "åıį", "åıįå¾©", "åıİ", "åıİæĿŁ", "åıĸ", "åıĸãĤĬ", "åıĸãĤĬåĩº", "åıĸãĤĭ", "åıĸå¾Ĺ", "åıĸå¾ĹãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ", "åıĹ", "åıĹãģĳ", "åĲ", "åĲ«", "åĲ«ãģ¾", "åĲ«ãģ¾ãĤĮãģªãģĦ", "åĲ«ãģ¾ãĤĮãĤĭ", "åĲ«ãĤģãĤĭ", "åĲĦ", "åĲĦä½įç½®", "åĲĦè¦ģç´łãģ¯", "åĲĪ", "åĲĪãģ¯", "åĲĪãĤıãģĽ", "åĲĮ", "åĲĮãģĺ", "åĲĮä¸Ģ", "åĲĮä¸Ģãģ®", "åĲĮä¸Ģãģ®æĮĻåĭķ", "åĲį", "åĲį:", "åĲįåīį", "åĲĳ", "åĳ", "åĳ¨æ³¢æķ°", "åĳ³", "åĳ¼", "åĳ¼ãģ³", "åĳ¼ãģ³åĩº", "åĳ¼ãģ³åĩºãģĹ", "åĳ¼ãģ³åĩºãģĹåģ´", "åĳ¼ãģ³åĩºãģĹåģ´ãģĮ", "åĴĮ", "åķ", "åķı", "åķıé¡", "åķıé¡Į", "åĸĦ", "åĽ", "åĽ²", "åĽ³", "åĽºå®ļ", "åĽºå®ļåŃ¦ç¿Ĵçİĩ", "åĽŀ", "åĽŀå¸°", "åĽŀæķ°", "åĽŀæķ°ãĢĤ", "åĽł", "åĽłæŀľ", "åĽłæŀľãĥŀãĤ¹ãĤ¯", "åľ", "åľ§ç¸®", "åľ§ç¸®åĬ¹çİĩ", "åľ¨", "åľ¨ãģ®", "åĿ", "åĿĩ", "åŀ", "åŀĭ", "åŁ", "åŁº", "åŁºãģ¥", "åŁºãģ¥ãģı", "åŁºæºĸ", "åŁĭãĤģ", "åŁĭãĤģè¾¼ãģ¿", "åŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°", "åŁĭãĤģè¾¼ãģ¿è¡ĮåĪĹ", "åŁŁ", "åł", "åł±", "åł´åĲĪ", "åł´åĲĪ(", "åł´åĲĪãĢģ", "åł´åĲĪãĢģ``", "åł´åĲĪãĢĤ", "åł´åĲĪãģ«", "åł´åĲĪãģ¯", "åŃ", "åŃ¦ç", "åŃ¦ç¿Ĵ", "åŃ¦ç¿ĴãģĮ", "åŃ¦ç¿ĴãģĹãģŁ", "åŃ¦ç¿ĴãĤ³ãĥ¼ãĥĳãĤ¹", "åŃ¦ç¿Ĵãĥ«ãĥ¼ãĥĹ", "åŃ¦ç¿Ĵçİĩ", "åŃ¦ç¿ĴçİĩãĤĴ", "åŃ¦ç¿ĴçĶ¨", "åŃ¦ç¿ĴçĶ¨ãĥ»æ¤ľè¨¼", "åŃĲ", "åŃĹ", "åŃĺ", "åŃĺåľ¨", "åŃĺåľ¨ãģĹãģªãģĦ", "æ", "æ¤ľè¨¼", "æ¤ľè¨¼ãĥĩãĥ¼ãĤ¿", "æ¥µ", "æ§ĭ", "æ§ĭæĪĲ", "æ§ĭç¯ī", "æ§ĺ", "æ¨", "æ¨ª", "æ¨Ļ", "æ¨Ļæºĸ", "æ¨Ļæºĸãģ®", "æ¨Ļæºĸåģıå·®", "æ¨ĻæºĸæŃ£è¦ıåĪĨå¸ĥ", "æ¨ĻæºĸçļĦãģª", "æ¨Ļæľ¬", "æ©Ł", "æ©Łæ§ĭ", "æ¬", "æ¬¡", "æ¬¡åħĥ", "æ¬¡åħĥ(", "æ¬¡åħĥãĢĤ", "æ¬²æ³ķ", "æ¬ł", "æ®Ĭ", "æ®ĭ", "æ®ĭãĤĬ", "æ®ĭå·®", "æ¯", "æ¯Ķ", "æ¯Ķä¾ĭ", "æ¯Ķä¾ĭãģĹãģ¦", "æ¯Ķçİĩ", "æ¯Ķè¼ĥ", "æ¯Ķè¼ĥãģĻãĤĭ", "æ¯Ķè¼ĥå¯¾è±¡", "æ±Ĥ", "æ±ĤãĤģ", "æ³", "æ³¨", "æ³¨åħ¥", "æ³¨æĦı", "æ³¨æĦıæ©Łæ§ĭ", "æ³ķ", "æ´»æĢ§åĮĸ", "æ´»æĢ§åĮĸéĸ¢æķ°", "æµ", "æµ®åĭķå°ıæķ°çĤ¹", "æ··", "æ¸", "æ¸¡", "æ¸¡ãģĻ", "æ¸¡ãģĻãĢĤ", "æ¸¡ãģĻæĥ³å®ļ", "æ¸¬", "æ¸¬ãĤĭ", "æ¸¬å®ļ", "æ¸¬å®ļåĢ¤ãģ®ç³»åĪĹ", "æ¸Ī", "æ¸Īãģ¿", "æ¸Īãģ¿ãģ®", "æ¸Ľ", "æ¸Ľç®Ĺ", "æ¸Ľè¡", "æ¸Ľè¡°", "æ¸Ľè¡°çİĩ", "æº", "æºĢ", "æºĢãģŁ", "æºĸ", "æ¼Ķç®Ĺ", "æĢ§", "æĢ§(", "æĢ§èĥ½", "æģ", "æģĴ", "æĥ³å®ļ", "æĥħåł±", "æĦı", "æĦıåĳ³", "æħ", "æħ®", "æħĭ", "æĪ", "æĪ»", "æĪĲ", "æĪĲåĬŁ", "æī", "æī±", "æī±ãģĨ", "æī±ãģĨ)ãĢĤ", "æī±ãģĨãĢĤ", "æīĢ", "æīĭ", "æīĭæ³ķ", "æĭ", "æĭ¡", "æĭ¡å¼µ", "æĮ", "æĮ¿", "æĮģ", "æĮģãģ¤", "æĮģãģŁ", "æĮģãģŁãģªãģĦ", "æĮĩ", "æĮĩå®ļ", "æĮĩå®ļãģķãĤĮãģŁ", "æĮĩå®ļãģķãĤĮãģŁåł´åĲĪãĢģ", "æĮĩå®ļãģĹãģŁ", "æĮĩå®ļãģĻãĤĭãģ¨", "æĮĩæ¨Ļ", "æĮĩæķ°", "æĮĩæķ°æİ¨å®ļ", "æĮĻ", "æĮĻåĭķ", "æį", "æİ", "æİ¢ç´¢", "æİ¥", "æİ¥ç¶ļ", "æİ¨", "æİ¨å®ļ", "æİ¨è«ĸ", "æı", "æıĥ", "æıĥãģĪãĤĭ", "æıı", "æııçĶ»", "æııçĶ»ãģ«ä½¿ãģ£ãģŁ", "æııçĶ»ãģĻãĤĭãĢĤ", "æııçĶ»åħĪãģ®", "æıĲ", "æıĽ", "æĲįå¤±", "æĶ", "æĶ¹", "æĶ¹åĸĦ", "æĶ¹è¡Į", "æķ", "æķ£", "æķ°", "æķ°ãĢĤ", "æķ°ãģĪ", "æķ°ãģĪæĸ¹", "æķ°ãģĪæĸ¹ãĢĮæľ¬ä½ĵ", "æķ°åĢ¤", "æķ°åĢ¤å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°", "æķ°åĢ¤çļĦãģ«", "æķ´", "æķ´æķ°", "æķĹ", "æĸ", "æĸ°", "æĸ°çīĪ", "æĸ°è¦ı", "æĸ°è¦ıä½ľæĪĲãģĻãĤĭãĢĤ", "æĸ¹", "æĸ¹äºĴæıĽæĢ§", "æĸ¹å¼ı", "æĸ¹åĲĳ", "æĸ¹éĩ", "æĸ¹éĩĿ", "æĸĩ", "æĸĩåŃĹ", "æĸĩåŃĹãģ®", "æĸĩåŃĹãģĮ", "æĸĩåŃĹãĤĴ", "æĸĩåŃĹãĥ¬ãĥĻãĥ«", "æĸĩåŃĹåĪĹ", "æĸĩåŃĹåĪĹãĤĴ", "æĸĩåŃĹåįĺä½įãģ®", "æĸĩåŃĹæķ°", "æĸĩè", "æĸĩèĦ", "æĸĩèĦĪ", "æĸŃ", "æĸŃçī", "æĸŃçīĩ", "æĹ", "æĹ¢", "æĹ¢ãģ«", "æĹ¢å®ļ", "æĹ¢å®ļåĢ¤", "æĹ¥æľ¬èªŀ", "æĺ", "æĺİ", "æĺİç¤º", "æĺİç¤ºçļĦãģª", "æĺİç¤ºçļĦãģ«", "æĻ", "æĻĤ", "æĻĤãģ®", "æĻĤçĤ¹", "æĻĤçĤ¹ãģ§", "æĽ", "æĽ²ç·ļ", "æĽ´", "æĽ´æĸ°", "æĽ¸", "æĽ¸ãģį", "æĽ¸ãģįåĩº", "æĽ¿", "æĽ¿ãģĪ", "æĽ¿ãģĪãĤĭ", "æľ", "æľª", "æľªæºĢ", "æľªçŁ¥", "æľªçŁ¥èªŀ", "æľ«å°¾", "æľ¬", "æľ¬ä½ĵ", "æľ¬æĸĩ", "æľ¬çķªå®Łè¡Įåīįãģ®ä¿®æŃ£", "æľĢ", "æľĢå¤§", "æľĢå¤§åĢ¤", "æľĢå°¤", "æľĢå°¤åĪĨåī²", "æľĢå°ı", "æľĢå°ıäºĮä¹Ĺ", "æľĢå°ıåĢ¤", "æľĢå°ıéĻĲ", "æľĢåĪĿ", "æľĢæĸ°çīĪ", "æľĢçµĤ", "æľĢçµĤæ¬¡åħĥ", "æľĢçµĤæŃ£è¦ıåĮĸå±¤", "æľĢèī", "æľĢèī¯", "æľĢéģ©", "æľĢéģ©åĮĸ", "æľī", "æľīåĬ¹", "æľīçĦ¡", "æľīéĻĲ", "æľīéĻĲåĢ¤", "æľŁ", "æĿ", "æĿ¡ä»¶", "æĿ¡ä»¶ãģĶãģ¨ãģ«", "æĿ¡ä»¶åĲį:", "æĿ¡ä»¶éĸĵãģ§", "æĿ¥", "æĿŁ", "æŀ", "æŀĲ", "æŀľ", "æł", "æł¹", "æŃ", "æŃ¢", "æŃ£", "æŃ£ãģĹ", "æŃ£ãģĹãģķ", "æŃ£å¼¦æ³¢", "æŃ£è¦ı", "æŃ£è¦ıåĪĨå¸ĥ", "æŃ£è¦ıåĮĸ", "æŃ£è¦ıåĮĸå±¤", "æŃ£è¦ıåĮĸå¾Į", "æŃ£è¦ıåĮĸå¾Įç½®", "æŃ£è¦ıåĮĸå¾Įç½®(", "æŃ£è¦ıåĮĸå¾Įç½®(Post-Layer", "æŃ£è¦ıåĮĸåīį", "æŃ£è¦ıåĮĸåīįç½®", "æŃ£è¦ıåĮĸåīįç½®(", "æŃ£è¦ıåĮĸåīįç½®(Pre-Layer", "æŃ£è¦ıåĮĸæ¸Īãģ¿", "æŃ»", "ç", "ç¢º", "ç¢ºç«ĭ", "ç¢ºçİĩ", "ç¢ºèªį", "ç¤º", "ç¤ºãģĻ", "ç§", "ç§»åĭķ", "ç¨", "ç¨®", "ç¨ĭ", "ç¨ĭåº¦", "ç©", "ç©º", "ç©ºçĻ½", "ç©ºçĻ½ãģ®", "ç©ºçĻ½ãģĮ", "ç©ºçĻ½ãĤĴ", "ç©ºéĸĵ", "ç©į", "ç©įãĤĴ", "çªĵ", "ç«", "ç«¯", "ç«ĭ", "ç¬¦åı·", "ç¬¦åı·åĮĸ", "ç¬¬", "ç®", "ç®Ĺ", "ç¯", "ç¯Ħ", "ç¯ĦåĽ²", "ç¯ī", "ç²", "ç³»åĪĹ", "ç³»åĪĹãģ®", "ç³»åĪĹåĲį:", "ç³»åĪĹéķ·", "ç´", "ç´¯ç©į", "ç´¯ç©įç¢ºçİĩ", "ç´°", "ç´°ãģ¯", "ç´Ħ", "ç´Ķ", "ç´ł", "çµ", "çµ±", "çµ±è¨Ī", "çµ±è¨Īéĩı", "çµ¶", "çµ¶å¯¾ä½įç½®", "çµ¶å¯¾ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹", "çµ¶å¯¾ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(å½¢çĬ¶", "çµĤ", "çµĤäºĨ", "çµĦ", "çµĦãģ¿", "çµĦãģ¿åĲĪãĤıãģĽ", "çµĮ", "çµĮçĶ±", "çµĮçĶ±ãģ§", "çµĮè·", "çµĮè·¯", "çµĲ", "çµĲæŀľ", "çµĲæŀľãģĮ", "ç¶", "ç¶ļ", "ç·", "ç·´", "ç·ı", "ç·ıåĴĮ", "ç·ļ", "ç·ļå½¢", "ç¸", "ç¸¦", "ç¸®", "ç¸®å°ı", "ç¹°ãĤĬ", "ç½®", "ç½®(", "ç½®ãģįæıĽãģĪ", "ç½®æıĽ", "ç¾", "ç¾¤", "çĤ¹", "çĦ", "çĦ¡", "çĦ¡ãģĦ", "çĦ¡åĬ¹", "çĦ¡è¦ĸ", "çī", "çī¹", "çī¹å¾´", "çī¹æ®Ĭ", "çīĪ", "çĬ", "çĬ¶", "çĬ¶æħĭ", "çĭ", "çĭ¬ç«ĭ", "çİĩ", "çı¾", "çĲĨ", "çĲĨçĶ±", "çĲĨè«ĸ", "çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³", "çĶ", "çĶ¨", "çĶ¨)ãĢĤ", "çĶ¨ãģ«", "çĶ¨ãģ®", "çĶ¨ãģĦãģ¦", "çĶ¨ãģĦãģŁ", "çĶ¨ãģĦãģŁéĢĲæ¬¡", "çĶ¨ãģĦãģŁéĢĲæ¬¡æİ¨è«ĸ", "çĶ±", "çĶ»", "çĶ»åĥı", "çĶŁ", "çĶŁãģĺ", "çĶŁæĪĲ", "çĶŁæĪĲãģ®åĲĦ", "çĶŁæĪĲãģĻãĤĭ", "çķ", "çķ¥", "çķª", "çķªå®Łè¡Įåīįãģ®", "çķªå®Łè¡Įåīįãģ®ä¿®æŃ£", "çķªåı·", "çķ°", "çķ°ãģªãĤĭ", "çķĮ", "çĻ", "çĻº", "çĻºåĭķ", "çĻºæķ£", "çĻºçĶŁ", "çĻ½", "çĻ¾", "çļĦ", "çļĦãģª", "çļĦãģ«", "çĽ", "çĽ®", "çĽ®æ¨Ļ", "çĽ®è¦ĸ", "çĽ®è¦ĸãģ§", "çĽ´", "çĽ´åīį", "çĽ´åīįãģ§", "çĽ´æİ¥", "çĽ¸", "çĽ¸å¯¾", "çĽ¸å¯¾ä½įç½®", "çĽ¸å¯¾è·ĿéĽ¢", "çľ", "çľģ", "çľģçķ¥", "çľģçķ¥åı¯", "çŁ", "çŁ¥", "çŃ", "çŃī", "çŃīä¾¡", "è", "è¡", "è¡¨", "è¡¨ç¤º", "è¡¨çı¾", "è¡Į", "è¡ĮãģĨ", "è¡ĮãģĮ", "è¡ĮãĤı", "è¡ĮãĤıãģªãģĦ", "è¡ĮãĤıãģļãĢģ", "è¡ĮåĪĹ", "è¢«è¦Ĩ", "è¢«è¦Ĩçİĩ", "è£", "è£ľ", "è¤ĩ", "è¤ĩæķ°", "è¤ĩæķ°ãģ®", "è¤ĩæķ°ãĤ·ãĥ¼ãĥī", "è¦", "è¦ģ", "è¦ģç´ł", "è¦ģç´łãģ¯", "è¦ĭ", "è¦ı", "è¦ıåīĩ", "è¦ıç´Ħ", "è¦ĸ", "è§", "è§£", "è§£æŀĲ", "è§Ĵ", "è§Ĵåĳ¨æ³¢æķ°", "è¨", "è¨±", "è¨±å®¹", "è¨¼", "è¨Ģ", "è¨Ģèªŀ", "è¨Ģèªŀãĥ¢ãĥĩãĥ«", "è¨Ģèªŀãĥ¢ãĥĩãĥ«(", "è¨Ģèªŀãĥ¢ãĥĩãĥ«(``forward(token_ids)", "è¨Ģèªŀãĥ¢ãĥĩãĥ«ãģ®", "è¨Ī", "è¨Īç®Ĺ", "è¨Īç®ĹãģĻãĤĭ(", "è¨Īç®ĹãģĻãĤĭãĢĤ", "è¨Īç®Ĺéĩı", "è¨Īç®ĹéĩıäºĪç®Ĺ", "è¨Īç®ĹéĩıæľĢéģ©", "è¨ĵ", "è¨ĵç·´", "è¨ĵç·´æĲįå¤±", "è¨ĺ", "è¨ĺäºĭ", "è¨ĺåı·", "è¨ĺéĮ²", "è¨Ń", "è¨Ńãģĳ", "è¨Ńå®ļ", "è¨Ńå®ļãģ§", "è¨Ńè¨Ī", "è©", "è©¦", "è©¦è¡Į", "è©²", "è©³", "è©³ç´°ãģ¯", "è©ķ", "è©ķä¾¡", "è©ķä¾¡å¯¾è±¡ãģ®", "èª", "èª¤", "èª¤ãĤĬ", "èª¬æĺİ", "èª¿", "èªį", "èªŀ", "èªŀå½Ļ", "èªŀå½Ļãģ«", "èªŀå½Ļãģ«åĲ«ãģ¾ãĤĮãģªãģĦ", "èªŀå½Ļãģ«åĲ«ãģ¾ãĤĮãĤĭ", "èªŀå½ĻãĤµãĤ¤ãĤº", "èªŀå½ĻåŃ¦ç¿Ĵ", "èªŃ", "èªŃãģ¿", "è«ĸ", "è²", "è²ª", "è²ªæ¬²æ³ķ", "è²ł", "è²łãģ®å¯¾æķ°å°¤åº¦", "è²łéłĺåŁŁ", "èµ·", "è¶", "è¶ħ", "è¶ħãģĪ", "è¶ħãģĪãģ¦ãģĦãĤĭ", "è¶ħãģĪãĤĭ", "è·", "è·ĿéĽ¢", "èº«", "èº«ãģ®", "è»", "è»¢", "è»¸", "è»¸ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ", "è»¸ãĥ©ãĥĻãĥ«ãĢĤ", "è¼ĥ", "è¾", "è¾¼", "è¾¼ãģ¿", "è¾¼ãĤĢ", "è¾ŀ", "è¾ŀæĽ¸", "è¾ŀæĽ¸ãĢĤ", "è¿", "è¿½", "è¿½åĬł", "è¿½åĬłãģĹãģŁ", "è¿ĳ", "è¿ĳä¼¼", "è¿Ķ", "è¿ĶãģĻ", "è¿ĶãģĻãĢĤ", "èĢ", "èĢĥ", "èĢĥæħ®", "èĥ", "èĥ½", "èĩ", "èĩª", "èĩªä½ĵãģ¯", "èĩªä½ľ", "èĩªä½ľãģ®", "èĩªå·±", "èĩªå·±åĽŀå¸°", "èĩªå·±æ³¨æĦı", "èĩªåĪĨ", "èĩªåĭķ", "èĩªåĭķçļĦãģ«", "èĪ", "èī", "èĭ", "èĭ±èªŀ", "èĲ", "èĲ½", "èĸĦ", "èĸĦãģĦãĥ©ãĥĥãĥĳãĥ¼ãĢĤ", "èŃ", "èŃĺ", "èŃĺåĪ¥", "èŃĺåĪ¥ä¸įèĥ½ãģª", "èŃĺåĪ¥ä¸įèĥ½ãģªãĥĩãĥ¼ãĤ¿", "é", "é¡", "é£", "é£½åĴĮãģ¹ãģįä¹Ĺåīĩ", "é«ĺ", "é«ĺãģĦ", "éĢ", "éĢ£", "éĢ£çµĲ", "éĢ£çµĲãģĹãģŁ", "éĢ£ç¶ļ", "éĢĢ", "éĢĨ", "éĢı", "éĢıéģİ", "éĢıéģİãģĻãĤĭ", "éĢĲ", "éĢĲæ¬¡", "éĢĶ", "éĢļ", "éĢļãĤĬ", "éĢļå¸¸", "éĢŁ", "éĢł", "éģ", "éģ©", "éģ©çĶ¨", "éģ©çĶ¨ãģĻãĤĭ", "éģ©çĶ¨ãģĻãĤĭãĢĤ", "éģ©çĶ¨åīį", "éģ¸", "éģ¿", "éģİ", "éģĶ", "éģĶãģĻãĤĭ", "éģķ", "éģķãģĦ", "éĥ", "éĥ¨", "éĥ¨åĪĨ", "éĥ¨åĪĨç©ºéĸĵ", "éĥ¨åĪĨèªŀ", "éĥ¨åĪĨèªŀãĤ·ãĥ³ãĥľãĥ«", "éħ", "éħį", "éĩ", "éĩį", "éĩįãģ¿", "éĩįãģ¿ãĤĴåħ±æľī", "éĩįãģ¿ä»ĺãģį", "éĩįãģ¿æ¸Ľè¡°", "éĩįãģ¿æ¸Ľè¡°ãģ®å®ŁåĬ¹", "éĩįãģŃ", "éĩįãģŃãģ¦", "éĩįãģŃãģ¦æııçĶ»ãģĻãĤĭãĢĤ", "éĩįè¤ĩ", "éĩı", "éĮ", "éĮ²", "éķ", "éķ·", "éķ·ãģĦ", "éķ·ãģķ", "éķ·ãģķãĥļãĥĬãĥ«ãĥĨãĤ£", "éĸ", "éĸ¢", "éĸ¢ä¿Ĥ", "éĸ¢æķ°", "éĸ¢æķ°ãģ®", "éĸ¾åĢ¤", "éĸ¾åĢ¤ãĢĤ", "éĸĵ", "éĸĵãģ§", "éĺ", "éĺ²", "éĺ²ãģĲ", "éĻ", "éĻ¤", "éĻ¤å¤ĸ", "éĻ¤åİ»", "éĻ¤ç®Ĺ", "éĻį", "éĻįéłĨãģ«", "éĻĲ", "éĻĲãģªãģı", "éļ", "éļİ", "éļĽ", "éļĽéĻĲãģªãģı", "éļł", "éļłãĤĮ", "éļłãĤĮçĬ¶æħĭ", "éĽ", "éĽ¢", "éĽĨ", "éĽĨåĲĪ", "éĽĨè¨Ī", "éĽĨè¨ĪãģĻãĤĭãĢĤ", "éĿ", "éĿŀ", "éĿŀåŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°", "éĿŀç·ļå½¢", "éĿŀç·ļå½¢æľĢå°ıäºĮä¹Ĺ", "éĿŀéĩįè¤ĩ", "éĿŀéĩįè¤ĩçªĵ", "éŁ", "éŁ¿", "éł", "éł»åº¦", "éłĤçĤ¹", "éłħ", "éłĨ", "éłĨãģ«", "éłĨä¼ĿæĴŃ", "éłĨä¼ĿæĴŃãĢĤ", "éłĨä¼ĿæĴŃãĥįãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯", "éłĨåºı", "éłĺ", "éłĺåŁŁ", "éłŃ", "ê", "ë", "ì", "í", "î", "ï", "ð", "ñ", "ò", "ó", "ô", "õ", "ö", "÷", "ø", "ù", "ú", "û", "ü", "ý", "þ", "ÿ", "Ā", "ā", "Ă", "ă", "Ą", "ą", "Ć", "ć", "Ĉ", "ĉ", "Ċ", "Ċ\"\"\"", "Ċ#", "Ċ(", "Ċ(``", "Ċ)", "Ċ-", "ĊICLR", "Ċ]", "Ċ_", "Ċ``", "Ċclass", "Ċfrom", "Ċimport", "ĊåŃ¦ç¿Ĵ", "ĊĊ", "ĊĊ\"\"\"", "ĊĊ\"\"\"theories/ãĥ»apps/", "ĊĊ#", "ĊĊ**", "ĊĊVaswani", "ĊĊ__all__", "ĊĊ``", "ĊĊfrom", "ĊĊimport", "ĊĊè¨ĺåı·", "ĊĊĊ", "ĊĊĊ#", "ĊĊĊ@", "ĊĊĊ@dataclass", "ĊĊĊclass", "ĊĊĊdef", "ĊĊĠĠĠĠ", "ĊĊĠĠĠĠ#", "ĊĊĠĠĠĠ..", "ĊĊĠĠĠĠ1.", "ĊĊĠĠĠĠ@", "ĊĊĠĠĠĠ@staticmethod", "ĊĊĠĠĠĠ@torch.no_grad()", "ĊĊĠĠĠĠA", "ĊĊĠĠĠĠArgs:", "ĊĊĠĠĠĠAttention", "ĊĊĠĠĠĠR", "ĊĊĠĠĠĠRaises:", "ĊĊĠĠĠĠReturns:", "ĊĊĠĠĠĠ``", "ĊĊĠĠĠĠax", "ĊĊĠĠĠĠax.set_", "ĊĊĠĠĠĠax.set_xlabel(xlabel)", "ĊĊĠĠĠĠax.set_xtick", "ĊĊĠĠĠĠcolor_cycle", "ĊĊĠĠĠĠdef", "ĊĊĠĠĠĠfor", "ĊĊĠĠĠĠif", "ĊĊĠĠĠĠreturn", "ĊĊĠĠĠĠåĲĦ", "ĊĊĠĠĠĠè¨ĺåı·", "ĊĊĠĠĠĠĠĠĠĠ", "ĊĊĠĠĠĠĠĠĠĠ#", "ĊĊĠĠĠĠĠĠĠĠArgs:", "ĊĊĠĠĠĠĠĠĠĠReturns:", "ĊĊĠĠĠĠĠĠĠĠ\\\\mathrm{", "ĊĊĠĠĠĠĠĠĠĠ``", "ĊĊĠĠĠĠĠĠĠĠfor", "ĊĊĠĠĠĠĠĠĠĠif", "ĊĊĠĠĠĠĠĠĠĠself.", "ĊĊĠĠĠĠĠĠĠĠx", "ĊĊĠĠĠĠĠĠĠĠĠĠĠĠ", "ĊĊĠĠĠĠĠĠĠĠĠĠĠĠm", "ĊĠĠ", "ĊĠĠ``", "ĊĠĠĠĠ", "ĊĠĠĠĠ\"", "ĊĠĠĠĠ\"\"\"", "ĊĠĠĠĠ\"\"\"``", "ĊĠĠĠĠ\"\"\"åŃ¦ç¿Ĵ", "ĊĠĠĠĠ\"compute_", "ĊĠĠĠĠ\"load_", "ĊĠĠĠĠ\"plot", "ĊĠĠĠĠ\"plot_", "ĊĠĠĠĠ#", "ĊĠĠĠĠ$", "ĊĠĠĠĠ(", "ĊĠĠĠĠ(``", "ĊĠĠĠĠ)", "ĊĠĠĠĠ2", "ĊĠĠĠĠ2.", "ĊĠĠĠĠ3", "ĊĠĠĠĠB", "ĊĠĠĠĠL", "ĊĠĠĠĠS", "ĊĠĠĠĠV", "ĊĠĠĠĠW", "ĊĠĠĠĠ]", "ĊĠĠĠĠ``", "ĊĠĠĠĠa", "ĊĠĠĠĠalpha:", "ĊĠĠĠĠalways_negative_", "ĊĠĠĠĠapi.", "ĊĠĠĠĠax", "ĊĠĠĠĠax.", "ĊĠĠĠĠax.grid(alpha=0.3", "ĊĠĠĠĠax.grid(alpha=0.3)", "ĊĠĠĠĠax.grid(alpha=0.3,", "ĊĠĠĠĠax.leg", "ĊĠĠĠĠax.legend(", "ĊĠĠĠĠax.legend(fontsize=8)", "ĊĠĠĠĠax.set_", "ĊĠĠĠĠax.set_xlabel(xlabel)", "ĊĠĠĠĠax.set_xtick", "ĊĠĠĠĠax.set_ylabel(", "ĊĠĠĠĠax.set_ylabel(ylabel)", "ĊĠĠĠĠax:", "ĊĠĠĠĠb", "ĊĠĠĠĠbatch_size:", "ĊĠĠĠĠbest", "ĊĠĠĠĠchunk_split_mode:", "ĊĠĠĠĠcompute_", "ĊĠĠĠĠconverged", "ĊĠĠĠĠcorpus_", "ĊĠĠĠĠd_", "ĊĠĠĠĠd_ff", "ĊĠĠĠĠd_k", "ĊĠĠĠĠd_model", "ĊĠĠĠĠdata", "ĊĠĠĠĠdef", "ĊĠĠĠĠdevice:", "ĊĠĠĠĠelse:", "ĊĠĠĠĠf", "ĊĠĠĠĠfit_", "ĊĠĠĠĠflops_per_token", "ĊĠĠĠĠfor", "ĊĠĠĠĠgenerator", "ĊĠĠĠĠh", "ĊĠĠĠĠhuber_delta:", "ĊĠĠĠĠi", "ĊĠĠĠĠif", "ĊĠĠĠĠload_", "ĊĠĠĠĠlog", "ĊĠĠĠĠlog_", "ĊĠĠĠĠlog_alpha", "ĊĠĠĠĠlog_scale:", "ĊĠĠĠĠm", "ĊĠĠĠĠmax_", "ĊĠĠĠĠmax_chunk_bytes:", "ĊĠĠĠĠmodel:", "ĊĠĠĠĠn", "ĊĠĠĠĠn_", "ĊĠĠĠĠnum", "ĊĠĠĠĠp", "ĊĠĠĠĠparams", "ĊĠĠĠĠpath", "ĊĠĠĠĠplot", "ĊĠĠĠĠplot_", "ĊĠĠĠĠprofiles:", "ĊĠĠĠĠresidual_fn:", "ĊĠĠĠĠresiduals:", "ĊĠĠĠĠresult", "ĊĠĠĠĠreturn", "ĊĠĠĠĠright", "ĊĠĠĠĠsorted_", "ĊĠĠĠĠtext", "ĊĠĠĠĠtext:", "ĊĠĠĠĠtitle:", "ĊĠĠĠĠtry:", "ĊĠĠĠĠv", "ĊĠĠĠĠvocab", "ĊĠĠĠĠvocab:", "ĊĠĠĠĠweights:", "ĊĠĠĠĠwhile", "ĊĠĠĠĠwith", "ĊĠĠĠĠx", "ĊĠĠĠĠxlabel:", "ĊĠĠĠĠylabel:", "ĊĠĠĠĠ}", "ĊĠĠĠĠÎ", "ĊĠĠĠĠãĢĮ", "ĊĠĠĠĠãģĮ", "ĊĠĠĠĠãģĿãģ®ãģ¾ãģ¾", "ĊĠĠĠĠãĤĴ", "ĊĠĠĠĠæĲįå¤±", "ĊĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠ\"", "ĊĠĠĠĠĠĠĠĠ\"\"\"", "ĊĠĠĠĠĠĠĠĠ\"\"\"éłĨä¼ĿæĴŃãĢĤ", "ĊĠĠĠĠĠĠĠĠ\"median", "ĊĠĠĠĠĠĠĠĠ\"p90", "ĊĠĠĠĠĠĠĠĠ#", "ĊĠĠĠĠĠĠĠĠ(", "ĊĠĠĠĠĠĠĠĠ)", "ĊĠĠĠĠĠĠĠĠ-", "ĊĠĠĠĠĠĠĠĠ1", "ĊĠĠĠĠĠĠĠĠValueError:", "ĊĠĠĠĠĠĠĠĠ_,", "ĊĠĠĠĠĠĠĠĠ``", "ĊĠĠĠĠĠĠĠĠa", "ĊĠĠĠĠĠĠĠĠactivation_fn:", "ĊĠĠĠĠĠĠĠĠalpha", "ĊĠĠĠĠĠĠĠĠax", "ĊĠĠĠĠĠĠĠĠax.plot", "ĊĠĠĠĠĠĠĠĠax.set_", "ĊĠĠĠĠĠĠĠĠax.set_title(title)", "ĊĠĠĠĠĠĠĠĠax.set_yscale(\"log\")", "ĊĠĠĠĠĠĠĠĠax:", "ĊĠĠĠĠĠĠĠĠb", "ĊĠĠĠĠĠĠĠĠbatch_size:", "ĊĠĠĠĠĠĠĠĠbeam", "ĊĠĠĠĠĠĠĠĠbeta", "ĊĠĠĠĠĠĠĠĠbeta:", "ĊĠĠĠĠĠĠĠĠbetas:", "ĊĠĠĠĠĠĠĠĠbia", "ĊĠĠĠĠĠĠĠĠbias:", "ĊĠĠĠĠĠĠĠĠbias_correction", "ĊĠĠĠĠĠĠĠĠbidirectional", "ĊĠĠĠĠĠĠĠĠbidirectional:", "ĊĠĠĠĠĠĠĠĠcache_dir:", "ĊĠĠĠĠĠĠĠĠchunk", "ĊĠĠĠĠĠĠĠĠchunk_split_mode:", "ĊĠĠĠĠĠĠĠĠcolor", "ĊĠĠĠĠĠĠĠĠcurv", "ĊĠĠĠĠĠĠĠĠd_ff:", "ĊĠĠĠĠĠĠĠĠd_k", "ĊĠĠĠĠĠĠĠĠd_model:", "ĊĠĠĠĠĠĠĠĠdevice:", "ĊĠĠĠĠĠĠĠĠdropout:", "ĊĠĠĠĠĠĠĠĠdtype:", "ĊĠĠĠĠĠĠĠĠe", "ĊĠĠĠĠĠĠĠĠelse:", "ĊĠĠĠĠĠĠĠĠeps:", "ĊĠĠĠĠĠĠĠĠexcept", "ĊĠĠĠĠĠĠĠĠf", "ĊĠĠĠĠĠĠĠĠfeed_forward_factory:", "ĊĠĠĠĠĠĠĠĠflops_per_token", "ĊĠĠĠĠĠĠĠĠfor", "ĊĠĠĠĠĠĠĠĠh", "ĊĠĠĠĠĠĠĠĠhistory[\"", "ĊĠĠĠĠĠĠĠĠif", "ĊĠĠĠĠĠĠĠĠk", "ĊĠĠĠĠĠĠĠĠkey_length:", "ĊĠĠĠĠĠĠĠĠlog", "ĊĠĠĠĠĠĠĠĠlog_n", "ĊĠĠĠĠĠĠĠĠlog_scale:", "ĊĠĠĠĠĠĠĠĠlr:", "ĊĠĠĠĠĠĠĠĠm", "ĊĠĠĠĠĠĠĠĠm,", "ĊĠĠĠĠĠĠĠĠmax_", "ĊĠĠĠĠĠĠĠĠmax_chunk_bytes:", "ĊĠĠĠĠĠĠĠĠmax_distance", "ĊĠĠĠĠĠĠĠĠmax_distance:", "ĊĠĠĠĠĠĠĠĠmerges", "ĊĠĠĠĠĠĠĠĠmodel:", "ĊĠĠĠĠĠĠĠĠn", "ĊĠĠĠĠĠĠĠĠn:", "ĊĠĠĠĠĠĠĠĠnew", "ĊĠĠĠĠĠĠĠĠnn.", "ĊĠĠĠĠĠĠĠĠnorm_first:", "ĊĠĠĠĠĠĠĠĠnorm_fn", "ĊĠĠĠĠĠĠĠĠnormalization_factory:", "ĊĠĠĠĠĠĠĠĠnum", "ĊĠĠĠĠĠĠĠĠnum_buckets:", "ĊĠĠĠĠĠĠĠĠnum_heads:", "ĊĠĠĠĠĠĠĠĠp", "ĊĠĠĠĠĠĠĠĠparam", "ĊĠĠĠĠĠĠĠĠparams:", "ĊĠĠĠĠĠĠĠĠparams_", "ĊĠĠĠĠĠĠĠĠpath", "ĊĠĠĠĠĠĠĠĠpositional_transform:", "ĊĠĠĠĠĠĠĠĠquery_length:", "ĊĠĠĠĠĠĠĠĠraise", "ĊĠĠĠĠĠĠĠĠratio", "ĊĠĠĠĠĠĠĠĠrelative_position", "ĊĠĠĠĠĠĠĠĠresiduals", "ĊĠĠĠĠĠĠĠĠresiduals=", "ĊĠĠĠĠĠĠĠĠreturn", "ĊĠĠĠĠĠĠĠĠs", "ĊĠĠĠĠĠĠĠĠseed:", "ĊĠĠĠĠĠĠĠĠself,", "ĊĠĠĠĠĠĠĠĠself.", "ĊĠĠĠĠĠĠĠĠself.beta", "ĊĠĠĠĠĠĠĠĠself.d_k", "ĊĠĠĠĠĠĠĠĠself.dropout", "ĊĠĠĠĠĠĠĠĠself.eps", "ĊĠĠĠĠĠĠĠĠself.lr", "ĊĠĠĠĠĠĠĠĠself.norm", "ĊĠĠĠĠĠĠĠĠself.num_heads", "ĊĠĠĠĠĠĠĠĠself.register", "ĊĠĠĠĠĠĠĠĠself.register_buffer(\"", "ĊĠĠĠĠĠĠĠĠself.step_count", "ĊĠĠĠĠĠĠĠĠself.w_", "ĊĠĠĠĠĠĠĠĠseq_len", "ĊĠĠĠĠĠĠĠĠsuper().__init__()", "ĊĠĠĠĠĠĠĠĠt", "ĊĠĠĠĠĠĠĠĠtext:", "ĊĠĠĠĠĠĠĠĠtitle:", "ĊĠĠĠĠĠĠĠĠtoken_ids:", "ĊĠĠĠĠĠĠĠĠtry:", "ĊĠĠĠĠĠĠĠĠv", "ĊĠĠĠĠĠĠĠĠvalues", "ĊĠĠĠĠĠĠĠĠvocab", "ĊĠĠĠĠĠĠĠĠvocab:", "ĊĠĠĠĠĠĠĠĠvocabulary_size:", "ĊĠĠĠĠĠĠĠĠweight_decay:", "ĊĠĠĠĠĠĠĠĠweights:", "ĊĠĠĠĠĠĠĠĠx", "ĊĠĠĠĠĠĠĠĠx:", "ĊĠĠĠĠĠĠĠĠx_c", "ĊĠĠĠĠĠĠĠĠxlabel:", "ĊĠĠĠĠĠĠĠĠylabel:", "ĊĠĠĠĠĠĠĠĠåıĸå¾ĹãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ", "ĊĠĠĠĠĠĠĠĠæııçĶ»ãģ«ä½¿ãģ£ãģŁ", "ĊĠĠĠĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ\"", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ#", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(004", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(``", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(output,", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ)", "ĊĠĠĠĠĠĠĠĠĠĠĠĠFalse", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ[", "ĊĠĠĠĠĠĠĠĠĠĠĠĠ``", "ĊĠĠĠĠĠĠĠĠĠĠĠĠax", "ĊĠĠĠĠĠĠĠĠĠĠĠĠb", "ĊĠĠĠĠĠĠĠĠĠĠĠĠbreak", "ĊĠĠĠĠĠĠĠĠĠĠĠĠcolor=\"tab:", "ĊĠĠĠĠĠĠĠĠĠĠĠĠcontinue", "ĊĠĠĠĠĠĠĠĠĠĠĠĠcurrent", "ĊĠĠĠĠĠĠĠĠĠĠĠĠf", "ĊĠĠĠĠĠĠĠĠĠĠĠĠf\"", "ĊĠĠĠĠĠĠĠĠĠĠĠĠfeed_forward_factory", "ĊĠĠĠĠĠĠĠĠĠĠĠĠfor", "ĊĠĠĠĠĠĠĠĠĠĠĠĠhistory", "ĊĠĠĠĠĠĠĠĠĠĠĠĠi", "ĊĠĠĠĠĠĠĠĠĠĠĠĠif", "ĊĠĠĠĠĠĠĠĠĠĠĠĠlabel=", "ĊĠĠĠĠĠĠĠĠĠĠĠĠlin", "ĊĠĠĠĠĠĠĠĠĠĠĠĠlog", "ĊĠĠĠĠĠĠĠĠĠĠĠĠm", "ĊĠĠĠĠĠĠĠĠĠĠĠĠn", "ĊĠĠĠĠĠĠĠĠĠĠĠĠnew", "ĊĠĠĠĠĠĠĠĠĠĠĠĠnp.log(warm_fit", "ĊĠĠĠĠĠĠĠĠĠĠĠĠnp.log(warm_fit.", "ĊĠĠĠĠĠĠĠĠĠĠĠĠoutput", "ĊĠĠĠĠĠĠĠĠĠĠĠĠp.grad", "ĊĠĠĠĠĠĠĠĠĠĠĠĠpositions:", "ĊĠĠĠĠĠĠĠĠĠĠĠĠraise", "ĊĠĠĠĠĠĠĠĠĠĠĠĠrelative_position", "ĊĠĠĠĠĠĠĠĠĠĠĠĠreturn", "ĊĠĠĠĠĠĠĠĠĠĠĠĠself.", "ĊĠĠĠĠĠĠĠĠĠĠĠĠself_", "ĊĠĠĠĠĠĠĠĠĠĠĠĠx", "ĊĠĠĠĠĠĠĠĠĠĠĠĠx:", "ĊĠĠĠĠĠĠĠĠĠĠĠĠx_hat", "ĊĠĠĠĠĠĠĠĠĠĠĠĠãģ®", "ĊĠĠĠĠĠĠĠĠĠĠĠĠå½¢çĬ¶", "ĊĠĠĠĠĠĠĠĠĠĠĠĠåł´åĲĪãģ¯", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ(", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ)", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ``", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠb", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠcontinue", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠf\"", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠfor", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠif", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠparam", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠparams,", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠs", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠself.", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠbest", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠj", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ", "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠnp.log(", "ċ", "Č", "č", "Ď", "ď", "Đ", "đ", "Ē", "ē", "Ĕ", "ĕ", "Ė", "ė", "Ę", "ę", "Ě", "ě", "Ĝ", "ĝ", "Ğ", "ğ", "Ġ", "Ġ!=", "Ġ\"", "Ġ\"\"", "Ġ\"\",", "Ġ\"\".join(", "Ġ\"Attention", "Ġ\"G", "Ġ\"Loss\",", "Ġ\"S", "Ġ\"T", "Ġ\"Value\",", "Ġ\"log", "Ġ$", "Ġ%", "Ġ&", "Ġ&=", "Ġ(", "Ġ(-", "Ġ(...,", "Ġ(0.", "Ġ(0.1,", "Ġ(009", "Ġ(1", "Ġ(2", "Ġ(20", "Ġ(B,", "Ġ(S_q,", "Ġ(b", "Ġ(h,", "Ġ(i", "Ġ(log", "Ġ(n", "Ġ(name,", "Ġ(np.", "Ġ(x", "Ġ(x_c", "Ġ({", "Ġ*", "Ġ**", "Ġ+", "Ġ+=", "Ġ-", "Ġ-(", "Ġ---------------------------------------------------------------", "Ġ-=", "Ġ->", "Ġ-inf", "Ġ...", "Ġ...,", "Ġ/", "Ġ//", "Ġ0", "Ġ0,", "Ġ0.", "Ġ0.0", "Ġ0.0,", "Ġ0.0:", "Ġ0.3", "Ġ0.5", "Ġ0.5,", "Ġ0.7", "Ġ0.8", "Ġ0.999", "Ġ003", "Ġ004", "Ġ005", "Ġ01", "Ġ010)", "Ġ0:", "Ġ1", "Ġ1)", "Ġ1))", "Ġ1):", "Ġ1,", "Ġ1.0", "Ġ1.0)", "Ġ1.0,", "Ġ10", "Ġ100", "Ġ10000", "Ġ10000^(2i", "Ġ1:", "Ġ1]", "Ġ1e", "Ġ1e-", "Ġ1e-1", "Ġ1e-12)", "Ġ1e-8", "Ġ2", "Ġ2)", "Ġ2,", "Ġ2-3", "Ġ2.0", "Ġ2.5", "Ġ20", "Ġ201", "Ġ2016", "Ġ2017", "Ġ2017)ãĢĤ", "Ġ2018", "Ġ2019", "Ġ202", "Ġ2020", "Ġ2024", "Ġ21)ãĢĤ", "Ġ2:", "Ġ2]", "Ġ3", "Ġ3:", "Ġ4", "Ġ4.0", "Ġ4.2))", "Ġ4.5))", "Ġ429", "Ġ5", "Ġ6", "Ġ:", "Ġ:,", "Ġ:]", "Ġ:math:`", "Ġ<", "Ġ<-", "Ġ<->", "Ġ<=", "Ġ=", "Ġ==", "Ġ>", "Ġ@", "ĠA", "ĠA,", "ĠA/N^alpha", "ĠABC", "ĠAPI", "ĠAdam", "ĠAdamW", "ĠAll", "ĠApproach", "ĠAttention", "ĠAx", "ĠAxes", "ĠAxes:", "ĠAxesãĢĤ", "ĠAxesãĢĤNone", "ĠB", "ĠB/D^beta", "ĠBPE", "ĠBPEIDTokenizer", "ĠBPETokenizer", "ĠBlock", "ĠBlock(", "ĠBãĢĤ", "ĠC", "ĠC^", "ĠCallab", "ĠCallable", "ĠCallable[", "ĠCallable[[", "ĠCallable[[Tensor],", "ĠCallable[[],", "ĠCallable[[int],", "ĠCallable[[np.ndarray],", "ĠCharacterLevelTokenizer", "ĠCharacterLevelTokenizer:", "ĠChinchilla", "ĠChinchillaParametricFit", "ĠCon", "ĠCounter", "ĠCounter[tuple[str,", "ĠCrossAttention", "ĠD", "ĠDecoder", "ĠDecoderBlock", "ĠDot-Product", "ĠE", "ĠEM", "ĠEmbedding", "ĠEncoder", "ĠEncoding", "ĠError", "ĠF", "ĠFace", "ĠFalse", "ĠFalse,", "ĠFeed", "ĠFeed-Forward", "ĠFeedForwardNetwork", "ĠFig", "ĠFunction", "ĠG", "ĠGPTLanguageModel", "ĠH", "ĠHub", "ĠHuber", "ĠI", "ĠICLR", "ĠID", "ĠIs", "ĠIsoFLOP", "ĠIsoFLOPParabolaFit", "ĠIter", "ĠJ", "ĠJSON", "ĠJaccard", "ĠK", "ĠKey", "ĠL", "ĠL2", "ĠLanguage", "ĠLaw", "ĠLayerNorm", "ĠLayerNorm(x", "ĠLayerNorm(x),", "ĠLayerNormalization", "ĠLearn", "ĠLevenberg-Marquardt", "ĠLinear", "ĠLongTensor", "ĠLongTensor(", "ĠLongTensor(``", "ĠLongTensorãĢĤ", "ĠM", "ĠMean", "ĠModel", "ĠMulti-Head", "ĠMultiHeadAttention", "ĠMultiHeadAttention(d_model,", "ĠN", "ĠN,", "ĠN_opt", "ĠNeed\",", "ĠNetwork", "ĠNeur", "ĠNeurIPS", "ĠNeural", "ĠNone", "ĠNone)", "ĠNone,", "ĠNone:", "ĠNone]", "ĠNormalization", "ĠNormalization\",", "ĠNotImplementedError", "ĠNotation:", "ĠOS", "ĠP", "ĠPair", "ĠPath", "ĠPath(", "ĠPath)", "ĠPath,", "ĠPowerLawFit", "ĠQuery", "ĠQueryKeyPositionalTransform", "ĠQueryãĥ»Key", "ĠR", "ĠRMS", "ĠRMSNorm", "ĠR^{", "ĠR^{d_ff", "ĠR^{d_model", "ĠReLU", "ĠS", "ĠS)``", "ĠS,", "ĠS_", "ĠS_0", "ĠS_k", "ĠS_k)", "ĠS_k)``", "ĠS_k,", "ĠS_q,", "ĠS_tgt", "ĠS_tgt,", "ĠSaturatingPowerLawFit", "ĠScal", "ĠSelfAttention", "ĠSennrich", "ĠSequence", "ĠSequence[", "ĠSequence[float", "ĠSequence[float])", "ĠSequence[float],", "ĠSequence[int]", "ĠSequence[str", "ĠSequence[str])", "ĠSequence[str],", "ĠShaw", "ĠSiLU", "ĠSquar", "ĠSquare", "ĠSt", "ĠSwiGLU", "ĠSwish", "ĠSãĢĤ", "ĠT", "ĠTensor", "ĠTensor)", "ĠTensor):", "ĠTensor,", "ĠTensor:", "ĠTensor]", "ĠTensor]:", "ĠText", "ĠTransformer", "ĠTrue", "ĠTrue(æĹ¢å®ļåĢ¤", "ĠTrue,", "ĠU", "ĠUTF-8", "ĠUn", "ĠUnicode", "ĠUnigram", "ĠUnigramTokenizer", "ĠUnit", "ĠV", "ĠV)", "ĠV)``", "ĠValue", "ĠValueError", "ĠValueError(", "ĠValueError(\"", "ĠValueError(f\"", "ĠViterbi", "ĠW", "ĠW^", "ĠW^O", "ĠW_2", "ĠWikipedia", "ĠYou", "Ġ[", "Ġ[]", "Ġ[],", "Ġ[p.", "Ġ\\\\", "Ġ\\\\\\\\", "Ġ\\\\beta", "Ġ\\\\beta_", "Ġ\\\\beta_1", "Ġ\\\\beta_2", "Ġ\\\\frac{", "Ġ\\\\lambda", "Ġ\\\\text{", "Ġ\\\\theta", "Ġ\\\\theta_{t-1}", "Ġ_", "Ġ_,", "Ġ_LOG_ALPHA_BETA_M", "Ġ_LOG_ALPHA_BETA_MAX", "Ġ_LOG_ALPHA_BETA_MAX)", "Ġ_LOG_ALPHA_BETA_MIN,", "Ġ_WIKIPEDIA_MANIFEST_DIR", "Ġ__future__", "Ġ__init__(", "Ġ__init__(self,", "Ġ_boundaries(", "Ġ_build_", "Ġ_chinchilla_residual", "Ġ_count_pair", "Ġ_count_pairs(", "Ġ_fit_nonlinear_huber", "Ġ_gnmt_length_penalty", "Ġ_levenberg_marquardt(", "Ġ_merge_", "Ġ_merge_pair_in_word(", "Ġ_power_of_2_slopes", "Ġ_strip_wikitext_links(", "Ġ_to_numpy", "Ġ``", "Ġ``\"", "Ġ``\"gradient_norm", "Ġ``\"loss_step_delta", "Ġ``\"loss_step_delta\"``", "Ġ``(", "Ġ``(...,", "Ġ``(1,", "Ġ``(B,", "Ġ``(N,", "Ġ``(S,", "Ġ``(S_q,", "Ġ``(num_windows,", "Ġ``-inf``", "Ġ``0", "Ġ``0.0", "Ġ``6", "Ġ``C", "Ġ``C``", "Ġ``DecoderBlock``", "Ġ``L", "Ġ``MultiHeadAttention``", "Ġ``None``", "Ġ``_", "Ġ``beam_size", "Ġ``beam_size``", "Ġ``d_model``", "Ġ``log", "Ġ``lp(Y", "Ġ``n", "Ġ``n``", "Ġ``nn.", "Ġ``p``", "Ġ``torch.", "Ġ``{", "Ġ``{æĿ¡ä»¶åĲį:", "Ġ``{ç³»åĪĹåĲį:", "Ġ``âĸģ", "Ġ``âĸģ``", "Ġa", "Ġa^K_mn", "Ġa_coef", "Ġal", "Ġal.", "Ġal.,", "Ġalpha", "Ġalpha=-", "Ġand", "Ġann", "Ġannotations", "Ġar", "ĠarXiv:", "Ġas", "Ġattempt", "Ġattn_weights", "Ġattn_weights)", "Ġax", "Ġaxis", "Ġaxis=\"y\")", "Ġb", "Ġb,", "Ġb_coef", "Ġbase", "Ġbatch", "Ġbeam", "Ġbest", "Ġbest_cost", "Ġbest_pair", "Ġbest_params", "Ġbest_score", "Ġbeta", "Ġbeta:", "Ġbias", "Ġbias(", "Ġbias=bias)", "Ġbias_correction", "Ġbits-per-byte", "Ġblock", "Ġbool", "Ġbool]:", "Ġboundaries", "Ġbpe.", "Ġbyt", "Ġbyte_", "Ġc", "Ġcache_dir", "Ġcache_dir,", "Ġcache_path", "Ġcallable", "Ġcenter", "Ġch", "Ġchar", "Ġchunk", "Ġchunk:", "Ġchunk_freq", "Ġchunks", "Ġclip", "Ġclipping", "Ġco", "Ġcoeff", "Ġcollection", "Ġcollections.abc", "Ġcolor", "Ġcolor=\"", "Ġcolor=color,", "Ġcolor_cycle[i", "Ġcompute_", "Ġcon", "Ġconverged", "Ġcos", "Ġcosine", "Ġcost", "Ġcount", "Ġcre", "Ġcreate_causal_mask", "Ġcross_", "Ġcross_attn_weights", "Ġcum_log_prob", "Ġcurrent", "Ġcurv", "Ġd", "Ġd_", "Ġd_ff", "Ġd_ff'", "Ġd_ff,", "Ġd_k", "Ġd_k)", "Ġd_k)``", "Ġd_k/", "Ġd_model", "Ġd_model)", "Ġd_model))", "Ġd_model)``", "Ġd_model)``ãĢģ", "Ġd_model)``ãĢĤ", "Ġd_model,", "Ġd_model:", "Ġd_model``", "Ġd_v", "Ġd_values,", "Ġdata", "Ġdata[\"", "Ġdataclass", "Ġdecay", "Ġdecod", "Ġdecode(self,", "Ġdecoder-only", "Ġdefault", "Ġdelta", "Ġdevice=", "Ġdevice=device)", "Ġdevice=query.device)", "Ġdi", "Ġdict", "Ġdict(", "Ġdict[float,", "Ġdict[int,", "Ġdict[str,", "Ġdim", "Ġdim=-1)", "Ġdocstring", "Ġdropout", "Ġdropout=", "Ġdropout=dropout)", "Ġdropout=dropout,", "Ġdtype=", "Ġdtype=float", "Ġdtype=float)", "Ġdtype=torch.", "Ġdtype=torch.bool", "Ġdtype=torch.float3", "Ġdtype=torch.float32)", "Ġe", "Ġe_mn", "Ġelse", "Ġen", "Ġencode(self,", "Ġencoding=\"utf-8\")", "Ġend", "Ġenumerate(", "Ġeps", "Ġet", "Ġevaluate_bits_per_byt", "Ġevaluation_window", "Ġexist_ok", "Ġexist_ok=True)", "Ġexp", "Ġextra", "Ġf", "Ġf\"", "Ġf\"{", "Ġfertility", "Ġfit_", "Ġfloat", "Ġfloat(", "Ġfloat(\"", "Ġfloat(\"-inf\"", "Ġfloat(\"-inf\"))", "Ġfloat(np.", "Ġfloat(np.exp(log", "Ġfloat(np.mean", "Ġfloat(np.median", "Ġfloat(np.sum(", "Ġfloat)", "Ġfloat,", "Ġfloat:", "Ġfloat]", "Ġfloat],", "Ġfloat]:", "Ġflops_per_token", "Ġfontsize=", "Ġfor", "Ġforward(", "Ġforward(self,", "Ġfreq", "Ġfreq)", "Ġfunctional", "Ġg", "Ġg_t", "Ġgenerator", "Ġget_random_batch", "Ġgradient", "Ġgradient_clip_threshold", "Ġgradient_norm", "Ġgrid", "Ġgrid],", "Ġh", "Ġh,", "Ġhead", "Ġhistory", "Ġhuber_delta", "Ġhuber_delta)", "ĠhãĢĤ", "Ġi", "Ġi,", "Ġid", "Ġid(p)", "Ġif", "Ġimport", "Ġin", "Ġinput", "Ġint", "Ġint)", "Ġint,", "Ġint]", "Ġinvalid=\"ignore\"):", "Ġis", "Ġisinstance", "Ġitem", "Ġitem:", "Ġj", "Ġjo", "Ġjson.loads(", "Ġjson.loads(Path", "Ġk", "Ġk,", "Ġk_clip", "Ġkeepdim", "Ġkeepdim=True)", "Ġkey", "Ġl", "Ġl_inf", "Ġlabel=", "Ġlabel=name,", "Ġlen", "Ġlen(", "Ġlen(color_cycle", "Ġlen(color_cycle)]", "Ġlength)", "Ġlin", "Ġline", "Ġlinewidth", "Ġlinewidth=1.", "Ġlist", "Ġlist(", "Ġlist[", "Ġlist[Tensor]", "Ġlist[float", "Ġlist[float]", "Ġlist[float]:", "Ġlist[float]],", "Ġlist[str", "Ġlist[str]", "Ġlist[str]:", "Ġlm_head", "Ġlo", "Ġload_", "Ġload_bpe_", "Ġload_wikipedia_corpus", "Ġlog", "Ġlog_", "Ġlog_alpha", "Ġlog_alpha_max", "Ġlog_n", "Ġlog_n_opt", "Ġlog_scale:", "Ġlogit", "Ġlogits", "Ġlogits``", "ĠlogitsãĢĤ", "Ġloss_", "Ġlp", "Ġlr", "Ġlr:", "Ġm", "Ġm,", "Ġmanifest_path", "Ġmanifest_path=", "Ġmask", "Ġmask)", "Ġmask.dim()", "Ġmath", "Ġmath.", "Ġmath.sqrt(", "Ġmath::", "Ġmatplotlib", "Ġmatrix", "Ġmatrix.ndim", "Ġmax(", "Ġmax_", "Ġmax_chunk", "Ġmax_chunk_bytes:", "Ġmax_exact", "Ġmax_exact)", "Ġmax_len", "Ġmax_relative_position", "Ġmean", "Ġmemory", "Ġmemory,", "Ġmemory_", "Ġmemory_mask)", "Ġmin", "Ġmin(", "Ġmodel", "Ġmodel.", "Ġmodel.parameters()", "Ġmulti-start", "Ġn", "Ġn-gram", "Ġn:", "Ġn_", "Ġn_ctx", "Ġn_layer", "Ġnam", "Ġname", "Ġneg", "Ġnew", "Ġnew_symbol", "Ġnn", "Ġnn.", "Ġnn.Dropout", "Ġnn.Dropout(dropout)", "Ġnn.Embedding", "Ġnn.Linear", "Ġnn.Linear(d_model,", "Ġnn.Module", "Ġnn.Module]", "Ġnn.Parameter(torch.", "Ġnoise_band", "Ġnoqa:", "Ġnorm", "Ġnorm_fn(d_model)", "Ġnormalization_factory", "Ġnormed,", "Ġnot", "Ġnp", "Ġnp.", "Ġnp.arange(", "Ġnp.array(", "Ġnp.array([", "Ġnp.array([p.", "Ġnp.as", "Ġnp.asarray", "Ġnp.asarray(", "Ġnp.errstate(over=\"ignore\",", "Ġnp.exp(log", "Ġnp.inf", "Ġnp.isfinite(", "Ġnp.log", "Ġnp.log(", "Ġnp.ndarray", "Ġnp.ndarray)", "Ġnp.ndarray,", "Ġnp.ndarray:", "Ġnp.ndarray],", "Ġnp.ones_like(", "Ġnum", "Ġnum_buckets", "Ġnum_heads", "Ġnum_heads,", "Ġnum_total", "Ġnumpy", "Ġo", "Ġobject", "Ġof", "Ġoptimizer", "Ġor", "Ġother_", "Ġp", "Ġp.grad", "Ġpair", "Ġpair,", "Ġpair_in", "Ġparabola.", "Ġparam", "Ġparams", "Ġparams)", "Ġparams,", "Ġpath", "Ġpersistent=False)", "Ġpiece", "Ġplot", "Ġplot_", "Ġplt", "Ġplt.rcParams[\"axes.prop_cycle\"].by_key()[\"color\"]", "Ġplt.subplots(", "Ġplt.subplots(figsize=(", "Ġplt.subplots(figsize=(6", "Ġplt.subplots(figsize=(6.5,", "Ġpoint", "Ġpoints]", "Ġpos", "Ġpositional_transform", "Ġpositions", "Ġpositions:", "Ġpre", "Ġpretokenize(", "Ġpro", "Ġq", "Ġq.", "Ġq_m", "Ġquery", "Ġr", "Ġran", "Ġrange(", "Ġrange(n", "Ġrange(num", "Ġraw_norm", "Ġre", "Ġre.sub", "Ġre.sub(r\"", "Ġre.sub(r\"(?", "Ġre.sub(r\"<", "Ġreconstruct_optimal_frontier(", "Ġrelative_", "Ġrelative_buckets", "Ġrelative_position", "Ġrelative_vector", "Ġresidual_fn", "Ġresidual_fn(param", "Ġresidual_fn(params)", "Ġrevid", "Ġs", "Ġs_", "Ġsampling", "Ġscal", "Ġscore", "Ġse", "Ġself.", "Ġself._", "Ġself._m", "Ġself._split_heads(self.w_", "Ġself.attention_score_bias", "Ġself.beta", "Ġself.beta2", "Ġself.dropout", "Ġself.dropout1(", "Ġself.dropout2", "Ġself.eps)", "Ġself.gamma", "Ġself.norm", "Ġself.norm1", "Ġself.norm2", "Ġself.params:", "Ġself.self_attn", "Ġself.w_", "Ġself.weight_decay", "Ġself_", "Ġsentencepiece", "Ġseq_len", "Ġseq_len,", "Ġsequence", "Ġsequence_length", "Ġsequence_length)", "Ġset", "Ġset()", "Ġset[str]", "Ġsin", "Ġsize", "Ġsize)", "Ġsoftmax", "Ġsorted", "Ġsorted_", "Ġsqrt(", "Ġsqrt(d_k)", "Ġsrc.", "Ġsrc.data.", "Ġsrc.layers.", "Ġsrc.layers.attention", "Ġsrc.layers.normalization", "Ġsrc.layers.positional_encoding", "Ġsrc.utils.", "Ġss_", "Ġstart", "Ġstep", "Ġstr", "Ġstr)", "Ġstr,", "Ġstr:", "Ġstr]", "Ġstr],", "Ġstr]]", "Ġstrict", "Ġstrict=True", "Ġstrict=True)", "Ġsum", "Ġswish", "Ġsymbol", "Ġsymbol_to_id", "Ġsymbols", "Ġsymbols[i", "Ġt", "Ġtarget", "Ġtargets", "Ġtemperature", "Ġtext", "Ġtext)", "Ġtext:", "Ġtgt_mask", "Ġth", "Ġthreshold", "Ġtitle", "Ġtitle:", "Ġto", "Ġtoken", "Ġtoken_id", "Ġtoken_ids", "Ġtokenizer.json", "Ġtop", "Ġtorch", "Ġtorch.", "Ġtorch.Generator", "Ġtorch.Tensor", "Ġtorch.arange(", "Ġtorch.arange(key_length,", "Ġtorch.arange(query_length,", "Ġtorch.cat", "Ġtorch.cat([", "Ġtorch.device", "Ġtorch.dtype,", "Ġtorch.m", "Ġtorch.s", "Ġtorch.sqrt(", "Ġtorch.t", "Ġtorch.tensor", "Ġtorch.zero", "Ġtorch.zeros(", "Ġtorch.zeros_like(", "Ġtorch.zeros_like(p)", "Ġtotal_bytes)", "Ġtrain", "Ġtry_decode_byte_level_symbol", "Ġtup", "Ġtuple", "Ġtuple[", "Ġtuple[Tensor,", "Ġtuple[float,", "Ġun", "Ġunit)", "Ġurllib", "Ġurllib.", "Ġuse_cross_attention", "Ġv", "Ġval", "Ġvalue", "Ġvalues", "Ġvalues)", "Ġvalues,", "Ġvalues_arr", "Ġvertex_log_n", "Ġvmax", "Ġvocab", "Ġvocab_size", "Ġvocabulary", "Ġw", "Ġwarmup", "Ġwarmup_step", "Ġwarmup_steps", "Ġweights", "Ġwikitext", "Ġx", "Ġx)", "Ġx,", "Ġx.", "Ġx:", "Ġx^", "Ġx_hat", "Ġy", "Ġzip", "Ġzip(", "Ġ{", "Ġ{\"", "Ġ{ch", "Ġ{}", "Ġ|", "ĠÃĹ", "ĠÎ", "ĠÎ²", "ĠÎ³", "ĠÎµ", "ĠÎµ)", "ĠÎµãĢĤ", "ĠÎ·", "ĠÎ¸", "ĠÎ»", "ĠÎ¼", "Ġâ", "ĠâĨ", "ĠâĨĴ", "ĠâĪ", "ĠâĪĪ", "ĠâĪĿ", "Ġâī", "ĠâīĪ", "Ġãģ¤", "Ġãģ¤ãģ®", "Ġãģ¤ãģ®åĪĨåī²", "Ġãģ¤ãģ®æĸĩåŃĹåĪĹ", "Ġãģ¤ãģ®æĸĩåŃĹåĪĹãĢĤ", "Ġãģ¤ãĤĤ", "Ġãģ§", "Ġãģ§ãĢģ", "Ġãģ§ãģ¯", "Ġãģ§ãģĤãĤĭ", "Ġãģ§è¿½åĬł", "Ġãģ§è¿½åĬł)", "Ġãģ§è¿½åĬł)ãĢĤ", "Ġãģ¨", "Ġãģ¨ãģ®åĨħç©įãĤĴ", "Ġãģ¨ãģĹãģ¦", "Ġãģ¨ãģĻãĤĭãĢĤ", "Ġãģ¨å®Įåħ¨ãģ«", "Ġãģ¨å®Įåħ¨ãģ«åĲĮä¸Ģãģ®", "Ġãģ¨åĲĮãģĺ", "Ġãģ¨åĲĮãģĺå½¢çĬ¶", "Ġãģ¨åĲĮãģĺå½¢çĬ¶ãģ®", "Ġãģ¨åĲĮãģĺå½¢çĬ¶ãģ®åĩºåĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ", "Ġãģ¨åĲĮãģĺéķ·ãģķ", "Ġãģ¨åĲĮãģĺéķ·ãģķ)ãĢĤ", "Ġãģ¨åĲĮä¸Ģãģ®æĮĻåĭķ", "Ġãģ¨åĲĮä¸Ģãģ®æĮĻåĭķ)ãĢĤ", "Ġãģª", "Ġãģªãģ©", "Ġãģªãģ©)ãĢĤ", "ĠãģªãĤī", "ĠãģªãĤīæĸ°è¦ıä½ľæĪĲãģĻãĤĭãĢĤ", "Ġãģ«", "Ġãģ«ãģªãĤĭ", "Ġãģ«ãģªãĤĭãĢĤ", "Ġãģ«ãģĻãĤĭãĢĤ", "Ġãģ«ãģĿãģ®ãģ¾ãģ¾æ¸¡ãģĻ", "Ġãģ«ãĤĪãĤĭ", "Ġãģ«ä½¿ãģĨ", "Ġãģ«å¯¾ãģĹãģ¦", "Ġãģ«å¯¾å¿ľãģĻãĤĭ", "Ġãģ«åŁºãģ¥ãģı", "Ġãģ«ç½®ãģįæıĽãģĪ", "Ġãģ«éģĶ", "Ġãģ®", "Ġãģ®ãģ¨ãģį", "Ġãģ®ãģ¨ãģįãģ¯", "Ġãģ®ãģ¿", "Ġãģ®ãĤµãĥĸãĤ¯ãĥ©ãĤ¹", "Ġãģ®ãĤµãĥĸãĤ¯ãĥ©ãĤ¹ãĢģä¾ĭ:", "Ġãģ®ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ", "Ġãģ®ãĤ¿ãĥĹ", "Ġãģ®ãĤ¿ãĥĹãĥ«ãĢĤ", "Ġãģ®ãĤ¿ãĥĹãĥ«ãĢĤãģĦãģļãĤĮãĤĤ", "Ġãģ®ãĥªãĤ¹ãĥĪ", "Ġãģ®ãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ", "Ġãģ®ãĥĺãĥĥãĥīæķ°", "Ġãģ®ãĥŀãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ", "Ġãģ®ä¸Ńéĸĵå±¤", "Ġãģ®ä¸Ńéĸĵå±¤æ¬¡åħĥ", "Ġãģ®ä¸Ńéĸĵå±¤æ¬¡åħĥãĢĤ", "Ġãģ®å®ļç¾©", "Ġãģ®å®Łè£ħæĸ¹éĩĿ", "Ġãģ®å®Łé¨ĵ", "Ġãģ®å®Łé¨ĵ5", "Ġãģ®å¯¾å¿ľ", "Ġãģ®åħ¥åĬĽãĢĤ", "Ġãģ®åĨħå®¹", "Ġãģ®åĩºåĬĽ", "Ġãģ®åĩºåĬĽ)ãĢĤ", "Ġãģ®åĪĿæľŁåĢ¤", "Ġãģ®åł´åĲĪãĢģ", "Ġãģ®åł´åĲĪãģ¯", "Ġãģ®æ¸¬å®ļåĢ¤ãģ®ç³»åĪĹ", "Ġãģ®æĹ¢å®ļ", "Ġãģ®ç¯ĦåĽ²", "Ġãģ®ç³»åĪĹ", "Ġãģ®çµĲæŀľ", "Ġãģ®çµĲæŀľãĢĤ", "Ġãģ®çµĲæŀľãĢĤL", "Ġãģ®çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³", "Ġãģ®çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³åıĤçħ§)ãĢĤ", "Ġãģ®è¨Ģèªŀ", "Ġãģ®è¾ŀæĽ¸ãĢĤ", "Ġãģ®éķ·ãģķ", "Ġãģ®éłĨä¼ĿæĴŃãĢĤ", "Ġãģ¯", "Ġãģ¯å½¢çĬ¶", "Ġãģ¸", "Ġãģ¾", "Ġãģ¾ãģ§ãģ®éĢ£çķª", "Ġãģ¾ãģ§ãģ®éĢ£çķªãģ¨ãģĹãģ¦æī±", "Ġãģ¾ãģ§ãģ®éĢ£çķªãģ¨ãģĹãģ¦æī±ãģĨãĢĤ", "Ġãģ¾ãģŁãģ¯", "ĠãģĭãĤī", "ĠãģĭãĤīãģ®åıĸå¾Ĺ", "ĠãģĭãĤīåĨįåĪ©çĶ¨ãģĻãĤĭ", "ĠãģĮ", "ĠãģĮãĥĳãĥĩãĤ£ãĥ³ãĤ°", "ĠãģĮå®ŁãĥĪãĥ¼ãĤ¯ãĥ³", "Ġãģĵãģ®", "ĠãģĶãģ¨ãģ«", "ĠãĤ«ãĥ©", "ĠãĤµãĥ³ãĥĹãĥªãĥ³ãĤ°", "ĠãĤ¹ãĤ±ãĤ¸ãĥ¥ãĥ¼ãĥ«", "ĠãĤ¹ãĤ³ãĤ¢", "ĠãĤ¹ãĤ³ãĤ¢ãģ«", "ĠãĤ¹ãĥĨãĥĥãĥĹ", "ĠãĤ¹ãĥĨãĥĥãĥĹãģĶãģ¨", "ĠãĤ¹ãĥĨãĥĥãĥĹãģĶãģ¨ãģ®åĭ¾éħįãĥİãĥ«ãĥł", "ĠãĤ¹ãĥĨãĥĥãĥĹçķªåı·", "ĠãĤĪãĤĬ", "ĠãĤĴ", "ĠãĤĴãĢģ", "ĠãĤĴä½¿ãģĨ", "ĠãĤĴåĬłç®Ĺ", "ĠãĤĴåıĤçħ§)ãĢĤ", "ĠãĤĴæĮģãģ¤", "ĠãĤĴæĮģãģ¤)ãĢĤ", "ĠãĤĴè¨Īç®ĹãģĻãĤĭãĢĤ", "ĠãĤĴè¶ħãģĪãģ¦ãģĦãĤĭ", "ĠãĤĴè¶ħãģĪãĤĭ", "ĠãĤĴè¿ĶãģĻ", "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥", "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥ãĤĴ", "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥åħĪãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯", "Ġãĥ", "Ġãĥ¢", "Ġãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ", "Ġãĥ¢ãĥĩãĥ«", "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮ", "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮæ¬¡åħĥ", "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮæ¬¡åħĥãĢĤ", "Ġãĥ©", "Ġãĥ©ãĤ¦ãĥ³ãĥīä¿®æŃ£", "Ġãĥ©ãĤ¦ãĥ³ãĥīä¿®æŃ£ãĢģ", "ĠãĥªãĥĿãĤ¸ãĥĪãĥª", "Ġãĥģ", "Ġãĥģãĥ£ãĥ³ãĤ¯", "ĠãĥĨãĥ³ãĤ½ãĥ«", "Ġãĥĩ", "ĠãĥĪãĥ¼ãĤ¯ãĥ³", "ĠãĥĪãĥ¼ãĤ¯ãĥ³ãģĤãģŁãĤĬ", "ĠãĥĪãĥ¼ãĤ¯ãĥ³ãģĤãģŁãĤĬãģ®è¨Īç®Ĺéĩı", "ĠãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶", "Ġãĥİ", "ĠãĥĲ", "ĠãĥĲãĤ¤ãĥĪ", "ĠãĥĲãĤ¤ãĥĪæķ°", "ĠãĥĲãĤ¤ãĥĪéķ·", "ĠãĥĲãĤ¤ãĥĪéķ·ãģ®ä¸ĬéĻĲ", "ĠãĥĲãĤ±ãĥĥãĥĪ", "ĠãĥĲãĥĥãĥģ", "ĠãĥĲãĥĥãĥģãĤµãĤ¤ãĤº", "Ġãĥĳ", "Ġãĥĳãĥ©ãĥ¡", "Ġãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿", "Ġãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ", "Ġãĥķ", "ĠãĥķãĤ", "ĠãĥĹ", "ĠãĥĹãĥŃãĥķãĤ¡ãĤ¤ãĥ«", "Ġãĥĺãĥĥãĥī", "ĠãĥĺãĥĥãĥīãģĤãģŁãĤĬãģ®", "Ġãĥĺãĥĥãĥīæķ°", "ĠãĥĻ", "ĠãĥĻãĥ¼ãĤ¹ãģ®åıįå¾©çļĦãģªç¸®å°ı", "Ġãĥŀ", "Ġä¸»å¼µ", "Ġä¸Ńéĸĵå±¤", "Ġä»¥", "Ġä»¥ä¸Ĭ", "Ġä»¥ä¸Ĭ)ãĢĤ", "Ġä»¥ä¸ĭ", "Ġä»¥ä¸ĭ)ãĢĤ", "Ġä»¥éĻį", "Ġä»»æĦıå½¢çĬ¶ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ", "Ġä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:", "Ġä¿Ĥæķ°", "Ġå", "Ġå¯¾", "Ġå¯¾å¿ľãģĻãĤĭæĲįå¤±", "Ġå±", "Ġå±¤", "Ġå±¤çĽ®", "Ġå±ŀæĢ§", "Ġå¼ķæķ°", "Ġå½¢å¼ı", "Ġå½¢çĬ¶", "ĠåĢ¤", "ĠåĢĭ", "ĠåĢĭãģ®", "Ġåģ´", "Ġåģ´)", "Ġåģ´ãĢģn", "Ġåģ´ãĢģn:", "Ġåģ´ãģ®", "Ġåģ´ãģ®ç³»åĪĹéķ·", "Ġåħ¥åĩºåĬĽ", "ĠåĨħ", "ĠåĪĨåī²", "ĠåĪĨåī²å¯¾è±¡ãģ®", "ĠåĪĨåī²çµĲæŀľ", "Ġåĭ¾éħį", "Ġåįĺ", "Ġåı", "ĠåıĤåĬł", "ĠåıĤçħ§", "ĠåıĤçħ§)ãĢĤ", "ĠåıĤçħ§ãĢģ006", "Ġåıĸå¾Ĺ", "ĠåĲĦ", "ĠåĽ", "ĠåĽ³", "ĠåĽ³ãģ®ãĤ¿ãĤ¤ãĥĪãĥ«ãĢĤ", "ĠåĽŀ", "ĠåŃ¦ç¿Ĵ", "ĠåŃ¦ç¿Ĵçİĩ", "Ġæ", "Ġæ¤ľè¨¼", "Ġæ§ĭæĪĲ", "Ġæ¨ª", "Ġæ¨ªè»¸ãĥ©ãĥĻãĥ«ãĢĤ", "Ġæ¬¡", "Ġæ¬¡ãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ", "Ġæ¬¡åħĥ", "Ġæ¬¡åħĥãģ®", "Ġæ³ķ", "ĠæııçĶ»åħĪãģ®", "ĠæĲįå¤±", "ĠæĲįå¤±ãģ®", "Ġæķ°åĢ¤å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°", "Ġæķ´æķ°", "Ġæĸ¹å¼ı", "Ġæĸ¹åĲĳ", "ĠæĸĩåŃĹ", "ĠæĸĩåŃĹãĤĴ", "ĠæĻĤçĤ¹", "ĠæľªæºĢ", "ĠæľªæºĢãģ§", "ĠæľĢå¤§", "ĠæŃ£", "ĠæŃ£ãģ®å®Łæķ°åĢ¤ãģ®ç³»åĪĹ", "ĠæŃ£åīĩåĮĸ", "ĠæŃ£è¦ıåĮĸ", "ĠæŃ£è¦ıåĮĸãĤĴè¡ĮãģĨçī¹å¾´", "Ġç", "Ġç¬¦åı·åĮĸ", "Ġç¬¬", "Ġç¯Ģ", "Ġç¯Ģ)ãĢĤ", "Ġç³»åĪĹ", "Ġç³»åĪĹéķ·", "Ġç³»åĪĹéķ·(sequence", "ĠçµĤäºĨ", "Ġç·ļå½¢", "Ġç¸¦", "Ġç¸¦è»¸ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ", "Ġç¸¦è»¸ãĥ©ãĥĻãĥ«ãĢĤ", "Ġç¾¤", "Ġç¾¤ãģĶãģ¨ãģ®", "Ġçİĩ", "ĠçİĩãĢĤ", "ĠçĶŁæĪĲ", "ĠçĶŁæĪĲãģĻãĤĭ", "Ġçķ", "ĠçķªçĽ®", "ĠçĽ", "ĠçĽ®æ¨Ļ", "ĠçĽ¸", "ĠçŃī", "Ġè¡Į", "Ġè¡ĮåĪĹ", "Ġè¨Ģèªŀãĥ¢ãĥĩãĥ«ãģ®", "Ġè¨ĵç·´", "Ġè¨ĺäºĭ", "Ġèªŀå½Ļ", "Ġèªŀå½ĻãĤµãĤ¤ãĤº", "Ġèªŀå½ĻãĤµãĤ¤ãĤº(vocabulary", "Ġè»¸", "Ġèĩª", "Ġèĩªå·±æ³¨æĦı", "Ġèĩªèº«ãģ®", "Ġèĩªèº«ãģ®åĪĨåī²çµĲæŀľ", "Ġéĩįãģ¿", "Ġéĩįãģ¿ãĥ»åĲĦãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ", "Ġéĩįãģ¿æ¸Ľè¡°", "Ġéĩįãģ¿æ¸Ľè¡°ä¿Ĥæķ°", "ĠéĿŀåŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°", "ĠĠ", "ĠĠ#", "ĠĠ:", "ĠĠĠ:", "ĠĠĠĠ", "ĠĠĠĠ:", "ĠĠĠĠĠ:", "ĠĠĠĠĠĠĠ:", "ĠĠĠĠĠĠĠĠ", "ĠĠĠĠĠĠĠĠĠ:", "ĠĠĠĠĠĠĠĠĠĠĠ:", "ĠĠĠĠĠĠĠĠĠĠĠĠĠ:", "ġ", "Ģ", "Ģ¤", "Ģ§", "ĢæīĢ", "Ģĭ", "ĢĻ", "ĢĻè£ľ", "ģ", "ģ´", "Ĥ", "Ĥ¹", "ĤåĬł", "Ĥç", "Ĥç³»åĪĹ", "ĤçĤ¹", "ĥ", "ĥ³å®ļ", "ĥçķĮ", "ĥħ", "ĥħåł±", "Ħ", "Ħå½±", "Ħı", "ħ", "ħ§", "Ĩ", "Ĩħ", "ĩ", "ĩãĤ£ãĥ³ãĤ°", "Ī", "ī", "Ĭ", "ĭ", "ĭ¾éħį", "ĭãĤģ", "Į", "į", "įå¤±", "įĺ", "İ", "İåĽŀ", "İç¤º", "İĩ", "ı", "ı¾", "Ĳ", "ĲĨ", "Ĳįå¤±", "ĳ", "ĳ¼ãģ³åĩºãģĹ", "Ĵ", "ĴĮ", "ĵ", "Ķ", "ķ", "ķæķ°", "ĸ", "ĸĦ", "Ĺ", "ĺ", "ĺãĥĥãĥī", "ĺİç¤º", "Ļ", "Ļæºĸ", "ĻĤçĤ¹", "ļ", "ļĦ", "Ľ", "ľ", "ľãĥ«", "Ŀ", "ĿåŃĺ", "ŀ", "ŀæĢ§", "Ł", "ł", "ł´", "ł´åĲĪ", "Ń", "Ńī" ], "byte_level": true, "chunk_split_mode": "whitespace", "max_chunk_bytes": 64, "symbol_to_id": { "!": 0, "!=": 1, "!r": 2, "!r}": 3, "\"": 4, "\"\"": 5, "\"\"\"": 6, "\"\".join(": 7, "\"\".join(current": 8, "\"\".join(current))": 9, "\")": 10, "\"))": 11, "\"),": 12, "\")]": 13, "\",": 14, "\".join(": 15, "\":": 16, "\"]": 17, "\"],": 18, "\"].append(": 19, "\"].append(step)": 20, "\"].b": 21, "\"].by_": 22, "\"].by_key": 23, "\"].by_key()": 24, "\"].by_key()[\"": 25, "\"].by_key()[\"color": 26, "\"].by_key()[\"color\"]": 27, "\"``": 28, "\"``:": 29, "\"``ãĥ»``": 30, "\"``ãĥ»``\"en": 31, "\"``ãĥ»``\"en\"``": 32, "\"axes": 33, "\"axes.": 34, "\"axes.prop": 35, "\"axes.prop_cycle": 36, "\"axes.prop_cycle\"].by_key()[\"color\"]": 37, "\"en": 38, "#": 39, "$": 40, "%": 41, "&": 42, "&oldid": 43, "&oldid=": 44, "&oldid=<": 45, "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³": 46, "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³ID": 47, "&oldid=<ãĥªãĥĵãĤ¸ãĥ§ãĥ³ID>``": 48, "&prop": 49, "&prop=": 50, "&prop=extra": 51, "&prop=extract": 52, "&prop=extracts``": 53, "'": 54, "''": 55, "(": 56, "(\"": 57, "(\".": 58, "(\"\\n": 59, "(\"en": 60, "(\"en\",": 61, "(\"ja": 62, "(\"ja\",": 63, "(\"nan": 64, "(\"nan\"),": 65, "()": 66, "())": 67, "(),": 68, "().": 69, "().__init__": 70, "().__init__()": 71, "().item": 72, "().item()": 73, "().item())": 74, "()``": 75, "(...,": 76, "(0": 77, "(0)}": 78, "(00": 79, "(001": 80, "(001ãĥ»002": 81, "(001ãĥ»002ãĥ»00": 82, "(001ãĥ»002ãĥ»003": 83, "(002": 84, "(003": 85, "(004": 86, "(005": 87, "(006": 88, "(006)ãĢĤ": 89, "(006ãĢģ": 90, "(007": 91, "(009": 92, "(1": 93, "(1)": 94, "(1-indexed": 95, "(1-indexed)ãĢĤ": 96, "(1.0": 97, "(2": 98, "(2)": 99, "(2i": 100, "(429": 101, "(5": 102, "(=": 103, "(?": 104, "(A": 105, "(Attention": 106, "(AttentionScoreBias": 107, "(AttentionScoreBias):": 108, "(B,": 109, "(Byte": 110, "(C": 111, "(CD": 112, "(CDF": 113, "(D": 114, "(De": 115, "(Encoder": 116, "(G": 117, "(GN": 118, "(GNM": 119, "(GNMT": 120, "(GNMT)ãĤ¹ãĤ¿ãĤ¤ãĥ«": 121, "(GNMT)ãĤ¹ãĤ¿ãĤ¤ãĥ«ãģ®éķ·ãģķ": 122, "(Gated": 123, "(Gaussian": 124, "(Key": 125, "(L": 126, "(L2": 127, "(Layer": 128, "(LayerNorm": 129, "(LayerNorm(x": 130, "(LayerNorm(x))": 131, "(LayerNorm(x),": 132, "(MediaWiki": 133, "(N": 134, "(N,": 135, "(N,D": 136, "(N,D)": 137, "(Out": 138, "(Out-": 139, "(Out-of": 140, "(Out-of-": 141, "(Out-of-V": 142, "(Out-of-Vocab": 143, "(Out-of-Vocabulary": 144, "(Q,": 145, "(Query": 146, "(R": 147, "(Radford": 148, "(Res": 149, "(Residual": 150, "(S": 151, "(Scaling": 152, "(T": 153, "(True": 154, "(V": 155, "(V)": 156, "(Vaswani": 157, "(W": 158, "(Wikipedia": 159, "(Wikipedia)": 160, "(Wikipedia)ãĢ": 161, "(Wikipedia)ãĢı": 162, "(Y": 163, "(Y)": 164, "([": 165, "([],": 166, "(`": 167, "(``": 168, "(``\"": 169, "(``\"ja\"``ãĥ»``\"en\"``": 170, "(``1": 171, "(``DecoderBlock``": 172, "(``GPTLanguageModel``": 173, "(``None``": 174, "(``QueryKeyPositionalTransform``": 175, "(``_le": 176, "(``_levenberg_m": 177, "(``_levenberg_marquardt": 178, "(``compute_": 179, "(``encod": 180, "(``for": 181, "(``forward(": 182, "(``forward(token_ids)": 183, "(``learn_bpe": 184, "(``list": 185, "(``list[int]": 186, "(``list[int]``": 187, "(``lm_head``)ãģ®": 188, "(``make_evaluation_windows``": 189, "(``n": 190, "(``nn.Parameter": 191, "(``nn.Parameter``": 192, "(``numpy": 193, "(``numpy.": 194, "(``numpy.linalg.lstsq": 195, "(``numpy.linalg.lstsq``)": 196, "(``pretokenize``": 197, "(``src/": 198, "(``src/data/": 199, "(``src/data/text": 200, "(``src/training/": 201, "(``train_language_model``": 202, "(a": 203, "(a)": 204, "(alpha": 205, "(alpha+": 206, "(alpha+beta": 207, "(alpha=0.3": 208, "(b": 209, "(batch": 210, "(batch_size,": 211, "(beam": 212, "(beam_sequences": 213, "(beam_sequences,": 214, "(best": 215, "(best_params)": 216, "(best_params),": 217, "(bpe": 218, "(bpe,": 219, "(bpe.": 220, "(c": 221, "(cache_dir:": 222, "(chunk": 223, "(chunk)": 224, "(cl": 225, "(clip": 226, "(con": 227, "(context)": 228, "(cross": 229, "(cross_": 230, "(d_": 231, "(d_ff,": 232, "(d_model": 233, "(d_model)": 234, "(d_model))": 235, "(d_model,": 236, "(data[\"": 237, "(de": 238, "(dead": 239, "(design": 240, "(design,": 241, "(dim": 242, "(dropout)": 243, "(exponent": 244, "(g": 245, "(g,": 246, "(gre": 247, "(greed": 248, "(greedy": 249, "(hidden": 250, "(hidden)": 251, "(htt": 252, "(i": 253, "(id": 254, "(ids,": 255, "(ids[i": 256, "(iter": 257, "(key": 258, "(l_inf": 259, "(len(": 260, "(length": 261, "(lengths)),": 262, "(lengths_without_leading_whitespace": 263, "(lengths_without_leading_whitespace)),": 264, "(log": 265, "(log_n": 266, "(log_n,": 267, "(logit": 268, "(logits": 269, "(logits,": 270, "(logits:": 271, "(m": 272, "(m:": 273, "(m_hat": 274, "(matrix": 275, "(memory": 276, "(model.parameters()": 277, "(model.parameters(),": 278, "(module": 279, "(n": 280, "(n,": 281, "(n-": 282, "(n\\\\text{-grams": 283, "(n\\\\text{-grams})": 284, "(n\\\\text{-grams})|}{|": 285, "(n\\\\text{-grams})|}{|n": 286, "(nats)ãĢĤ": 287, "(new": 288, "(new_symbol": 289, "(new_symbol)": 290, "(next": 291, "(next(iter": 292, "(ngrams": 293, "(ngrams)": 294, "(ngrams))": 295, "(nn.": 296, "(nn.Module": 297, "(nn.Module):": 298, "(nn.Module,": 299, "(noisy_l": 300, "(non-": 301, "(norm": 302, "(normed,": 303, "(np.": 304, "(np.arange(": 305, "(np.as": 306, "(np.asarray(": 307, "(np.exp(": 308, "(np.exp(log": 309, "(num": 310, "(num_heads)": 311, "(num_window": 312, "(num_windows,": 313, "(p": 314, "(p)": 315, "(p)}": 316, "(param": 317, "(params)": 318, "(params,": 319, "(parent": 320, "(parents": 321, "(parents=True,": 322, "(path": 323, "(path)": 324, "(path).read_text(encoding=\"utf-8\"))": 325, "(path:": 326, "(pos,": 327, "(positions.": 328, "(positions.max": 329, "(positions.max().item())": 330, "(pre-": 331, "(pre-token": 332, "(pre-tokenization": 333, "(query": 334, "(r\"": 335, "(re-": 336, "(re-centering": 337, "(re-scaling": 338, "(repo_id=repo_id,": 339, "(residual_fn": 340, "(residual_fn,": 341, "(retry_wait_second": 342, "(retry_wait_seconds": 343, "(s)": 344, "(se": 345, "(seed)": 346, "(self)": 347, "(self,": 348, "(self.": 349, "(self.beta": 350, "(self.dropout": 351, "(self.f": 352, "(self.feed_forwar": 353, "(self.feed_forward": 354, "(self.feed_forward(": 355, "(self.feed_forward(self.": 356, "(self.feed_forward(self.norm": 357, "(self.feed_forward(x)))": 358, "(seq_len,": 359, "(sequence": 360, "(set": 361, "(symbol": 362, "(text": 363, "(text:": 364, "(title": 365, "(token": 366, "(token_ids,": 367, "(torch.": 368, "(values_by_condition": 369, "(vocabulary": 370, "(weight": 371, "(weights)": 372, "(x": 373, "(x)": 374, "(x))": 375, "(x)))": 376, "(x,": 377, "(x:": 378, "(x_arr": 379, "(x_n": 380, "(xlabel)": 381, "({": 382, "(ãģ¾ãģŁãģ¯": 383, "(ãģĦãģļãĤĮãĤĤ": 384, "(ãĤªãĥ¼ãĥĲãĥ¼ãĥķãĥŃãĥ¼": 385, "(ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°": 386, "(ãĤ¹": 387, "(ãĤ¹ãĥĨãĥĥãĥĹ": 388, "(ãĤ¹ãĥļ": 389, "(ãĤ¹ãĥļãĥ¼ãĤ¹": 390, "(ãĤ¹ãĥļãĥ¼ãĤ¹ãĥ»": 391, "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶": 392, "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®": 393, "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²": 394, "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²Ĵ": 395, "(ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãģ®ç²Ĵåº¦": 396, "(ãĥĲãĥĥãĥģ": 397, "(äºĭåīįåĪĨåī²": 398, "(äºĭåīįåĪĨåī²ãģ§å¾ĹãģŁ": 399, "(ä¾ĭ:": 400, "(å¤ļ": 401, "(å®Łé¨ĵ": 402, "(å¯¾æķ°": 403, "(å¯¾æķ°å°¤åº¦": 404, "(å¯¾æķ°ç©ºéĸĵ": 405, "(å¹³åĿĩ": 406, "(å¹³åĿĩæ¸Ľç®Ĺ": 407, "(å½¢çĬ¶": 408, "(åĢĻè£ľèªŀå½ĻãģĭãĤīãģ®": 409, "(åħ¨ãĥĺãĥĥãĥī": 410, "(åħ¨ãĥĺãĥĥãĥīåĪĨãĤĴ": 411, "(åĭ¾éħį": 412, "(åįĺ": 413, "(åŃ¦ç¿Ĵ": 414, "(æ¬¡": 415, "(æ¯Ķè¼ĥå¯¾è±¡": 416, "(æµ®åĭķå°ıæķ°çĤ¹": 417, "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®Ĺ": 418, "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®ĹåĽŀæķ°": 419, "(æµ®åĭķå°ıæķ°çĤ¹æ¼Ķç®ĹåĽŀæķ°)": 420, "(æĸĩ": 421, "(æĹ¢å®ļåĢ¤": 422, "(æĹ¢å®ļåĢ¤)": 423, "(æĹ¢å®ļåĢ¤)ãģ®": 424, "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪ": 425, "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãĢģ": 426, "(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯": 427, "(æĹ¢å®ļåĢ¤ãĢģ": 428, "(æľĢ": 429, "(ç©ºçĻ½ãģ®": 430, "(çĽ´åīį": 431, "(çĽ´åīįãĤ¹ãĥĨãĥĥãĥĹãģ¨ãģ®": 432, "(çľģçķ¥åı¯": 433, "(çľģçķ¥åı¯)ãĢĤ": 434, "(è©³ç´°ãģ¯": 435, "(èªŀå½Ļ": 436, "(èªŀå½Ļãģ¸": 437, "(èªŀå½Ļãģ¸ãģ®å°Ħå½±": 438, "(èªŀå½ĻãĤµãĤ¤ãĤº": 439, "(éĩįãģ¿": 440, ")": 441, ")\"": 442, ")\")": 443, ")\",": 444, "))": 445, ")))": 446, ")),": 447, ")):": 448, "),": 449, ").": 450, "):": 451, ")]": 452, ")],": 453, ")^": 454, ")^2": 455, ")^T": 456, ")^alpha": 457, ")``": 458, ")``)ãĢĤ": 459, ")``ãĢģ": 460, ")``ãĢĤ": 461, ")}": 462, ")ãĢģ": 463, ")ãĢģFalse": 464, ")ãĢģæİ¥ç¶ļ": 465, ")ãĢģæİ¥ç¶ļç¢ºç«ĭ": 466, ")ãĢĤ": 467, ")ãĢĤ\"\"\"": 468, ")ãĢĤ**": 469, ")ãĢĤFalse": 470, ")ãĢĤ``": 471, ")ãĢĤ``None``": 472, ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)": 473, ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®": 474, ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯": 475, ")ãĢĤ``None``(æĹ¢å®ļåĢ¤)ãģ®åł´åĲĪãģ¯å¾ĵæĿ¥éĢļãĤĬ": 476, ")ãģ§": 477, ")ãģ§ãĢģ": 478, ")ãģ§ãģ¯": 479, ")ãģ¨": 480, ")ãģ¨ãģĹãģ¦": 481, ")ãģªãģ©": 482, ")ãģ«": 483, ")ãģ«ãģª": 484, ")ãģ«ãĤĪãĤĬ": 485, ")ãģ«ä¾ĿåŃĺ": 486, ")ãģ«å¯¾ãģĹãģ¦": 487, ")ãģ®": 488, ")ãģ®ãģ¿": 489, ")ãģ®ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãĢĤ": 490, ")ãģ®ãĤĪãģĨãģ«": 491, ")ãģ®ãĤĪãģĨãģ«ãĢģ": 492, ")ãģ®ãĥ©ãĥĻãĥ«ãĢĤ": 493, ")ãģ®åħ±éĢļ": 494, ")ãģ®åħ±éĢļãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹": 495, ")ãģ®åħ±éĢļãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹ãĢĤ": 496, ")ãģ®åł´åĲĪãĢģ": 497, ")ãģ¯": 498, ")ãģ¯ãĢģ": 499, ")ãģ¸": 500, ")ãģĭãĤī": 501, ")ãģĮ": 502, ")ãģĵãģ¨ãģĮæľŁå¾ħãģķãĤĮãĤĭãĢĤ": 503, ")ãģĵãģ¨ãĤĴæĦıåĳ³": 504, ")ãģĵãģ¨ãĤĴæĦıåĳ³ãģĻãĤĭãĢĤ": 505, ")ãģŁãĤģãĢģ": 506, ")ãĤ¹ãĤ¿ãĤ¤ãĥ«": 507, ")ãĤĦ": 508, ")ãĤĴ": 509, ")ãĤĴä½¿ãģĨãĢĤ": 510, ")ãĤĴä½ľ": 511, ")ãĤĴå": 512, ")ãĤĴå¤±": 513, ")ãĤĴæĮģãģ¤": 514, ")ãĤĴè¨Īç®Ĺ": 515, ")ãĤĴè¨Īç®ĹãģĻãĤĭãĢĤ": 516, ")ãĤĴéģ¿": 517, ")ãĥ»": 518, ")åł´åĲĪãģ¯": 519, "*": 520, "**": 521, "**(": 522, "**2": 523, "**8": 524, "**:": 525, "**alpha": 526, "**beta": 527, "**t": 528, "*?": 529, "+": 530, "+\\s*": 531, "+åĩºåĬĽå±¤": 532, "+åĩºåĬĽå±¤ãĢįãģ®": 533, ",": 534, ",)": 535, "-": 536, "-\",": 537, "--": 538, "---": 539, "----": 540, "--------": 541, "------------": 542, "---------------": 543, "----------------": 544, "-------------------------------": 545, "--------------------------------": 546, "---------------------------------------------------------------": 547, "-1": 548, "-1,": 549, "-1]": 550, "-1}": 551, "-1})": 552, "-2": 553, "-2)": 554, "-3": 555, "-8": 556, "-8h": 557, "-8h/": 558, "-8h/H": 559, "-8h/H)": 560, "-A": 561, "-Ag": 562, "-Agent": 563, "-Decoder": 564, "-Forward": 565, "-Head": 566, "-Layer": 567, "-M": 568, "-Marquardt": 569, "-N": 570, "-P": 571, "-Product": 572, "-S": 573, "-SA": 574, "-attention": 575, "-attention)": 576, "-byt": 577, "-gram": 578, "-grams": 579, "-in": 580, "-index": 581, "-indexed": 582, "-inf": 583, "-inf\"": 584, "-inf``": 585, "-level": 586, "-max": 587, "-n": 588, "-n_val": 589, "-n_val:": 590, "-n_val:]": 591, "-n_val],": 592, "-negative": 593, "-onl": 594, "-only": 595, "-per": 596, "-per-byt": 597, "-rn": 598, "-rnn": 599, "-start": 600, "-theories": 601, "-w": 602, "-wise": 603, "-x": 604, "-x2,": 605, "-}": 606, "-}n": 607, "-ç¶": 608, "-ç¶Ļ": 609, "-ç¶Ļæī": 610, "-ç¶Ļæī¿": 611, ".": 612, ".*?": 613, ".,": 614, "..": 615, "...": 616, "...)``": 617, "...,": 618, "...],": 619, "...]]": 620, ".0": 621, ".0,": 622, ".2": 623, ".2))": 624, ".3": 625, ".4": 626, ".5": 627, ".5))": 628, ".5,": 629, ".6": 630, ".D": 631, ".SentencePiece": 632, ".T": 633, ".ad": 634, ".add": 635, ".add_": 636, ".append(": 637, ".append(float": 638, ".ax": 639, ".b": 640, ".bias": 641, ".co": 642, ".cross_": 643, ".decod": 644, ".default": 645, ".detach().": 646, ".device)": 647, ".dim": 648, ".dim()": 649, ".dtype,": 650, ".encod": 651, ".encode(\"utf-8": 652, ".encode(\"utf-8\"))": 653, ".exponent": 654, ".fig": 655, ".forward``": 656, ".g": 657, ".get": 658, ".grad": 659, ".ipyn": 660, ".items()": 661, ".items()):": 662, ".items():": 663, ".items()}": 664, ".jo": 665, ".join(": 666, ".json": 667, ".json\"": 668, ".key": 669, ".keys()": 670, ".l": 671, ".law": 672, ".log(": 673, ".m": 674, ".max": 675, ".mean": 676, ".mean(dim": 677, ".mean(dim=-1,": 678, ".min": 679, ".mk": 680, ".mkdir": 681, ".mkdir(parents=True,": 682, ".mul_": 683, ".mul_(self.beta": 684, ".n": 685, ".ndim": 686, ".norm": 687, ".normal": 688, ".org/": 689, ".p": 690, ".plot": 691, ".plot(": 692, ".pow(2)": 693, ".power_law_fit": 694, ".py": 695, ".py``": 696, ".rc": 697, ".rcParam": 698, ".rcParams[": 699, ".rcParams[\"axes.prop_cycle\"].by_key()[\"color\"]": 700, ".re": 701, ".read_text(encoding=\"utf-8": 702, ".read_text(encoding=\"utf-8\"))": 703, ".replace": 704, ".replace(\"": 705, ".request": 706, ".request.": 707, ".s": 708, ".sc": 709, ".scal": 710, ".scatter": 711, ".set": 712, ".set_": 713, ".set_t": 714, ".set_title": 715, ".set_title(title": 716, ".set_title(title)": 717, ".shape": 718, ".size(": 719, ".size(-1": 720, ".size(1)": 721, ".sle": 722, ".sleep": 723, ".sqrt(": 724, ".sub": 725, ".subplot": 726, ".subplots(": 727, ".t": 728, ".to(": 729, ".to(dtype=": 730, ".to(dtype=dtype)": 731, ".trans": 732, ".transpo": 733, ".transpose": 734, ".transpose(": 735, ".tx": 736, ".txt": 737, ".txt\"": 738, ".values": 739, ".view": 740, ".vocab``": 741, ".weight": 742, ".wikipedia": 743, ".wikipedia.org/": 744, "/": 745, "//": 746, "/D": 747, "/D^": 748, "/D^beta": 749, "/N": 750, "/N^": 751, "/N^alpha": 752, "/sqrt(": 753, "0": 754, "0)": 755, "0,": 756, "0.": 757, "0.0": 758, "0.0,": 759, "00": 760, "001": 761, "002": 762, "003": 763, "004": 764, "004_norm": 765, "005": 766, "006": 767, "006ãĥ»00": 768, "006ãĥ»008": 769, "007": 770, "009": 771, "1": 772, "1(": 773, "1)": 774, "1)ãĢĤ": 775, "1,": 776, "1.": 777, "1.0": 778, "12,": 779, "16": 780, "1:": 781, "1_f": 782, "1_found": 783, "1_foundation": 784, "1_foundations/": 785, "2": 786, "2)": 787, "2)ãĢĤ": 788, "2,": 789, "2.0": 790, "20": 791, "25": 792, "29": 793, "2]": 794, "3": 795, "3)": 796, "3,": 797, "4": 798, "4,": 799, "429": 800, "44": 801, "5": 802, "5,": 803, "56": 804, "6": 805, "6,": 806, "7": 807, "7,": 808, "71": 809, "8": 810, "8)": 811, "8,": 812, "9": 813, "90": 814, "99": 815, "999": 816, ":": 817, ":2]": 818, ":math:`": 819, ";": 820, "<": 821, "": 823, "": 825, "=": 826, "=\"": 827, "=\"ident": 828, "=\"identity": 829, "=\"ign": 830, "=\"ignor": 831, "=\"ignore\"):": 832, "=\"ignore\",": 833, "=\"utf-8": 834, "=\"y": 835, "=\"y\")": 836, "=-": 837, "=-1": 838, "=-1)": 839, "=-1,": 840, "=0": 841, "=0.": 842, "=0.0": 843, "=1": 844, "=1)": 845, "=1.": 846, "=1.0": 847, "=1e": 848, "=False": 849, "=False)": 850, "=False``": 851, "=None": 852, "=True": 853, "=True)": 854, "=True,": 855, "=True``": 856, "=b": 857, "=bias)": 858, "=color": 859, "=color,": 860, "=data[\"": 861, "=float": 862, "=float(np.exp(log": 863, "=lambda": 864, "=repo_id": 865, "=repo_id,": 866, "=str": 867, "=str(": 868, "={": 869, ">": 870, ">\",": 871, ">]": 872, ">``": 873, "?": 874, "@": 875, "A": 876, "AACL": 877, "AC": 878, "ACE_": 879, "ACL": 880, "AL": 881, "ALP": 882, "ALPH": 883, "ALPHA_": 884, "ALPHA_BE": 885, "ALPHA_BET": 886, "ALPHA_BETA_M": 887, "ALiBi": 888, "ALiBiPositionBias": 889, "AN": 890, "ANI": 891, "ANIF": 892, "ANIFES": 893, "ANIFEST": 894, "ANIFEST_": 895, "ANIFEST_DI": 896, "ANIFEST_DIR": 897, "AX": 898, "AX_": 899, "AX_CHUNK_": 900, "AX_CHUNK_CA": 901, "AX_CHUNK_CACH": 902, "AX_CHUNK_CACHE_": 903, "AX_CHUNK_CACHE_E": 904, "AX_CHUNK_CACHE_EN": 905, "AX_CHUNK_CACHE_ENT": 906, "AX_CHUNK_CACHE_ENTR": 907, "AX_CHUNK_CACHE_ENTRI": 908, "AX_CHUNK_CACHE_ENTRIES": 909, "A_": 910, "A_M": 911, "A_MANIFEST_DIR": 912, "Absolute": 913, "AbsolutePositional": 914, "AbsolutePositionalEmbedding": 915, "Adam": 916, "AdamW": 917, "AdamW``": 918, "AdamWithL2Regularization": 919, "Arg": 920, "Args:": 921, "Attention": 922, "AttentionScoreBias": 923, "B": 924, "B,": 925, "BC": 926, "BE": 927, "BPE": 928, "BPEIDTokenizer": 929, "BPETokenizer": 930, "BY": 931, "BYT": 932, "BYTE": 933, "BYTE_TO_": 934, "BYTE_TO_UNICOD": 935, "BYTE_TO_UNICODE": 936, "Bi": 937, "Bias": 938, "Block": 939, "Byt": 940, "Byte": 941, "C": 942, "CA": 943, "CH": 944, "CHUN": 945, "CHUNK_": 946, "CL": 947, "CLR": 948, "CO": 949, "COD": 950, "CharacterLevelTokenizer": 951, "Chinchilla": 952, "ChinchillaParametricFit": 953, "D": 954, "DI": 955, "DIA_MANIFEST_DIR": 956, "D``": 957, "De": 958, "Decoder": 959, "DecoderBlock": 960, "DecoderBlock``": 961, "Dropout": 962, "E": 963, "ELU": 964, "ES": 965, "ESP": 966, "ESPACE_": 967, "E_": 968, "E_T": 969, "E_TO": 970, "E_TO_": 971, "E_TO_BYTE": 972, "Embedding": 973, "En": 974, "Encoder": 975, "EncoderBlock": 976, "EncoderBlock``": 977, "Encoding": 978, "Error": 979, "Error,": 980, "Error:": 981, "F": 982, "F-8": 983, "FF": 984, "FLO": 985, "FLOP": 986, "Fal": 987, "False": 988, "Feed": 989, "FeedForwardNetwork": 990, "FeedForwardNetwork``": 991, "File": 992, "Fit": 993, "For": 994, "Forward": 995, "ForwardN": 996, "ForwardNetwork": 997, "FrontierResult": 998, "G": 999, "GELU": 1000, "GLU": 1001, "GP": 1002, "GPT": 1003, "GPT-2": 1004, "GPTLanguage": 1005, "GPTLanguageModel": 1006, "GPTLanguageModel``": 1007, "G_ALPHA_BETA_M": 1008, "Generator": 1009, "GridPoint": 1010, "H": 1011, "Head": 1012, "Hoff": 1013, "Hoffm": 1014, "Hoffmann": 1015, "Hol": 1016, "Holt": 1017, "Holtz": 1018, "Holtzman": 1019, "I": 1020, "ICLR": 1021, "ICOD": 1022, "ID": 1023, "IDTokenizer": 1024, "IK": 1025, "IKI": 1026, "IN": 1027, "IN,": 1028, "IP": 1029, "IPS": 1030, "IRLS": 1031, "Imp": 1032, "Imple": 1033, "Implement": 1034, "Implemented": 1035, "ImplementedError": 1036, "IsoFLOP": 1037, "IsoFLOPParabolaFit": 1038, "J": 1039, "K": 1040, "KV": 1041, "K_": 1042, "Kap": 1043, "Kaplan": 1044, "Key": 1045, "KeyPositional": 1046, "KeyPositionalTransform": 1047, "L": 1048, "L2": 1049, "L2R": 1050, "L2Regularization": 1051, "LO": 1052, "LOG_ALPHA_BETA_M": 1053, "LU": 1054, "Language": 1055, "Layer": 1056, "LayerNorm": 1057, "LayerNormalization": 1058, "LayerNormalization``": 1059, "Learn": 1060, "LearnedAbsolutePositionalEmbedding": 1061, "Levenberg-Marquardt": 1062, "Lik": 1063, "Linear": 1064, "Lo": 1065, "Loss": 1066, "Loss\",": 1067, "M": 1068, "MAX_CHUNK_CACHE_ENTRIES": 1069, "MS": 1070, "MediaWiki": 1071, "Mod": 1072, "Model": 1073, "Module": 1074, "Module``": 1075, "Multi": 1076, "Multi-Head": 1077, "MultiHead": 1078, "MultiHeadAttention": 1079, "MultiHeadAttention``": 1080, "N": 1081, "N,": 1082, "None": 1083, "None,": 1084, "None``": 1085, "Norm": 1086, "Not": 1087, "Note:": 1088, "O": 1089, "ON": 1090, "OS": 1091, "Op": 1092, "Optim": 1093, "Optimizer": 1094, "OptimizerLik": 1095, "Optimizer``": 1096, "Out": 1097, "P": 1098, "PE": 1099, "PE(pos,": 1100, "PEDIA_MANIFEST_DIR": 1101, "PI": 1102, "Par": 1103, "Parabola": 1104, "ParabolaFit": 1105, "Param": 1106, "Parameter": 1107, "Parameter(torch.": 1108, "Parametric": 1109, "ParametricFit": 1110, "Path": 1111, "Piece": 1112, "Po": 1113, "Point": 1114, "Position": 1115, "PositionBias": 1116, "PositionEmbedding": 1117, "Positional": 1118, "PositionalEncoding": 1119, "Post": 1120, "Post-Layer": 1121, "Pow": 1122, "PowerL": 1123, "PowerLaw": 1124, "PowerLawFit": 1125, "Pre-": 1126, "Pre-Layer": 1127, "Press": 1128, "Pro": 1129, "Processor": 1130, "Py": 1131, "Q": 1132, "Q,": 1133, "Qu": 1134, "Query": 1135, "QueryKeyPositionalTransform": 1136, "QueryKeyPositionalTransform``": 1137, "Queryãĥ»Key": 1138, "R": 1139, "RE": 1140, "RL": 1141, "RLS": 1142, "RMS": 1143, "RMSNorm": 1144, "ReLU": 1145, "Rel": 1146, "Relative": 1147, "RelativePositionBias": 1148, "Res": 1149, "Result": 1150, "Return": 1151, "Returns:": 1152, "RoPE": 1153, "RotaryPositionEmbedding": 1154, "S": 1155, "S+": 1156, "S,": 1157, "SON": 1158, "S_": 1159, "S_q,": 1160, "Scal": 1161, "Scaled": 1162, "Scaling": 1163, "Score": 1164, "ScoreBias": 1165, "SentencePiece": 1166, "Sequence": 1167, "Shaw": 1168, "ShawRelativePositionBias": 1169, "Split": 1170, "SplitMod": 1171, "SwiGLUFeedForwardNetwork": 1172, "Swish": 1173, "T": 1174, "T5": 1175, "T5RelativePositionBias": 1176, "TF-8": 1177, "Tensor": 1178, "Tensor,": 1179, "Tensor]": 1180, "Tensor],": 1181, "Tokenizer": 1182, "Transform": 1183, "True": 1184, "U": 1185, "UN": 1186, "UNICOD": 1187, "UNICODE_TO_BYTE": 1188, "URL": 1189, "UTF-8": 1190, "Un": 1191, "Unigram": 1192, "UnigramTokenizer": 1193, "V": 1194, "V^{": 1195, "V^{(p)}": 1196, "Value": 1197, "Value\",": 1198, "ValueError:": 1199, "Vas": 1200, "Vasw": 1201, "Vaswan": 1202, "Vaswani": 1203, "W": 1204, "WIKI": 1205, "WIKIPEDIA_MANIFEST_DIR": 1206, "Wik": 1207, "Wiki": 1208, "Wikim": 1209, "Wikimedia": 1210, "Wikipedia": 1211, "X": 1212, "Xiv": 1213, "Xiv:": 1214, "Y": 1215, "Y_": 1216, "You": 1217, "Z": 1218, "[": 1219, "[\"": 1220, "[0": 1221, "[0]": 1222, "[1": 1223, "[1]": 1224, "[:": 1225, "[:,": 1226, "[None,": 1227, "[[": 1228, "[^": 1229, "[^>]": 1230, "[b": 1231, "[b]": 1232, "[ch": 1233, "[ch]": 1234, "[d_model": 1235, "[d_model]": 1236, "[end": 1237, "[float": 1238, "[float,": 1239, "[i": 1240, "[i]": 1241, "[int": 1242, "[int,": 1243, "[int]": 1244, "[int],": 1245, "[int]:": 1246, "[j": 1247, "[key": 1248, "[key]": 1249, "[start": 1250, "[str": 1251, "[str,": 1252, "[str]": 1253, "[tuple[str,": 1254, "\\": 1255, "\\S+": 1256, "\\\\": 1257, "\\\\_lr": 1258, "\\\\_lr}": 1259, "\\\\alpha": 1260, "\\\\eta": 1261, "\\\\hat{": 1262, "\\\\hat{m}_t": 1263, "\\\\hat{v": 1264, "\\\\hat{v}_t": 1265, "\\\\math": 1266, "\\\\mathrm": 1267, "\\\\mathrm{": 1268, "\\\\text{": 1269, "\\\\text{-grams": 1270, "\\\\theta_i^{": 1271, "\\n": 1272, "\\s": 1273, "\\s*": 1274, "\\{": 1275, "]": 1276, "])": 1277, "],": 1278, "]:": 1279, "]]": 1280, "]],": 1281, "^": 1282, "^(": 1283, "^(2i": 1284, "^2": 1285, "^K": 1286, "^K_mn": 1287, "^T": 1288, "^t)": 1289, "^{": 1290, "_": 1291, "_,": 1292, "_1": 1293, "_2": 1294, "_2_slopes": 1295, "_ALPHA_BETA_M": 1296, "_BYTE_TO_UNICODE": 1297, "__": 1298, "__all": 1299, "__all__": 1300, "__init__": 1301, "_a": 1302, "_ag": 1303, "_agre": 1304, "_agreement": 1305, "_agreement_rat": 1306, "_alpha": 1307, "_an": 1308, "_anal": 1309, "_analys": 1310, "_and_": 1311, "_and_activation": 1312, "_and_activation.ipyn": 1313, "_ap": 1314, "_appro": 1315, "_approx": 1316, "_approxim": 1317, "_approximation": 1318, "_arr": 1319, "_arr,": 1320, "_attempt": 1321, "_ax": 1322, "_axis_": 1323, "_axis_curv": 1324, "_b": 1325, "_b)": 1326, "_b,": 1327, "_b:": 1328, "_bar": 1329, "_bar_by_": 1330, "_bar_by_layer": 1331, "_batch": 1332, "_beta": 1333, "_beta3": 1334, "_bits_per_byt": 1335, "_bits_per_byte": 1336, "_block": 1337, "_bo": 1338, "_bod": 1339, "_body": 1340, "_body:": 1341, "_body_": 1342, "_body_output": 1343, "_bootstrap": 1344, "_boundaries": 1345, "_boundaries(": 1346, "_bu": 1347, "_bucket": 1348, "_buckets": 1349, "_buckets:": 1350, "_buffer": 1351, "_buffer(\"": 1352, "_buil": 1353, "_build_": 1354, "_by": 1355, "_by_": 1356, "_by_condition": 1357, "_by_d_model": 1358, "_by_d_model:": 1359, "_by_max": 1360, "_by_max_bytes": 1361, "_by_unit_": 1362, "_by_unit_group": 1363, "_byt": 1364, "_bytes": 1365, "_bytes)": 1366, "_bytes:": 1367, "_c": 1368, "_candidates": 1369, "_clip": 1370, "_co": 1371, "_code": 1372, "_coe": 1373, "_coef": 1374, "_coef_init": 1375, "_coeff": 1376, "_col": 1377, "_cols,": 1378, "_compute_": 1379, "_compute_b": 1380, "_compute_budget": 1381, "_compute_budgets:": 1382, "_con": 1383, "_condition": 1384, "_conditions": 1385, "_converged": 1386, "_corp": 1387, "_corpus": 1388, "_cost": 1389, "_count": 1390, "_count:": 1391, "_counts[": 1392, "_counts[pair]": 1393, "_cover": 1394, "_curv": 1395, "_curves": 1396, "_curves:": 1397, "_cy": 1398, "_cyc": 1399, "_cycle": 1400, "_cycle[": 1401, "_cycle[i": 1402, "_decay": 1403, "_den": 1404, "_denom": 1405, "_dir": 1406, "_embedding": 1407, "_exact": 1408, "_exponent": 1409, "_f": 1410, "_factory": 1411, "_fet": 1412, "_fetch": 1413, "_fetch_w": 1414, "_fetch_wikipedia": 1415, "_fetch_wikipedia_re": 1416, "_fetch_wikipedia_revis": 1417, "_fetch_wikipedia_revision": 1418, "_file": 1419, "_filter": 1420, "_fir": 1421, "_first": 1422, "_fit": 1423, "_fit_n": 1424, "_fit_nonl": 1425, "_fit_nonlinear": 1426, "_fit_nonlinear_h": 1427, "_fit_nonlinear_huber": 1428, "_fits": 1429, "_fn": 1430, "_fn(d_model)": 1431, "_fraction": 1432, "_freq": 1433, "_function": 1434, "_function_curves": 1435, "_fut": 1436, "_future_": 1437, "_future__": 1438, "_grad": 1439, "_grad()": 1440, "_grad()``": 1441, "_grid": 1442, "_grid:": 1443, "_h": 1444, "_hat": 1445, "_he": 1446, "_head": 1447, "_head_": 1448, "_head_attention": 1449, "_heads": 1450, "_heads(": 1451, "_heads)": 1452, "_heads,": 1453, "_heads:": 1454, "_heat": 1455, "_heatmap": 1456, "_hub": 1457, "_huber_": 1458, "_huber_weight": 1459, "_i": 1460, "_i)": 1461, "_i^": 1462, "_id": 1463, "_if": 1464, "_if_l": 1465, "_if_lar": 1466, "_if_large": 1467, "_in": 1468, "_inf": 1469, "_init": 1470, "_init_": 1471, "_init__": 1472, "_init__(": 1473, "_inter": 1474, "_interval": 1475, "_k": 1476, "_k_filter": 1477, "_l": 1478, "_labels": 1479, "_language_model": 1480, "_layer": 1481, "_layer:": 1482, "_le": 1483, "_lead": 1484, "_leading": 1485, "_leading_w": 1486, "_leading_whitespace": 1487, "_learning_rat": 1488, "_len": 1489, "_length_penalty": 1490, "_lik": 1491, "_like(": 1492, "_likelihood": 1493, "_likelihood_n": 1494, "_lin": 1495, "_link": 1496, "_links(": 1497, "_list": 1498, "_list,": 1499, "_list.append(": 1500, "_log": 1501, "_log_prob": 1502, "_log_probs": 1503, "_log_probs,": 1504, "_loss_": 1505, "_loss_per_token": 1506, "_lr": 1507, "_m": 1508, "_manifest": 1509, "_mask": 1510, "_match_rat": 1511, "_max": 1512, "_median": 1513, "_min": 1514, "_min,": 1515, "_mn": 1516, "_model": 1517, "_multi": 1518, "_multi_head_attention": 1519, "_n": 1520, "_nam": 1521, "_names": 1522, "_negative_": 1523, "_negative_log": 1524, "_negative_log_likelihood_n": 1525, "_new": 1526, "_new_token": 1527, "_ngrams": 1528, "_non_embedding_param": 1529, "_non_embedding_parameter": 1530, "_norm": 1531, "_num": 1532, "_numpy": 1533, "_ok": 1534, "_op": 1535, "_opt": 1536, "_opt(": 1537, "_opt(C": 1538, "_opt(C)": 1539, "_or": 1540, "_p": 1541, "_p_filter": 1542, "_pair": 1543, "_parabola": 1544, "_param": 1545, "_params": 1546, "_params)": 1547, "_params,": 1548, "_pen": 1549, "_penal": 1550, "_penalty": 1551, "_per_": 1552, "_per_layer": 1553, "_pla": 1554, "_plain": 1555, "_plaintext(": 1556, "_position": 1557, "_pow": 1558, "_power_law": 1559, "_pred": 1560, "_prefix": 1561, "_pret": 1562, "_pretraining": 1563, "_prob": 1564, "_prod": 1565, "_product_": 1566, "_product_attention": 1567, "_r": 1568, "_rat": 1569, "_ratio": 1570, "_ratio:": 1571, "_re": 1572, "_residual": 1573, "_rm": 1574, "_rms_": 1575, "_rms_ratio": 1576, "_ro": 1577, "_root": 1578, "_rul": 1579, "_rule_": 1580, "_rule_name": 1581, "_rule_name=\"identity": 1582, "_s": 1583, "_saturating": 1584, "_saturating_power_law": 1585, "_scal": 1586, "_scale:": 1587, "_scaling": 1588, "_scaling_analys": 1589, "_score": 1590, "_se": 1591, "_search": 1592, "_seed": 1593, "_seed_": 1594, "_seed_sc": 1595, "_seed_scatter": 1596, "_segment": 1597, "_seq": 1598, "_seq_term": 1599, "_seq_term:": 1600, "_sequences": 1601, "_series": 1602, "_set": 1603, "_shakespear": 1604, "_sing": 1605, "_singl": 1606, "_single_": 1607, "_single_step_": 1608, "_size": 1609, "_size)": 1610, "_size,": 1611, "_size:": 1612, "_slopes": 1613, "_spl": 1614, "_split": 1615, "_split_": 1616, "_split_chunk": 1617, "_split_chunk_by_max_bytes": 1618, "_split_heads(": 1619, "_split_heads(self.w_": 1620, "_split_mod": 1621, "_sqrt": 1622, "_squar": 1623, "_square": 1624, "_stat": 1625, "_states:": 1626, "_statistic": 1627, "_step": 1628, "_su": 1629, "_suc": 1630, "_succes": 1631, "_success": 1632, "_successful": 1633, "_symbol": 1634, "_t": 1635, "_tanh": 1636, "_tanh_approximation": 1637, "_ter": 1638, "_term": 1639, "_text": 1640, "_text,": 1641, "_tgt": 1642, "_threshold": 1643, "_time": 1644, "_to": 1645, "_to_id": 1646, "_to_rms_ratio": 1647, "_token": 1648, "_tokenizer_json": 1649, "_tot": 1650, "_total": 1651, "_tr": 1652, "_train": 1653, "_train_text": 1654, "_train_val": 1655, "_transform": 1656, "_trig": 1657, "_trigg": 1658, "_trigger": 1659, "_triggered": 1660, "_ty": 1661, "_type": 1662, "_type=\"": 1663, "_type=\"model": 1664, "_type=\"model\",": 1665, "_v": 1666, "_v[key]": 1667, "_val": 1668, "_values": 1669, "_values,": 1670, "_w": 1671, "_weight": 1672, "_width": 1673, "_wikitext": 1674, "_wikitext_links(": 1675, "_window": 1676, "_with": 1677, "_without_leading_whitespace\":": 1678, "_word(": 1679, "_x": 1680, "_ylabel:": 1681, "_{": 1682, "_{t": 1683, "_{t-1}": 1684, "`": 1685, "``": 1686, "``(": 1687, "``)": 1688, "``)ãĢĤ": 1689, "``)ãģ§": 1690, "``)ãģ®": 1691, "``)ãģ®åł´åĲĪ": 1692, "``)ãģ®åł´åĲĪ:": 1693, "``)ãģ¯": 1694, "``)ãĤĴ": 1695, "``ãĢģ": 1696, "``ãĢģ``": 1697, "``ãĢĤ": 1698, "``ãĤĴ": 1699, "``ãĥ»": 1700, "``ãĥ»``": 1701, "``ãĥ»``zero": 1702, "a": 1703, "a,": 1704, "a3": 1705, "a:": 1706, "a_bo": 1707, "a_body_": 1708, "a_body_list": 1709, "a_coef_init": 1710, "ab": 1711, "ab:": 1712, "abc": 1713, "able": 1714, "able,": 1715, "abo": 1716, "abola": 1717, "abs(": 1718, "abs_": 1719, "abs_r": 1720, "abstra": 1721, "abstract": 1722, "abstractmethod": 1723, "ac": 1724, "acc": 1725, "accar": 1726, "accard": 1727, "accept": 1728, "ace": 1729, "ach": 1730, "ack": 1731, "ack([": 1732, "act": 1733, "acterL": 1734, "acterLe": 1735, "acterLevel": 1736, "acterLevelTokenizer": 1737, "action": 1738, "action=": 1739, "action=parse": 1740, "activation": 1741, "activation_fn": 1742, "activation_fn:": 1743, "activation_fn=": 1744, "activations:": 1745, "actory": 1746, "ad": 1747, "ad_": 1748, "ad_am": 1749, "ad_amount": 1750, "ad_text(": 1751, "ad_text(encoding=\"utf-8": 1752, "add(id": 1753, "add(id(": 1754, "add_": 1755, "added_": 1756, "adding": 1757, "adding_mask": 1758, "adfor": 1759, "adford": 1760, "ag": 1761, "age": 1762, "air": 1763, "ais": 1764, "aises:": 1765, "ak": 1766, "ake_": 1767, "ake_evaluation_window": 1768, "al": 1769, "al_": 1770, "al_allo": 1771, "al_allocation": 1772, "al_allocation_exponent": 1773, "al_encoding": 1774, "al_f": 1775, "al_norm": 1776, "al_transform": 1777, "al_values": 1778, "alf": 1779, "alg": 1780, "alization": 1781, "all": 1782, "allab": 1783, "allable": 1784, "allback": 1785, "allback_": 1786, "allback_train_text": 1787, "allo": 1788, "alp": 1789, "alpha": 1790, "alpha:": 1791, "alpha=0": 1792, "alpha=0.3": 1793, "alue": 1794, "always_negative_": 1795, "always_negative_unit_": 1796, "always_negative_unit_ratio": 1797, "am": 1798, "amma": 1799, "ampl": 1800, "ampling": 1801, "an": 1802, "ance": 1803, "and": 1804, "andid": 1805, "andidat": 1806, "andidates": 1807, "andidates:": 1808, "anes": 1809, "ang": 1810, "ange(": 1811, "anguage": 1812, "anh": 1813, "anif": 1814, "anifest": 1815, "anifest_p": 1816, "anifest_path": 1817, "anifest_path)": 1818, "ann": 1819, "annotat": 1820, "annotate:": 1821, "anon": 1822, "anoni": 1823, "anonical_": 1824, "anonical_tokenizer": 1825, "anonical_tokenizers.": 1826, "ap": 1827, "apanes": 1828, "api": 1829, "api.": 1830, "app": 1831, "append(": 1832, "appl": 1833, "apply(": 1834, "apps/": 1835, "ar": 1836, "ar_width": 1837, "aracterLevelTokenizer": 1838, "arange(": 1839, "arch": 1840, "arget": 1841, "aries": 1842, "arization": 1843, "ark": 1844, "arker": 1845, "arker=\"": 1846, "arn": 1847, "arpath": 1848, "arpathy": 1849, "arquar": 1850, "arquard": 1851, "arquardt": 1852, "arquardt(": 1853, "arr": 1854, "arr,": 1855, "arra": 1856, "array": 1857, "array(": 1858, "art": 1859, "arti": 1860, "articl": 1861, "article_": 1862, "articles": 1863, "articles_dir": 1864, "ary": 1865, "as": 1866, "ase": 1867, "asel": 1868, "ask": 1869, "ast": 1870, "at": 1871, "ata": 1872, "atch": 1873, "atch_rat": 1874, "atch_size,": 1875, "atch_size:": 1876, "atches": 1877, "ate": 1878, "ate_": 1879, "ate_b": 1880, "ate_bits_per_byt": 1881, "ate_causal_mask": 1882, "ate_f": 1883, "ate_flops_per_token": 1884, "ate_h": 1885, "ate_half": 1886, "ate_p": 1887, "ate_padding_mask": 1888, "ated": 1889, "ateg": 1890, "ategory": 1891, "ath": 1892, "ath:": 1893, "ath::": 1894, "ath:`": 1895, "ating": 1896, "ation": 1897, "ativ": 1898, "ative": 1899, "ative_": 1900, "atplot": 1901, "atplotlib": 1902, "atr": 1903, "atrix": 1904, "ats": 1905, "ats)ãĢĤ": 1906, "att": 1907, "attemp": 1908, "attempt": 1909, "attention": 1910, "attention_heatmap": 1911, "attention_s": 1912, "attention_score_bias": 1913, "attention_seq_term:": 1914, "atter": 1915, "attn": 1916, "attn_": 1917, "attn_weight": 1918, "attn_weights": 1919, "attn_weights)": 1920, "attn_weights,": 1921, "attr(": 1922, "atur": 1923, "aturating": 1924, "aturatingPowerLawFit": 1925, "aturating_fit": 1926, "aturating_fits": 1927, "at{": 1928, "au": 1929, "auss": 1930, "aussian": 1931, "avi": 1932, "aw": 1933, "ax": 1934, "axes": 1935, "ayer": 1936, "ayerNorm": 1937, "b": 1938, "b``": 1939, "ba": 1940, "back": 1941, "bar": 1942, "base": 1943, "beam": 1944, "beam_search": 1945, "beam_size": 1946, "berg": 1947, "best": 1948, "best_param": 1949, "beta": 1950, "bi": 1951, "bian": 1952, "bits_per_byt": 1953, "bits_per_byte": 1954, "bj": 1955, "bject": 1956, "block": 1957, "bool": 1958, "bootstrap_scaling_analysis": 1959, "boundaries_a": 1960, "bpe_": 1961, "bs": 1962, "bsol": 1963, "bsolut": 1964, "bsolute": 1965, "bucket": 1966, "buckets": 1967, "byt": 1968, "byte_": 1969, "byte_fallback": 1970, "byte_level": 1971, "byte_level:": 1972, "byte_level_symbol": 1973, "byte_to_": 1974, "c": 1975, "ca": 1976, "cab": 1977, "cach": 1978, "cache": 1979, "cache(": 1980, "cache_": 1981, "cache_dir": 1982, "cache_dir:": 1983, "cache_path": 1984, "cached": 1985, "cached_": 1986, "cached_pro": 1987, "cached_proper": 1988, "cached_property": 1989, "cal": 1990, "cal_": 1991, "cal_jacobian": 1992, "candidates": 1993, "canonical_tokenizers.": 1994, "cat": 1995, "categ": 1996, "categories": 1997, "cation": 1998, "caus": 1999, "causal": 2000, "causal_mask": 2001, "ce": 2002, "ce_": 2003, "ce_length": 2004, "cent": 2005, "center": 2006, "center:": 2007, "centering": 2008, "centile": 2009, "cept": 2010, "ces": 2011, "ces:": 2012, "cessor": 2013, "ch": 2014, "ch().": 2015, "chan": 2016, "char": 2017, "char_to_id": 2018, "charact": 2019, "character_": 2020, "character_coverag": 2021, "character_level": 2022, "character_level_tokenizer_json": 2023, "chars": 2024, "chars)": 2025, "ched": 2026, "chedule": 2027, "chil": 2028, "chilla": 2029, "chinchilla": 2030, "chinchilla_param": 2031, "chinchilla_parametric": 2032, "chinchilla_residual": 2033, "chunk": 2034, "chunk_": 2035, "chunk_length": 2036, "chunk_length_statistic": 2037, "chunk_split_mod": 2038, "chunk_split_mode": 2039, "chunk_split_mode,": 2040, "chunk_split_mode:": 2041, "chunk_split_mode=\"": 2042, "chunk_split_mode=\"whitespace": 2043, "chunk_split_mode=\"whitespace\"``": 2044, "ci": 2045, "ck": 2046, "cket": 2047, "cl": 2048, "cla": 2049, "class": 2050, "cle": 2051, "clean": 2052, "clip": 2053, "clip(log": 2054, "clip(log_alpha": 2055, "clip(log_alpha,": 2056, "clip(log_beta": 2057, "clip_threshold": 2058, "clip_threshold:": 2059, "clud": 2060, "clude_": 2061, "cmap": 2062, "cmap:": 2063, "co": 2064, "cobian": 2065, "cod": 2066, "code": 2067, "coder": 2068, "coder-only": 2069, "coding": 2070, "coeff": 2071, "coefficient": 2072, "color": 2073, "color=\"": 2074, "color=\"tab:": 2075, "color_cycle": 2076, "colorbar": 2077, "comput": 2078, "compute_": 2079, "con": 2080, "cond": 2081, "constr": 2082, "construct_": 2083, "construct_optimal_f": 2084, "construct_optimal_frontier": 2085, "construct_optimal_frontier(": 2086, "context": 2087, "contin": 2088, "continue": 2089, "converged": 2090, "cor": 2091, "core": 2092, "core_b": 2093, "core_bias": 2094, "cores": 2095, "corp": 2096, "corpu": 2097, "corpus": 2098, "corpus,": 2099, "corpus_": 2100, "correction": 2101, "cos": 2102, "cosine": 2103, "cost": 2104, "count": 2105, "count_non_embedding_parameter": 2106, "count_pair": 2107, "couple": 2108, "coupled": 2109, "cover": 2110, "coverag": 2111, "cre": 2112, "crease": 2113, "cross_": 2114, "cross_attention": 2115, "cross_attn": 2116, "cross_attn_weights": 2117, "cstr": 2118, "cstring": 2119, "ct": 2120, "ction": 2121, "cto": 2122, "ctool": 2123, "ctx": 2124, "ctx:": 2125, "cur": 2126, "current": 2127, "cursor": 2128, "curv": 2129, "d": 2130, "d(": 2131, "d(id": 2132, "d(idx": 2133, "d(idx)": 2134, "d,": 2135, "d_": 2136, "d_bar": 2137, "d_factory": 2138, "d_ff": 2139, "d_ff,": 2140, "d_ff:": 2141, "d_i": 2142, "d_k": 2143, "d_k)": 2144, "d_model": 2145, "d_model)": 2146, "d_model,": 2147, "d_model:": 2148, "d_model``": 2149, "d_values": 2150, "da": 2151, "dam": 2152, "darray": 2153, "data": 2154, "data.": 2155, "data/": 2156, "data[\"": 2157, "dataclass": 2158, "de": 2159, "deca": 2160, "decay": 2161, "decay_": 2162, "decay_fraction": 2163, "decod": 2164, "decode_": 2165, "decode_byte_level_symbol": 2166, "decoder-only": 2167, "ded_": 2168, "ded_param": 2169, "ded_param_id": 2170, "def": 2171, "defa": 2172, "default": 2173, "del": 2174, "delta": 2175, "den": 2176, "dep": 2177, "depth": 2178, "der": 2179, "design": 2180, "deta": 2181, "detach().": 2182, "devi": 2183, "device": 2184, "device)": 2185, "device,": 2186, "device:": 2187, "device=": 2188, "dex": 2189, "dex[": 2190, "dex[pair]": 2191, "de{": 2192, "de{g": 2193, "de{g}_t": 2194, "di": 2195, "dict": 2196, "dient": 2197, "dim": 2198, "ding": 2199, "dir": 2200, "dist": 2201, "distance": 2202, "distin": 2203, "distinct": 2204, "dition": 2205, "divergence": 2206, "do": 2207, "dom": 2208, "dot": 2209, "dot_product_attention": 2210, "dow": 2211, "down": 2212, "dropout": 2213, "dropout)": 2214, "dropout,": 2215, "dropout:": 2216, "dtype": 2217, "dtype)": 2218, "dtype,": 2219, "dtype:": 2220, "dtype=": 2221, "dual": 2222, "dual_axis_curv": 2223, "e": 2224, "e\"": 2225, "e\"):": 2226, "e\",": 2227, "e(": 2228, "e(\"": 2229, "e(\"log": 2230, "e(\"log\")": 2231, "e(\"utf-8": 2232, "e()``": 2233, "e(over": 2234, "e(over=\"ignore\",": 2235, "e(self,": 2236, "e(æĹ¢å®ļåĢ¤": 2237, "e)": 2238, "e,": 2239, "e:": 2240, "e=": 2241, "eLU": 2242, "e_": 2243, "e_b": 2244, "e_band": 2245, "e_bpe_": 2246, "e_bpe_id_tokenizer_json": 2247, "e_canonical_tokenizers.": 2248, "e_corpus": 2249, "e_cross_attention": 2250, "e_decay_": 2251, "e_embedding": 2252, "e_init": 2253, "e_m": 2254, "e_mn": 2255, "e_model": 2256, "e_per_": 2257, "e_per_head": 2258, "e_s": 2259, "e_schedule": 2260, "e_std": 2261, "e_x": 2262, "e``": 2263, "ead": 2264, "ead_": 2265, "eam": 2266, "ean": 2267, "earn": 2268, "ect": 2269, "ectiv": 2270, "ective_decay_": 2271, "ector": 2272, "ed": 2273, "ed)": 2274, "ed,": 2275, "ed.": 2276, "ed:": 2277, "ed=": 2278, "edAbsolutePositionalEmbedding": 2279, "ed_": 2280, "ed_dot_product_attention": 2281, "edding": 2282, "edi": 2283, "edia": 2284, "ediaWiki": 2285, "edian": 2286, "edict": 2287, "eed": 2288, "eed\",": 2289, "eed_": 2290, "eed_for": 2291, "eed_forwar": 2292, "eed_forward": 2293, "eed_forward_factory": 2294, "eed_forward_factory:": 2295, "eep": 2296, "eepdim": 2297, "eff": 2298, "effective_decay_": 2299, "eg": 2300, "egative": 2301, "egative_": 2302, "egul": 2303, "egularization": 2304, "eight": 2305, "el": 2306, "el()": 2307, "elative_": 2308, "elf": 2309, "elfAttention": 2310, "eli": 2311, "elif": 2312, "eliho": 2313, "elihood": 2314, "else": 2315, "else:": 2316, "elt": 2317, "elta": 2318, "elu": 2319, "elu_exact": 2320, "elu_tanh_approximation": 2321, "em": 2322, "embedding": 2323, "ement": 2324, "emor": 2325, "emory": 2326, "emory_": 2327, "emp": 2328, "emperat": 2329, "emperatur": 2330, "emperature": 2331, "en": 2332, "en_": 2333, "enberg": 2334, "ence": 2335, "encod": 2336, "encode(": 2337, "encode(self,": 2338, "encode(text)": 2339, "encode_": 2340, "encode_chunk": 2341, "encode_corpus": 2342, "encoding": 2343, "encoding=\"utf-8": 2344, "encoding=\"utf-8\")": 2345, "end": 2346, "end(": 2347, "ener": 2348, "enerat": 2349, "eneration": 2350, "eneration\",": 2351, "enerator": 2352, "enn": 2353, "ennr": 2354, "ennri": 2355, "ennrich": 2356, "ensor": 2357, "ent": 2358, "ent=False)": 2359, "entence": 2360, "entencePiece": 2361, "entencepiece": 2362, "enter": 2363, "entro": 2364, "entropy": 2365, "eories": 2366, "ep": 2367, "eps": 2368, "eps)": 2369, "eps:": 2370, "equence": 2371, "er": 2372, "erL": 2373, "er_": 2374, "er_clean": 2375, "erat": 2376, "erate(": 2377, "ere(": 2378, "erf": 2379, "erg": 2380, "eri": 2381, "erical_jacobian": 2382, "eries": 2383, "eries_names": 2384, "erior": 2385, "ero": 2386, "err": 2387, "error": 2388, "errstat": 2389, "errstate(over=\"ignore\",": 2390, "ers": 2391, "ers\",": 2392, "ersist": 2393, "ersistent=False)": 2394, "ertil": 2395, "ertility": 2396, "ery": 2397, "es": 2398, "es(": 2399, "es:": 2400, "esign": 2401, "esp": 2402, "espe": 2403, "espear": 2404, "est": 2405, "est_p": 2406, "est_pair": 2407, "estim": 2408, "estimate_flops_per_token": 2409, "esty": 2410, "estyle": 2411, "estyle=\"": 2412, "estyle=\"--": 2413, "et": 2414, "et_": 2415, "et_learning_rat": 2416, "et_random": 2417, "et_random_batch": 2418, "eta": 2419, "eter": 2420, "eth": 2421, "ethod": 2422, "etr": 2423, "etric": 2424, "etur": 2425, "eturn": 2426, "etwor": 2427, "etwork": 2428, "eur": 2429, "eval": 2430, "eval()": 2431, "eval_bits_per_byte": 2432, "eval_interval": 2433, "eval_step": 2434, "evalu": 2435, "evaluate_bits_per_byt": 2436, "evaluate_bits_per_byte``": 2437, "evaluation": 2438, "evaluation_mask": 2439, "evaluation_mask:": 2440, "evaluation_window": 2441, "evaluation_windows:": 2442, "evenberg": 2443, "evenberg-Marquardt": 2444, "ew": 2445, "ex": 2446, "exact": 2447, "exact_match_rat": 2448, "except": 2449, "excl": 2450, "exclu": 2451, "excluded_param_id": 2452, "excluded_param_ids.": 2453, "excluded_param_ids.add(id(": 2454, "exist": 2455, "exist_ok": 2456, "exists()": 2457, "exists():": 2458, "exp": 2459, "exp(": 2460, "exp(log": 2461, "expon": 2462, "exponent": 2463, "ext": 2464, "ext_code": 2465, "ext_token": 2466, "ext_token_log": 2467, "extra": 2468, "ey": 2469, "ez": 2470, "eze(": 2471, "f": 2472, "f\"": 2473, "f\"{": 2474, "f-8": 2475, "factory": 2476, "fallback": 2477, "feed_forward_factory": 2478, "feed_forward_factory``": 2479, "fertility": 2480, "ff": 2481, "ffer": 2482, "fi": 2483, "fig": 2484, "figsize=(": 2485, "fil": 2486, "file": 2487, "file(": 2488, "file[": 2489, "file[\"": 2490, "filter": 2491, "fin": 2492, "finit": 2493, "finite(": 2494, "fit_": 2495, "fit_chinchilla_parametric": 2496, "fit_chinchilla_parametric``": 2497, "fit_isoflop_parabola": 2498, "fit_power_law": 2499, "fit_saturating_power_law": 2500, "float": 2501, "float3": 2502, "fontsize=": 2503, "fontsize=8)": 2504, "for": 2505, "form": 2506, "forward": 2507, "forward``": 2508, "frac": 2509, "frac{": 2510, "from": 2511, "frontier_": 2512, "frontier_log_c": 2513, "ft": 2514, "ful": 2515, "full": 2516, "full_like(": 2517, "g": 2518, "g/": 2519, "gamma": 2520, "ge": 2521, "ge(": 2522, "gelu_exact": 2523, "generator": 2524, "get": 2525, "get_random_batch": 2526, "gg": 2527, "gging": 2528, "gist": 2529, "gister": 2530, "gment": 2531, "gmentation": 2532, "gn": 2533, "gnmt": 2534, "gnmt_length_penalty": 2535, "gpt": 2536, "gra": 2537, "grad": 2538, "gradient": 2539, "gradient_clip": 2540, "gradient_clip_threshold": 2541, "gradient_clip_triggered": 2542, "gradient_norm": 2543, "gradient_norm_by_unit_group": 2544, "gradient_norm_per_layer": 2545, "gradient_norms)": 2546, "gradient_norms:": 2547, "gram": 2548, "grams": 2549, "grams)": 2550, "gray": 2551, "gray\",": 2552, "gres": 2553, "grid": 2554, "grid(alpha=0.3": 2555, "grid:": 2556, "gro": 2557, "grou": 2558, "group": 2559, "groupe": 2560, "grouped_bar": 2561, "gt": 2562, "gt_mask": 2563, "gth": 2564, "gu": 2565, "guag": 2566, "guage": 2567, "h": 2568, "h,": 2569, "ha": 2570, "hak": 2571, "hakespear": 2572, "hape": 2573, "hape(": 2574, "haracterLevelTokenizer": 2575, "hat{": 2576, "haw": 2577, "head": 2578, "hidden": 2579, "hinchilla": 2580, "history": 2581, "history[\"": 2582, "hit": 2583, "hitesp": 2584, "hitespace": 2585, "ho": 2586, "hold": 2587, "ht": 2588, "htt": 2589, "http": 2590, "https://": 2591, "hunk": 2592, "hunkSplitMod": 2593, "hunkSplitMode": 2594, "i": 2595, "i)": 2596, "i-theories": 2597, "i:": 2598, "iBi": 2599, "iGLU": 2600, "iLU": 2601, "ia": 2602, "ian": 2603, "iance": 2604, "ias": 2605, "ias)": 2606, "ias_correction": 2607, "ic": 2608, "ical": 2609, "icient": 2610, "icmethod": 2611, "icode": 2612, "ict": 2613, "id": 2614, "id=-1,": 2615, "id_": 2616, "id_to": 2617, "id_to_": 2618, "id_to_char": 2619, "id_tokenizer_json": 2620, "idal": 2621, "idden": 2622, "ident": 2623, "idirection": 2624, "idirectional": 2625, "ids": 2626, "idth": 2627, "idual": 2628, "idx": 2629, "idx]": 2630, "ie": 2631, "iece": 2632, "ieces": 2633, "ieces.append(": 2634, "ient": 2635, "ies": 2636, "ies:": 2637, "if": 2638, "iform": 2639, "ig": 2640, "ight": 2641, "ign": 2642, "igra": 2643, "igram": 2644, "igsize_per_head": 2645, "ik": 2646, "ikipedia": 2647, "ikitext": 2648, "il": 2649, "ile": 2650, "ilenam": 2651, "ilter": 2652, "im": 2653, "ime": 2654, "import": 2655, "in": 2656, "in(": 2657, "inal_values": 2658, "inchilla": 2659, "include_": 2660, "increase": 2661, "indi": 2662, "indices": 2663, "indow": 2664, "ine": 2665, "ine_": 2666, "inear": 2667, "ined_": 2668, "inewidth": 2669, "inf": 2670, "ing": 2671, "init": 2672, "init.": 2673, "init_alpha": 2674, "init_beta": 2675, "initi": 2676, "initial_values": 2677, "inn": 2678, "inner_clean": 2679, "input": 2680, "inputs,": 2681, "instance": 2682, "int": 2683, "interior": 2684, "inus": 2685, "inuso": 2686, "inusoidal": 2687, "inusoidalPositionalEncoding": 2688, "inv": 2689, "inv_freq": 2690, "io": 2691, "ion": 2692, "ip": 2693, "ipedia": 2694, "ipt": 2695, "ipts/": 2696, "ipts/p": 2697, "ipts/promot": 2698, "ipts/promote_canonical_tokenizers.": 2699, "ipy": 2700, "ipyn": 2701, "ipynb``": 2702, "ir": 2703, "irection": 2704, "is": 2705, "is_": 2706, "is_interior": 2707, "is_interior:": 2708, "ise": 2709, "isfinite(": 2710, "ish": 2711, "ism": 2712, "ismatches": 2713, "isof": 2714, "isoflop": 2715, "isoflop_parabola": 2716, "ist": 2717, "istic": 2718, "istor": 2719, "istories": 2720, "istories:": 2721, "istories_by_condition": 2722, "istory": 2723, "it": 2724, "it)": 2725, "it.": 2726, "it_": 2727, "it_alpha": 2728, "it_saturating_power_law": 2729, "it_se": 2730, "it_second": 2731, "ite_": 2732, "item": 2733, "items()": 2734, "iter": 2735, "iter:": 2736, "iterbi": 2737, "iterbi_segment": 2738, "itext": 2739, "ith": 2740, "ithL2Regularization": 2741, "ition": 2742, "itle": 2743, "its": 2744, "its-per-byt": 2745, "its-per-byte": 2746, "its-per-byte(": 2747, "its_": 2748, "its_per": 2749, "its_per_byt": 2750, "ity": 2751, "iv": 2752, "ivation": 2753, "ive": 2754, "ix": 2755, "iz": 2756, "ization": 2757, "ize": 2758, "ize(": 2759, "ize=": 2760, "ize_per_head": 2761, "ize``": 2762, "izer": 2763, "izer_": 2764, "izer_json": 2765, "j": 2766, "ja": 2767, "ja\"``ãĥ»``\"en\"``": 2768, "jacobian": 2769, "japanes": 2770, "jo": 2771, "joined_": 2772, "joined_a": 2773, "js": 2774, "json": 2775, "json.": 2776, "json.load": 2777, "json.loads(": 2778, "k": 2779, "k,": 2780, "ke": 2781, "ken": 2782, "key": 2783, "key:": 2784, "key_": 2785, "key_embedding": 2786, "key_length,": 2787, "key_length:": 2788, "l": 2789, "l_inf": 2790, "l_values": 2791, "la": 2792, "lab": 2793, "label": 2794, "label)": 2795, "label:": 2796, "label=": 2797, "label=\"": 2798, "labels": 2799, "labels(": 2800, "labels:": 2801, "lace": 2802, "lamb": 2803, "lambda": 2804, "lambda_": 2805, "lan": 2806, "languag": 2807, "language": 2808, "language:": 2809, "language_model": 2810, "language}": 2811, "lat": 2812, "lat_mask": 2813, "lation": 2814, "law": 2815, "lay": 2816, "layer": 2817, "layer_": 2818, "layer_output": 2819, "layers.": 2820, "le": 2821, "le[": 2822, "learn": 2823, "learn_b": 2824, "learn_bpe": 2825, "learning": 2826, "learning_curves": 2827, "learning_curves_multi": 2828, "learning_rat": 2829, "learning_rate": 2830, "learning_rate(": 2831, "learning_rate_schedule": 2832, "left": 2833, "leg": 2834, "len": 2835, "len(": 2836, "lenam": 2837, "length": 2838, "length)": 2839, "length_penalty": 2840, "level": 2841, "levenberg_m": 2842, "levenberg_marquardt(": 2843, "lex": 2844, "lexity": 2845, "li": 2846, "lib": 2847, "lik": 2848, "like(": 2849, "lin": 2850, "linalg": 2851, "linalg.": 2852, "linalg.lst": 2853, "linalg.lsts": 2854, "linalg.lstsq": 2855, "linalg.norm": 2856, "line": 2857, "line(": 2858, "linear": 2859, "linear1": 2860, "lines": 2861, "linewidth": 2862, "lip": 2863, "list": 2864, "ll": 2865, "lle": 2866, "llection": 2867, "lm_head": 2868, "lm_head``)ãģ®": 2869, "lo": 2870, "load": 2871, "load_": 2872, "load_wikipedia_corpus": 2873, "loat": 2874, "lock": 2875, "log": 2876, "log(": 2877, "log_c": 2878, "log_n": 2879, "log_n_arr": 2880, "log_n_opt": 2881, "log_x": 2882, "logit": 2883, "logits": 2884, "logits,": 2885, "logits``": 2886, "long": 2887, "loor": 2888, "lop": 2889, "lopen": 2890, "lopes": 2891, "lops": 2892, "lops_": 2893, "lops_per_token": 2894, "lor": 2895, "los": 2896, "lose": 2897, "losest": 2898, "losest_pow": 2899, "loss": 2900, "loss)": 2901, "loss,": 2902, "loss.": 2903, "loss:": 2904, "loss_": 2905, "loss_step_": 2906, "loss_step_delta": 2907, "loss_step_deltas:": 2908, "lot": 2909, "low": 2910, "lp": 2911, "lp(Y": 2912, "lr": 2913, "lr)": 2914, "lr:": 2915, "lst": 2916, "lt": 2917, "m": 2918, "m,": 2919, "m_h": 2920, "m_head": 2921, "ma": 2922, "make_evaluation_window": 2923, "make_evaluation_windows``": 2924, "man": 2925, "map": 2926, "mask": 2927, "mask)": 2928, "mask:": 2929, "mask]": 2930, "math": 2931, "math:`": 2932, "matrix": 2933, "max": 2934, "max(": 2935, "max_": 2936, "max_chunk": 2937, "max_chunk_bytes": 2938, "max_chunk_bytes,": 2939, "max_chunk_bytes:": 2940, "max_chunk_bytes``": 2941, "max_distance": 2942, "max_len": 2943, "max_len,": 2944, "max_new_token": 2945, "max_new_tokens)": 2946, "max_new_tokens:": 2947, "max_position": 2948, "max_position,": 2949, "max_relative_position": 2950, "max_sequence_length": 2951, "mb": 2952, "mbedding": 2953, "mbol": 2954, "mean": 2955, "mean_loss_per_token": 2956, "mean_square": 2957, "mean_to_rms_ratio": 2958, "median": 2959, "ment": 2960, "merg": 2961, "merge_": 2962, "merges": 2963, "merges:": 2964, "method": 2965, "min": 2966, "min(": 2967, "min,": 2968, "min\\\\_lr}": 2969, "min_learning_rat": 2970, "minus": 2971, "mo": 2972, "mod": 2973, "model": 2974, "model,": 2975, "model.": 2976, "model.parameters()": 2977, "model:": 2978, "model_prefix": 2979, "model_prefix.": 2980, "module": 2981, "move_": 2982, "mp": 2983, "mport": 2984, "mpt": 2985, "mpt_id": 2986, "mput": 2987, "mpute_": 2988, "mt": 2989, "mul_": 2990, "m}_t": 2991, "n": 2992, "n.": 2993, "n_": 2994, "n_key": 2995, "n_query": 2996, "n_rat": 2997, "nam": 2998, "name": 2999, "name,": 3000, "name}": 3001, "nan": 3002, "ndarray": 3003, "negative_": 3004, "new": 3005, "nn.": 3006, "nn.Parameter": 3007, "no": 3008, "no_grad()": 3009, "none": 3010, "norm": 3011, "norm_first": 3012, "norm_first:": 3013, "norm_first=True": 3014, "norm_fn": 3015, "normalization": 3016, "normalization_factory": 3017, "normalization_factory:": 3018, "normalization_rule_name=\"identity": 3019, "normed": 3020, "norms": 3021, "now": 3022, "nown_rat": 3023, "np.": 3024, "np.log(": 3025, "np.log(warm_fit": 3026, "np.ndarray": 3027, "np.ndarray],": 3028, "num": 3029, "num_buckets": 3030, "num_heads": 3031, "num_heads,": 3032, "numel()": 3033, "o": 3034, "oPE": 3035, "ocab": 3036, "ocstring": 3037, "od": 3038, "of": 3039, "of_2_slopes": 3040, "off": 3041, "oft": 3042, "oftmax": 3043, "oftmax(": 3044, "ois": 3045, "oise_band": 3046, "oise_std": 3047, "oisy_": 3048, "oisy_grid": 3049, "oisy_l": 3050, "ok": 3051, "oken": 3052, "okenizer": 3053, "ol": 3054, "old": 3055, "old_": 3056, "oldid": 3057, "om": 3058, "on": 3059, "on-": 3060, "on_embedding": 3061, "on_embedding_param": 3062, "on_embedding_params": 3063, "on_embedding_params:": 3064, "one": 3065, "ones": 3066, "ones_like(": 3067, "ong": 3068, "ongTensor": 3069, "onl": 3070, "ont": 3071, "onti": 3072, "ontsize=": 3073, "ool": 3074, "oot": 3075, "ootstra": 3076, "ootstrap": 3077, "ootstrapResult": 3078, "ootstrap_scaling_analys": 3079, "ootstrap_scaling_analysis": 3080, "op": 3081, "optim": 3082, "optim.": 3083, "optimal_allocation_exponent": 3084, "optimal_f": 3085, "optimizer": 3086, "optimizer.": 3087, "optimizer:": 3088, "optimizer``": 3089, "oq": 3090, "oqa:": 3091, "or": 3092, "orch": 3093, "orch.": 3094, "ord(": 3095, "ord(\"": 3096, "ord(\"Â": 3097, "order": 3098, "org/": 3099, "ories": 3100, "orm": 3101, "ormalization": 3102, "ormalization_factory": 3103, "orp": 3104, "ort": 3105, "ory": 3106, "ot": 3107, "ot-Product": 3108, "otImplementedError": 3109, "otary": 3110, "otaryPositionEmbedding": 3111, "otat": 3112, "otation": 3113, "otation:": 3114, "otations": 3115, "ou": 3116, "oun": 3117, "ound": 3118, "oundaries": 3119, "oundaries_a": 3120, "ount": 3121, "ounter": 3122, "ous_": 3123, "out": 3124, "out)": 3125, "out))": 3126, "out,": 3127, "out_leading_whitespace": 3128, "out_leading_whitespace\":": 3129, "output": 3130, "output,": 3131, "output_layer:": 3132, "over": 3133, "p": 3134, "p)": 3135, "p.": 3136, "p.add_": 3137, "p.grad": 3138, "p90": 3139, "p``": 3140, "pair": 3141, "pair]": 3142, "pair_counts[pair]": 3143, "pair_in": 3144, "pair_in_word(": 3145, "pair_index[pair]": 3146, "par": 3147, "parabola": 3148, "parabola.": 3149, "parabola_coeff": 3150, "param": 3151, "parameter": 3152, "parameters()": 3153, "params": 3154, "params0": 3155, "params0:": 3156, "params:": 3157, "parent": 3158, "parent.mkdir(parents=True,": 3159, "parse": 3160, "path": 3161, "path.": 3162, "path.exists():": 3163, "path.w": 3164, "path.wr": 3165, "path.write_": 3166, "path.write_text(": 3167, "pe": 3168, "pe(": 3169, "pe.": 3170, "pe[:,": 3171, "pe_": 3172, "peak": 3173, "peak\\\\_lr}": 3174, "peak_learning_rat": 3175, "peak_learning_rate": 3176, "per": 3177, "per_": 3178, "per_token": 3179, "percentile": 3180, "perplexity": 3181, "pi": 3182, "pie": 3183, "piece": 3184, "ping": 3185, "pl": 3186, "place": 3187, "plexity": 3188, "plit": 3189, "plit_train_val": 3190, "plot": 3191, "plot_": 3192, "pm": 3193, "pm.SentencePiece": 3194, "po": 3195, "point": 3196, "pointer": 3197, "pos": 3198, "pos,": 3199, "position": 3200, "position:": 3201, "position_embedding": 3202, "positional_encoding": 3203, "positional_transform": 3204, "positional_transform:": 3205, "positional_transform``": 3206, "positions": 3207, "positions)": 3208, "positions.": 3209, "positions:": 3210, "positions]": 3211, "pout": 3212, "pow": 3213, "pow(2)": 3214, "power_": 3215, "power_law": 3216, "power_law_fit": 3217, "power_of_2_slopes": 3218, "pproach": 3219, "pred": 3220, "predicted_": 3221, "pretoken": 3222, "pretokenize``": 3223, "pro": 3224, "proach": 3225, "prob": 3226, "probs": 3227, "processor": 3228, "profil": 3229, "profiles": 3230, "profiles:": 3231, "progres": 3232, "prompt_id": 3233, "prop": 3234, "pt": 3235, "put": 3236, "py": 3237, "q": 3238, "q,": 3239, "qr": 3240, "qrt": 3241, "qrt(": 3242, "qu": 3243, "quar": 3244, "que": 3245, "queeze(": 3246, "quen": 3247, "quence": 3248, "quer": 3249, "query": 3250, "query.device)": 3251, "query:": 3252, "query_length,": 3253, "query_length:": 3254, "quest": 3255, "quir": 3256, "r": 3257, "r\"": 3258, "r(": 3259, "r_squar": 3260, "ra": 3261, "rac": 3262, "raction": 3263, "rain": 3264, "raining": 3265, "raise": 3266, "ran": 3267, "random": 3268, "rang": 3269, "range(": 3270, "rank": 3271, "rans": 3272, "ransform": 3273, "ransformer": 3274, "ransformer_block": 3275, "rat": 3276, "ratio": 3277, "ratios": 3278, "raw": 3279, "raw_norm": 3280, "rc": 3281, "rc.": 3282, "re": 3283, "re-": 3284, "reak": 3285, "reconstruct_optimal_frontier": 3286, "rection": 3287, "red": 3288, "redict": 3289, "redicted_": 3290, "ref": 3291, "refix": 3292, "register": 3293, "relative_": 3294, "relative_buckets": 3295, "relative_key_embedding": 3296, "relative_position": 3297, "relative_position)": 3298, "relative_position_bucket": 3299, "relative_position_bucket(": 3300, "relative_vector": 3301, "remove_": 3302, "rent": 3303, "rep": 3304, "repo": 3305, "repo_id": 3306, "repo_id:": 3307, "repo_id=repo_id,": 3308, "repo_root": 3309, "req": 3310, "request": 3311, "request_time": 3312, "requir": 3313, "res": 3314, "reshape(": 3315, "reshold": 3316, "residual": 3317, "residual_fn": 3318, "residual_fn:": 3319, "residuals": 3320, "residuals)": 3321, "residuals**2": 3322, "residuals**2))": 3323, "residuals,": 3324, "residuals:": 3325, "ress": 3326, "result": 3327, "result_bo": 3328, "result_body": 3329, "ret": 3330, "retoken": 3331, "retokenize(": 3332, "retr": 3333, "retries": 3334, "retry_": 3335, "retry_wa": 3336, "retry_wait_second": 3337, "return": 3338, "revid": 3339, "rg": 3340, "rid": 3341, "ridPoint": 3342, "right": 3343, "rint": 3344, "ripts/promote_canonical_tokenizers.": 3345, "rm": 3346, "rn": 3347, "rng": 3348, "ro": 3349, "rob": 3350, "rod": 3351, "roduct": 3352, "rom": 3353, "romot": 3354, "ronti": 3355, "rontier": 3356, "rontierResult": 3357, "rontier_": 3358, "rontier_log_n_opt": 3359, "ropout": 3360, "ror": 3361, "ross": 3362, "rossAttention": 3363, "ross_": 3364, "rot": 3365, "rotate_half": 3366, "row": 3367, "rows": 3368, "rows.": 3369, "rror": 3370, "ru": 3371, "rue": 3372, "rue(æĹ¢å®ļåĢ¤": 3373, "ry_decode_byte_level_symbol": 3374, "r|": 3375, "s": 3376, "s\"": 3377, "s\",": 3378, "s\":": 3379, "s(": 3380, "s()": 3381, "s)": 3382, "s)),": 3383, "s,": 3384, "s.": 3385, "s.abc": 3386, "s.items()": 3387, "s.reshape(": 3388, "s.to": 3389, "s.tolist": 3390, "s.tolist()": 3391, "s/": 3392, "s:": 3393, "s://": 3394, "s=": 3395, "s[": 3396, "s[i": 3397, "s[idx]": 3398, "s]": 3399, "s])": 3400, "s_": 3401, "s_correction": 3402, "s_like(": 3403, "s_negative_": 3404, "s_q,": 3405, "s_train": 3406, "s_training": 3407, "s_with": 3408, "s_without_leading_whitespace": 3409, "s``": 3410, "sa": 3411, "saturating_fit": 3412, "sc": 3413, "scal": 3414, "scale(\"log\")": 3415, "scaling": 3416, "score": 3417, "scripts/promote_canonical_tokenizers.": 3418, "se": 3419, "seed:": 3420, "segmentation": 3421, "segmentation_a": 3422, "segmentation_agreement_rat": 3423, "sel": 3424, "self": 3425, "self)": 3426, "self,": 3427, "self.": 3428, "self.w_": 3429, "self_": 3430, "self_attn": 3431, "self_attn``": 3432, "sentencepiece": 3433, "seq": 3434, "seq_len": 3435, "seq_len,": 3436, "sequen": 3437, "sequence_length": 3438, "sequence_length:": 3439, "sequences": 3440, "sequences_a": 3441, "ser": 3442, "ser-Agent": 3443, "series_names": 3444, "set": 3445, "sh": 3446, "sion": 3447, "sition": 3448, "size=": 3449, "size=(": 3450, "size_per_head": 3451, "slopes": 3452, "so": 3453, "soFLOP": 3454, "softmax": 3455, "sor": 3456, "sort": 3457, "sorted": 3458, "sorted_": 3459, "split_train_val": 3460, "split_train_val``": 3461, "sqrt(": 3462, "squeeze(": 3463, "src": 3464, "src.": 3465, "src.training.": 3466, "src.training.optimizer``": 3467, "src/": 3468, "src/data/": 3469, "src/data/wikipedia_manifests/": 3470, "ss": 3471, "ss_": 3472, "st": 3473, "stack([": 3474, "stance": 3475, "start": 3476, "stat": 3477, "staticmethod": 3478, "statistic": 3479, "std": 3480, "step": 3481, "step()": 3482, "step()``ãĥ»``zero": 3483, "step)": 3484, "step:": 3485, "step_": 3486, "step_accept": 3487, "step_count": 3488, "str": 3489, "stra": 3490, "strip": 3491, "strip_wikitext_links(": 3492, "sum": 3493, "sum(": 3494, "symbol": 3495, "symbol_to_id": 3496, "symbols": 3497, "symbols)": 3498, "symbols,": 3499, "t": 3500, "t(": 3501, "t)": 3502, "tab:": 3503, "target": 3504, "target_compute_budgets:": 3505, "temperature": 3506, "temperature=0.0": 3507, "tent": 3508, "tention": 3509, "ter": 3510, "ter,": 3511, "tex": 3512, "text": 3513, "text(": 3514, "text)": 3515, "text,": 3516, "text:": 3517, "text{": 3518, "th": 3519, "theories": 3520, "theories/": 3521, "theories/0": 3522, "theories/01_foundations/": 3523, "theories/01_foundations/004_norm": 3524, "theories/01_foundations/004_normalization": 3525, "theories/ãĥ»": 3526, "theories/ãĥ»apps/": 3527, "ther_": 3528, "theta": 3529, "theta_i": 3530, "theta_i^{": 3531, "thon": 3532, "threshold": 3533, "ti": 3534, "tick": 3535, "tie_embedding": 3536, "til": 3537, "tilde{g}_t": 3538, "tim": 3539, "time": 3540, "time.sleep": 3541, "tin": 3542, "tiny": 3543, "tiny_shakespear": 3544, "title": 3545, "title:": 3546, "tj": 3547, "to": 3548, "to(": 3549, "to_": 3550, "to_numpy": 3551, "token": 3552, "token_embedding": 3553, "token_embedding.weight": 3554, "token_id": 3555, "token_ids": 3556, "token_ids)": 3557, "token_ids,": 3558, "token_ids.": 3559, "token_ids:": 3560, "token_ids``": 3561, "tokenizer": 3562, "tokenizer.": 3563, "tokenizer.char_to_id": 3564, "tokenizer.json": 3565, "tokenizer:": 3566, "tokenizer_json": 3567, "tokenizer_json_p": 3568, "tokens:": 3569, "tokens_train": 3570, "top": 3571, "top-": 3572, "top-k": 3573, "topk": 3574, "torch.": 3575, "torch.nn.Parameter": 3576, "torch.no_grad()": 3577, "torch.optim.": 3578, "torch.optim.Optimizer``": 3579, "tot": 3580, "total": 3581, "total_": 3582, "total_bytes)": 3583, "total_bytes:": 3584, "total_eval": 3585, "total_negative_log_likelihood_n": 3586, "total_negative_log_likelihood_nat": 3587, "total_negative_log_likelihood_nats": 3588, "total_step": 3589, "total_steps": 3590, "total_steps``": 3591, "tr": 3592, "train": 3593, "train(": 3594, "train_": 3595, "train_language_model": 3596, "train_language_model``": 3597, "train_loss": 3598, "train_unigram_model": 3599, "trainer": 3600, "training": 3601, "training.": 3602, "training/": 3603, "transformer_block": 3604, "try:": 3605, "try_decode_byte_level_symbol": 3606, "ttention": 3607, "tuple": 3608, "tuple[str,": 3609, "ty": 3610, "type": 3611, "type=": 3612, "u": 3613, "ual": 3614, "ual_seed": 3615, "ual_seed(seed)": 3616, "ub": 3617, "uber": 3618, "uber_": 3619, "uber_delta": 3620, "uber_delta:": 3621, "ucket": 3622, "ucle": 3623, "ucleus": 3624, "uct": 3625, "uct_": 3626, "ud": 3627, "udget": 3628, "ue": 3629, "ugging": 3630, "ul": 3631, "ul_": 3632, "ulary": 3633, "ule": 3634, "ult": 3635, "ult(": 3636, "ulti": 3637, "ulti-start": 3638, "um": 3639, "um_log_prob": 3640, "umerate(": 3641, "ump": 3642, "umpy": 3643, "un": 3644, "unction": 3645, "unctional": 3646, "unctool": 3647, "uniform": 3648, "unigram": 3649, "unigram_model": 3650, "unit_": 3651, "unk": 3652, "unk_token": 3653, "unk_token:": 3654, "unknown_rat": 3655, "unsqueeze(": 3656, "unt": 3657, "up": 3658, "uper": 3659, "uper().__init__()": 3660, "uple": 3661, "upload_": 3662, "ur": 3663, "ure": 3664, "ure_": 3665, "url": 3666, "urllib": 3667, "urlopen": 3668, "urrent": 3669, "urv": 3670, "us": 3671, "us)": 3672, "use_cross_attention": 3673, "use_cross_attention:": 3674, "use_cross_attention=False": 3675, "use_cross_attention=False``": 3676, "ut": 3677, "utf-8": 3678, "util": 3679, "utils.": 3680, "v": 3681, "v_hat": 3682, "v_term": 3683, "val": 3684, "valid": 3685, "valid=\"ignore\"):": 3686, "validation": 3687, "valu": 3688, "value": 3689, "value:": 3690, "value_": 3691, "values": 3692, "values,": 3693, "values_by": 3694, "values_by_condition": 3695, "values_by_condition:": 3696, "values_by_series": 3697, "var": 3698, "ve": 3699, "ve_": 3700, "ve_bpe_id_tokenizer_json": 3701, "ve_character_level_tokenizer_json": 3702, "vector": 3703, "vel": 3704, "venberg": 3705, "venberg_m": 3706, "ver": 3707, "verg": 3708, "verged": 3709, "vergence": 3710, "verse": 3711, "vertex": 3712, "vertex_": 3713, "vertex_log_n": 3714, "vertex_log_n,": 3715, "vertex_loss": 3716, "vi": 3717, "vid": 3718, "view": 3719, "vious_": 3720, "vious_loss": 3721, "vis": 3722, "viterbi_segment": 3723, "vmax": 3724, "vmax:": 3725, "vocab": 3726, "vocab)": 3727, "vocab_size": 3728, "vocab``": 3729, "vocabulary": 3730, "vocabulary_size:": 3731, "w": 3732, "w_": 3733, "wa": 3734, "war": 3735, "ward": 3736, "ward(": 3737, "warm": 3738, "warm_fit": 3739, "warmup": 3740, "warmup_step": 3741, "warmup_steps": 3742, "was_training": 3743, "way": 3744, "ways_negative_": 3745, "weight": 3746, "weight_decay": 3747, "weight_decay:": 3748, "weight_grad": 3749, "weights": 3750, "weights:": 3751, "wh": 3752, "where(": 3753, "while": 3754, "whitespace": 3755, "wiGLU": 3756, "wiGLUFeedForwardNetwork": 3757, "width": 3758, "wik": 3759, "wikipedia": 3760, "wikipedia_corpus": 3761, "wikipedia_manifest": 3762, "wikipedia_manifests/": 3763, "wikitext": 3764, "window": 3765, "wish": 3766, "with": 3767, "wor": 3768, "word": 3769, "x": 3770, "x:": 3771, "x_c": 3772, "x_c,": 3773, "x``": 3774, "xlabel": 3775, "xlabel(xlabel)": 3776, "xlabel:": 3777, "xtick": 3778, "y": 3779, "y(": 3780, "y,": 3781, "y:": 3782, "y_": 3783, "y_arr": 3784, "y_decode_byte_level_symbol": 3785, "y_labels": 3786, "y_length": 3787, "y_length,": 3788, "y_length:": 3789, "yer": 3790, "ylabel": 3791, "ylabel(": 3792, "ylabel)": 3793, "ylabel:": 3794, "ymbol": 3795, "ys": 3796, "yscale(\"log\")": 3797, "yt": 3798, "z": 3799, "zero": 3800, "zeros(": 3801, "zip": 3802, "zorder": 3803, "zorder=": 3804, "{": 3805, "{\"": 3806, "{|": 3807, "|": 3808, "|}": 3809, "|}{|": 3810, "}": 3811, "}\"": 3812, "}\")": 3813, "}\",": 3814, "})": 3815, "},": 3816, "}_t": 3817, "}``": 3818, "}{": 3819, "}}``": 3820, "~": 3821, "¡": 3822, "¡ãĤ¤ãĥ«": 3823, "¡ãĥ³ãĥĪ": 3824, "¡ä»¶": 3825, "¡ä¾ĭ": 3826, "¡çĶ¨": 3827, "¡çĶ¨ãģĻãĤĭ": 3828, "¢": 3829, "¢«": 3830, "¢«è¦": 3831, "¢«è¦Ĩ": 3832, "¢º": 3833, "¢ãĤ¹": 3834, "¢æķ°": 3835, "¢ç´": 3836, "¢ç´¢": 3837, "¢ĥçķĮ": 3838, "£": 3839, "£Ĵ": 3840, "¤": 3841, "¤º": 3842, "¤ĩ": 3843, "¤ľ": 3844, "¤ľè¨¼": 3845, "¥": 3846, "¥µ": 3847, "¥æľ¬": 3848, "¥æľ¬èªŀ": 3849, "¦": 3850, "¦(x)": 3851, "¦åı·": 3852, "¦æ³": 3853, "¦æ³¢": 3854, "¦ç": 3855, "¦çĲĨ": 3856, "§": 3857, "§ãĥ³": 3858, "§ãĥ¼ãĤ¹": 3859, "§ç¸®": 3860, "§ĭ": 3861, "§ĭæĪĲ": 3862, "§ĭç¯ī": 3863, "§Ķ": 3864, "§ĺ": 3865, "¨": 3866, "¨ãģ¦": 3867, "¨æ³": 3868, "¨æ³¢æķ°": 3869, "¨ĵ": 3870, "©": 3871, "©ãĥ¼ãĥ«": 3872, "©ãĥ¼ãĥ«ãĥĲ": 3873, "©ãĥ¼ãĥ«ãĥĲãĥĥãĤ¯": 3874, "©ç·ļ": 3875, "©Ł": 3876, "ª": 3877, "ªçĽ®": 3878, "ªè»¸": 3879, "ªĵ": 3880, "«": 3881, "«å°": 3882, "«å°¾": 3883, "«ĸ": 3884, "«ĺ": 3885, "¬": 3886, "¬¦åı·": 3887, "¬¦åı·åĮĸ": 3888, "¬¬": 3889, "¬æĺİ": 3890, "¬ç«ĭ": 3891, "®": 3892, "®µ": 3893, "®µéļİ": 3894, "®µéļİãģ§": 3895, "®å°ı": 3896, "®å°ıå®ļæķ°": 3897, "®åĪĨ": 3898, "®åĭķ": 3899, "®åĭķå°ı": 3900, "®åĭķå°ıæķ°": 3901, "®åĭķå°ıæķ°çĤ¹": 3902, "®æ¨Ļ": 3903, "®æŃ£": 3904, "®Ĭ": 3905, "®ĭ": 3906, "®ĭå·®": 3907, "¯": 3908, "¯¾": 3909, "¯å±¤": 3910, "¯ç©į": 3911, "¯Ģ": 3912, "¯Ķ": 3913, "¯Ķè¼ĥ": 3914, "°": 3915, "°´": 3916, "°ãĤĬ": 3917, "±": 3918, "±¡": 3919, "±¤": 3920, "±º": 3921, "±ãģĦ": 3922, "±æķ°": 3923, "±èªŀ": 3924, "±éŁ¿": 3925, "±éŁ¿ãĤĴåıĹãģĳ": 3926, "±Ĥ": 3927, "±ĤãĤģ": 3928, "²": 3929, "²åĲĪ": 3930, "²æ³ķ": 3931, "²ç·ļ": 3932, "³": 3933, "³»": 3934, "³»åĪĹ": 3935, "³ãĥ¼ãĥ": 3936, "³å®ļ": 3937, "³å¯": 3938, "³å¯Ĩ": 3939, "³åĿĩ": 3940, "´": 3941, "´»": 3942, "´»æĢ§": 3943, "´»æĢ§åĮĸ": 3944, "´ãĥª": 3945, "µ": 3946, "µ·": 3947, "¶": 3948, "¶æķ°": 3949, "·": 3950, "··": 3951, "·ãģķ": 3952, "·åº¦": 3953, "·ļ": 3954, "·ļå½¢": 3955, "·Ŀ": 3956, "·ĿéĽ¢": 3957, "¸": 3958, "¸¦": 3959, "¹": 3960, "¹°ãĤĬ": 3961, "¹³åĿĩ": 3962, "º": 3963, "º«": 3964, "ºå®ļ": 3965, "»": 3966, "»åº¦": 3967, "»åĭķ": 3968, "»æĦı": 3969, "¼": 3970, "¼ãģ³": 3971, "¼ãģ³åĩº": 3972, "¼ãģ³åĩºãģĹ": 3973, "¼ĥ": 3974, "¼Ķ": 3975, "¼Ķç®Ĺ": 3976, "¼ķæķ°": 3977, "½": 3978, "½®": 3979, "½åĴĮ": 3980, "½åĴĮãģ¹ãģįä¹Ĺåīĩ": 3981, "¾": 3982, "¾¤": 3983, "¾ãģį": 3984, "¾åĢ¤": 3985, "¾çī": 3986, "¾çī©ç·ļ": 3987, "¾éħį": 3988, "¿": 3989, "¿Ĵ": 3990, "À": 3991, "Á": 3992, "Â": 3993, "Ã": 3994, "ÃĹ": 3995, "Ä": 3996, "Å": 3997, "Æ": 3998, "Ç": 3999, "È": 4000, "É": 4001, "Ê": 4002, "Ë": 4003, "Ì": 4004, "Í": 4005, "Î": 4006, "Î²": 4007, "Î²1,": 4008, "Î¸": 4009, "Ï": 4010, "Ïĥ": 4011, "Ïĥ(": 4012, "Ïĥ^2": 4013, "Ð": 4014, "Ñ": 4015, "Ò": 4016, "Ó": 4017, "Ô": 4018, "Õ": 4019, "Ö": 4020, "×": 4021, "Ø": 4022, "Ù": 4023, "Ú": 4024, "Û": 4025, "Ü": 4026, "Ý": 4027, "Þ": 4028, "ß": 4029, "à": 4030, "á": 4031, "â": 4032, "âĸ": 4033, "âĸģ": 4034, "ã": 4035, "ãĢ": 4036, "ãĢģ": 4037, "ãĢģ006": 4038, "ãĢģ007": 4039, "ãĢģ009": 4040, "ãĢģFalse": 4041, "ãĢģ``": 4042, "ãĢģn": 4043, "ãĢģãģĬãĤĪãģ³": 4044, "ãĢģãģĻ": 4045, "ãĢģãģĿãģ®": 4046, "ãĢģãģĿãĤĮ": 4047, "ãĢģä¾ĭ:": 4048, "ãĢģå½¢çĬ¶": 4049, "ãĢģå¾Į": 4050, "ãĢģåĲĮãģĺ": 4051, "ãĢģæľ¬": 4052, "ãĢģæľ¬çķªå®Łè¡Įåīįãģ®ä¿®æŃ£": 4053, "ãĢģéłĤçĤ¹": 4054, "ãĢĤ": 4055, "ãĢĤ\"\"\"": 4056, "ãĢĤ``": 4057, "ãĢĤ``None``": 4058, "ãĢĤåŃĺåľ¨ãģĹãģªãģĦ": 4059, "ãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯": 4060, "ãĢĤæĹ¢å®ļåĢ¤": 4061, "ãĢĮ": 4062, "ãĢĮæľ¬ä½ĵ": 4063, "ãĢį": 4064, "ãĢįãģ¨ãģĦãģĨ": 4065, "ãĢįãģ®": 4066, "ãĢįãĤĴ": 4067, "ãĢľ": 4068, "ãĢľ004": 4069, "ãģ": 4070, "ãģ¡": 4071, "ãģ¡ãĢģ": 4072, "ãģ¡ãĤī": 4073, "ãģ£": 4074, "ãģ£ãģ¦": 4075, "ãģ£ãģ¦ãģĦ": 4076, "ãģ£ãģ¦ãģĦãĤĭ": 4077, "ãģ£ãģ¦ãģĹãģ¾": 4078, "ãģ£ãģ¦ãĤĤ": 4079, "ãģ£ãģŁ": 4080, "ãģ£ãģŁãģ¾ãģ¾": 4081, "ãģ¤": 4082, "ãģ¤ãģ¤": 4083, "ãģ¤ãģĦãģ¦": 4084, "ãģ¤ãģĦãģ¦ãĢģ": 4085, "ãģ¤ãģį": 4086, "ãģ¥": 4087, "ãģ¦": 4088, "ãģ¦ãģ¯ãĤģ": 4089, "ãģ¦ãģĦ": 4090, "ãģ¦ãģĦãĤĭ": 4091, "ãģ¦ãĤĤ": 4092, "ãģ§": 4093, "ãģ§ãĢģ": 4094, "ãģ§ãģªãģĦ": 4095, "ãģ§ãģ®": 4096, "ãģ§ãģ®éĢ£": 4097, "ãģ§ãģ®éĢ£çķª": 4098, "ãģ§ãģ¯": 4099, "ãģ§ãģ¯ãĢģ": 4100, "ãģ§ãģ¯ãģªãģı": 4101, "ãģ§ãģ¯ãģªãģıãĢģ": 4102, "ãģ§ãģĤ": 4103, "ãģ§ãģĤãĤĬãĢģ": 4104, "ãģ§ãģĤãĤĭ": 4105, "ãģ§ãģĤãĤĭãĢĤ": 4106, "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭ": 4107, "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭ:": 4108, "ãģ§ãģĤãĤĭå¿ħè¦ģãģĮãģĤãĤĭãĢĤ": 4109, "ãģ§ãģĤãĤĮãģ°": 4110, "ãģ§ãģį": 4111, "ãģ§ãģįãĢģ": 4112, "ãģ§ãģįãģªãģĦ": 4113, "ãģ§ãģįãĤĭ": 4114, "ãģ§ãģįãĤĭãĤĪãģĨãģ«": 4115, "ãģ§ãĤĤ": 4116, "ãģ§ãĤĪ": 4117, "ãģ§åº": 4118, "ãģ§åºĥ": 4119, "ãģ§å¾ĹãģŁ": 4120, "ãģ§æ¯Ķè¼ĥãģĻãĤĭ": 4121, "ãģ§æııçĶ»ãģĻãĤĭãĢĤ": 4122, "ãģ§è¡ĮãģĨ": 4123, "ãģ¨": 4124, "ãģ¨ãĢģ": 4125, "ãģ¨ãģªãĤĭ": 4126, "ãģ¨ãģ«": 4127, "ãģ¨ãģ®": 4128, "ãģ¨ãģ®ä½µ": 4129, "ãģ¨ãģ¿": 4130, "ãģ¨ãģ¿ãģª": 4131, "ãģ¨ãģĦãģĨ": 4132, "ãģ¨ãģį": 4133, "ãģ¨ãģįãģ®": 4134, "ãģ¨ãģįãģ¯": 4135, "ãģ¨ãģĹ": 4136, "ãģ¨ãģĹãĢģ": 4137, "ãģ¨ãģĹãģ¦": 4138, "ãģ¨ãģĹãģ¦æī±": 4139, "ãģ¨ãĤģ": 4140, "ãģ¨ãĤģãģ¦": 4141, "ãģ¨åĩºåĬĽå±¤": 4142, "ãģ¨åĲĮãģĺ": 4143, "ãģ¨åĲĮä¸Ģ": 4144, "ãģ©": 4145, "ãģ©ãģ¡ãĤī": 4146, "ãģ©ãģ®": 4147, "ãģ©ãģĨ": 4148, "ãģ©ãģĨãģĹ": 4149, "ãģ©ãĤĮ": 4150, "ãģ©ãĤĮãģłãģĳ": 4151, "ãģª": 4152, "ãģªãģ©": 4153, "ãģªãģ©)": 4154, "ãģªãģ©)ãĢĤ": 4155, "ãģªãģ©)ãĤĴ": 4156, "ãģªãģ©ãĤĴ": 4157, "ãģªãģ®": 4158, "ãģªãģĦ": 4159, "ãģªãģĦ(": 4160, "ãģªãģĦ)ãĢĤ": 4161, "ãģªãģĦãĢĤ": 4162, "ãģªãģĦãģŁãĤģãĢģ": 4163, "ãģªãģı": 4164, "ãģªãģıãģ¦ãĤĤ": 4165, "ãģªãģĳãĤĮãģ°": 4166, "ãģªãģĹ": 4167, "ãģªãģĹãģ¨": 4168, "ãģªãģĹãģ®": 4169, "ãģªãĤĬ": 4170, "ãģªãĤĬãĢģ": 4171, "ãģªãĤĭ": 4172, "ãģªãĤĭãĢĤ": 4173, "ãģªãĤı": 4174, "ãģªãĤıãģ¡": 4175, "ãģ«": 4176, "ãģ«ãĢģ": 4177, "ãģ«ãģ¤ãģĦãģ¦": 4178, "ãģ«ãģª": 4179, "ãģ«ãģªãĤĬãĢģ": 4180, "ãģ«ãģªãĤĭ": 4181, "ãģ«ãģªãĤĭ(": 4182, "ãģ«ãģªãĤĭãĢĤ": 4183, "ãģ«ãģ¯": 4184, "ãģ«ãģ¯ãĢģ": 4185, "ãģ«ãģĹãģŁ": 4186, "ãģ«ãģĹãģŁæĿ¡ä»¶": 4187, "ãģ«ãģĻãĤĭ": 4188, "ãģ«ãģĻãĤĭãĢĤ": 4189, "ãģ«ãģĻãĤĭãģĭãĢĤ": 4190, "ãģ«ãģĿãģ®ãģ¾ãģ¾": 4191, "ãģ«ãĤĤ": 4192, "ãģ«ãĤĪ": 4193, "ãģ«ãĤĪãģ£ãģ¦": 4194, "ãģ«ãĤĪãĤīãģļ": 4195, "ãģ«ãĤĪãĤĬ": 4196, "ãģ«ãĤĪãĤĬãĢģ": 4197, "ãģ«ãĤĪãĤĭ": 4198, "ãģ«ãĤı": 4199, "ãģ«ãĤıãģŁ": 4200, "ãģ«ä½¿": 4201, "ãģ«ä½¿ãģ£ãģŁ": 4202, "ãģ«ä½¿ãģĨ": 4203, "ãģ«ä½¿ãģĨ(005": 4204, "ãģ«ä½¿ãģĨ)ãĢĤ": 4205, "ãģ«ä½¿ãģĨãĢĤ": 4206, "ãģ«ä½įç½®": 4207, "ãģ«ä¾ĿåŃĺ": 4208, "ãģ«ä¾ĿåŃĺãģĻãĤĭ": 4209, "ãģ«å¤±æķĹ": 4210, "ãģ«å¤īæıĽ": 4211, "ãģ«å¤īæıĽãģĻãĤĭãĢĤ\"\"\"": 4212, "ãģ«å¯¾": 4213, "ãģ«å¯¾ãģĹãģ¦": 4214, "ãģ«å¯¾ãģĻãĤĭ": 4215, "ãģ«å¯¾å¿ľ": 4216, "ãģ«å¯¾å¿ľãģĻãĤĭ": 4217, "ãģ«å¯¾å¿ľãģĻãĤĭãĢĤ": 4218, "ãģ«å¾ĵ": 4219, "ãģ«å¿ľ": 4220, "ãģ«å¿ľãģĺ": 4221, "ãģ«å¿ľãģĺãģ¦": 4222, "ãģ«åĩºçı¾": 4223, "ãģ«åĩºçı¾ãģĻãĤĭ": 4224, "ãģ«åĪĨåī²": 4225, "ãģ«åĬłç®Ĺ": 4226, "ãģ«åĬłç®ĹãģĻãĤĭ": 4227, "ãģ«åŁºãģ¥ãģı": 4228, "ãģ«åŁºãģ¥ãģıãĢĤ": 4229, "ãģ«æ": 4230, "ãģ«æ³¨æĦı": 4231, "ãģ«æ¸¡ãģĻ": 4232, "ãģ«æ¸¡ãģĻãĥŀãĤ¹ãĤ¯": 4233, "ãģ«éģ©çĶ¨ãģĻãĤĭ": 4234, "ãģ«éģĶãģĻãĤĭ": 4235, "ãģ®": 4236, "ãģ®ãģ¾ãģ¾": 4237, "ãģ®ãģ¿": 4238, "ãģ®ãģ¿ãģ«ä¾ĿåŃĺãģĻãĤĭ": 4239, "ãģ®ãģĨ": 4240, "ãģ®ãģŁãĤģ": 4241, "ãģ®ãģŁãĤģãĢģ": 4242, "ãģ®ãģŁãĤģãģ®å¾": 4243, "ãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°": 4244, "ãģ®ãĤ·ãĥ³ãĥľãĥ«": 4245, "ãģ®ãĤ¿ãĤ¤ãĥĪ": 4246, "ãģ®ãĤ¿ãĤ¤ãĥĪãĥ«ãĢĤ": 4247, "ãģ®ãĥ¦ãĥĭãĥĥãĥĪ": 4248, "ãģ®ãĥªãĤ¹ãĥĪ": 4249, "ãģ®ãĥªãĤ¹ãĥĪ}``": 4250, "ãģ®ãĥªãĤ¹ãĥĪãĢĤ": 4251, "ãģ®ãĥªãĥ³ãĤ¯": 4252, "ãģ®ä¸ĬéĻĲ": 4253, "ãģ®äºĮä¹Ĺ": 4254, "ãģ®ä½įç½®": 4255, "ãģ®å": 4256, "ãģ®å¢ĥçķĮ": 4257, "ãģ®å¤": 4258, "ãģ®å¤§": 4259, "ãģ®å¤§ãģį": 4260, "ãģ®å¤§ãģįãģķ": 4261, "ãģ®å¤ī": 4262, "ãģ®å®ļç¾©": 4263, "ãģ®å®Ł": 4264, "ãģ®å®ŁåĬ¹": 4265, "ãģ®å®Łæķ°åĢ¤": 4266, "ãģ®å®Łæķ°åĢ¤ãģ®ç³»åĪĹ": 4267, "ãģ®å®Łè£ħ": 4268, "ãģ®å¯¾": 4269, "ãģ®å¯¾å¿ľ": 4270, "ãģ®å¯¾æķ°": 4271, "ãģ®å¯¾æķ°å°¤åº¦": 4272, "ãģ®å°": 4273, "ãģ®å°Ħå½±": 4274, "ãģ®å°İ": 4275, "ãģ®å°İåĩº": 4276, "ãģ®å·": 4277, "ãģ®å·®": 4278, "ãģ®å¸": 4279, "ãģ®å¹³åĿĩ": 4280, "ãģ®åº": 4281, "ãģ®å¼": 4282, "ãģ®å¼ķæķ°": 4283, "ãģ®å½": 4284, "ãģ®å½±éŁ¿ãĤĴåıĹãģĳ": 4285, "ãģ®å¾": 4286, "ãģ®å¾Į": 4287, "ãģ®åĢ¤": 4288, "ãģ®åĢ¤ãĤĴ": 4289, "ãģ®åħ¥åĬĽ": 4290, "ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ": 4291, "ãģ®åħ¨": 4292, "ãģ®åĨħ": 4293, "ãģ®åĨħç©įãĤĴ": 4294, "ãģ®åĩ": 4295, "ãģ®åĩ¦çĲĨ": 4296, "ãģ®åĩº": 4297, "ãģ®åĩºåĬĽ": 4298, "ãģ®åĩºåĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ": 4299, "ãģ®åĪĿæľŁåĢ¤": 4300, "ãģ®åī": 4301, "ãģ®åī¯å±¤": 4302, "ãģ®åī²åĲĪ": 4303, "ãģ®åīį": 4304, "ãģ®åĬ": 4305, "ãģ®åĬ¹": 4306, "ãģ®åĭ¾éħį": 4307, "ãģ®åĭ¾éħįãĤĴ": 4308, "ãģ®åĭ¾éħįãĥİãĥ«ãĥł": 4309, "ãģ®åįĺ": 4310, "ãģ®åıįå¾©": 4311, "ãģ®åıįå¾©çļĦãģª": 4312, "ãģ®åıįå¾©çļĦãģªç¸®å°ı": 4313, "ãģ®åıĸå¾Ĺ": 4314, "ãģ®åĲĦ": 4315, "ãģ®åĳ¼ãģ³åĩºãģĹ": 4316, "ãģ®åĴĮ": 4317, "ãģ®åĴĮãģĮ": 4318, "ãģ®åĽ": 4319, "ãģ®åŁ": 4320, "ãģ®åŁº": 4321, "ãģ®åł´åĲĪ": 4322, "ãģ®åł´åĲĪãĢģ": 4323, "ãģ®åł´åĲĪãģ¯": 4324, "ãģ®åŃ¦ç¿Ĵ": 4325, "ãģ®æ": 4326, "ãģ®æ¤ľè¨¼": 4327, "ãģ®æ§ĭç¯ī": 4328, "ãģ®æ¨": 4329, "ãģ®æ¨Ļæºĸ": 4330, "ãģ®æ¨Ļæºĸåģıå·®": 4331, "ãģ®æ¬¡åħĥãĢĤ": 4332, "ãģ®æ®ĭå·®": 4333, "ãģ®æ¯Ķ": 4334, "ãģ®æ¯Ķè¼ĥ": 4335, "ãģ®æ¸Ľè¡°çİĩ": 4336, "ãģ®æī": 4337, "ãģ®æīĭ": 4338, "ãģ®æĮĩæķ°": 4339, "ãģ®æĮĩæķ°ç§»åĭķ": 4340, "ãģ®æĮĩæķ°ç§»åĭķå¹³åĿĩ": 4341, "ãģ®æĲįå¤±": 4342, "ãģ®æķ°": 4343, "ãģ®æĹ": 4344, "ãģ®æĹ¢": 4345, "ãģ®æĹ¢å®ļ": 4346, "ãģ®æľĢå¤§åĢ¤": 4347, "ãģ®æĿ": 4348, "ãģ®æĿ¡ä»¶": 4349, "ãģ®ç³»åĪĹ": 4350, "ãģ®ç³»åĪĹéķ·": 4351, "ãģ®ç³»åĪĹéķ·ä¾ĿåŃĺ": 4352, "ãģ®ç³»åĪĹéķ·ä¾ĿåŃĺéłħ": 4353, "ãģ®çµĦãģ¿åĲĪãĤıãģĽ": 4354, "ãģ®ç·ıåĴĮ": 4355, "ãģ®ç·ļå½¢": 4356, "ãģ®çĲĨè«ĸ": 4357, "ãģ®è§£æŀĲ": 4358, "ãģ®è§£æŀĲãĥĳãĤ¤ãĥĹãĥ©ãĤ¤ãĥ³": 4359, "ãģ®è¨ĺäºĭ": 4360, "ãģ®è¾ŀæĽ¸": 4361, "ãģ®éģķãģĦ": 4362, "ãģ®éķ·ãģķ": 4363, "ãģ®éļłãĤĮ": 4364, "ãģ®éĽĨåĲĪ": 4365, "ãģ®éłĨåºı": 4366, "ãģ¯": 4367, "ãģ¯ãĢģ": 4368, "ãģ¯ãĤģ": 4369, "ãģ¯çĦ¡è¦ĸ": 4370, "ãģ°": 4371, "ãģ°ãĤī": 4372, "ãģ°ãĤīãģ¤ãģį": 4373, "ãģ³": 4374, "ãģ¶": 4375, "ãģ¸": 4376, "ãģ¸ãģ®": 4377, "ãģ¹": 4378, "ãģ¹ãģ¦": 4379, "ãģ¹ãģį": 4380, "ãģ¹ãģįä¹Ĺ": 4381, "ãģ¹ãģįä¹Ĺåīĩ": 4382, "ãģ¹ãģįä¹ĹåīĩãģĤãģ¦ãģ¯ãĤģ": 4383, "ãģ¹ãģŁ": 4384, "ãģ»": 4385, "ãģ»ãģ©": 4386, "ãģ»ãģ¼": 4387, "ãģ¼": 4388, "ãģ¾": 4389, "ãģ¾ãģ¨ãĤģãģ¦": 4390, "ãģ¾ãģ¾": 4391, "ãģ¾ãģŁ": 4392, "ãģ¾ãģŁãģ¯": 4393, "ãģ¾ãĤĭ": 4394, "ãģ¾ãĤĭãģĶãģ¨": 4395, "ãģ¿": 4396, "ãģ¿ãĤĴ": 4397, "ãģĤ": 4398, "ãģĤãģ¦ãģ¯ãĤģ": 4399, "ãģĤãģ¦ãģ¯ãĤģ(": 4400, "ãģĤãģ¦ãģ¯ãĤģãĢģéłĤçĤ¹": 4401, "ãģĤãģŁ": 4402, "ãģĤãģŁãĤĬ": 4403, "ãģĤãģŁãĤĬãģ®": 4404, "ãģĤãģŁãĤĬãģ®è¨Īç®Ĺéĩı": 4405, "ãģĤãĤĭ": 4406, "ãģĦ": 4407, "ãģĦ)ãĢĤ": 4408, "ãģĦãĢģ": 4409, "ãģĦãĢĤ": 4410, "ãģĦãģ¦": 4411, "ãģĦãģĨ": 4412, "ãģĦãģļãĤĮ": 4413, "ãģĦãģļãĤĮãĤĤ": 4414, "ãģĦãģŁ": 4415, "ãģĦãĥ©ãĥĥãĥĳãĥ¼ãĢĤ": 4416, "ãģĦåł´åĲĪãģ¯": 4417, "ãģĨ": 4418, "ãģĨãĢĤ": 4419, "ãģĨãĤĭ": 4420, "ãģĪ": 4421, "ãģĪãĢģ": 4422, "ãģĪãģ°": 4423, "ãģĪãģŁ": 4424, "ãģĪãĤĭ": 4425, "ãģĪãĤĭãĤĪãģĨãģ«": 4426, "ãģĬ": 4427, "ãģĬãĤĪ": 4428, "ãģĬãĤĪãģ³": 4429, "ãģĭ": 4430, "ãģĭ(": 4431, "ãģĭãĢĤ": 4432, "ãģĭãģ£ãģŁ": 4433, "ãģĭãģ¤": 4434, "ãģĭãģ©ãģĨ": 4435, "ãģĭãģ©ãģĨãģĭ": 4436, "ãģĭãĤī": 4437, "ãģĭãĤīãģ®": 4438, "ãģĭãĤīéĻ¤å¤ĸ": 4439, "ãģĭãĤīéłĨãģ«": 4440, "ãģĭãĤĴ": 4441, "ãģĮ": 4442, "ãģĮãĢģ": 4443, "ãģĮãĢĮ": 4444, "ãģĮãģĤ": 4445, "ãģĮãģĤãĤĭãĢĤ": 4446, "ãģĮãĤŃãĥ£ãĥĥãĤ·ãĥ¥": 4447, "ãģĮä¸Ģ": 4448, "ãģĮä¸Ģèĩ´": 4449, "ãģĮä¸Ģèĩ´ãģĹãģ¾": 4450, "ãģĮä¸Ģèĩ´ãģĹãģ¾ãģĽ": 4451, "ãģĮä¸Ģèĩ´ãģĹãģ¾ãģĽãĤĵ": 4452, "ãģĮå¢Ĺ": 4453, "ãģĮå¤§ãģįãģĦ": 4454, "ãģĮå°ıãģķãģĦ": 4455, "ãģĮå°ıãģķãģĦãģ»ãģ©": 4456, "ãģĮçķ°ãģªãĤĭ": 4457, "ãģĮé«ĺãģĦ": 4458, "ãģĮéļĽéĻĲãģªãģı": 4459, "ãģį": 4460, "ãģįæıĽ": 4461, "ãģįæıĽãģĪ": 4462, "ãģı": 4463, "ãģı)ãĢĤ": 4464, "ãģıãĢģ": 4465, "ãģıä½¿ãĤı": 4466, "ãģıä½¿ãĤıãĤĮãĤĭ": 4467, "ãģĲ": 4468, "ãģĳ": 4469, "ãģĳãĤĭ": 4470, "ãģĳãĤĮãģ°": 4471, "ãģĴ": 4472, "ãģĵ": 4473, "ãģĵãģ¨": 4474, "ãģĵãģ¨ãģ§": 4475, "ãģĵãģ¨ãģ¯": 4476, "ãģĵãģ¨ãģĮ": 4477, "ãģĵãģ¨ãģĮæľŁ": 4478, "ãģĵãģ¨ãģĮæľŁå¾ħ": 4479, "ãģĵãģ¨ãģĮæľŁå¾ħãģķãĤĮãĤĭãĢĤ": 4480, "ãģĵãģ¨ãĤĴ": 4481, "ãģĵãģ¨ãĤĴæĦıåĳ³": 4482, "ãģĵãģ®": 4483, "ãģĵãģ®å¼ķæķ°": 4484, "ãģĵãģ®åĢ¤": 4485, "ãģĵãģ®åĢ¤ãĤĴ": 4486, "ãģĵãģ®æ": 4487, "ãģĵãģ®æ¯Ķ": 4488, "ãģĵãģ®éĸ¢æķ°": 4489, "ãģĵãģĵ": 4490, "ãģĵãģĵãģ«": 4491, "ãģĵãĤĮ": 4492, "ãģĵãĤĮãģĮ": 4493, "ãģĵãĤĮãĤĴ": 4494, "ãģĵãĤĮãĤĴè¶ħãģĪãĤĭ": 4495, "ãģĶ": 4496, "ãģĶãģ¨": 4497, "ãģĶãģ¨ãģ«": 4498, "ãģĶãģ¨ãģ®": 4499, "ãģķ": 4500, "ãģķãģªãģĦ": 4501, "ãģķãģĽ": 4502, "ãģķãģĽãĤĭ": 4503, "ãģķãģĽãĤĭ(": 4504, "ãģķãģĽãĤĭãĢĤ": 4505, "ãģķãĤĮ": 4506, "ãģķãĤĮãģŁ": 4507, "ãģķãĤĮãģŁãĥĪãĥ¼ãĤ¯ãĥ³": 4508, "ãģķãĤĮãĤĭ": 4509, "ãģķãĤĮãĤĭãĢĤ": 4510, "ãģĹ": 4511, "ãģĹãĢģ": 4512, "ãģĹãĢģ``": 4513, "ãģĹãĢģæ®ĭãĤĬ": 4514, "ãģĹãģ¦": 4515, "ãģĹãģ¦ãģĦ": 4516, "ãģĹãģ¦ãģĦãĤĭ": 4517, "ãģĹãģ¦ãģĦãĤĭãģŁãĤģãĢģ": 4518, "ãģĹãģ¦ãģĬ": 4519, "ãģĹãģ¦ãģĬãģı": 4520, "ãģĹãģ¦ãģĭãĤī": 4521, "ãģĹãģ¦ãĤĤ": 4522, "ãģĹãģ¦ãĤŃãĥ£ãĥĥãĤ·ãĥ¥": 4523, "ãģĹãģª": 4524, "ãģĹãģªãģĦ": 4525, "ãģĹãģªãģĦ)ãĢĤ": 4526, "ãģĹãģªãģĦãĢĤ": 4527, "ãģĹãģ¾": 4528, "ãģĹãģĨãĤĭ": 4529, "ãģĹãģŁ": 4530, "ãģĹãģŁ)ãĢĤ": 4531, "ãģĹãģŁãģĦ": 4532, "ãģĹãģŁåıįå¾©": 4533, "ãģĹãģŁè¨ĺäºĭ": 4534, "ãģĹãģŁè¨ĺäºĭæľ¬æĸĩ": 4535, "ãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ": 4536, "ãģĹåģ´": 4537, "ãģĺ": 4538, "ãģĻ": 4539, "ãģĻ)ãĢĤ": 4540, "ãģĻãĢĤ": 4541, "ãģĻãģ¹ãģ¦": 4542, "ãģĻãģĵãģ¨": 4543, "ãģĻãĤĭ": 4544, "ãģĻãĤĭ(": 4545, "ãģĻãĤĭ(00": 4546, "ãģĻãĤĭ(006": 4547, "ãģĻãĤĭ(007": 4548, "ãģĻãĤĭ(009": 4549, "ãģĻãĤĭ(``": 4550, "ãģĻãĤĭ)": 4551, "ãģĻãĤĭ)ãĢĤ": 4552, "ãģĻãĤĭãĢĤ": 4553, "ãģĻãĤĭãĢĤ\"\"\"": 4554, "ãģĻãĤĭãģ¨": 4555, "ãģĻãĤĭãģ®": 4556, "ãģĻãĤĭãģĵãģ¨": 4557, "ãģĻãĤĭãģĵãģ¨ãģ§": 4558, "ãģĻãĤĭãģĵãģ¨ãģ§ãĢģ": 4559, "ãģĻãĤĭãģŁãĤģ": 4560, "ãģĻãĤĭãģŁãĤģãĢģ": 4561, "ãģĻãĤĭãģŁãĤģãģ«ä½¿ãģĨãĢĤ": 4562, "ãģĻãĤĭãģŁãĤģãģ®": 4563, "ãģĻãĤĭåŃ¦ç¿Ĵçİĩ": 4564, "ãģĻãĤĭéļĽ": 4565, "ãģĻãĤĮãģ°": 4566, "ãģļ": 4567, "ãģļãĢģ": 4568, "ãģļãģ«": 4569, "ãģļãĤĮ": 4570, "ãģĽ": 4571, "ãģĽãģļ": 4572, "ãģĽãĤĭ": 4573, "ãģľ": 4574, "ãģĿ": 4575, "ãģĿãģ®": 4576, "ãģĿãģ®ãģ¾ãģ¾": 4577, "ãģĿãģ®ãģ¾ãģ¾æ¸¡ãģĻ": 4578, "ãģĿãģ®ãĤĤãģ®": 4579, "ãģĿãĤĮ": 4580, "ãģŁ": 4581, "ãģŁãĤģ": 4582, "ãģŁãĤģ)ãĢĤ": 4583, "ãģŁãĤģãĢģ": 4584, "ãģŁãĤģãĢģ``": 4585, "ãģł": 4586, "ãģłãģĮãĢģ": 4587, "ãģłãģĳ": 4588, "ãģłãģĳãģ§": 4589, "ãģłãģĳãģĮ": 4590, "ãģŃ": 4591, "ãĤ": 4592, "ãĤ¢": 4593, "ãĤ¢ãĤ¦": 4594, "ãĤ¢ãĥ³": 4595, "ãĤ¢ãĥĥãĥĹ": 4596, "ãĤ¢ãĥĥãĥĹãĥŃ": 4597, "ãĤ¢ãĥĥãĥĹãĥŃãĥ¼ãĥī": 4598, "ãĤ£": 4599, "ãĤ£ãĥª": 4600, "ãĤ£ãĥªãĥĨãĤ£": 4601, "ãĤ£ãĥ«": 4602, "ãĤ£ãĥ«ãĤ¿": 4603, "ãĤ£ãĥ³ãĤ°": 4604, "ãĤ¤": 4605, "ãĤ¤ãĤ": 4606, "ãĤ¤ãĤ¢ãĤ¹": 4607, "ãĤ¤ãĤ¶": 4608, "ãĤ¤ãĤ¹ãĢĤ": 4609, "ãĤ¤ãĤº": 4610, "ãĤ¤ãĥ¤": 4611, "ãĤ¤ãĥ¤ãĥ¼": 4612, "ãĤ¤ãĥ«": 4613, "ãĤ¤ãĥ³": 4614, "ãĤ¤ãĥ³ãĤ¹ãĤ¿ãĥ³ãĤ¹": 4615, "ãĤ¤ãĥ³ãĤ¿": 4616, "ãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ": 4617, "ãĤ¤ãĥ³ãĤ¿ãĥ¼ãĥķãĤ§ãĥ¼ãĤ¹": 4618, "ãĤ¤ãĥ³ãĥĩ": 4619, "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯": 4620, "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹": 4621, "ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:": 4622, "ãĤ¤ãĥĪ": 4623, "ãĤ¦": 4624, "ãĤ¦ãĥ³ãĥ": 4625, "ãĤ¦ãĥ³ãĥī": 4626, "ãĤ¦ãĥ³ãĥīä¿®æŃ£": 4627, "ãĤ¨": 4628, "ãĤ¨ãĤ¤": 4629, "ãĤ¨ãĤ¹ãĥĪ": 4630, "ãĤ¨ãĥ©": 4631, "ãĤ¨ãĥ©ãĥ¼": 4632, "ãĤ¨ãĥ³ãĤ": 4633, "ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥ": 4634, "ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°": 4635, "ãĤª": 4636, "ãĤªãĥ¼ãĥĲ": 4637, "ãĤªãĥ¼ãĥĲãĥ¼ãĥķãĥŃãĥ¼": 4638, "ãĤ«": 4639, "ãĤ«ãĥ©": 4640, "ãĤ«ãĥ©ãĥ¼": 4641, "ãĤ«ãĥĨ": 4642, "ãĤ«ãĥĨãĤ´ãĥª": 4643, "ãĤ¯": 4644, "ãĤ¯ãĤ¨ãĤ¹ãĥĪ": 4645, "ãĤ¯ãĤ·ãĥ§ãĥ³": 4646, "ãĤ¯ãĥ©": 4647, "ãĤ¯ãĥ©ãĤ¹": 4648, "ãĤ¯ãĥª": 4649, "ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°": 4650, "ãĤ¯ãĥªãĥĥãĥĶãĥ³ãĤ°éģ©çĶ¨åīį": 4651, "ãĤ¯ãĥªãĥĥãĥĹ": 4652, "ãĤ¯ãĥĪ": 4653, "ãĤ¯ãĥĪãĥª": 4654, "ãĤ¯ãĥĪãĥ«": 4655, "ãĤ°": 4656, "ãĤ°ãĥ©ãĥ": 4657, "ãĤ°ãĥ©ãĥķ": 4658, "ãĤ°ãĥª": 4659, "ãĤ°ãĥªãĥĥãĥī": 4660, "ãĤ±": 4661, "ãĤ±ãĥĥãĥĪ": 4662, "ãĤ³": 4663, "ãĤ³ãĤ¹ãĥĪ": 4664, "ãĤ³ãĤ¹ãĥĪãģĮ": 4665, "ãĤ³ãĥ¼ãĥ": 4666, "ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°": 4667, "ãĤ³ãĥ¼ãĥī": 4668, "ãĤ³ãĥ¼ãĥĳ": 4669, "ãĤ³ãĥ¼ãĥĳãĤ¹": 4670, "ãĤ³ãĥ¼ãĥĳãĤ¹ãĤĴåıĸå¾Ĺ": 4671, "ãĤ³ãĥĵ": 4672, "ãĤ³ãĥĵãĤ¢ãĥ³": 4673, "ãĤ´ãĥª": 4674, "ãĤµ": 4675, "ãĤµãĤ¤ãĤº": 4676, "ãĤµãĥ³ãĥĹ": 4677, "ãĤµãĥ³ãĥĹãĥªãĥ³ãĤ°": 4678, "ãĤµãĥ¼ãĥ": 4679, "ãĤµãĥ¼ãĥģ": 4680, "ãĤµãĥĸ": 4681, "ãĤµãĥĸãĤ¯ãĥ©ãĤ¹": 4682, "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼": 4683, "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽ": 4684, "ãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ": 4685, "ãĤµãĥĸãĥ¯": 4686, "ãĤµãĥĸãĥ¯ãĥ¼ãĥī": 4687, "ãĤµãĥĸãĥ¯ãĥ¼ãĥīåĪĨåī²": 4688, "ãĤ·": 4689, "ãĤ·ãĥ¥": 4690, "ãĤ·ãĥ§ãĥ³": 4691, "ãĤ·ãĥ³ãĥ": 4692, "ãĤ·ãĥ³ãĥľãĥ«": 4693, "ãĤ·ãĥ³ãĥľãĥ«ãģ®": 4694, "ãĤ·ãĥ¼ãĥī": 4695, "ãĤ·ãĥ¼ãĥīãģĶãģ¨ãģ®": 4696, "ãĤ·ãĥ¼ãĥīéĸĵ": 4697, "ãĤ¸": 4698, "ãĤ¸ãĥ¥": 4699, "ãĤ¸ãĥ¥ãĥ¼ãĥ«": 4700, "ãĤ¸ãĥ§ãĥ³": 4701, "ãĤ¹": 4702, "ãĤ¹ãĢĤ": 4703, "ãĤ¹ãĤ¯": 4704, "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ": 4705, "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãĢĤ": 4706, "ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħãģĻãĤĭãĢĤ": 4707, "ãĤ¹ãĤ±": 4708, "ãĤ¹ãĤ±ãĤ¸ãĥ¥ãĥ¼ãĥ«": 4709, "ãĤ¹ãĤ±ãĥ¼": 4710, "ãĤ¹ãĤ±ãĥ¼ãĥªãĥ³ãĤ°": 4711, "ãĤ¹ãĤ±ãĥ¼ãĥªãĥ³ãĤ°åīĩ": 4712, "ãĤ¹ãĤ±ãĥ¼ãĥ«": 4713, "ãĤ¹ãĤ³": 4714, "ãĤ¹ãĤ³ãĤ¢": 4715, "ãĤ¹ãĤ¿": 4716, "ãĤ¹ãĤ¿ãĤ¤ãĥ«": 4717, "ãĤ¹ãĤ¿ãĥ³ãĤ¹": 4718, "ãĤ¹ãĤŃ": 4719, "ãĤ¹ãĥĨ": 4720, "ãĤ¹ãĥĨãĥĥãĥĹ": 4721, "ãĤ¹ãĥĨãĥĥãĥĹãģ§": 4722, "ãĤ¹ãĥĨãĥĥãĥĹãģ¨ãģ®": 4723, "ãĤ¹ãĥĨãĥĥãĥĹæĲįå¤±": 4724, "ãĤ¹ãĥĨãĥĥãĥĹæķ°": 4725, "ãĤ¹ãĥĨãĥĥãĥĹçķªåı·": 4726, "ãĤ¹ãĥĪ": 4727, "ãĤ¹ãĥĪãĥ©": 4728, "ãĤ¹ãĥĪãĥ©ãĥĥãĥĹ": 4729, "ãĤº": 4730, "ãĤ»": 4731, "ãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³": 4732, "ãĤ»ãĥ«": 4733, "ãĤ»ãĥĥãĥĪ": 4734, "ãĤ½": 4735, "ãĤ¿": 4736, "ãĤ¿ãĤ¤": 4737, "ãĤ¿ãĤ¤ãĥĪ": 4738, "ãĤ¿ãĤ¤ãĥĪãĥ«": 4739, "ãĤ¿ãĤ¤ãĥł": 4740, "ãĤ¿ãĤ¤ãĥłãĤ¢ãĤ¦": 4741, "ãĤ¿ãĤ¤ãĥłãĤ¢ãĤ¦ãĥĪ": 4742, "ãĤ¿ãĥĸãĥ»": 4743, "ãĤ¿ãĥĸãĥ»æĶ¹è¡Į": 4744, "ãĤ¿ãĥĹ": 4745, "ãĤĢ": 4746, "ãĤģ": 4747, "ãĤģãĤĭ": 4748, "ãĤĤ": 4749, "ãĤĤãģ®": 4750, "ãĤĦ": 4751, "ãĤĦãģĻ": 4752, "ãĤĦãģĻãģĦ": 4753, "ãĤĪ": 4754, "ãĤĪãģĨ": 4755, "ãĤĪãģĨãģª": 4756, "ãĤĪãģĨãģ«": 4757, "ãĤĪãĤĬ": 4758, "ãĤĪãĤĬä¸Ĭä½į": 4759, "ãĤĪãĤĬä¸Ĭä½įãģ®": 4760, "ãĤĪãĤĭ": 4761, "ãĤī": 4762, "ãĤīãģªãģĦ)ãĢĤ": 4763, "ãĤīãģļ": 4764, "ãĤīãĤĮ": 4765, "ãĤīãĤĮãģŁ": 4766, "ãĤīãĤĮãĤĭ": 4767, "ãĤīãĤĮãĤĭãĢĤ": 4768, "ãĤĬ": 4769, "ãĤĬãĢģ": 4770, "ãĤĬå½ĵ": 4771, "ãĤĬå½ĵãģ¦": 4772, "ãĤĭ": 4773, "ãĤĭ(": 4774, "ãĤĭ(Hoffmann": 4775, "ãĤĭ(``": 4776, "ãĤĭãĢĤ": 4777, "ãĤĭãĢĤãģĵãģ®": 4778, "ãĤĭãģĵãģ¨": 4779, "ãĤĭãģĵãģ¨ãģ§ãĢģ": 4780, "ãĤĮ": 4781, "ãĤĮãģªãģĦ": 4782, "ãĤĮãģªãģĦãĢĤ": 4783, "ãĤĮãģ°": 4784, "ãĤĮãĤĭ": 4785, "ãĤĮãĤĭ)ãĢĤ": 4786, "ãĤĮåĢ¤": 4787, "ãĤı": 4788, "ãĤıãģªãģĦ": 4789, "ãĤıãģĽ": 4790, "ãĤıãĤĮ": 4791, "ãĤĴ": 4792, "ãĤĴãĢģ": 4793, "ãĤĴãģ¾ãģŁ": 4794, "ãĤĴãģ¾ãģŁãģĦ": 4795, "ãĤĴãģ¾ãģŁãģĲ": 4796, "ãĤĴãģĿãģ®ãģ¾ãģ¾": 4797, "ãĤĴä¸ĢåĪĩ": 4798, "ãĤĴä¸İ": 4799, "ãĤĴä¸İãģĪãĤĭ": 4800, "ãĤĴä½¿": 4801, "ãĤĴä½¿ãģĨ": 4802, "ãĤĴä½ľ": 4803, "ãĤĴä¿Ŀ": 4804, "ãĤĴä¿Ŀãģ£ãģŁãģ¾ãģ¾": 4805, "ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«": 4806, "ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ": 4807, "ãĤĴåħ±æľī": 4808, "ãĤĴåĬłãģĪãĤĭ": 4809, "ãĤĴåıĤçħ§)ãĢĤ": 4810, "ãĤĴåıĸå¾Ĺ": 4811, "ãĤĴåıĹãģĳ": 4812, "ãĤĴåĲ«": 4813, "ãĤĴåĲ«ãĤĢ": 4814, "ãĤĴåĲ«ãĤģ": 4815, "ãĤĴåĽºå®ļ": 4816, "ãĤĴæ": 4817, "ãĤĴæ¤ľè¨¼": 4818, "ãĤĴæ®ĭ": 4819, "ãĤĴæ¯Ķè¼ĥ": 4820, "ãĤĴæ±ĤãĤģ": 4821, "ãĤĴæ¸¡": 4822, "ãĤĴæ¸¡ãģĻãĢĤ": 4823, "ãĤĴæĥ³å®ļ": 4824, "ãĤĴæĥ³å®ļ)ãĢĤ": 4825, "ãĤĴæĦı": 4826, "ãĤĴæĦıåĳ³": 4827, "ãĤĴæĬ": 4828, "ãĤĴæĬĳ": 4829, "ãĤĴæĮģãģ¤": 4830, "ãĤĴæĮģãģŁ": 4831, "ãĤĴæĮģãģŁãģªãģĦ": 4832, "ãĤĴæİ": 4833, "ãĤĴæĽ": 4834, "ãĤĴçĶŁæĪĲ": 4835, "ãĤĴè¡¨": 4836, "ãĤĴè¡¨ãģĻ": 4837, "ãĤĴè¡Į": 4838, "ãĤĴè¡ĮãģĨ": 4839, "ãĤĴè¡ĮãģĨãĢĤ": 4840, "ãĤĴè¡ĮãģĨçī¹å¾´": 4841, "ãĤĴè¡ĮãĤıãģªãģĦ": 4842, "ãĤĴè¦ĭ": 4843, "ãĤĴè¨Īç®Ĺ": 4844, "ãĤĴè¨Īç®ĹãģĻãĤĭ(007": 4845, "ãĤĴè¨Īç®ĹãģĻãĤĭãĢĤ": 4846, "ãĤĴè©¦": 4847, "ãĤĴè¿ĶãģĻ": 4848, "ãĤĴè¿ĶãģĻãĢĤ": 4849, "ãĤĴéĢ£çµĲãģĹãģŁ": 4850, "ãĤĴéĺ²ãģĲ": 4851, "ãĤĴéĻ¤": 4852, "ãĤĵ": 4853, "ãĤĵãģł": 4854, "ãĤŃ": 4855, "ãĤŃãĥ£": 4856, "ãĤŃãĥ£ãĤ¹ãĥĪ": 4857, "ãĤŃãĥ£ãĥĥãĤ·ãĥ¥": 4858, "ãĤŃãĥ¼": 4859, "ãĥ": 4860, "ãĥ¡": 4861, "ãĥ¡ãĤ¤ãĥ³": 4862, "ãĥ¡ãĥ¢": 4863, "ãĥ¢": 4864, "ãĥ¢ãĤ¸ãĥ¥ãĥ¼ãĥ«": 4865, "ãĥ¢ãĤ¸ãĥ¥ãĥ¼ãĥ«ãģ¯": 4866, "ãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ": 4867, "ãĥ¢ãĥĩãĥ«": 4868, "ãĥ¢ãĥĩãĥ«ãģ®": 4869, "ãĥ£": 4870, "ãĥ£ãĥ³ãĤ¯": 4871, "ãĥ¤": 4872, "ãĥ¤ãĤ³ãĥĵãĤ¢ãĥ³": 4873, "ãĥ¥": 4874, "ãĥ¦": 4875, "ãĥ¦ãĥ¼ãĥ": 4876, "ãĥ¦ãĥ¼ãĥĨ": 4877, "ãĥ¦ãĥ¼ãĥĨãĤ£ãĥªãĥĨãĤ£": 4878, "ãĥ¦ãĥĭ": 4879, "ãĥ¦ãĥĭãĥ¼ãĤ¯": 4880, "ãĥ¦ãĥĭãĥĥãĥĪ": 4881, "ãĥ§ãĥ³": 4882, "ãĥ©": 4883, "ãĥ©ãĤ¤ãĥ³": 4884, "ãĥ©ãĥ": 4885, "ãĥ©ãĥ¡": 4886, "ãĥ©ãĥ³ãĥĢ": 4887, "ãĥ©ãĥ³ãĥĢãĥł": 4888, "ãĥ©ãĥĥãĥģ": 4889, "ãĥ©ãĥĥãĥģå®Łè£ħ": 4890, "ãĥ©ãĥĥãĥĳ": 4891, "ãĥ©ãĥĥãĥĳãĥ¼ãĢĤ": 4892, "ãĥ©ãĥĻ": 4893, "ãĥ©ãĥĻãĥ«ãĢĤ": 4894, "ãĥª": 4895, "ãĥªãĤ¢": 4896, "ãĥªãĤ¯ãĤ¨ãĤ¹ãĥĪ": 4897, "ãĥªãĤ¹ãĥĪ": 4898, "ãĥªãĥ": 4899, "ãĥªãĥ³ãĤ¯": 4900, "ãĥªãĥ³ãĤ°": 4901, "ãĥªãĥĵ": 4902, "ãĥªãĥĵãĤ¸ãĥ§ãĥ³": 4903, "ãĥªãĥĿ": 4904, "ãĥªãĥĿãĤ¸": 4905, "ãĥªãĥĿãĤ¸ãĥĪãĥª": 4906, "ãĥ«": 4907, "ãĥ«ãĢĤ": 4908, "ãĥ«ãĥ»": 4909, "ãĥ«ãĥ¼ãĥ": 4910, "ãĥ«ãĥ¼ãĥĹ": 4911, "ãĥ«ãĥģ": 4912, "ãĥ«ãĥĨãĤ£": 4913, "ãĥ«ãĥł": 4914, "ãĥ¬": 4915, "ãĥ¬ãĤ¤ãĥ¤ãĥ¼": 4916, "ãĥ¬ãĤ¯ãĥĪãĥª": 4917, "ãĥ¬ãĥ¼ãĥĪ": 4918, "ãĥ¬ãĥĻ": 4919, "ãĥ¬ãĥĻãĥ«": 4920, "ãĥ¯": 4921, "ãĥ¯ãĥ¼ãĤ¯": 4922, "ãĥ³": 4923, "ãĥ³ãĤ": 4924, "ãĥ³ãĤ¯": 4925, "ãĥ³ãĤ°": 4926, "ãĥ³ãĤ¹": 4927, "ãĥ³ãĤ½": 4928, "ãĥ³ãĥ": 4929, "ãĥ³ãĥĢ": 4930, "ãĥ³ãĥĨãĤ£": 4931, "ãĥ³ãĥĪ": 4932, "ãĥ³ãĥĹ": 4933, "ãĥ³ãĥĹãĥ¬ãĥ¼ãĥĪ": 4934, "ãĥ»": 4935, "ãĥ»00": 4936, "ãĥ»002": 4937, "ãĥ»2": 4938, "ãĥ»ALiBi": 4939, "ãĥ»D": 4940, "ãĥ»Key": 4941, "ãĥ»``": 4942, "ãĥ»å®Łé¨ĵ": 4943, "ãĥ»åĲĦ": 4944, "ãĥ»åĲĦãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ": 4945, "ãĥ»æ¤ľè¨¼": 4946, "ãĥ¼": 4947, "ãĥ¼ãĢĤ": 4948, "ãĥ¼ãĤ¯": 4949, "ãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ": 4950, "ãĥ¼ãĤ¸": 4951, "ãĥ¼ãĤ¹": 4952, "ãĥ¼ãĤ¹ãģ®åıįå¾©çļĦãģªç¸®å°ı": 4953, "ãĥ¼ãĤ¿": 4954, "ãĥ¼ãĥ": 4955, "ãĥ¼ãĥ¡ãĥ³ãĥĪ": 4956, "ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ": 4957, "ãĥ¼ãĥ«": 4958, "ãĥ¼ãĥĪ": 4959, "ãĥ¼ãĥĪãĤ¹ãĥĪãĥ©ãĥĥãĥĹ": 4960, "ãĥ¼ãĥĪãĥĸ": 4961, "ãĥ¼ãĥĪãĥĸãĥĥãĤ¯": 4962, "ãĥ¼ãĥī": 4963, "ãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪ": 4964, "ãĥ¼ãĥĲ": 4965, "ãĥ¼ãĥĲãĥ¼": 4966, "ãĥ¼ãĥķ": 4967, "ãĥ¼ãĥķãĤ": 4968, "ãĥ¼ãĥķãĥŃ": 4969, "ãĥ¼ãĥķãĥŃãĥ¼": 4970, "ãĥ¼ãĥŀ": 4971, "ãĥ¼ãĥł": 4972, "ãĥ¼ãĥłãĤµãĥ¼ãĥģ": 4973, "ãĥĢ": 4974, "ãĥģ": 4975, "ãĥģãĥ£ãĥ³ãĤ¯": 4976, "ãĥģãĥ£ãĥ³ãĤ¯ãģ®": 4977, "ãĥģãĥ£ãĥ³ãĤ¯éķ·": 4978, "ãĥĥ": 4979, "ãĥĥãĤ¯": 4980, "ãĥĥãĤ·ãĥ¥": 4981, "ãĥĥãĥ": 4982, "ãĥĥãĥģ": 4983, "ãĥĥãĥĪ": 4984, "ãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯": 4985, "ãĥĥãĥī": 4986, "ãĥĥãĥĳ": 4987, "ãĥĥãĥĶ": 4988, "ãĥĥãĥĶãĥ³ãĤ°": 4989, "ãĥĥãĥĹ": 4990, "ãĥĨ": 4991, "ãĥĨãĤ£": 4992, "ãĥĨãĤŃ": 4993, "ãĥĨãĤŃãĤ¹ãĥĪ": 4994, "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵ": 4995, "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵãģ®": 4996, "ãĥĨãĤŃãĤ¹ãĥĪåħ¨ä½ĵãĤĴ": 4997, "ãĥĨãĥ³ãĤ½": 4998, "ãĥĨãĥ³ãĤ½ãĥ«": 4999, "ãĥĨãĥ³ãĤ½ãĥ«ãĢĤ": 5000, "ãĥĨãĥ³ãĥĹãĥ¬ãĥ¼ãĥĪ": 5001, "ãĥĩ": 5002, "ãĥĩãĤ£": 5003, "ãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥª": 5004, "ãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯": 5005, "ãĥĩãĤ£ãĥ³ãĤ°": 5006, "ãĥĩãĤ³ãĥ¼ãĥī": 5007, "ãĥĩãĥ«": 5008, "ãĥĩãĥ¼ãĤ¿": 5009, "ãĥĩãĥĲ": 5010, "ãĥĩãĥĲãĤ¤ãĤ¹ãĢĤ": 5011, "ãĥĪ": 5012, "ãĥĪãĥª": 5013, "ãĥĪãĥªãĥĥãĤ¯": 5014, "ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ": 5015, "ãĥĪãĥ¼ãĤ¯": 5016, "ãĥĪãĥ¼ãĤ¯ãĥ³": 5017, "ãĥĪãĥ¼ãĤ¯ãĥ³ãģ®": 5018, "ãĥĪãĥ¼ãĤ¯ãĥ³ãĤĴ": 5019, "ãĥĪãĥ¼ãĤ¯ãĥ³å¢ĥçķĮ": 5020, "ãĥĪãĥ¼ãĤ¯ãĥ³åĮĸ": 5021, "ãĥĪãĥ¼ãĤ¯ãĥ³åŁĭãĤģè¾¼ãģ¿": 5022, "ãĥĪãĥ¼ãĤ¯ãĥ³æķ°": 5023, "ãĥĪãĥ¼ãĤ¯ãĥ³æķ°ãĢĤ": 5024, "ãĥĪãĥ¼ãĤ¯ãĥĬ": 5025, "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶": 5026, "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶ãĢĤ": 5027, "ãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶æĿ¡ä»¶": 5028, "ãĥĪãĥĶ": 5029, "ãĥĪãĥĶãĥĥãĤ¯": 5030, "ãĥī": 5031, "ãĥīãĥ¡ãĤ¤ãĥ³": 5032, "ãĥĬ": 5033, "ãĥĬãĥ«ãĥĨãĤ£": 5034, "ãĥĭ": 5035, "ãĥį": 5036, "ãĥįãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯": 5037, "ãĥİ": 5038, "ãĥİãĤ¤ãĤº": 5039, "ãĥİãĤ¤ãĤºåº": 5040, "ãĥİãĤ¤ãĤºåºĬ": 5041, "ãĥİãĥ«ãĥł": 5042, "ãĥİãĥ¼ãĥĪãĥĸãĥĥãĤ¯": 5043, "ãĥı": 5044, "ãĥĲ": 5045, "ãĥĲãĤ¤ãĤ¢ãĤ¹": 5046, "ãĥĲãĤ¤ãĤ¢ãĤ¹ãĤĴåĬłãģĪãĤĭ": 5047, "ãĥĲãĤ¤ãĤ¢ãĤ¹éłħ": 5048, "ãĥĲãĤ¤ãĥĪ": 5049, "ãĥĲãĤ¤ãĥĪãĥ¬ãĥĻãĥ«": 5050, "ãĥĲãĤ¤ãĥĪåĪĹ": 5051, "ãĥĲãĤ±ãĥĥãĥĪ": 5052, "ãĥĲãĤ±ãĥĥãĥĪåĮĸ": 5053, "ãĥĲãĥĥãĥģ": 5054, "ãĥĲãĥĥãĥģãĤµãĤ¤ãĤº": 5055, "ãĥĳ": 5056, "ãĥĳãĤ¤": 5057, "ãĥĳãĤ¤ãĥĹ": 5058, "ãĥĳãĤ¤ãĥĹãĥ©ãĤ¤ãĥ³": 5059, "ãĥĳãĤ¹": 5060, "ãĥĳãĥ©ãĥ¡": 5061, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿": 5062, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ãĢģå½¢çĬ¶": 5063, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ãĤĴæĮģãģŁãģªãģĦ": 5064, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°": 5065, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æĽ´æĸ°": 5066, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æĽ´æĸ°å¼ı": 5067, "ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ç¾¤": 5068, "ãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ": 5069, "ãĥĳãĥĩãĤ£ãĥ³ãĤ°": 5070, "ãĥĳãĥĩãĤ£ãĥ³ãĤ°ãĥŀãĤ¹ãĤ¯": 5071, "ãĥĳãĥĩãĤ£ãĥ³ãĤ°ä½įç½®": 5072, "ãĥĵ": 5073, "ãĥĵãĥ¼ãĥłãĤµãĥ¼ãĥģ": 5074, "ãĥĶ": 5075, "ãĥķ": 5076, "ãĥķãĤ": 5077, "ãĥķãĤ¡ãĤ¤ãĥ«": 5078, "ãĥķãĤ¡ãĤ¤ãĥ«åĲį": 5079, "ãĥķãĤ£ãĥ«ãĤ¿": 5080, "ãĥķãĤ©ãĥ¼ãĥ«ãĥĲãĥĥãĤ¯": 5081, "ãĥķãĥŃãĥ³ãĥĨãĤ£ãĤ¢": 5082, "ãĥĸ": 5083, "ãĥĸãĥ»": 5084, "ãĥĸãĥ¼ãĥĪãĤ¹ãĥĪãĥ©ãĥĥãĥĹ": 5085, "ãĥĸãĥŃ": 5086, "ãĥĸãĥŃãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪ": 5087, "ãĥĸãĥŃãĥ¼ãĥīãĤŃãĥ£ãĤ¹ãĥĪåı¯èĥ½": 5088, "ãĥĸãĥŃãĥĥãĤ¯": 5089, "ãĥĹ": 5090, "ãĥĺãĥĥãĥī": 5091, "ãĥĺãĥĥãĥīæķ°": 5092, "ãĥĻ": 5093, "ãĥĻãĤ¯ãĥĪãĥ«": 5094, "ãĥļ": 5095, "ãĥļãĤ¢": 5096, "ãĥļãĥĬãĥ«ãĥĨãĤ£": 5097, "ãĥŀ": 5098, "ãĥŀãĤ¹ãĤ¯": 5099, "ãĥŀãĥ«ãĥģ": 5100, "ãĥŀãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ": 5101, "ãĥŀãĥ¼ãĤ¸": 5102, "ãĥŀãĥ¼ãĤ¸è¦ıåīĩ": 5103, "ãĥŀãĥĥãĥĹ": 5104, "ãĥŁ": 5105, "ãĥł": 5106, "ãĥŃ": 5107, "ãĥŃãĥ³ãĥĨãĤ£": 5108, "ãĥŃãĥ³ãĥĨãĤ£ãĤ¢": 5109, "ãĥŃãĥķãĤ¡ãĤ¤ãĥ«": 5110, "ä": 5111, "ä¸": 5112, "ä¸¡": 5113, "ä¸¡æĸ¹": 5114, "ä¸¦": 5115, "ä¸¦ãģ¹ãģŁ": 5116, "ä¸¸": 5117, "ä¸¸ãĤģ": 5118, "ä¸»": 5119, "ä¸»å¼µ": 5120, "ä¸Ģ": 5121, "ä¸ĢåĪĩ": 5122, "ä¸Ģèĩ": 5123, "ä¸Ģèĩ´": 5124, "ä¸Ģèĩ´ãģĻãĤĭ": 5125, "ä¸ĢèĪ": 5126, "ä¸ĢèĪ¬": 5127, "ä¸Ĭ": 5128, "ä¸Ĭãģ®": 5129, "ä¸ĬãģĴ": 5130, "ä¸Ĭä½į": 5131, "ä¸Ĭæĺ": 5132, "ä¸Ĭæĺĩ": 5133, "ä¸Ĭæĺĩå¹ħ": 5134, "ä¸ĬéĻĲ": 5135, "ä¸ĭ": 5136, "ä¸ĭéĻĲ": 5137, "ä¸į": 5138, "ä¸įä¸Ģèĩ´": 5139, "ä¸įå¤ī": 5140, "ä¸įæŃ£": 5141, "ä¸įæŃ£ãģª": 5142, "ä¸įè¦ģ": 5143, "ä¸įèĥ½": 5144, "ä¸įèĥ½ãģª": 5145, "ä¸İ": 5146, "ä¸İãģĪ": 5147, "ä¸Ń": 5148, "ä¸Ńãģ®": 5149, "ä¸Ńå¿": 5150, "ä¸Ńå¿ĥ": 5151, "ä¸Ńå¿ĥå·®åĪĨ": 5152, "ä¸Ńå¿ĥå·®åĪĨãģ«ãĤĪãĤĭ": 5153, "ä¸Ńéĸĵ": 5154, "ä¸Ńéĸĵå±¤": 5155, "ä¸Ńéĸĵæ¬¡åħĥ": 5156, "ä¹": 5157, "ä¹±æķ°": 5158, "ä¹±æķ°ãĤ·ãĥ¼ãĥī": 5159, "ä¹ĸ": 5160, "ä¹ĸéĽ¢": 5161, "ä¹Ĺ": 5162, "äº": 5163, "äº¤": 5164, "äº¤å·®": 5165, "äº¤å·®æ³¨æĦı": 5166, "äºĨ": 5167, "äºĪ": 5168, "äºĪæ¸¬": 5169, "äºĪç®Ĺ": 5170, "äºĭ": 5171, "äºĭåīį": 5172, "äºĭåīįåĪĨåī²": 5173, "äºĭåīįåĪĨåī²(pre-tokenization": 5174, "äºĮ": 5175, "äºĮä¹Ĺ": 5176, "äºĮä¹Ĺå¹³åĿĩ": 5177, "äºĴ": 5178, "äºĴæıĽ": 5179, "äºĴæıĽæĢ§": 5180, "ä»": 5181, "ä»£": 5182, "ä»¥": 5183, "ä»¥ä¸ĭ": 5184, "ä»¥å¤ĸ": 5185, "ä»¥å¤ĸãģ®": 5186, "ä»¥éĻį": 5187, "ä»¶": 5188, "ä»»æĦı": 5189, "ä»»æĦıå½¢çĬ¶": 5190, "ä»»æĦıå½¢çĬ¶ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ": 5191, "ä»ĺ": 5192, "ä»ĺãģį": 5193, "ä»ĺãģĳãĤĭ": 5194, "ä»ĺä¸İ": 5195, "ä¼": 5196, "ä¼¼": 5197, "ä¼Ŀ": 5198, "ä¼ĿãĤı": 5199, "ä¼Ŀæ": 5200, "ä¼ĿæĴ": 5201, "ä¼ĿæĴŃ": 5202, "ä½": 5203, "ä½µ": 5204, "ä½¿": 5205, "ä½¿ãģĨ": 5206, "ä½¿ãģĨ(": 5207, "ä½¿ãģĨ)ãĢĤ": 5208, "ä½¿ãģĨãĢĤ": 5209, "ä½¿ãĤı": 5210, "ä½¿ãĤıãģªãģĦ": 5211, "ä½į": 5212, "ä½įç½®": 5213, "ä½įç½®ãģ§": 5214, "ä½įç½®ãģ®ãģ¿ãģ«ä¾ĿåŃĺãģĻãĤĭ": 5215, "ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:": 5216, "ä½įç½®ãĤ¨ãĥ³ãĤ³ãĥ¼ãĥĩãĤ£ãĥ³ãĤ°": 5217, "ä½įç½®å¤īæıĽ": 5218, "ä½İ": 5219, "ä½ĵ": 5220, "ä½ĵãģ¯": 5221, "ä½ķ": 5222, "ä½ķãĤĤ": 5223, "ä½ľ": 5224, "ä½ľæĪĲ": 5225, "ä½ľæĪĲãģĻãĤĭãĢĤ": 5226, "ä¾": 5227, "ä¾¡": 5228, "ä¾ĭ": 5229, "ä¾ĭ:": 5230, "ä¾ĭãģĪãģ°": 5231, "ä¾ĭå¤ĸ": 5232, "ä¾ĿåŃĺ": 5233, "ä¿": 5234, "ä¿®æŃ£": 5235, "ä¿Ĥ": 5236, "ä¿Ĥæķ°": 5237, "ä¿Ŀ": 5238, "ä¿ĿåŃĺ": 5239, "ä¿ĿæĮģ": 5240, "å": 5241, "å¢": 5242, "å¢ĥçķĮ": 5243, "å¢Ĺ": 5244, "å¤": 5245, "å¤§": 5246, "å¤§ãģį": 5247, "å¤§ãģįãģª": 5248, "å¤§ãģįãģĦ": 5249, "å¤§ãģįãģı": 5250, "å¤§ãģįãģıãģªãĤĬ": 5251, "å¤±": 5252, "å¤±æķĹ": 5253, "å¤ī": 5254, "å¤īãĤı": 5255, "å¤īæıĽ": 5256, "å¤īæıĽãģĻãĤĭ": 5257, "å¤īæĽ´": 5258, "å¤ĸ": 5259, "å¤ĸãĤĮåĢ¤": 5260, "å¤ĸéĥ¨": 5261, "å¤ĸéĥ¨ãģĭãĤī": 5262, "å¤ļ": 5263, "å¤ļãģĦ": 5264, "å§Ķ": 5265, "å§ĶãģŃ": 5266, "å®": 5267, "å®¹": 5268, "å®ī": 5269, "å®īå®ļ": 5270, "å®īå®ļåĮĸ": 5271, "å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°": 5272, "å®Į": 5273, "å®Įåħ¨": 5274, "å®Įåħ¨ãģª": 5275, "å®Įåħ¨ãģ«": 5276, "å®Įåħ¨ä¸Ģèĩ´": 5277, "å®ļ": 5278, "å®ļåĢ¤": 5279, "å®ļåĢ¤ãģ®ç³»åĪĹ": 5280, "å®ļæķ°": 5281, "å®ļç¾": 5282, "å®ļç¾©": 5283, "å®Ł": 5284, "å®ŁãĥĪãĥ¼ãĤ¯ãĥ³": 5285, "å®Łæ¸¬": 5286, "å®Łè¡Į": 5287, "å®Łè¡Įåīįãģ®": 5288, "å®Łè£": 5289, "å®Łè£ħ": 5290, "å®Łè£ħãģ§ãģ¯": 5291, "å®Łè£ħãģĹãģªãģĦ": 5292, "å®Łè£ħãģĻãĤĭãĢĤ": 5293, "å®Łè£ħæĸ¹éĩĿ": 5294, "å®Łé": 5295, "å®Łé¨ĵ": 5296, "å®ŁéļĽ": 5297, "å®ŁéļĽãģ«": 5298, "å¯": 5299, "å¯¾": 5300, "å¯¾ãģĹãģ¦": 5301, "å¯¾ä½įç½®": 5302, "å¯¾å¿ľ": 5303, "å¯¾å¿ľãģĻãĤĭ": 5304, "å¯¾å¿ľãģĻãĤĭæĲįå¤±": 5305, "å¯¾æķ°": 5306, "å¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«": 5307, "å¯¾æķ°ç¢ºçİĩ": 5308, "å¯¾æķ°ç©ºéĸĵ": 5309, "å¯¾æķ°ç©ºéĸĵãģ§": 5310, "å¯¾è": 5311, "å¯¾è±¡": 5312, "å¯¾è±¡ãģ®": 5313, "å¯¾è·ĿéĽ¢": 5314, "å°": 5315, "å°¤": 5316, "å°¤åº¦": 5317, "å°Ħå½±": 5318, "å°İ": 5319, "å°İåħ¥": 5320, "å°İåĩº": 5321, "å°İéĸ¢æķ°": 5322, "å°ı": 5323, "å°ıãģķ": 5324, "å°ıãģķãģĦ": 5325, "å°ĳ": 5326, "å±": 5327, "å±¤": 5328, "å±¤ãģ®": 5329, "å±¤ãģĶãģ¨": 5330, "å±¤æķ°": 5331, "å±¤æŃ£è¦ıåĮĸ": 5332, "å±¤æŃ£è¦ıåĮĸ(Layer": 5333, "å±ĢæīĢ": 5334, "å±ķ": 5335, "å±ķéĸ": 5336, "å±ķéĸĭ": 5337, "å·": 5338, "å·¦": 5339, "å·®": 5340, "å·®ãģĹ": 5341, "å·®åĪĨ": 5342, "å·±": 5343, "å¸": 5344, "å¸°": 5345, "å¸¸": 5346, "å¸¸ãģ«": 5347, "å¸¸ãģ«è²ł": 5348, "å¸¸ãģ«è²łãģ®ãĥ¦ãĥĭãĥĥãĥĪ": 5349, "å¸ĥ": 5350, "å¹": 5351, "å¹³": 5352, "å¹³åĿĩ": 5353, "å¹³åĿĩæ¸Ľç®Ĺ": 5354, "å¹³åĿĩæ¸Ľç®Ĺ(re-centering": 5355, "å¹³æĸ¹": 5356, "å¹³æĸĩ": 5357, "å¹³æĸĩãģ«å¤īæıĽ": 5358, "å¹³è¡Į": 5359, "å¹³è¡Įç§»åĭķ": 5360, "å¹ħ": 5361, "åº": 5362, "åº¦": 5363, "åº¦ãĤĤ": 5364, "åºı": 5365, "å¼": 5366, "å¼¦æ³¢": 5367, "å¼µ": 5368, "å¼·åº¦": 5369, "å¼·åº¦ãģĮ": 5370, "å¼ı": 5371, "å¼ķæķ°": 5372, "å½": 5373, "å½¢": 5374, "å½¢å¼ı": 5375, "å½¢çĬ¶": 5376, "å½¢çĬ¶ãģ¯": 5377, "å½±": 5378, "å½±éŁ¿": 5379, "å½ĵ": 5380, "å½ĵãģĻãĤĭ": 5381, "å½Ļ": 5382, "å¾": 5383, "å¾©": 5384, "å¾©åħĥ": 5385, "å¾©åħĥãģ§ãģįãĤĭ": 5386, "å¾®åĪĨ": 5387, "å¾´": 5388, "å¾ħ": 5389, "å¾Į": 5390, "å¾Įãģ«": 5391, "å¾Įãģ®": 5392, "å¾ĮåįĬ": 5393, "å¾ĵ": 5394, "å¾ĵæĿ¥": 5395, "å¾ĵæĿ¥éĢļãĤĬ": 5396, "å¾Ĺ": 5397, "å¾ĹãģŁ": 5398, "å¿": 5399, "å¿ħ": 5400, "å¿ħãģļ": 5401, "å¿ħè¦ģ": 5402, "å¿ħè¦ģãģĮãģĤ": 5403, "å¿ħè¦ģãģĮãģĤãĤĭ": 5404, "å¿ħè¦ģãģĮãģĤãĤĭ(": 5405, "å¿ħè¦ģãģĮãģĤãĤĭãĢĤ": 5406, "å¿ľ": 5407, "åĢ¤": 5408, "åĢ¤ãģ¨ãģĹãģ¦": 5409, "åĢ¤ãģ¯": 5410, "åĢ¤ãĤĴ": 5411, "åĢĻè£ľ": 5412, "åĢĻè£ľèªŀå½Ļ": 5413, "åĢĻè£ľèªŀå½ĻãģĭãĤīãģ®": 5414, "åģ": 5415, "åģ´": 5416, "åģ´ãģ§": 5417, "åģ¶æķ°": 5418, "åģı": 5419, "åģıå·®": 5420, "åĤ": 5421, "åĤ¾ãģį": 5422, "åĥ": 5423, "åĥı": 5424, "åĦ": 5425, "åħ": 5426, "åħ¥": 5427, "åħ¥ãĤĮ": 5428, "åħ¥åĩºåĬĽ": 5429, "åħ¥åĬĽ": 5430, "åħ¥åĬĽãĢĤ": 5431, "åħ¥åĬĽåŁĭãĤģè¾¼ãģ¿": 5432, "åħ¨": 5433, "åħ¨ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿": 5434, "åħ¨ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿ãģ®åĭ¾éħįãĤĴ": 5435, "åħ¨ãĥĺãĥĥãĥī": 5436, "åħ¨ä½ĵ": 5437, "åħ¨ä½ĵãģ®": 5438, "åħ¨ä½ĵãĤĴ": 5439, "åħ¬": 5440, "åħ¬å¹³": 5441, "åħ¬å¹³ãģ«": 5442, "åħ±": 5443, "åħ±æľī": 5444, "åħ±éĢļ": 5445, "åħ±éĢļãģ®": 5446, "åħ¸": 5447, "åħĥ": 5448, "åħĥãģ®": 5449, "åħĪ": 5450, "åħĪãģ®": 5451, "åħĪãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯": 5452, "åħĪéłŃ": 5453, "åħĪéłŃãģ«": 5454, "åħĪéłŃãģ«ä»ĺ": 5455, "åĨ": 5456, "åĨħ": 5457, "åĨħãģ®": 5458, "åĨħå®¹": 5459, "åĨħéĥ¨": 5460, "åĨį": 5461, "åĨįåĪ©çĶ¨": 5462, "åĨįåĪ©çĶ¨ãģĻãĤĭ": 5463, "åĨįåŃ¦ç¿Ĵ": 5464, "åĨįæ§ĭæĪĲ": 5465, "åĨįçı¾": 5466, "åĨįè©¦è¡Į": 5467, "åĩ": 5468, "åĩ¡ä¾ĭ": 5469, "åĩ¦çĲĨ": 5470, "åĩº": 5471, "åĩºåĬĽ": 5472, "åĩºåĬĽå°Ħå½±": 5473, "åĩºåĬĽå±¤": 5474, "åĩºçı¾": 5475, "åĩºçı¾ãģĹãģŁ": 5476, "åĪ": 5477, "åĪ¤": 5478, "åĪ¤å®ļ": 5479, "åĪ¥": 5480, "åĪ©": 5481, "åĪ©çĶ¨": 5482, "åĪ¶": 5483, "åĪ¶ç´Ħ": 5484, "åĪ¶éĻĲ": 5485, "åĪĨ": 5486, "åĪĨãĤĴ": 5487, "åĪĨå¸ĥ": 5488, "åĪĨåī²": 5489, "åĪĨåī²çµĲæŀľ": 5490, "åĪĨæ¯": 5491, "åĪĨæ¯į": 5492, "åĪĨæķ£": 5493, "åĪĩ": 5494, "åĪĩãĤĬ": 5495, "åĪĩãĤĬåĩº": 5496, "åĪĩãĤĬæį": 5497, "åĪĩãĤĬæį¨ãģ¦": 5498, "åĪĹ": 5499, "åĪĹ)ãĢĤ": 5500, "åĪĹãĤĴ": 5501, "åĪĿ": 5502, "åĪĿæľŁ": 5503, "åĪĿæľŁåĢ¤": 5504, "åĪĿæľŁåĮĸ": 5505, "åĪĿæľŁèªŀå½Ļ": 5506, "åī": 5507, "åī²": 5508, "åī²ãĤĬå½ĵãģ¦": 5509, "åī²ãĤĬå½ĵãģ¦ãĤĭ": 5510, "åī²åĲĪ": 5511, "åīĩ": 5512, "åīĩåĮĸ": 5513, "åīĬ": 5514, "åīį": 5515, "åīįãģ«": 5516, "åīįãģ®": 5517, "åīįåįĬ": 5518, "åīįæıĲ": 5519, "åīįæıĲæĿ¡ä»¶": 5520, "åĬ": 5521, "åĬ¹": 5522, "åĬ¹æŀľ": 5523, "åĬ¹çİĩ": 5524, "åĬĽ": 5525, "åĬŁ": 5526, "åĬł": 5527, "åĬłãģĪãĤĭ": 5528, "åĬłç®Ĺ": 5529, "åĭ": 5530, "åĭ¾éħį": 5531, "åĭ¾éħįãģĮ": 5532, "åĭ¾éħįãĥİãĥ«ãĥł": 5533, "åĭķ": 5534, "åĮ": 5535, "åĮº": 5536, "åĮºåĪĩ": 5537, "åĮºéĸĵ": 5538, "åĮĸ": 5539, "åĮĸãģĹãģŁ": 5540, "åį": 5541, "åįĬ": 5542, "åįĺ": 5543, "åįĺä¸Ģ": 5544, "åįĺä¸ĢãĤ¹ãĥĨãĥĥãĥĹæĲįå¤±": 5545, "åįĺä¸ĢãĥĪãĥ¼ãĤ¯ãĥ³": 5546, "åįĺä½į": 5547, "åįĺä½įãģ®": 5548, "åįĺèªŀ": 5549, "åİ": 5550, "åİ³å¯Ĩ": 5551, "åİ»": 5552, "åİŁ": 5553, "åİŁè«ĸ": 5554, "åİŁè«ĸæĸĩ": 5555, "åİŁè«ĸæĸĩãģ®": 5556, "åİŁè«ĸæĸĩãģ¯": 5557, "åı": 5558, "åı¯": 5559, "åı¯èĥ½": 5560, "åı¯èĥ½ãģª": 5561, "åı¯éĢĨ": 5562, "åı¯éĢĨæĢ§": 5563, "åı³": 5564, "åı·": 5565, "åıĤç": 5566, "åıĤçħ§": 5567, "åıĤçħ§)ãĢĤ": 5568, "åıį": 5569, "åıįå¾©": 5570, "åıİ": 5571, "åıİæĿŁ": 5572, "åıĸ": 5573, "åıĸãĤĬ": 5574, "åıĸãĤĬåĩº": 5575, "åıĸãĤĭ": 5576, "åıĸå¾Ĺ": 5577, "åıĸå¾ĹãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ": 5578, "åıĹ": 5579, "åıĹãģĳ": 5580, "åĲ": 5581, "åĲ«": 5582, "åĲ«ãģ¾": 5583, "åĲ«ãģ¾ãĤĮãģªãģĦ": 5584, "åĲ«ãģ¾ãĤĮãĤĭ": 5585, "åĲ«ãĤģãĤĭ": 5586, "åĲĦ": 5587, "åĲĦä½įç½®": 5588, "åĲĦè¦ģç´łãģ¯": 5589, "åĲĪ": 5590, "åĲĪãģ¯": 5591, "åĲĪãĤıãģĽ": 5592, "åĲĮ": 5593, "åĲĮãģĺ": 5594, "åĲĮä¸Ģ": 5595, "åĲĮä¸Ģãģ®": 5596, "åĲĮä¸Ģãģ®æĮĻåĭķ": 5597, "åĲį": 5598, "åĲį:": 5599, "åĲįåīį": 5600, "åĲĳ": 5601, "åĳ": 5602, "åĳ¨æ³¢æķ°": 5603, "åĳ³": 5604, "åĳ¼": 5605, "åĳ¼ãģ³": 5606, "åĳ¼ãģ³åĩº": 5607, "åĳ¼ãģ³åĩºãģĹ": 5608, "åĳ¼ãģ³åĩºãģĹåģ´": 5609, "åĳ¼ãģ³åĩºãģĹåģ´ãģĮ": 5610, "åĴĮ": 5611, "åķ": 5612, "åķı": 5613, "åķıé¡": 5614, "åķıé¡Į": 5615, "åĸĦ": 5616, "åĽ": 5617, "åĽ²": 5618, "åĽ³": 5619, "åĽºå®ļ": 5620, "åĽºå®ļåŃ¦ç¿Ĵçİĩ": 5621, "åĽŀ": 5622, "åĽŀå¸°": 5623, "åĽŀæķ°": 5624, "åĽŀæķ°ãĢĤ": 5625, "åĽł": 5626, "åĽłæŀľ": 5627, "åĽłæŀľãĥŀãĤ¹ãĤ¯": 5628, "åľ": 5629, "åľ§ç¸®": 5630, "åľ§ç¸®åĬ¹çİĩ": 5631, "åľ¨": 5632, "åľ¨ãģ®": 5633, "åĿ": 5634, "åĿĩ": 5635, "åŀ": 5636, "åŀĭ": 5637, "åŁ": 5638, "åŁº": 5639, "åŁºãģ¥": 5640, "åŁºãģ¥ãģı": 5641, "åŁºæºĸ": 5642, "åŁĭãĤģ": 5643, "åŁĭãĤģè¾¼ãģ¿": 5644, "åŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°": 5645, "åŁĭãĤģè¾¼ãģ¿è¡ĮåĪĹ": 5646, "åŁŁ": 5647, "åł": 5648, "åł±": 5649, "åł´åĲĪ": 5650, "åł´åĲĪ(": 5651, "åł´åĲĪãĢģ": 5652, "åł´åĲĪãĢģ``": 5653, "åł´åĲĪãĢĤ": 5654, "åł´åĲĪãģ«": 5655, "åł´åĲĪãģ¯": 5656, "åŃ": 5657, "åŃ¦ç": 5658, "åŃ¦ç¿Ĵ": 5659, "åŃ¦ç¿ĴãģĮ": 5660, "åŃ¦ç¿ĴãģĹãģŁ": 5661, "åŃ¦ç¿ĴãĤ³ãĥ¼ãĥĳãĤ¹": 5662, "åŃ¦ç¿Ĵãĥ«ãĥ¼ãĥĹ": 5663, "åŃ¦ç¿Ĵçİĩ": 5664, "åŃ¦ç¿ĴçİĩãĤĴ": 5665, "åŃ¦ç¿ĴçĶ¨": 5666, "åŃ¦ç¿ĴçĶ¨ãĥ»æ¤ľè¨¼": 5667, "åŃĲ": 5668, "åŃĹ": 5669, "åŃĺ": 5670, "åŃĺåľ¨": 5671, "åŃĺåľ¨ãģĹãģªãģĦ": 5672, "æ": 5673, "æ¤ľè¨¼": 5674, "æ¤ľè¨¼ãĥĩãĥ¼ãĤ¿": 5675, "æ¥µ": 5676, "æ§ĭ": 5677, "æ§ĭæĪĲ": 5678, "æ§ĭç¯ī": 5679, "æ§ĺ": 5680, "æ¨": 5681, "æ¨ª": 5682, "æ¨Ļ": 5683, "æ¨Ļæºĸ": 5684, "æ¨Ļæºĸãģ®": 5685, "æ¨Ļæºĸåģıå·®": 5686, "æ¨ĻæºĸæŃ£è¦ıåĪĨå¸ĥ": 5687, "æ¨ĻæºĸçļĦãģª": 5688, "æ¨Ļæľ¬": 5689, "æ©Ł": 5690, "æ©Łæ§ĭ": 5691, "æ¬": 5692, "æ¬¡": 5693, "æ¬¡åħĥ": 5694, "æ¬¡åħĥ(": 5695, "æ¬¡åħĥãĢĤ": 5696, "æ¬²æ³ķ": 5697, "æ¬ł": 5698, "æ®Ĭ": 5699, "æ®ĭ": 5700, "æ®ĭãĤĬ": 5701, "æ®ĭå·®": 5702, "æ¯": 5703, "æ¯Ķ": 5704, "æ¯Ķä¾ĭ": 5705, "æ¯Ķä¾ĭãģĹãģ¦": 5706, "æ¯Ķçİĩ": 5707, "æ¯Ķè¼ĥ": 5708, "æ¯Ķè¼ĥãģĻãĤĭ": 5709, "æ¯Ķè¼ĥå¯¾è±¡": 5710, "æ±Ĥ": 5711, "æ±ĤãĤģ": 5712, "æ³": 5713, "æ³¨": 5714, "æ³¨åħ¥": 5715, "æ³¨æĦı": 5716, "æ³¨æĦıæ©Łæ§ĭ": 5717, "æ³ķ": 5718, "æ´»æĢ§åĮĸ": 5719, "æ´»æĢ§åĮĸéĸ¢æķ°": 5720, "æµ": 5721, "æµ®åĭķå°ıæķ°çĤ¹": 5722, "æ··": 5723, "æ¸": 5724, "æ¸¡": 5725, "æ¸¡ãģĻ": 5726, "æ¸¡ãģĻãĢĤ": 5727, "æ¸¡ãģĻæĥ³å®ļ": 5728, "æ¸¬": 5729, "æ¸¬ãĤĭ": 5730, "æ¸¬å®ļ": 5731, "æ¸¬å®ļåĢ¤ãģ®ç³»åĪĹ": 5732, "æ¸Ī": 5733, "æ¸Īãģ¿": 5734, "æ¸Īãģ¿ãģ®": 5735, "æ¸Ľ": 5736, "æ¸Ľç®Ĺ": 5737, "æ¸Ľè¡": 5738, "æ¸Ľè¡°": 5739, "æ¸Ľè¡°çİĩ": 5740, "æº": 5741, "æºĢ": 5742, "æºĢãģŁ": 5743, "æºĸ": 5744, "æ¼Ķç®Ĺ": 5745, "æĢ§": 5746, "æĢ§(": 5747, "æĢ§èĥ½": 5748, "æģ": 5749, "æģĴ": 5750, "æĥ³å®ļ": 5751, "æĥħåł±": 5752, "æĦı": 5753, "æĦıåĳ³": 5754, "æħ": 5755, "æħ®": 5756, "æħĭ": 5757, "æĪ": 5758, "æĪ»": 5759, "æĪĲ": 5760, "æĪĲåĬŁ": 5761, "æī": 5762, "æī±": 5763, "æī±ãģĨ": 5764, "æī±ãģĨ)ãĢĤ": 5765, "æī±ãģĨãĢĤ": 5766, "æīĢ": 5767, "æīĭ": 5768, "æīĭæ³ķ": 5769, "æĭ": 5770, "æĭ¡": 5771, "æĭ¡å¼µ": 5772, "æĮ": 5773, "æĮ¿": 5774, "æĮģ": 5775, "æĮģãģ¤": 5776, "æĮģãģŁ": 5777, "æĮģãģŁãģªãģĦ": 5778, "æĮĩ": 5779, "æĮĩå®ļ": 5780, "æĮĩå®ļãģķãĤĮãģŁ": 5781, "æĮĩå®ļãģķãĤĮãģŁåł´åĲĪãĢģ": 5782, "æĮĩå®ļãģĹãģŁ": 5783, "æĮĩå®ļãģĻãĤĭãģ¨": 5784, "æĮĩæ¨Ļ": 5785, "æĮĩæķ°": 5786, "æĮĩæķ°æİ¨å®ļ": 5787, "æĮĻ": 5788, "æĮĻåĭķ": 5789, "æį": 5790, "æİ": 5791, "æİ¢ç´¢": 5792, "æİ¥": 5793, "æİ¥ç¶ļ": 5794, "æİ¨": 5795, "æİ¨å®ļ": 5796, "æİ¨è«ĸ": 5797, "æı": 5798, "æıĥ": 5799, "æıĥãģĪãĤĭ": 5800, "æıı": 5801, "æııçĶ»": 5802, "æııçĶ»ãģ«ä½¿ãģ£ãģŁ": 5803, "æııçĶ»ãģĻãĤĭãĢĤ": 5804, "æııçĶ»åħĪãģ®": 5805, "æıĲ": 5806, "æıĽ": 5807, "æĲįå¤±": 5808, "æĶ": 5809, "æĶ¹": 5810, "æĶ¹åĸĦ": 5811, "æĶ¹è¡Į": 5812, "æķ": 5813, "æķ£": 5814, "æķ°": 5815, "æķ°ãĢĤ": 5816, "æķ°ãģĪ": 5817, "æķ°ãģĪæĸ¹": 5818, "æķ°ãģĪæĸ¹ãĢĮæľ¬ä½ĵ": 5819, "æķ°åĢ¤": 5820, "æķ°åĢ¤å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°": 5821, "æķ°åĢ¤çļĦãģ«": 5822, "æķ´": 5823, "æķ´æķ°": 5824, "æķĹ": 5825, "æĸ": 5826, "æĸ°": 5827, "æĸ°çīĪ": 5828, "æĸ°è¦ı": 5829, "æĸ°è¦ıä½ľæĪĲãģĻãĤĭãĢĤ": 5830, "æĸ¹": 5831, "æĸ¹äºĴæıĽæĢ§": 5832, "æĸ¹å¼ı": 5833, "æĸ¹åĲĳ": 5834, "æĸ¹éĩ": 5835, "æĸ¹éĩĿ": 5836, "æĸĩ": 5837, "æĸĩåŃĹ": 5838, "æĸĩåŃĹãģ®": 5839, "æĸĩåŃĹãģĮ": 5840, "æĸĩåŃĹãĤĴ": 5841, "æĸĩåŃĹãĥ¬ãĥĻãĥ«": 5842, "æĸĩåŃĹåĪĹ": 5843, "æĸĩåŃĹåĪĹãĤĴ": 5844, "æĸĩåŃĹåįĺä½įãģ®": 5845, "æĸĩåŃĹæķ°": 5846, "æĸĩè": 5847, "æĸĩèĦ": 5848, "æĸĩèĦĪ": 5849, "æĸŃ": 5850, "æĸŃçī": 5851, "æĸŃçīĩ": 5852, "æĹ": 5853, "æĹ¢": 5854, "æĹ¢ãģ«": 5855, "æĹ¢å®ļ": 5856, "æĹ¢å®ļåĢ¤": 5857, "æĹ¥æľ¬èªŀ": 5858, "æĺ": 5859, "æĺİ": 5860, "æĺİç¤º": 5861, "æĺİç¤ºçļĦãģª": 5862, "æĺİç¤ºçļĦãģ«": 5863, "æĻ": 5864, "æĻĤ": 5865, "æĻĤãģ®": 5866, "æĻĤçĤ¹": 5867, "æĻĤçĤ¹ãģ§": 5868, "æĽ": 5869, "æĽ²ç·ļ": 5870, "æĽ´": 5871, "æĽ´æĸ°": 5872, "æĽ¸": 5873, "æĽ¸ãģį": 5874, "æĽ¸ãģįåĩº": 5875, "æĽ¿": 5876, "æĽ¿ãģĪ": 5877, "æĽ¿ãģĪãĤĭ": 5878, "æľ": 5879, "æľª": 5880, "æľªæºĢ": 5881, "æľªçŁ¥": 5882, "æľªçŁ¥èªŀ": 5883, "æľ«å°¾": 5884, "æľ¬": 5885, "æľ¬ä½ĵ": 5886, "æľ¬æĸĩ": 5887, "æľ¬çķªå®Łè¡Įåīįãģ®ä¿®æŃ£": 5888, "æľĢ": 5889, "æľĢå¤§": 5890, "æľĢå¤§åĢ¤": 5891, "æľĢå°¤": 5892, "æľĢå°¤åĪĨåī²": 5893, "æľĢå°ı": 5894, "æľĢå°ıäºĮä¹Ĺ": 5895, "æľĢå°ıåĢ¤": 5896, "æľĢå°ıéĻĲ": 5897, "æľĢåĪĿ": 5898, "æľĢæĸ°çīĪ": 5899, "æľĢçµĤ": 5900, "æľĢçµĤæ¬¡åħĥ": 5901, "æľĢçµĤæŃ£è¦ıåĮĸå±¤": 5902, "æľĢèī": 5903, "æľĢèī¯": 5904, "æľĢéģ©": 5905, "æľĢéģ©åĮĸ": 5906, "æľī": 5907, "æľīåĬ¹": 5908, "æľīçĦ¡": 5909, "æľīéĻĲ": 5910, "æľīéĻĲåĢ¤": 5911, "æľŁ": 5912, "æĿ": 5913, "æĿ¡ä»¶": 5914, "æĿ¡ä»¶ãģĶãģ¨ãģ«": 5915, "æĿ¡ä»¶åĲį:": 5916, "æĿ¡ä»¶éĸĵãģ§": 5917, "æĿ¥": 5918, "æĿŁ": 5919, "æŀ": 5920, "æŀĲ": 5921, "æŀľ": 5922, "æł": 5923, "æł¹": 5924, "æŃ": 5925, "æŃ¢": 5926, "æŃ£": 5927, "æŃ£ãģĹ": 5928, "æŃ£ãģĹãģķ": 5929, "æŃ£å¼¦æ³¢": 5930, "æŃ£è¦ı": 5931, "æŃ£è¦ıåĪĨå¸ĥ": 5932, "æŃ£è¦ıåĮĸ": 5933, "æŃ£è¦ıåĮĸå±¤": 5934, "æŃ£è¦ıåĮĸå¾Į": 5935, "æŃ£è¦ıåĮĸå¾Įç½®": 5936, "æŃ£è¦ıåĮĸå¾Įç½®(": 5937, "æŃ£è¦ıåĮĸå¾Įç½®(Post-Layer": 5938, "æŃ£è¦ıåĮĸåīį": 5939, "æŃ£è¦ıåĮĸåīįç½®": 5940, "æŃ£è¦ıåĮĸåīįç½®(": 5941, "æŃ£è¦ıåĮĸåīįç½®(Pre-Layer": 5942, "æŃ£è¦ıåĮĸæ¸Īãģ¿": 5943, "æŃ»": 5944, "ç": 5945, "ç¢º": 5946, "ç¢ºç«ĭ": 5947, "ç¢ºçİĩ": 5948, "ç¢ºèªį": 5949, "ç¤º": 5950, "ç¤ºãģĻ": 5951, "ç§": 5952, "ç§»åĭķ": 5953, "ç¨": 5954, "ç¨®": 5955, "ç¨ĭ": 5956, "ç¨ĭåº¦": 5957, "ç©": 5958, "ç©º": 5959, "ç©ºçĻ½": 5960, "ç©ºçĻ½ãģ®": 5961, "ç©ºçĻ½ãģĮ": 5962, "ç©ºçĻ½ãĤĴ": 5963, "ç©ºéĸĵ": 5964, "ç©į": 5965, "ç©įãĤĴ": 5966, "çªĵ": 5967, "ç«": 5968, "ç«¯": 5969, "ç«ĭ": 5970, "ç¬¦åı·": 5971, "ç¬¦åı·åĮĸ": 5972, "ç¬¬": 5973, "ç®": 5974, "ç®Ĺ": 5975, "ç¯": 5976, "ç¯Ħ": 5977, "ç¯ĦåĽ²": 5978, "ç¯ī": 5979, "ç²": 5980, "ç³»åĪĹ": 5981, "ç³»åĪĹãģ®": 5982, "ç³»åĪĹåĲį:": 5983, "ç³»åĪĹéķ·": 5984, "ç´": 5985, "ç´¯ç©į": 5986, "ç´¯ç©įç¢ºçİĩ": 5987, "ç´°": 5988, "ç´°ãģ¯": 5989, "ç´Ħ": 5990, "ç´Ķ": 5991, "ç´ł": 5992, "çµ": 5993, "çµ±": 5994, "çµ±è¨Ī": 5995, "çµ±è¨Īéĩı": 5996, "çµ¶": 5997, "çµ¶å¯¾ä½įç½®": 5998, "çµ¶å¯¾ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹": 5999, "çµ¶å¯¾ä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(å½¢çĬ¶": 6000, "çµĤ": 6001, "çµĤäºĨ": 6002, "çµĦ": 6003, "çµĦãģ¿": 6004, "çµĦãģ¿åĲĪãĤıãģĽ": 6005, "çµĮ": 6006, "çµĮçĶ±": 6007, "çµĮçĶ±ãģ§": 6008, "çµĮè·": 6009, "çµĮè·¯": 6010, "çµĲ": 6011, "çµĲæŀľ": 6012, "çµĲæŀľãģĮ": 6013, "ç¶": 6014, "ç¶ļ": 6015, "ç·": 6016, "ç·´": 6017, "ç·ı": 6018, "ç·ıåĴĮ": 6019, "ç·ļ": 6020, "ç·ļå½¢": 6021, "ç¸": 6022, "ç¸¦": 6023, "ç¸®": 6024, "ç¸®å°ı": 6025, "ç¹°ãĤĬ": 6026, "ç½®": 6027, "ç½®(": 6028, "ç½®ãģįæıĽãģĪ": 6029, "ç½®æıĽ": 6030, "ç¾": 6031, "ç¾¤": 6032, "çĤ¹": 6033, "çĦ": 6034, "çĦ¡": 6035, "çĦ¡ãģĦ": 6036, "çĦ¡åĬ¹": 6037, "çĦ¡è¦ĸ": 6038, "çī": 6039, "çī¹": 6040, "çī¹å¾´": 6041, "çī¹æ®Ĭ": 6042, "çīĪ": 6043, "çĬ": 6044, "çĬ¶": 6045, "çĬ¶æħĭ": 6046, "çĭ": 6047, "çĭ¬ç«ĭ": 6048, "çİĩ": 6049, "çı¾": 6050, "çĲĨ": 6051, "çĲĨçĶ±": 6052, "çĲĨè«ĸ": 6053, "çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³": 6054, "çĶ": 6055, "çĶ¨": 6056, "çĶ¨)ãĢĤ": 6057, "çĶ¨ãģ«": 6058, "çĶ¨ãģ®": 6059, "çĶ¨ãģĦãģ¦": 6060, "çĶ¨ãģĦãģŁ": 6061, "çĶ¨ãģĦãģŁéĢĲæ¬¡": 6062, "çĶ¨ãģĦãģŁéĢĲæ¬¡æİ¨è«ĸ": 6063, "çĶ±": 6064, "çĶ»": 6065, "çĶ»åĥı": 6066, "çĶŁ": 6067, "çĶŁãģĺ": 6068, "çĶŁæĪĲ": 6069, "çĶŁæĪĲãģ®åĲĦ": 6070, "çĶŁæĪĲãģĻãĤĭ": 6071, "çķ": 6072, "çķ¥": 6073, "çķª": 6074, "çķªå®Łè¡Įåīįãģ®": 6075, "çķªå®Łè¡Įåīįãģ®ä¿®æŃ£": 6076, "çķªåı·": 6077, "çķ°": 6078, "çķ°ãģªãĤĭ": 6079, "çķĮ": 6080, "çĻ": 6081, "çĻº": 6082, "çĻºåĭķ": 6083, "çĻºæķ£": 6084, "çĻºçĶŁ": 6085, "çĻ½": 6086, "çĻ¾": 6087, "çļĦ": 6088, "çļĦãģª": 6089, "çļĦãģ«": 6090, "çĽ": 6091, "çĽ®": 6092, "çĽ®æ¨Ļ": 6093, "çĽ®è¦ĸ": 6094, "çĽ®è¦ĸãģ§": 6095, "çĽ´": 6096, "çĽ´åīį": 6097, "çĽ´åīįãģ§": 6098, "çĽ´æİ¥": 6099, "çĽ¸": 6100, "çĽ¸å¯¾": 6101, "çĽ¸å¯¾ä½įç½®": 6102, "çĽ¸å¯¾è·ĿéĽ¢": 6103, "çľ": 6104, "çľģ": 6105, "çľģçķ¥": 6106, "çľģçķ¥åı¯": 6107, "çŁ": 6108, "çŁ¥": 6109, "çŃ": 6110, "çŃī": 6111, "çŃīä¾¡": 6112, "è": 6113, "è¡": 6114, "è¡¨": 6115, "è¡¨ç¤º": 6116, "è¡¨çı¾": 6117, "è¡Į": 6118, "è¡ĮãģĨ": 6119, "è¡ĮãģĮ": 6120, "è¡ĮãĤı": 6121, "è¡ĮãĤıãģªãģĦ": 6122, "è¡ĮãĤıãģļãĢģ": 6123, "è¡ĮåĪĹ": 6124, "è¢«è¦Ĩ": 6125, "è¢«è¦Ĩçİĩ": 6126, "è£": 6127, "è£ľ": 6128, "è¤ĩ": 6129, "è¤ĩæķ°": 6130, "è¤ĩæķ°ãģ®": 6131, "è¤ĩæķ°ãĤ·ãĥ¼ãĥī": 6132, "è¦": 6133, "è¦ģ": 6134, "è¦ģç´ł": 6135, "è¦ģç´łãģ¯": 6136, "è¦ĭ": 6137, "è¦ı": 6138, "è¦ıåīĩ": 6139, "è¦ıç´Ħ": 6140, "è¦ĸ": 6141, "è§": 6142, "è§£": 6143, "è§£æŀĲ": 6144, "è§Ĵ": 6145, "è§Ĵåĳ¨æ³¢æķ°": 6146, "è¨": 6147, "è¨±": 6148, "è¨±å®¹": 6149, "è¨¼": 6150, "è¨Ģ": 6151, "è¨Ģèªŀ": 6152, "è¨Ģèªŀãĥ¢ãĥĩãĥ«": 6153, "è¨Ģèªŀãĥ¢ãĥĩãĥ«(": 6154, "è¨Ģèªŀãĥ¢ãĥĩãĥ«(``forward(token_ids)": 6155, "è¨Ģèªŀãĥ¢ãĥĩãĥ«ãģ®": 6156, "è¨Ī": 6157, "è¨Īç®Ĺ": 6158, "è¨Īç®ĹãģĻãĤĭ(": 6159, "è¨Īç®ĹãģĻãĤĭãĢĤ": 6160, "è¨Īç®Ĺéĩı": 6161, "è¨Īç®ĹéĩıäºĪç®Ĺ": 6162, "è¨Īç®ĹéĩıæľĢéģ©": 6163, "è¨ĵ": 6164, "è¨ĵç·´": 6165, "è¨ĵç·´æĲįå¤±": 6166, "è¨ĺ": 6167, "è¨ĺäºĭ": 6168, "è¨ĺåı·": 6169, "è¨ĺéĮ²": 6170, "è¨Ń": 6171, "è¨Ńãģĳ": 6172, "è¨Ńå®ļ": 6173, "è¨Ńå®ļãģ§": 6174, "è¨Ńè¨Ī": 6175, "è©": 6176, "è©¦": 6177, "è©¦è¡Į": 6178, "è©²": 6179, "è©³": 6180, "è©³ç´°ãģ¯": 6181, "è©ķ": 6182, "è©ķä¾¡": 6183, "è©ķä¾¡å¯¾è±¡ãģ®": 6184, "èª": 6185, "èª¤": 6186, "èª¤ãĤĬ": 6187, "èª¬æĺİ": 6188, "èª¿": 6189, "èªį": 6190, "èªŀ": 6191, "èªŀå½Ļ": 6192, "èªŀå½Ļãģ«": 6193, "èªŀå½Ļãģ«åĲ«ãģ¾ãĤĮãģªãģĦ": 6194, "èªŀå½Ļãģ«åĲ«ãģ¾ãĤĮãĤĭ": 6195, "èªŀå½ĻãĤµãĤ¤ãĤº": 6196, "èªŀå½ĻåŃ¦ç¿Ĵ": 6197, "èªŃ": 6198, "èªŃãģ¿": 6199, "è«ĸ": 6200, "è²": 6201, "è²ª": 6202, "è²ªæ¬²æ³ķ": 6203, "è²ł": 6204, "è²łãģ®å¯¾æķ°å°¤åº¦": 6205, "è²łéłĺåŁŁ": 6206, "èµ·": 6207, "è¶": 6208, "è¶ħ": 6209, "è¶ħãģĪ": 6210, "è¶ħãģĪãģ¦ãģĦãĤĭ": 6211, "è¶ħãģĪãĤĭ": 6212, "è·": 6213, "è·ĿéĽ¢": 6214, "èº«": 6215, "èº«ãģ®": 6216, "è»": 6217, "è»¢": 6218, "è»¸": 6219, "è»¸ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ": 6220, "è»¸ãĥ©ãĥĻãĥ«ãĢĤ": 6221, "è¼ĥ": 6222, "è¾": 6223, "è¾¼": 6224, "è¾¼ãģ¿": 6225, "è¾¼ãĤĢ": 6226, "è¾ŀ": 6227, "è¾ŀæĽ¸": 6228, "è¾ŀæĽ¸ãĢĤ": 6229, "è¿": 6230, "è¿½": 6231, "è¿½åĬł": 6232, "è¿½åĬłãģĹãģŁ": 6233, "è¿ĳ": 6234, "è¿ĳä¼¼": 6235, "è¿Ķ": 6236, "è¿ĶãģĻ": 6237, "è¿ĶãģĻãĢĤ": 6238, "èĢ": 6239, "èĢĥ": 6240, "èĢĥæħ®": 6241, "èĥ": 6242, "èĥ½": 6243, "èĩ": 6244, "èĩª": 6245, "èĩªä½ĵãģ¯": 6246, "èĩªä½ľ": 6247, "èĩªä½ľãģ®": 6248, "èĩªå·±": 6249, "èĩªå·±åĽŀå¸°": 6250, "èĩªå·±æ³¨æĦı": 6251, "èĩªåĪĨ": 6252, "èĩªåĭķ": 6253, "èĩªåĭķçļĦãģ«": 6254, "èĪ": 6255, "èī": 6256, "èĭ": 6257, "èĭ±èªŀ": 6258, "èĲ": 6259, "èĲ½": 6260, "èĸĦ": 6261, "èĸĦãģĦãĥ©ãĥĥãĥĳãĥ¼ãĢĤ": 6262, "èŃ": 6263, "èŃĺ": 6264, "èŃĺåĪ¥": 6265, "èŃĺåĪ¥ä¸įèĥ½ãģª": 6266, "èŃĺåĪ¥ä¸įèĥ½ãģªãĥĩãĥ¼ãĤ¿": 6267, "é": 6268, "é¡": 6269, "é£": 6270, "é£½åĴĮãģ¹ãģįä¹Ĺåīĩ": 6271, "é«ĺ": 6272, "é«ĺãģĦ": 6273, "éĢ": 6274, "éĢ£": 6275, "éĢ£çµĲ": 6276, "éĢ£çµĲãģĹãģŁ": 6277, "éĢ£ç¶ļ": 6278, "éĢĢ": 6279, "éĢĨ": 6280, "éĢı": 6281, "éĢıéģİ": 6282, "éĢıéģİãģĻãĤĭ": 6283, "éĢĲ": 6284, "éĢĲæ¬¡": 6285, "éĢĶ": 6286, "éĢļ": 6287, "éĢļãĤĬ": 6288, "éĢļå¸¸": 6289, "éĢŁ": 6290, "éĢł": 6291, "éģ": 6292, "éģ©": 6293, "éģ©çĶ¨": 6294, "éģ©çĶ¨ãģĻãĤĭ": 6295, "éģ©çĶ¨ãģĻãĤĭãĢĤ": 6296, "éģ©çĶ¨åīį": 6297, "éģ¸": 6298, "éģ¿": 6299, "éģİ": 6300, "éģĶ": 6301, "éģĶãģĻãĤĭ": 6302, "éģķ": 6303, "éģķãģĦ": 6304, "éĥ": 6305, "éĥ¨": 6306, "éĥ¨åĪĨ": 6307, "éĥ¨åĪĨç©ºéĸĵ": 6308, "éĥ¨åĪĨèªŀ": 6309, "éĥ¨åĪĨèªŀãĤ·ãĥ³ãĥľãĥ«": 6310, "éħ": 6311, "éħį": 6312, "éĩ": 6313, "éĩį": 6314, "éĩįãģ¿": 6315, "éĩįãģ¿ãĤĴåħ±æľī": 6316, "éĩįãģ¿ä»ĺãģį": 6317, "éĩįãģ¿æ¸Ľè¡°": 6318, "éĩįãģ¿æ¸Ľè¡°ãģ®å®ŁåĬ¹": 6319, "éĩįãģŃ": 6320, "éĩįãģŃãģ¦": 6321, "éĩįãģŃãģ¦æııçĶ»ãģĻãĤĭãĢĤ": 6322, "éĩįè¤ĩ": 6323, "éĩı": 6324, "éĮ": 6325, "éĮ²": 6326, "éķ": 6327, "éķ·": 6328, "éķ·ãģĦ": 6329, "éķ·ãģķ": 6330, "éķ·ãģķãĥļãĥĬãĥ«ãĥĨãĤ£": 6331, "éĸ": 6332, "éĸ¢": 6333, "éĸ¢ä¿Ĥ": 6334, "éĸ¢æķ°": 6335, "éĸ¢æķ°ãģ®": 6336, "éĸ¾åĢ¤": 6337, "éĸ¾åĢ¤ãĢĤ": 6338, "éĸĵ": 6339, "éĸĵãģ§": 6340, "éĺ": 6341, "éĺ²": 6342, "éĺ²ãģĲ": 6343, "éĻ": 6344, "éĻ¤": 6345, "éĻ¤å¤ĸ": 6346, "éĻ¤åİ»": 6347, "éĻ¤ç®Ĺ": 6348, "éĻį": 6349, "éĻįéłĨãģ«": 6350, "éĻĲ": 6351, "éĻĲãģªãģı": 6352, "éļ": 6353, "éļİ": 6354, "éļĽ": 6355, "éļĽéĻĲãģªãģı": 6356, "éļł": 6357, "éļłãĤĮ": 6358, "éļłãĤĮçĬ¶æħĭ": 6359, "éĽ": 6360, "éĽ¢": 6361, "éĽĨ": 6362, "éĽĨåĲĪ": 6363, "éĽĨè¨Ī": 6364, "éĽĨè¨ĪãģĻãĤĭãĢĤ": 6365, "éĿ": 6366, "éĿŀ": 6367, "éĿŀåŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°": 6368, "éĿŀç·ļå½¢": 6369, "éĿŀç·ļå½¢æľĢå°ıäºĮä¹Ĺ": 6370, "éĿŀéĩįè¤ĩ": 6371, "éĿŀéĩįè¤ĩçªĵ": 6372, "éŁ": 6373, "éŁ¿": 6374, "éł": 6375, "éł»åº¦": 6376, "éłĤçĤ¹": 6377, "éłħ": 6378, "éłĨ": 6379, "éłĨãģ«": 6380, "éłĨä¼ĿæĴŃ": 6381, "éłĨä¼ĿæĴŃãĢĤ": 6382, "éłĨä¼ĿæĴŃãĥįãĥĥãĥĪãĥ¯ãĥ¼ãĤ¯": 6383, "éłĨåºı": 6384, "éłĺ": 6385, "éłĺåŁŁ": 6386, "éłŃ": 6387, "ê": 6388, "ë": 6389, "ì": 6390, "í": 6391, "î": 6392, "ï": 6393, "ð": 6394, "ñ": 6395, "ò": 6396, "ó": 6397, "ô": 6398, "õ": 6399, "ö": 6400, "÷": 6401, "ø": 6402, "ù": 6403, "ú": 6404, "û": 6405, "ü": 6406, "ý": 6407, "þ": 6408, "ÿ": 6409, "Ā": 6410, "ā": 6411, "Ă": 6412, "ă": 6413, "Ą": 6414, "ą": 6415, "Ć": 6416, "ć": 6417, "Ĉ": 6418, "ĉ": 6419, "Ċ": 6420, "Ċ\"\"\"": 6421, "Ċ#": 6422, "Ċ(": 6423, "Ċ(``": 6424, "Ċ)": 6425, "Ċ-": 6426, "ĊICLR": 6427, "Ċ]": 6428, "Ċ_": 6429, "Ċ``": 6430, "Ċclass": 6431, "Ċfrom": 6432, "Ċimport": 6433, "ĊåŃ¦ç¿Ĵ": 6434, "ĊĊ": 6435, "ĊĊ\"\"\"": 6436, "ĊĊ\"\"\"theories/ãĥ»apps/": 6437, "ĊĊ#": 6438, "ĊĊ**": 6439, "ĊĊVaswani": 6440, "ĊĊ__all__": 6441, "ĊĊ``": 6442, "ĊĊfrom": 6443, "ĊĊimport": 6444, "ĊĊè¨ĺåı·": 6445, "ĊĊĊ": 6446, "ĊĊĊ#": 6447, "ĊĊĊ@": 6448, "ĊĊĊ@dataclass": 6449, "ĊĊĊclass": 6450, "ĊĊĊdef": 6451, "ĊĊĠĠĠĠ": 6452, "ĊĊĠĠĠĠ#": 6453, "ĊĊĠĠĠĠ..": 6454, "ĊĊĠĠĠĠ1.": 6455, "ĊĊĠĠĠĠ@": 6456, "ĊĊĠĠĠĠ@staticmethod": 6457, "ĊĊĠĠĠĠ@torch.no_grad()": 6458, "ĊĊĠĠĠĠA": 6459, "ĊĊĠĠĠĠArgs:": 6460, "ĊĊĠĠĠĠAttention": 6461, "ĊĊĠĠĠĠR": 6462, "ĊĊĠĠĠĠRaises:": 6463, "ĊĊĠĠĠĠReturns:": 6464, "ĊĊĠĠĠĠ``": 6465, "ĊĊĠĠĠĠax": 6466, "ĊĊĠĠĠĠax.set_": 6467, "ĊĊĠĠĠĠax.set_xlabel(xlabel)": 6468, "ĊĊĠĠĠĠax.set_xtick": 6469, "ĊĊĠĠĠĠcolor_cycle": 6470, "ĊĊĠĠĠĠdef": 6471, "ĊĊĠĠĠĠfor": 6472, "ĊĊĠĠĠĠif": 6473, "ĊĊĠĠĠĠreturn": 6474, "ĊĊĠĠĠĠåĲĦ": 6475, "ĊĊĠĠĠĠè¨ĺåı·": 6476, "ĊĊĠĠĠĠĠĠĠĠ": 6477, "ĊĊĠĠĠĠĠĠĠĠ#": 6478, "ĊĊĠĠĠĠĠĠĠĠArgs:": 6479, "ĊĊĠĠĠĠĠĠĠĠReturns:": 6480, "ĊĊĠĠĠĠĠĠĠĠ\\\\mathrm{": 6481, "ĊĊĠĠĠĠĠĠĠĠ``": 6482, "ĊĊĠĠĠĠĠĠĠĠfor": 6483, "ĊĊĠĠĠĠĠĠĠĠif": 6484, "ĊĊĠĠĠĠĠĠĠĠself.": 6485, "ĊĊĠĠĠĠĠĠĠĠx": 6486, "ĊĊĠĠĠĠĠĠĠĠĠĠĠĠ": 6487, "ĊĊĠĠĠĠĠĠĠĠĠĠĠĠm": 6488, "ĊĠĠ": 6489, "ĊĠĠ``": 6490, "ĊĠĠĠĠ": 6491, "ĊĠĠĠĠ\"": 6492, "ĊĠĠĠĠ\"\"\"": 6493, "ĊĠĠĠĠ\"\"\"``": 6494, "ĊĠĠĠĠ\"\"\"åŃ¦ç¿Ĵ": 6495, "ĊĠĠĠĠ\"compute_": 6496, "ĊĠĠĠĠ\"load_": 6497, "ĊĠĠĠĠ\"plot": 6498, "ĊĠĠĠĠ\"plot_": 6499, "ĊĠĠĠĠ#": 6500, "ĊĠĠĠĠ$": 6501, "ĊĠĠĠĠ(": 6502, "ĊĠĠĠĠ(``": 6503, "ĊĠĠĠĠ)": 6504, "ĊĠĠĠĠ2": 6505, "ĊĠĠĠĠ2.": 6506, "ĊĠĠĠĠ3": 6507, "ĊĠĠĠĠB": 6508, "ĊĠĠĠĠL": 6509, "ĊĠĠĠĠS": 6510, "ĊĠĠĠĠV": 6511, "ĊĠĠĠĠW": 6512, "ĊĠĠĠĠ]": 6513, "ĊĠĠĠĠ``": 6514, "ĊĠĠĠĠa": 6515, "ĊĠĠĠĠalpha:": 6516, "ĊĠĠĠĠalways_negative_": 6517, "ĊĠĠĠĠapi.": 6518, "ĊĠĠĠĠax": 6519, "ĊĠĠĠĠax.": 6520, "ĊĠĠĠĠax.grid(alpha=0.3": 6521, "ĊĠĠĠĠax.grid(alpha=0.3)": 6522, "ĊĠĠĠĠax.grid(alpha=0.3,": 6523, "ĊĠĠĠĠax.leg": 6524, "ĊĠĠĠĠax.legend(": 6525, "ĊĠĠĠĠax.legend(fontsize=8)": 6526, "ĊĠĠĠĠax.set_": 6527, "ĊĠĠĠĠax.set_xlabel(xlabel)": 6528, "ĊĠĠĠĠax.set_xtick": 6529, "ĊĠĠĠĠax.set_ylabel(": 6530, "ĊĠĠĠĠax.set_ylabel(ylabel)": 6531, "ĊĠĠĠĠax:": 6532, "ĊĠĠĠĠb": 6533, "ĊĠĠĠĠbatch_size:": 6534, "ĊĠĠĠĠbest": 6535, "ĊĠĠĠĠchunk_split_mode:": 6536, "ĊĠĠĠĠcompute_": 6537, "ĊĠĠĠĠconverged": 6538, "ĊĠĠĠĠcorpus_": 6539, "ĊĠĠĠĠd_": 6540, "ĊĠĠĠĠd_ff": 6541, "ĊĠĠĠĠd_k": 6542, "ĊĠĠĠĠd_model": 6543, "ĊĠĠĠĠdata": 6544, "ĊĠĠĠĠdef": 6545, "ĊĠĠĠĠdevice:": 6546, "ĊĠĠĠĠelse:": 6547, "ĊĠĠĠĠf": 6548, "ĊĠĠĠĠfit_": 6549, "ĊĠĠĠĠflops_per_token": 6550, "ĊĠĠĠĠfor": 6551, "ĊĠĠĠĠgenerator": 6552, "ĊĠĠĠĠh": 6553, "ĊĠĠĠĠhuber_delta:": 6554, "ĊĠĠĠĠi": 6555, "ĊĠĠĠĠif": 6556, "ĊĠĠĠĠload_": 6557, "ĊĠĠĠĠlog": 6558, "ĊĠĠĠĠlog_": 6559, "ĊĠĠĠĠlog_alpha": 6560, "ĊĠĠĠĠlog_scale:": 6561, "ĊĠĠĠĠm": 6562, "ĊĠĠĠĠmax_": 6563, "ĊĠĠĠĠmax_chunk_bytes:": 6564, "ĊĠĠĠĠmodel:": 6565, "ĊĠĠĠĠn": 6566, "ĊĠĠĠĠn_": 6567, "ĊĠĠĠĠnum": 6568, "ĊĠĠĠĠp": 6569, "ĊĠĠĠĠparams": 6570, "ĊĠĠĠĠpath": 6571, "ĊĠĠĠĠplot": 6572, "ĊĠĠĠĠplot_": 6573, "ĊĠĠĠĠprofiles:": 6574, "ĊĠĠĠĠresidual_fn:": 6575, "ĊĠĠĠĠresiduals:": 6576, "ĊĠĠĠĠresult": 6577, "ĊĠĠĠĠreturn": 6578, "ĊĠĠĠĠright": 6579, "ĊĠĠĠĠsorted_": 6580, "ĊĠĠĠĠtext": 6581, "ĊĠĠĠĠtext:": 6582, "ĊĠĠĠĠtitle:": 6583, "ĊĠĠĠĠtry:": 6584, "ĊĠĠĠĠv": 6585, "ĊĠĠĠĠvocab": 6586, "ĊĠĠĠĠvocab:": 6587, "ĊĠĠĠĠweights:": 6588, "ĊĠĠĠĠwhile": 6589, "ĊĠĠĠĠwith": 6590, "ĊĠĠĠĠx": 6591, "ĊĠĠĠĠxlabel:": 6592, "ĊĠĠĠĠylabel:": 6593, "ĊĠĠĠĠ}": 6594, "ĊĠĠĠĠÎ": 6595, "ĊĠĠĠĠãĢĮ": 6596, "ĊĠĠĠĠãģĮ": 6597, "ĊĠĠĠĠãģĿãģ®ãģ¾ãģ¾": 6598, "ĊĠĠĠĠãĤĴ": 6599, "ĊĠĠĠĠæĲįå¤±": 6600, "ĊĠĠĠĠĠĠ": 6601, "ĊĠĠĠĠĠĠĠ": 6602, "ĊĠĠĠĠĠĠĠĠ": 6603, "ĊĠĠĠĠĠĠĠĠ\"": 6604, "ĊĠĠĠĠĠĠĠĠ\"\"\"": 6605, "ĊĠĠĠĠĠĠĠĠ\"\"\"éłĨä¼ĿæĴŃãĢĤ": 6606, "ĊĠĠĠĠĠĠĠĠ\"median": 6607, "ĊĠĠĠĠĠĠĠĠ\"p90": 6608, "ĊĠĠĠĠĠĠĠĠ#": 6609, "ĊĠĠĠĠĠĠĠĠ(": 6610, "ĊĠĠĠĠĠĠĠĠ)": 6611, "ĊĠĠĠĠĠĠĠĠ-": 6612, "ĊĠĠĠĠĠĠĠĠ1": 6613, "ĊĠĠĠĠĠĠĠĠValueError:": 6614, "ĊĠĠĠĠĠĠĠĠ_,": 6615, "ĊĠĠĠĠĠĠĠĠ``": 6616, "ĊĠĠĠĠĠĠĠĠa": 6617, "ĊĠĠĠĠĠĠĠĠactivation_fn:": 6618, "ĊĠĠĠĠĠĠĠĠalpha": 6619, "ĊĠĠĠĠĠĠĠĠax": 6620, "ĊĠĠĠĠĠĠĠĠax.plot": 6621, "ĊĠĠĠĠĠĠĠĠax.set_": 6622, "ĊĠĠĠĠĠĠĠĠax.set_title(title)": 6623, "ĊĠĠĠĠĠĠĠĠax.set_yscale(\"log\")": 6624, "ĊĠĠĠĠĠĠĠĠax:": 6625, "ĊĠĠĠĠĠĠĠĠb": 6626, "ĊĠĠĠĠĠĠĠĠbatch_size:": 6627, "ĊĠĠĠĠĠĠĠĠbeam": 6628, "ĊĠĠĠĠĠĠĠĠbeta": 6629, "ĊĠĠĠĠĠĠĠĠbeta:": 6630, "ĊĠĠĠĠĠĠĠĠbetas:": 6631, "ĊĠĠĠĠĠĠĠĠbia": 6632, "ĊĠĠĠĠĠĠĠĠbias:": 6633, "ĊĠĠĠĠĠĠĠĠbias_correction": 6634, "ĊĠĠĠĠĠĠĠĠbidirectional": 6635, "ĊĠĠĠĠĠĠĠĠbidirectional:": 6636, "ĊĠĠĠĠĠĠĠĠcache_dir:": 6637, "ĊĠĠĠĠĠĠĠĠchunk": 6638, "ĊĠĠĠĠĠĠĠĠchunk_split_mode:": 6639, "ĊĠĠĠĠĠĠĠĠcolor": 6640, "ĊĠĠĠĠĠĠĠĠcurv": 6641, "ĊĠĠĠĠĠĠĠĠd_ff:": 6642, "ĊĠĠĠĠĠĠĠĠd_k": 6643, "ĊĠĠĠĠĠĠĠĠd_model:": 6644, "ĊĠĠĠĠĠĠĠĠdevice:": 6645, "ĊĠĠĠĠĠĠĠĠdropout:": 6646, "ĊĠĠĠĠĠĠĠĠdtype:": 6647, "ĊĠĠĠĠĠĠĠĠe": 6648, "ĊĠĠĠĠĠĠĠĠelse:": 6649, "ĊĠĠĠĠĠĠĠĠeps:": 6650, "ĊĠĠĠĠĠĠĠĠexcept": 6651, "ĊĠĠĠĠĠĠĠĠf": 6652, "ĊĠĠĠĠĠĠĠĠfeed_forward_factory:": 6653, "ĊĠĠĠĠĠĠĠĠflops_per_token": 6654, "ĊĠĠĠĠĠĠĠĠfor": 6655, "ĊĠĠĠĠĠĠĠĠh": 6656, "ĊĠĠĠĠĠĠĠĠhistory[\"": 6657, "ĊĠĠĠĠĠĠĠĠif": 6658, "ĊĠĠĠĠĠĠĠĠk": 6659, "ĊĠĠĠĠĠĠĠĠkey_length:": 6660, "ĊĠĠĠĠĠĠĠĠlog": 6661, "ĊĠĠĠĠĠĠĠĠlog_n": 6662, "ĊĠĠĠĠĠĠĠĠlog_scale:": 6663, "ĊĠĠĠĠĠĠĠĠlr:": 6664, "ĊĠĠĠĠĠĠĠĠm": 6665, "ĊĠĠĠĠĠĠĠĠm,": 6666, "ĊĠĠĠĠĠĠĠĠmax_": 6667, "ĊĠĠĠĠĠĠĠĠmax_chunk_bytes:": 6668, "ĊĠĠĠĠĠĠĠĠmax_distance": 6669, "ĊĠĠĠĠĠĠĠĠmax_distance:": 6670, "ĊĠĠĠĠĠĠĠĠmerges": 6671, "ĊĠĠĠĠĠĠĠĠmodel:": 6672, "ĊĠĠĠĠĠĠĠĠn": 6673, "ĊĠĠĠĠĠĠĠĠn:": 6674, "ĊĠĠĠĠĠĠĠĠnew": 6675, "ĊĠĠĠĠĠĠĠĠnn.": 6676, "ĊĠĠĠĠĠĠĠĠnorm_first:": 6677, "ĊĠĠĠĠĠĠĠĠnorm_fn": 6678, "ĊĠĠĠĠĠĠĠĠnormalization_factory:": 6679, "ĊĠĠĠĠĠĠĠĠnum": 6680, "ĊĠĠĠĠĠĠĠĠnum_buckets:": 6681, "ĊĠĠĠĠĠĠĠĠnum_heads:": 6682, "ĊĠĠĠĠĠĠĠĠp": 6683, "ĊĠĠĠĠĠĠĠĠparam": 6684, "ĊĠĠĠĠĠĠĠĠparams:": 6685, "ĊĠĠĠĠĠĠĠĠparams_": 6686, "ĊĠĠĠĠĠĠĠĠpath": 6687, "ĊĠĠĠĠĠĠĠĠpositional_transform:": 6688, "ĊĠĠĠĠĠĠĠĠquery_length:": 6689, "ĊĠĠĠĠĠĠĠĠraise": 6690, "ĊĠĠĠĠĠĠĠĠratio": 6691, "ĊĠĠĠĠĠĠĠĠrelative_position": 6692, "ĊĠĠĠĠĠĠĠĠresiduals": 6693, "ĊĠĠĠĠĠĠĠĠresiduals=": 6694, "ĊĠĠĠĠĠĠĠĠreturn": 6695, "ĊĠĠĠĠĠĠĠĠs": 6696, "ĊĠĠĠĠĠĠĠĠseed:": 6697, "ĊĠĠĠĠĠĠĠĠself,": 6698, "ĊĠĠĠĠĠĠĠĠself.": 6699, "ĊĠĠĠĠĠĠĠĠself.beta": 6700, "ĊĠĠĠĠĠĠĠĠself.d_k": 6701, "ĊĠĠĠĠĠĠĠĠself.dropout": 6702, "ĊĠĠĠĠĠĠĠĠself.eps": 6703, "ĊĠĠĠĠĠĠĠĠself.lr": 6704, "ĊĠĠĠĠĠĠĠĠself.norm": 6705, "ĊĠĠĠĠĠĠĠĠself.num_heads": 6706, "ĊĠĠĠĠĠĠĠĠself.register": 6707, "ĊĠĠĠĠĠĠĠĠself.register_buffer(\"": 6708, "ĊĠĠĠĠĠĠĠĠself.step_count": 6709, "ĊĠĠĠĠĠĠĠĠself.w_": 6710, "ĊĠĠĠĠĠĠĠĠseq_len": 6711, "ĊĠĠĠĠĠĠĠĠsuper().__init__()": 6712, "ĊĠĠĠĠĠĠĠĠt": 6713, "ĊĠĠĠĠĠĠĠĠtext:": 6714, "ĊĠĠĠĠĠĠĠĠtitle:": 6715, "ĊĠĠĠĠĠĠĠĠtoken_ids:": 6716, "ĊĠĠĠĠĠĠĠĠtry:": 6717, "ĊĠĠĠĠĠĠĠĠv": 6718, "ĊĠĠĠĠĠĠĠĠvalues": 6719, "ĊĠĠĠĠĠĠĠĠvocab": 6720, "ĊĠĠĠĠĠĠĠĠvocab:": 6721, "ĊĠĠĠĠĠĠĠĠvocabulary_size:": 6722, "ĊĠĠĠĠĠĠĠĠweight_decay:": 6723, "ĊĠĠĠĠĠĠĠĠweights:": 6724, "ĊĠĠĠĠĠĠĠĠx": 6725, "ĊĠĠĠĠĠĠĠĠx:": 6726, "ĊĠĠĠĠĠĠĠĠx_c": 6727, "ĊĠĠĠĠĠĠĠĠxlabel:": 6728, "ĊĠĠĠĠĠĠĠĠylabel:": 6729, "ĊĠĠĠĠĠĠĠĠåıĸå¾ĹãģĹãģŁè¨ĺäºĭæľ¬æĸĩãĤĴéĢ£çµĲãģĹãģŁ": 6730, "ĊĠĠĠĠĠĠĠĠæııçĶ»ãģ«ä½¿ãģ£ãģŁ": 6731, "ĊĠĠĠĠĠĠĠĠĠĠ": 6732, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ": 6733, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ\"": 6734, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ#": 6735, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(": 6736, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(004": 6737, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(``": 6738, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ(output,": 6739, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ)": 6740, "ĊĠĠĠĠĠĠĠĠĠĠĠĠFalse": 6741, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ[": 6742, "ĊĠĠĠĠĠĠĠĠĠĠĠĠ``": 6743, "ĊĠĠĠĠĠĠĠĠĠĠĠĠax": 6744, "ĊĠĠĠĠĠĠĠĠĠĠĠĠb": 6745, "ĊĠĠĠĠĠĠĠĠĠĠĠĠbreak": 6746, "ĊĠĠĠĠĠĠĠĠĠĠĠĠcolor=\"tab:": 6747, "ĊĠĠĠĠĠĠĠĠĠĠĠĠcontinue": 6748, "ĊĠĠĠĠĠĠĠĠĠĠĠĠcurrent": 6749, "ĊĠĠĠĠĠĠĠĠĠĠĠĠf": 6750, "ĊĠĠĠĠĠĠĠĠĠĠĠĠf\"": 6751, "ĊĠĠĠĠĠĠĠĠĠĠĠĠfeed_forward_factory": 6752, "ĊĠĠĠĠĠĠĠĠĠĠĠĠfor": 6753, "ĊĠĠĠĠĠĠĠĠĠĠĠĠhistory": 6754, "ĊĠĠĠĠĠĠĠĠĠĠĠĠi": 6755, "ĊĠĠĠĠĠĠĠĠĠĠĠĠif": 6756, "ĊĠĠĠĠĠĠĠĠĠĠĠĠlabel=": 6757, "ĊĠĠĠĠĠĠĠĠĠĠĠĠlin": 6758, "ĊĠĠĠĠĠĠĠĠĠĠĠĠlog": 6759, "ĊĠĠĠĠĠĠĠĠĠĠĠĠm": 6760, "ĊĠĠĠĠĠĠĠĠĠĠĠĠn": 6761, "ĊĠĠĠĠĠĠĠĠĠĠĠĠnew": 6762, "ĊĠĠĠĠĠĠĠĠĠĠĠĠnp.log(warm_fit": 6763, "ĊĠĠĠĠĠĠĠĠĠĠĠĠnp.log(warm_fit.": 6764, "ĊĠĠĠĠĠĠĠĠĠĠĠĠoutput": 6765, "ĊĠĠĠĠĠĠĠĠĠĠĠĠp.grad": 6766, "ĊĠĠĠĠĠĠĠĠĠĠĠĠpositions:": 6767, "ĊĠĠĠĠĠĠĠĠĠĠĠĠraise": 6768, "ĊĠĠĠĠĠĠĠĠĠĠĠĠrelative_position": 6769, "ĊĠĠĠĠĠĠĠĠĠĠĠĠreturn": 6770, "ĊĠĠĠĠĠĠĠĠĠĠĠĠself.": 6771, "ĊĠĠĠĠĠĠĠĠĠĠĠĠself_": 6772, "ĊĠĠĠĠĠĠĠĠĠĠĠĠx": 6773, "ĊĠĠĠĠĠĠĠĠĠĠĠĠx:": 6774, "ĊĠĠĠĠĠĠĠĠĠĠĠĠx_hat": 6775, "ĊĠĠĠĠĠĠĠĠĠĠĠĠãģ®": 6776, "ĊĠĠĠĠĠĠĠĠĠĠĠĠå½¢çĬ¶": 6777, "ĊĠĠĠĠĠĠĠĠĠĠĠĠåł´åĲĪãģ¯": 6778, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ": 6779, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ(": 6780, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ)": 6781, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ``": 6782, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠb": 6783, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠcontinue": 6784, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠf\"": 6785, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠfor": 6786, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠif": 6787, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠparam": 6788, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠparams,": 6789, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠs": 6790, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠself.": 6791, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ": 6792, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠbest": 6793, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠj": 6794, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ": 6795, "ĊĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠnp.log(": 6796, "ċ": 6797, "Č": 6798, "č": 6799, "Ď": 6800, "ď": 6801, "Đ": 6802, "đ": 6803, "Ē": 6804, "ē": 6805, "Ĕ": 6806, "ĕ": 6807, "Ė": 6808, "ė": 6809, "Ę": 6810, "ę": 6811, "Ě": 6812, "ě": 6813, "Ĝ": 6814, "ĝ": 6815, "Ğ": 6816, "ğ": 6817, "Ġ": 6818, "Ġ!=": 6819, "Ġ\"": 6820, "Ġ\"\"": 6821, "Ġ\"\",": 6822, "Ġ\"\".join(": 6823, "Ġ\"Attention": 6824, "Ġ\"G": 6825, "Ġ\"Loss\",": 6826, "Ġ\"S": 6827, "Ġ\"T": 6828, "Ġ\"Value\",": 6829, "Ġ\"log": 6830, "Ġ$": 6831, "Ġ%": 6832, "Ġ&": 6833, "Ġ&=": 6834, "Ġ(": 6835, "Ġ(-": 6836, "Ġ(...,": 6837, "Ġ(0.": 6838, "Ġ(0.1,": 6839, "Ġ(009": 6840, "Ġ(1": 6841, "Ġ(2": 6842, "Ġ(20": 6843, "Ġ(B,": 6844, "Ġ(S_q,": 6845, "Ġ(b": 6846, "Ġ(h,": 6847, "Ġ(i": 6848, "Ġ(log": 6849, "Ġ(n": 6850, "Ġ(name,": 6851, "Ġ(np.": 6852, "Ġ(x": 6853, "Ġ(x_c": 6854, "Ġ({": 6855, "Ġ*": 6856, "Ġ**": 6857, "Ġ+": 6858, "Ġ+=": 6859, "Ġ-": 6860, "Ġ-(": 6861, "Ġ---------------------------------------------------------------": 6862, "Ġ-=": 6863, "Ġ->": 6864, "Ġ-inf": 6865, "Ġ...": 6866, "Ġ...,": 6867, "Ġ/": 6868, "Ġ//": 6869, "Ġ0": 6870, "Ġ0,": 6871, "Ġ0.": 6872, "Ġ0.0": 6873, "Ġ0.0,": 6874, "Ġ0.0:": 6875, "Ġ0.3": 6876, "Ġ0.5": 6877, "Ġ0.5,": 6878, "Ġ0.7": 6879, "Ġ0.8": 6880, "Ġ0.999": 6881, "Ġ003": 6882, "Ġ004": 6883, "Ġ005": 6884, "Ġ01": 6885, "Ġ010)": 6886, "Ġ0:": 6887, "Ġ1": 6888, "Ġ1)": 6889, "Ġ1))": 6890, "Ġ1):": 6891, "Ġ1,": 6892, "Ġ1.0": 6893, "Ġ1.0)": 6894, "Ġ1.0,": 6895, "Ġ10": 6896, "Ġ100": 6897, "Ġ10000": 6898, "Ġ10000^(2i": 6899, "Ġ1:": 6900, "Ġ1]": 6901, "Ġ1e": 6902, "Ġ1e-": 6903, "Ġ1e-1": 6904, "Ġ1e-12)": 6905, "Ġ1e-8": 6906, "Ġ2": 6907, "Ġ2)": 6908, "Ġ2,": 6909, "Ġ2-3": 6910, "Ġ2.0": 6911, "Ġ2.5": 6912, "Ġ20": 6913, "Ġ201": 6914, "Ġ2016": 6915, "Ġ2017": 6916, "Ġ2017)ãĢĤ": 6917, "Ġ2018": 6918, "Ġ2019": 6919, "Ġ202": 6920, "Ġ2020": 6921, "Ġ2024": 6922, "Ġ21)ãĢĤ": 6923, "Ġ2:": 6924, "Ġ2]": 6925, "Ġ3": 6926, "Ġ3:": 6927, "Ġ4": 6928, "Ġ4.0": 6929, "Ġ4.2))": 6930, "Ġ4.5))": 6931, "Ġ429": 6932, "Ġ5": 6933, "Ġ6": 6934, "Ġ:": 6935, "Ġ:,": 6936, "Ġ:]": 6937, "Ġ:math:`": 6938, "Ġ<": 6939, "Ġ<-": 6940, "Ġ<->": 6941, "Ġ<=": 6942, "Ġ=": 6943, "Ġ==": 6944, "Ġ>": 6945, "Ġ@": 6946, "ĠA": 6947, "ĠA,": 6948, "ĠA/N^alpha": 6949, "ĠABC": 6950, "ĠAPI": 6951, "ĠAdam": 6952, "ĠAdamW": 6953, "ĠAll": 6954, "ĠApproach": 6955, "ĠAttention": 6956, "ĠAx": 6957, "ĠAxes": 6958, "ĠAxes:": 6959, "ĠAxesãĢĤ": 6960, "ĠAxesãĢĤNone": 6961, "ĠB": 6962, "ĠB/D^beta": 6963, "ĠBPE": 6964, "ĠBPEIDTokenizer": 6965, "ĠBPETokenizer": 6966, "ĠBlock": 6967, "ĠBlock(": 6968, "ĠBãĢĤ": 6969, "ĠC": 6970, "ĠC^": 6971, "ĠCallab": 6972, "ĠCallable": 6973, "ĠCallable[": 6974, "ĠCallable[[": 6975, "ĠCallable[[Tensor],": 6976, "ĠCallable[[],": 6977, "ĠCallable[[int],": 6978, "ĠCallable[[np.ndarray],": 6979, "ĠCharacterLevelTokenizer": 6980, "ĠCharacterLevelTokenizer:": 6981, "ĠChinchilla": 6982, "ĠChinchillaParametricFit": 6983, "ĠCon": 6984, "ĠCounter": 6985, "ĠCounter[tuple[str,": 6986, "ĠCrossAttention": 6987, "ĠD": 6988, "ĠDecoder": 6989, "ĠDecoderBlock": 6990, "ĠDot-Product": 6991, "ĠE": 6992, "ĠEM": 6993, "ĠEmbedding": 6994, "ĠEncoder": 6995, "ĠEncoding": 6996, "ĠError": 6997, "ĠF": 6998, "ĠFace": 6999, "ĠFalse": 7000, "ĠFalse,": 7001, "ĠFeed": 7002, "ĠFeed-Forward": 7003, "ĠFeedForwardNetwork": 7004, "ĠFig": 7005, "ĠFunction": 7006, "ĠG": 7007, "ĠGPTLanguageModel": 7008, "ĠH": 7009, "ĠHub": 7010, "ĠHuber": 7011, "ĠI": 7012, "ĠICLR": 7013, "ĠID": 7014, "ĠIs": 7015, "ĠIsoFLOP": 7016, "ĠIsoFLOPParabolaFit": 7017, "ĠIter": 7018, "ĠJ": 7019, "ĠJSON": 7020, "ĠJaccard": 7021, "ĠK": 7022, "ĠKey": 7023, "ĠL": 7024, "ĠL2": 7025, "ĠLanguage": 7026, "ĠLaw": 7027, "ĠLayerNorm": 7028, "ĠLayerNorm(x": 7029, "ĠLayerNorm(x),": 7030, "ĠLayerNormalization": 7031, "ĠLearn": 7032, "ĠLevenberg-Marquardt": 7033, "ĠLinear": 7034, "ĠLongTensor": 7035, "ĠLongTensor(": 7036, "ĠLongTensor(``": 7037, "ĠLongTensorãĢĤ": 7038, "ĠM": 7039, "ĠMean": 7040, "ĠModel": 7041, "ĠMulti-Head": 7042, "ĠMultiHeadAttention": 7043, "ĠMultiHeadAttention(d_model,": 7044, "ĠN": 7045, "ĠN,": 7046, "ĠN_opt": 7047, "ĠNeed\",": 7048, "ĠNetwork": 7049, "ĠNeur": 7050, "ĠNeurIPS": 7051, "ĠNeural": 7052, "ĠNone": 7053, "ĠNone)": 7054, "ĠNone,": 7055, "ĠNone:": 7056, "ĠNone]": 7057, "ĠNormalization": 7058, "ĠNormalization\",": 7059, "ĠNotImplementedError": 7060, "ĠNotation:": 7061, "ĠOS": 7062, "ĠP": 7063, "ĠPair": 7064, "ĠPath": 7065, "ĠPath(": 7066, "ĠPath)": 7067, "ĠPath,": 7068, "ĠPowerLawFit": 7069, "ĠQuery": 7070, "ĠQueryKeyPositionalTransform": 7071, "ĠQueryãĥ»Key": 7072, "ĠR": 7073, "ĠRMS": 7074, "ĠRMSNorm": 7075, "ĠR^{": 7076, "ĠR^{d_ff": 7077, "ĠR^{d_model": 7078, "ĠReLU": 7079, "ĠS": 7080, "ĠS)``": 7081, "ĠS,": 7082, "ĠS_": 7083, "ĠS_0": 7084, "ĠS_k": 7085, "ĠS_k)": 7086, "ĠS_k)``": 7087, "ĠS_k,": 7088, "ĠS_q,": 7089, "ĠS_tgt": 7090, "ĠS_tgt,": 7091, "ĠSaturatingPowerLawFit": 7092, "ĠScal": 7093, "ĠSelfAttention": 7094, "ĠSennrich": 7095, "ĠSequence": 7096, "ĠSequence[": 7097, "ĠSequence[float": 7098, "ĠSequence[float])": 7099, "ĠSequence[float],": 7100, "ĠSequence[int]": 7101, "ĠSequence[str": 7102, "ĠSequence[str])": 7103, "ĠSequence[str],": 7104, "ĠShaw": 7105, "ĠSiLU": 7106, "ĠSquar": 7107, "ĠSquare": 7108, "ĠSt": 7109, "ĠSwiGLU": 7110, "ĠSwish": 7111, "ĠSãĢĤ": 7112, "ĠT": 7113, "ĠTensor": 7114, "ĠTensor)": 7115, "ĠTensor):": 7116, "ĠTensor,": 7117, "ĠTensor:": 7118, "ĠTensor]": 7119, "ĠTensor]:": 7120, "ĠText": 7121, "ĠTransformer": 7122, "ĠTrue": 7123, "ĠTrue(æĹ¢å®ļåĢ¤": 7124, "ĠTrue,": 7125, "ĠU": 7126, "ĠUTF-8": 7127, "ĠUn": 7128, "ĠUnicode": 7129, "ĠUnigram": 7130, "ĠUnigramTokenizer": 7131, "ĠUnit": 7132, "ĠV": 7133, "ĠV)": 7134, "ĠV)``": 7135, "ĠValue": 7136, "ĠValueError": 7137, "ĠValueError(": 7138, "ĠValueError(\"": 7139, "ĠValueError(f\"": 7140, "ĠViterbi": 7141, "ĠW": 7142, "ĠW^": 7143, "ĠW^O": 7144, "ĠW_2": 7145, "ĠWikipedia": 7146, "ĠYou": 7147, "Ġ[": 7148, "Ġ[]": 7149, "Ġ[],": 7150, "Ġ[p.": 7151, "Ġ\\\\": 7152, "Ġ\\\\\\\\": 7153, "Ġ\\\\beta": 7154, "Ġ\\\\beta_": 7155, "Ġ\\\\beta_1": 7156, "Ġ\\\\beta_2": 7157, "Ġ\\\\frac{": 7158, "Ġ\\\\lambda": 7159, "Ġ\\\\text{": 7160, "Ġ\\\\theta": 7161, "Ġ\\\\theta_{t-1}": 7162, "Ġ_": 7163, "Ġ_,": 7164, "Ġ_LOG_ALPHA_BETA_M": 7165, "Ġ_LOG_ALPHA_BETA_MAX": 7166, "Ġ_LOG_ALPHA_BETA_MAX)": 7167, "Ġ_LOG_ALPHA_BETA_MIN,": 7168, "Ġ_WIKIPEDIA_MANIFEST_DIR": 7169, "Ġ__future__": 7170, "Ġ__init__(": 7171, "Ġ__init__(self,": 7172, "Ġ_boundaries(": 7173, "Ġ_build_": 7174, "Ġ_chinchilla_residual": 7175, "Ġ_count_pair": 7176, "Ġ_count_pairs(": 7177, "Ġ_fit_nonlinear_huber": 7178, "Ġ_gnmt_length_penalty": 7179, "Ġ_levenberg_marquardt(": 7180, "Ġ_merge_": 7181, "Ġ_merge_pair_in_word(": 7182, "Ġ_power_of_2_slopes": 7183, "Ġ_strip_wikitext_links(": 7184, "Ġ_to_numpy": 7185, "Ġ``": 7186, "Ġ``\"": 7187, "Ġ``\"gradient_norm": 7188, "Ġ``\"loss_step_delta": 7189, "Ġ``\"loss_step_delta\"``": 7190, "Ġ``(": 7191, "Ġ``(...,": 7192, "Ġ``(1,": 7193, "Ġ``(B,": 7194, "Ġ``(N,": 7195, "Ġ``(S,": 7196, "Ġ``(S_q,": 7197, "Ġ``(num_windows,": 7198, "Ġ``-inf``": 7199, "Ġ``0": 7200, "Ġ``0.0": 7201, "Ġ``6": 7202, "Ġ``C": 7203, "Ġ``C``": 7204, "Ġ``DecoderBlock``": 7205, "Ġ``L": 7206, "Ġ``MultiHeadAttention``": 7207, "Ġ``None``": 7208, "Ġ``_": 7209, "Ġ``beam_size": 7210, "Ġ``beam_size``": 7211, "Ġ``d_model``": 7212, "Ġ``log": 7213, "Ġ``lp(Y": 7214, "Ġ``n": 7215, "Ġ``n``": 7216, "Ġ``nn.": 7217, "Ġ``p``": 7218, "Ġ``torch.": 7219, "Ġ``{": 7220, "Ġ``{æĿ¡ä»¶åĲį:": 7221, "Ġ``{ç³»åĪĹåĲį:": 7222, "Ġ``âĸģ": 7223, "Ġ``âĸģ``": 7224, "Ġa": 7225, "Ġa^K_mn": 7226, "Ġa_coef": 7227, "Ġal": 7228, "Ġal.": 7229, "Ġal.,": 7230, "Ġalpha": 7231, "Ġalpha=-": 7232, "Ġand": 7233, "Ġann": 7234, "Ġannotations": 7235, "Ġar": 7236, "ĠarXiv:": 7237, "Ġas": 7238, "Ġattempt": 7239, "Ġattn_weights": 7240, "Ġattn_weights)": 7241, "Ġax": 7242, "Ġaxis": 7243, "Ġaxis=\"y\")": 7244, "Ġb": 7245, "Ġb,": 7246, "Ġb_coef": 7247, "Ġbase": 7248, "Ġbatch": 7249, "Ġbeam": 7250, "Ġbest": 7251, "Ġbest_cost": 7252, "Ġbest_pair": 7253, "Ġbest_params": 7254, "Ġbest_score": 7255, "Ġbeta": 7256, "Ġbeta:": 7257, "Ġbias": 7258, "Ġbias(": 7259, "Ġbias=bias)": 7260, "Ġbias_correction": 7261, "Ġbits-per-byte": 7262, "Ġblock": 7263, "Ġbool": 7264, "Ġbool]:": 7265, "Ġboundaries": 7266, "Ġbpe.": 7267, "Ġbyt": 7268, "Ġbyte_": 7269, "Ġc": 7270, "Ġcache_dir": 7271, "Ġcache_dir,": 7272, "Ġcache_path": 7273, "Ġcallable": 7274, "Ġcenter": 7275, "Ġch": 7276, "Ġchar": 7277, "Ġchunk": 7278, "Ġchunk:": 7279, "Ġchunk_freq": 7280, "Ġchunks": 7281, "Ġclip": 7282, "Ġclipping": 7283, "Ġco": 7284, "Ġcoeff": 7285, "Ġcollection": 7286, "Ġcollections.abc": 7287, "Ġcolor": 7288, "Ġcolor=\"": 7289, "Ġcolor=color,": 7290, "Ġcolor_cycle[i": 7291, "Ġcompute_": 7292, "Ġcon": 7293, "Ġconverged": 7294, "Ġcos": 7295, "Ġcosine": 7296, "Ġcost": 7297, "Ġcount": 7298, "Ġcre": 7299, "Ġcreate_causal_mask": 7300, "Ġcross_": 7301, "Ġcross_attn_weights": 7302, "Ġcum_log_prob": 7303, "Ġcurrent": 7304, "Ġcurv": 7305, "Ġd": 7306, "Ġd_": 7307, "Ġd_ff": 7308, "Ġd_ff'": 7309, "Ġd_ff,": 7310, "Ġd_k": 7311, "Ġd_k)": 7312, "Ġd_k)``": 7313, "Ġd_k/": 7314, "Ġd_model": 7315, "Ġd_model)": 7316, "Ġd_model))": 7317, "Ġd_model)``": 7318, "Ġd_model)``ãĢģ": 7319, "Ġd_model)``ãĢĤ": 7320, "Ġd_model,": 7321, "Ġd_model:": 7322, "Ġd_model``": 7323, "Ġd_v": 7324, "Ġd_values,": 7325, "Ġdata": 7326, "Ġdata[\"": 7327, "Ġdataclass": 7328, "Ġdecay": 7329, "Ġdecod": 7330, "Ġdecode(self,": 7331, "Ġdecoder-only": 7332, "Ġdefault": 7333, "Ġdelta": 7334, "Ġdevice=": 7335, "Ġdevice=device)": 7336, "Ġdevice=query.device)": 7337, "Ġdi": 7338, "Ġdict": 7339, "Ġdict(": 7340, "Ġdict[float,": 7341, "Ġdict[int,": 7342, "Ġdict[str,": 7343, "Ġdim": 7344, "Ġdim=-1)": 7345, "Ġdocstring": 7346, "Ġdropout": 7347, "Ġdropout=": 7348, "Ġdropout=dropout)": 7349, "Ġdropout=dropout,": 7350, "Ġdtype=": 7351, "Ġdtype=float": 7352, "Ġdtype=float)": 7353, "Ġdtype=torch.": 7354, "Ġdtype=torch.bool": 7355, "Ġdtype=torch.float3": 7356, "Ġdtype=torch.float32)": 7357, "Ġe": 7358, "Ġe_mn": 7359, "Ġelse": 7360, "Ġen": 7361, "Ġencode(self,": 7362, "Ġencoding=\"utf-8\")": 7363, "Ġend": 7364, "Ġenumerate(": 7365, "Ġeps": 7366, "Ġet": 7367, "Ġevaluate_bits_per_byt": 7368, "Ġevaluation_window": 7369, "Ġexist_ok": 7370, "Ġexist_ok=True)": 7371, "Ġexp": 7372, "Ġextra": 7373, "Ġf": 7374, "Ġf\"": 7375, "Ġf\"{": 7376, "Ġfertility": 7377, "Ġfit_": 7378, "Ġfloat": 7379, "Ġfloat(": 7380, "Ġfloat(\"": 7381, "Ġfloat(\"-inf\"": 7382, "Ġfloat(\"-inf\"))": 7383, "Ġfloat(np.": 7384, "Ġfloat(np.exp(log": 7385, "Ġfloat(np.mean": 7386, "Ġfloat(np.median": 7387, "Ġfloat(np.sum(": 7388, "Ġfloat)": 7389, "Ġfloat,": 7390, "Ġfloat:": 7391, "Ġfloat]": 7392, "Ġfloat],": 7393, "Ġfloat]:": 7394, "Ġflops_per_token": 7395, "Ġfontsize=": 7396, "Ġfor": 7397, "Ġforward(": 7398, "Ġforward(self,": 7399, "Ġfreq": 7400, "Ġfreq)": 7401, "Ġfunctional": 7402, "Ġg": 7403, "Ġg_t": 7404, "Ġgenerator": 7405, "Ġget_random_batch": 7406, "Ġgradient": 7407, "Ġgradient_clip_threshold": 7408, "Ġgradient_norm": 7409, "Ġgrid": 7410, "Ġgrid],": 7411, "Ġh": 7412, "Ġh,": 7413, "Ġhead": 7414, "Ġhistory": 7415, "Ġhuber_delta": 7416, "Ġhuber_delta)": 7417, "ĠhãĢĤ": 7418, "Ġi": 7419, "Ġi,": 7420, "Ġid": 7421, "Ġid(p)": 7422, "Ġif": 7423, "Ġimport": 7424, "Ġin": 7425, "Ġinput": 7426, "Ġint": 7427, "Ġint)": 7428, "Ġint,": 7429, "Ġint]": 7430, "Ġinvalid=\"ignore\"):": 7431, "Ġis": 7432, "Ġisinstance": 7433, "Ġitem": 7434, "Ġitem:": 7435, "Ġj": 7436, "Ġjo": 7437, "Ġjson.loads(": 7438, "Ġjson.loads(Path": 7439, "Ġk": 7440, "Ġk,": 7441, "Ġk_clip": 7442, "Ġkeepdim": 7443, "Ġkeepdim=True)": 7444, "Ġkey": 7445, "Ġl": 7446, "Ġl_inf": 7447, "Ġlabel=": 7448, "Ġlabel=name,": 7449, "Ġlen": 7450, "Ġlen(": 7451, "Ġlen(color_cycle": 7452, "Ġlen(color_cycle)]": 7453, "Ġlength)": 7454, "Ġlin": 7455, "Ġline": 7456, "Ġlinewidth": 7457, "Ġlinewidth=1.": 7458, "Ġlist": 7459, "Ġlist(": 7460, "Ġlist[": 7461, "Ġlist[Tensor]": 7462, "Ġlist[float": 7463, "Ġlist[float]": 7464, "Ġlist[float]:": 7465, "Ġlist[float]],": 7466, "Ġlist[str": 7467, "Ġlist[str]": 7468, "Ġlist[str]:": 7469, "Ġlm_head": 7470, "Ġlo": 7471, "Ġload_": 7472, "Ġload_bpe_": 7473, "Ġload_wikipedia_corpus": 7474, "Ġlog": 7475, "Ġlog_": 7476, "Ġlog_alpha": 7477, "Ġlog_alpha_max": 7478, "Ġlog_n": 7479, "Ġlog_n_opt": 7480, "Ġlog_scale:": 7481, "Ġlogit": 7482, "Ġlogits": 7483, "Ġlogits``": 7484, "ĠlogitsãĢĤ": 7485, "Ġloss_": 7486, "Ġlp": 7487, "Ġlr": 7488, "Ġlr:": 7489, "Ġm": 7490, "Ġm,": 7491, "Ġmanifest_path": 7492, "Ġmanifest_path=": 7493, "Ġmask": 7494, "Ġmask)": 7495, "Ġmask.dim()": 7496, "Ġmath": 7497, "Ġmath.": 7498, "Ġmath.sqrt(": 7499, "Ġmath::": 7500, "Ġmatplotlib": 7501, "Ġmatrix": 7502, "Ġmatrix.ndim": 7503, "Ġmax(": 7504, "Ġmax_": 7505, "Ġmax_chunk": 7506, "Ġmax_chunk_bytes:": 7507, "Ġmax_exact": 7508, "Ġmax_exact)": 7509, "Ġmax_len": 7510, "Ġmax_relative_position": 7511, "Ġmean": 7512, "Ġmemory": 7513, "Ġmemory,": 7514, "Ġmemory_": 7515, "Ġmemory_mask)": 7516, "Ġmin": 7517, "Ġmin(": 7518, "Ġmodel": 7519, "Ġmodel.": 7520, "Ġmodel.parameters()": 7521, "Ġmulti-start": 7522, "Ġn": 7523, "Ġn-gram": 7524, "Ġn:": 7525, "Ġn_": 7526, "Ġn_ctx": 7527, "Ġn_layer": 7528, "Ġnam": 7529, "Ġname": 7530, "Ġneg": 7531, "Ġnew": 7532, "Ġnew_symbol": 7533, "Ġnn": 7534, "Ġnn.": 7535, "Ġnn.Dropout": 7536, "Ġnn.Dropout(dropout)": 7537, "Ġnn.Embedding": 7538, "Ġnn.Linear": 7539, "Ġnn.Linear(d_model,": 7540, "Ġnn.Module": 7541, "Ġnn.Module]": 7542, "Ġnn.Parameter(torch.": 7543, "Ġnoise_band": 7544, "Ġnoqa:": 7545, "Ġnorm": 7546, "Ġnorm_fn(d_model)": 7547, "Ġnormalization_factory": 7548, "Ġnormed,": 7549, "Ġnot": 7550, "Ġnp": 7551, "Ġnp.": 7552, "Ġnp.arange(": 7553, "Ġnp.array(": 7554, "Ġnp.array([": 7555, "Ġnp.array([p.": 7556, "Ġnp.as": 7557, "Ġnp.asarray": 7558, "Ġnp.asarray(": 7559, "Ġnp.errstate(over=\"ignore\",": 7560, "Ġnp.exp(log": 7561, "Ġnp.inf": 7562, "Ġnp.isfinite(": 7563, "Ġnp.log": 7564, "Ġnp.log(": 7565, "Ġnp.ndarray": 7566, "Ġnp.ndarray)": 7567, "Ġnp.ndarray,": 7568, "Ġnp.ndarray:": 7569, "Ġnp.ndarray],": 7570, "Ġnp.ones_like(": 7571, "Ġnum": 7572, "Ġnum_buckets": 7573, "Ġnum_heads": 7574, "Ġnum_heads,": 7575, "Ġnum_total": 7576, "Ġnumpy": 7577, "Ġo": 7578, "Ġobject": 7579, "Ġof": 7580, "Ġoptimizer": 7581, "Ġor": 7582, "Ġother_": 7583, "Ġp": 7584, "Ġp.grad": 7585, "Ġpair": 7586, "Ġpair,": 7587, "Ġpair_in": 7588, "Ġparabola.": 7589, "Ġparam": 7590, "Ġparams": 7591, "Ġparams)": 7592, "Ġparams,": 7593, "Ġpath": 7594, "Ġpersistent=False)": 7595, "Ġpiece": 7596, "Ġplot": 7597, "Ġplot_": 7598, "Ġplt": 7599, "Ġplt.rcParams[\"axes.prop_cycle\"].by_key()[\"color\"]": 7600, "Ġplt.subplots(": 7601, "Ġplt.subplots(figsize=(": 7602, "Ġplt.subplots(figsize=(6": 7603, "Ġplt.subplots(figsize=(6.5,": 7604, "Ġpoint": 7605, "Ġpoints]": 7606, "Ġpos": 7607, "Ġpositional_transform": 7608, "Ġpositions": 7609, "Ġpositions:": 7610, "Ġpre": 7611, "Ġpretokenize(": 7612, "Ġpro": 7613, "Ġq": 7614, "Ġq.": 7615, "Ġq_m": 7616, "Ġquery": 7617, "Ġr": 7618, "Ġran": 7619, "Ġrange(": 7620, "Ġrange(n": 7621, "Ġrange(num": 7622, "Ġraw_norm": 7623, "Ġre": 7624, "Ġre.sub": 7625, "Ġre.sub(r\"": 7626, "Ġre.sub(r\"(?": 7627, "Ġre.sub(r\"<": 7628, "Ġreconstruct_optimal_frontier(": 7629, "Ġrelative_": 7630, "Ġrelative_buckets": 7631, "Ġrelative_position": 7632, "Ġrelative_vector": 7633, "Ġresidual_fn": 7634, "Ġresidual_fn(param": 7635, "Ġresidual_fn(params)": 7636, "Ġrevid": 7637, "Ġs": 7638, "Ġs_": 7639, "Ġsampling": 7640, "Ġscal": 7641, "Ġscore": 7642, "Ġse": 7643, "Ġself.": 7644, "Ġself._": 7645, "Ġself._m": 7646, "Ġself._split_heads(self.w_": 7647, "Ġself.attention_score_bias": 7648, "Ġself.beta": 7649, "Ġself.beta2": 7650, "Ġself.dropout": 7651, "Ġself.dropout1(": 7652, "Ġself.dropout2": 7653, "Ġself.eps)": 7654, "Ġself.gamma": 7655, "Ġself.norm": 7656, "Ġself.norm1": 7657, "Ġself.norm2": 7658, "Ġself.params:": 7659, "Ġself.self_attn": 7660, "Ġself.w_": 7661, "Ġself.weight_decay": 7662, "Ġself_": 7663, "Ġsentencepiece": 7664, "Ġseq_len": 7665, "Ġseq_len,": 7666, "Ġsequence": 7667, "Ġsequence_length": 7668, "Ġsequence_length)": 7669, "Ġset": 7670, "Ġset()": 7671, "Ġset[str]": 7672, "Ġsin": 7673, "Ġsize": 7674, "Ġsize)": 7675, "Ġsoftmax": 7676, "Ġsorted": 7677, "Ġsorted_": 7678, "Ġsqrt(": 7679, "Ġsqrt(d_k)": 7680, "Ġsrc.": 7681, "Ġsrc.data.": 7682, "Ġsrc.layers.": 7683, "Ġsrc.layers.attention": 7684, "Ġsrc.layers.normalization": 7685, "Ġsrc.layers.positional_encoding": 7686, "Ġsrc.utils.": 7687, "Ġss_": 7688, "Ġstart": 7689, "Ġstep": 7690, "Ġstr": 7691, "Ġstr)": 7692, "Ġstr,": 7693, "Ġstr:": 7694, "Ġstr]": 7695, "Ġstr],": 7696, "Ġstr]]": 7697, "Ġstrict": 7698, "Ġstrict=True": 7699, "Ġstrict=True)": 7700, "Ġsum": 7701, "Ġswish": 7702, "Ġsymbol": 7703, "Ġsymbol_to_id": 7704, "Ġsymbols": 7705, "Ġsymbols[i": 7706, "Ġt": 7707, "Ġtarget": 7708, "Ġtargets": 7709, "Ġtemperature": 7710, "Ġtext": 7711, "Ġtext)": 7712, "Ġtext:": 7713, "Ġtgt_mask": 7714, "Ġth": 7715, "Ġthreshold": 7716, "Ġtitle": 7717, "Ġtitle:": 7718, "Ġto": 7719, "Ġtoken": 7720, "Ġtoken_id": 7721, "Ġtoken_ids": 7722, "Ġtokenizer.json": 7723, "Ġtop": 7724, "Ġtorch": 7725, "Ġtorch.": 7726, "Ġtorch.Generator": 7727, "Ġtorch.Tensor": 7728, "Ġtorch.arange(": 7729, "Ġtorch.arange(key_length,": 7730, "Ġtorch.arange(query_length,": 7731, "Ġtorch.cat": 7732, "Ġtorch.cat([": 7733, "Ġtorch.device": 7734, "Ġtorch.dtype,": 7735, "Ġtorch.m": 7736, "Ġtorch.s": 7737, "Ġtorch.sqrt(": 7738, "Ġtorch.t": 7739, "Ġtorch.tensor": 7740, "Ġtorch.zero": 7741, "Ġtorch.zeros(": 7742, "Ġtorch.zeros_like(": 7743, "Ġtorch.zeros_like(p)": 7744, "Ġtotal_bytes)": 7745, "Ġtrain": 7746, "Ġtry_decode_byte_level_symbol": 7747, "Ġtup": 7748, "Ġtuple": 7749, "Ġtuple[": 7750, "Ġtuple[Tensor,": 7751, "Ġtuple[float,": 7752, "Ġun": 7753, "Ġunit)": 7754, "Ġurllib": 7755, "Ġurllib.": 7756, "Ġuse_cross_attention": 7757, "Ġv": 7758, "Ġval": 7759, "Ġvalue": 7760, "Ġvalues": 7761, "Ġvalues)": 7762, "Ġvalues,": 7763, "Ġvalues_arr": 7764, "Ġvertex_log_n": 7765, "Ġvmax": 7766, "Ġvocab": 7767, "Ġvocab_size": 7768, "Ġvocabulary": 7769, "Ġw": 7770, "Ġwarmup": 7771, "Ġwarmup_step": 7772, "Ġwarmup_steps": 7773, "Ġweights": 7774, "Ġwikitext": 7775, "Ġx": 7776, "Ġx)": 7777, "Ġx,": 7778, "Ġx.": 7779, "Ġx:": 7780, "Ġx^": 7781, "Ġx_hat": 7782, "Ġy": 7783, "Ġzip": 7784, "Ġzip(": 7785, "Ġ{": 7786, "Ġ{\"": 7787, "Ġ{ch": 7788, "Ġ{}": 7789, "Ġ|": 7790, "ĠÃĹ": 7791, "ĠÎ": 7792, "ĠÎ²": 7793, "ĠÎ³": 7794, "ĠÎµ": 7795, "ĠÎµ)": 7796, "ĠÎµãĢĤ": 7797, "ĠÎ·": 7798, "ĠÎ¸": 7799, "ĠÎ»": 7800, "ĠÎ¼": 7801, "Ġâ": 7802, "ĠâĨ": 7803, "ĠâĨĴ": 7804, "ĠâĪ": 7805, "ĠâĪĪ": 7806, "ĠâĪĿ": 7807, "Ġâī": 7808, "ĠâīĪ": 7809, "Ġãģ¤": 7810, "Ġãģ¤ãģ®": 7811, "Ġãģ¤ãģ®åĪĨåī²": 7812, "Ġãģ¤ãģ®æĸĩåŃĹåĪĹ": 7813, "Ġãģ¤ãģ®æĸĩåŃĹåĪĹãĢĤ": 7814, "Ġãģ¤ãĤĤ": 7815, "Ġãģ§": 7816, "Ġãģ§ãĢģ": 7817, "Ġãģ§ãģ¯": 7818, "Ġãģ§ãģĤãĤĭ": 7819, "Ġãģ§è¿½åĬł": 7820, "Ġãģ§è¿½åĬł)": 7821, "Ġãģ§è¿½åĬł)ãĢĤ": 7822, "Ġãģ¨": 7823, "Ġãģ¨ãģ®åĨħç©įãĤĴ": 7824, "Ġãģ¨ãģĹãģ¦": 7825, "Ġãģ¨ãģĻãĤĭãĢĤ": 7826, "Ġãģ¨å®Įåħ¨ãģ«": 7827, "Ġãģ¨å®Įåħ¨ãģ«åĲĮä¸Ģãģ®": 7828, "Ġãģ¨åĲĮãģĺ": 7829, "Ġãģ¨åĲĮãģĺå½¢çĬ¶": 7830, "Ġãģ¨åĲĮãģĺå½¢çĬ¶ãģ®": 7831, "Ġãģ¨åĲĮãģĺå½¢çĬ¶ãģ®åĩºåĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ": 7832, "Ġãģ¨åĲĮãģĺéķ·ãģķ": 7833, "Ġãģ¨åĲĮãģĺéķ·ãģķ)ãĢĤ": 7834, "Ġãģ¨åĲĮä¸Ģãģ®æĮĻåĭķ": 7835, "Ġãģ¨åĲĮä¸Ģãģ®æĮĻåĭķ)ãĢĤ": 7836, "Ġãģª": 7837, "Ġãģªãģ©": 7838, "Ġãģªãģ©)ãĢĤ": 7839, "ĠãģªãĤī": 7840, "ĠãģªãĤīæĸ°è¦ıä½ľæĪĲãģĻãĤĭãĢĤ": 7841, "Ġãģ«": 7842, "Ġãģ«ãģªãĤĭ": 7843, "Ġãģ«ãģªãĤĭãĢĤ": 7844, "Ġãģ«ãģĻãĤĭãĢĤ": 7845, "Ġãģ«ãģĿãģ®ãģ¾ãģ¾æ¸¡ãģĻ": 7846, "Ġãģ«ãĤĪãĤĭ": 7847, "Ġãģ«ä½¿ãģĨ": 7848, "Ġãģ«å¯¾ãģĹãģ¦": 7849, "Ġãģ«å¯¾å¿ľãģĻãĤĭ": 7850, "Ġãģ«åŁºãģ¥ãģı": 7851, "Ġãģ«ç½®ãģįæıĽãģĪ": 7852, "Ġãģ«éģĶ": 7853, "Ġãģ®": 7854, "Ġãģ®ãģ¨ãģį": 7855, "Ġãģ®ãģ¨ãģįãģ¯": 7856, "Ġãģ®ãģ¿": 7857, "Ġãģ®ãĤµãĥĸãĤ¯ãĥ©ãĤ¹": 7858, "Ġãģ®ãĤµãĥĸãĤ¯ãĥ©ãĤ¹ãĢģä¾ĭ:": 7859, "Ġãģ®ãĤ¹ãĤ¯ãĥ©ãĥĥãĥģå®Łè£ħ": 7860, "Ġãģ®ãĤ¿ãĥĹ": 7861, "Ġãģ®ãĤ¿ãĥĹãĥ«ãĢĤ": 7862, "Ġãģ®ãĤ¿ãĥĹãĥ«ãĢĤãģĦãģļãĤĮãĤĤ": 7863, "Ġãģ®ãĥªãĤ¹ãĥĪ": 7864, "Ġãģ®ãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ": 7865, "Ġãģ®ãĥĺãĥĥãĥīæķ°": 7866, "Ġãģ®ãĥŀãĥ¼ãĤ¯ãĤ¢ãĥĥãĥĹ": 7867, "Ġãģ®ä¸Ńéĸĵå±¤": 7868, "Ġãģ®ä¸Ńéĸĵå±¤æ¬¡åħĥ": 7869, "Ġãģ®ä¸Ńéĸĵå±¤æ¬¡åħĥãĢĤ": 7870, "Ġãģ®å®ļç¾©": 7871, "Ġãģ®å®Łè£ħæĸ¹éĩĿ": 7872, "Ġãģ®å®Łé¨ĵ": 7873, "Ġãģ®å®Łé¨ĵ5": 7874, "Ġãģ®å¯¾å¿ľ": 7875, "Ġãģ®åħ¥åĬĽãĢĤ": 7876, "Ġãģ®åĨħå®¹": 7877, "Ġãģ®åĩºåĬĽ": 7878, "Ġãģ®åĩºåĬĽ)ãĢĤ": 7879, "Ġãģ®åĪĿæľŁåĢ¤": 7880, "Ġãģ®åł´åĲĪãĢģ": 7881, "Ġãģ®åł´åĲĪãģ¯": 7882, "Ġãģ®æ¸¬å®ļåĢ¤ãģ®ç³»åĪĹ": 7883, "Ġãģ®æĹ¢å®ļ": 7884, "Ġãģ®ç¯ĦåĽ²": 7885, "Ġãģ®ç³»åĪĹ": 7886, "Ġãģ®çµĲæŀľ": 7887, "Ġãģ®çµĲæŀľãĢĤ": 7888, "Ġãģ®çµĲæŀľãĢĤL": 7889, "Ġãģ®çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³": 7890, "Ġãģ®çĲĨè«ĸãĤ»ãĤ¯ãĤ·ãĥ§ãĥ³åıĤçħ§)ãĢĤ": 7891, "Ġãģ®è¨Ģèªŀ": 7892, "Ġãģ®è¾ŀæĽ¸ãĢĤ": 7893, "Ġãģ®éķ·ãģķ": 7894, "Ġãģ®éłĨä¼ĿæĴŃãĢĤ": 7895, "Ġãģ¯": 7896, "Ġãģ¯å½¢çĬ¶": 7897, "Ġãģ¸": 7898, "Ġãģ¾": 7899, "Ġãģ¾ãģ§ãģ®éĢ£çķª": 7900, "Ġãģ¾ãģ§ãģ®éĢ£çķªãģ¨ãģĹãģ¦æī±": 7901, "Ġãģ¾ãģ§ãģ®éĢ£çķªãģ¨ãģĹãģ¦æī±ãģĨãĢĤ": 7902, "Ġãģ¾ãģŁãģ¯": 7903, "ĠãģĭãĤī": 7904, "ĠãģĭãĤīãģ®åıĸå¾Ĺ": 7905, "ĠãģĭãĤīåĨįåĪ©çĶ¨ãģĻãĤĭ": 7906, "ĠãģĮ": 7907, "ĠãģĮãĥĳãĥĩãĤ£ãĥ³ãĤ°": 7908, "ĠãģĮå®ŁãĥĪãĥ¼ãĤ¯ãĥ³": 7909, "Ġãģĵãģ®": 7910, "ĠãģĶãģ¨ãģ«": 7911, "ĠãĤ«ãĥ©": 7912, "ĠãĤµãĥ³ãĥĹãĥªãĥ³ãĤ°": 7913, "ĠãĤ¹ãĤ±ãĤ¸ãĥ¥ãĥ¼ãĥ«": 7914, "ĠãĤ¹ãĤ³ãĤ¢": 7915, "ĠãĤ¹ãĤ³ãĤ¢ãģ«": 7916, "ĠãĤ¹ãĥĨãĥĥãĥĹ": 7917, "ĠãĤ¹ãĥĨãĥĥãĥĹãģĶãģ¨": 7918, "ĠãĤ¹ãĥĨãĥĥãĥĹãģĶãģ¨ãģ®åĭ¾éħįãĥİãĥ«ãĥł": 7919, "ĠãĤ¹ãĥĨãĥĥãĥĹçķªåı·": 7920, "ĠãĤĪãĤĬ": 7921, "ĠãĤĴ": 7922, "ĠãĤĴãĢģ": 7923, "ĠãĤĴä½¿ãģĨ": 7924, "ĠãĤĴåĬłç®Ĺ": 7925, "ĠãĤĴåıĤçħ§)ãĢĤ": 7926, "ĠãĤĴæĮģãģ¤": 7927, "ĠãĤĴæĮģãģ¤)ãĢĤ": 7928, "ĠãĤĴè¨Īç®ĹãģĻãĤĭãĢĤ": 7929, "ĠãĤĴè¶ħãģĪãģ¦ãģĦãĤĭ": 7930, "ĠãĤĴè¶ħãģĪãĤĭ": 7931, "ĠãĤĴè¿ĶãģĻ": 7932, "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥": 7933, "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥ãĤĴ": 7934, "ĠãĤŃãĥ£ãĥĥãĤ·ãĥ¥åħĪãĥĩãĤ£ãĥ¬ãĤ¯ãĥĪãĥªãĢĤåŃĺåľ¨ãģĹãģªãģĦåł´åĲĪãģ¯": 7935, "Ġãĥ": 7936, "Ġãĥ¢": 7937, "Ġãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ": 7938, "Ġãĥ¢ãĥĩãĥ«": 7939, "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮ": 7940, "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮæ¬¡åħĥ": 7941, "Ġãĥ¢ãĥĩãĥ«ãģ®éļłãĤĮæ¬¡åħĥãĢĤ": 7942, "Ġãĥ©": 7943, "Ġãĥ©ãĤ¦ãĥ³ãĥīä¿®æŃ£": 7944, "Ġãĥ©ãĤ¦ãĥ³ãĥīä¿®æŃ£ãĢģ": 7945, "ĠãĥªãĥĿãĤ¸ãĥĪãĥª": 7946, "Ġãĥģ": 7947, "Ġãĥģãĥ£ãĥ³ãĤ¯": 7948, "ĠãĥĨãĥ³ãĤ½ãĥ«": 7949, "Ġãĥĩ": 7950, "ĠãĥĪãĥ¼ãĤ¯ãĥ³": 7951, "ĠãĥĪãĥ¼ãĤ¯ãĥ³ãģĤãģŁãĤĬ": 7952, "ĠãĥĪãĥ¼ãĤ¯ãĥ³ãģĤãģŁãĤĬãģ®è¨Īç®Ĺéĩı": 7953, "ĠãĥĪãĥ¼ãĤ¯ãĥĬãĤ¤ãĤ¶": 7954, "Ġãĥİ": 7955, "ĠãĥĲ": 7956, "ĠãĥĲãĤ¤ãĥĪ": 7957, "ĠãĥĲãĤ¤ãĥĪæķ°": 7958, "ĠãĥĲãĤ¤ãĥĪéķ·": 7959, "ĠãĥĲãĤ¤ãĥĪéķ·ãģ®ä¸ĬéĻĲ": 7960, "ĠãĥĲãĤ±ãĥĥãĥĪ": 7961, "ĠãĥĲãĥĥãĥģ": 7962, "ĠãĥĲãĥĥãĥģãĤµãĤ¤ãĤº": 7963, "Ġãĥĳ": 7964, "Ġãĥĳãĥ©ãĥ¡": 7965, "Ġãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿": 7966, "Ġãĥĳãĥ©ãĥ¡ãĥĪãĥªãĥĥãĤ¯ãģĤãģ¦ãģ¯ãĤģ": 7967, "Ġãĥķ": 7968, "ĠãĥķãĤ": 7969, "ĠãĥĹ": 7970, "ĠãĥĹãĥŃãĥķãĤ¡ãĤ¤ãĥ«": 7971, "Ġãĥĺãĥĥãĥī": 7972, "ĠãĥĺãĥĥãĥīãģĤãģŁãĤĬãģ®": 7973, "Ġãĥĺãĥĥãĥīæķ°": 7974, "ĠãĥĻ": 7975, "ĠãĥĻãĥ¼ãĤ¹ãģ®åıįå¾©çļĦãģªç¸®å°ı": 7976, "Ġãĥŀ": 7977, "Ġä¸»å¼µ": 7978, "Ġä¸Ńéĸĵå±¤": 7979, "Ġä»¥": 7980, "Ġä»¥ä¸Ĭ": 7981, "Ġä»¥ä¸Ĭ)ãĢĤ": 7982, "Ġä»¥ä¸ĭ": 7983, "Ġä»¥ä¸ĭ)ãĢĤ": 7984, "Ġä»¥éĻį": 7985, "Ġä»»æĦıå½¢çĬ¶ãģ®åħ¥åĬĽãĥĨãĥ³ãĤ½ãĥ«ãĢĤ": 7986, "Ġä½įç½®ãĤ¤ãĥ³ãĥĩãĥĥãĤ¯ãĤ¹(m:": 7987, "Ġä¿Ĥæķ°": 7988, "Ġå": 7989, "Ġå¯¾": 7990, "Ġå¯¾å¿ľãģĻãĤĭæĲįå¤±": 7991, "Ġå±": 7992, "Ġå±¤": 7993, "Ġå±¤çĽ®": 7994, "Ġå±ŀæĢ§": 7995, "Ġå¼ķæķ°": 7996, "Ġå½¢å¼ı": 7997, "Ġå½¢çĬ¶": 7998, "ĠåĢ¤": 7999, "ĠåĢĭ": 8000, "ĠåĢĭãģ®": 8001, "Ġåģ´": 8002, "Ġåģ´)": 8003, "Ġåģ´ãĢģn": 8004, "Ġåģ´ãĢģn:": 8005, "Ġåģ´ãģ®": 8006, "Ġåģ´ãģ®ç³»åĪĹéķ·": 8007, "Ġåħ¥åĩºåĬĽ": 8008, "ĠåĨħ": 8009, "ĠåĪĨåī²": 8010, "ĠåĪĨåī²å¯¾è±¡ãģ®": 8011, "ĠåĪĨåī²çµĲæŀľ": 8012, "Ġåĭ¾éħį": 8013, "Ġåįĺ": 8014, "Ġåı": 8015, "ĠåıĤåĬł": 8016, "ĠåıĤçħ§": 8017, "ĠåıĤçħ§)ãĢĤ": 8018, "ĠåıĤçħ§ãĢģ006": 8019, "Ġåıĸå¾Ĺ": 8020, "ĠåĲĦ": 8021, "ĠåĽ": 8022, "ĠåĽ³": 8023, "ĠåĽ³ãģ®ãĤ¿ãĤ¤ãĥĪãĥ«ãĢĤ": 8024, "ĠåĽŀ": 8025, "ĠåŃ¦ç¿Ĵ": 8026, "ĠåŃ¦ç¿Ĵçİĩ": 8027, "Ġæ": 8028, "Ġæ¤ľè¨¼": 8029, "Ġæ§ĭæĪĲ": 8030, "Ġæ¨ª": 8031, "Ġæ¨ªè»¸ãĥ©ãĥĻãĥ«ãĢĤ": 8032, "Ġæ¬¡": 8033, "Ġæ¬¡ãĥ¢ãĥ¼ãĥ¡ãĥ³ãĥĪæİ¨å®ļ": 8034, "Ġæ¬¡åħĥ": 8035, "Ġæ¬¡åħĥãģ®": 8036, "Ġæ³ķ": 8037, "ĠæııçĶ»åħĪãģ®": 8038, "ĠæĲįå¤±": 8039, "ĠæĲįå¤±ãģ®": 8040, "Ġæķ°åĢ¤å®īå®ļåĮĸãģ®ãģŁãĤģãģ®å¾®å°ıå®ļæķ°": 8041, "Ġæķ´æķ°": 8042, "Ġæĸ¹å¼ı": 8043, "Ġæĸ¹åĲĳ": 8044, "ĠæĸĩåŃĹ": 8045, "ĠæĸĩåŃĹãĤĴ": 8046, "ĠæĻĤçĤ¹": 8047, "ĠæľªæºĢ": 8048, "ĠæľªæºĢãģ§": 8049, "ĠæľĢå¤§": 8050, "ĠæŃ£": 8051, "ĠæŃ£ãģ®å®Łæķ°åĢ¤ãģ®ç³»åĪĹ": 8052, "ĠæŃ£åīĩåĮĸ": 8053, "ĠæŃ£è¦ıåĮĸ": 8054, "ĠæŃ£è¦ıåĮĸãĤĴè¡ĮãģĨçī¹å¾´": 8055, "Ġç": 8056, "Ġç¬¦åı·åĮĸ": 8057, "Ġç¬¬": 8058, "Ġç¯Ģ": 8059, "Ġç¯Ģ)ãĢĤ": 8060, "Ġç³»åĪĹ": 8061, "Ġç³»åĪĹéķ·": 8062, "Ġç³»åĪĹéķ·(sequence": 8063, "ĠçµĤäºĨ": 8064, "Ġç·ļå½¢": 8065, "Ġç¸¦": 8066, "Ġç¸¦è»¸ãĤĴå¯¾æķ°ãĤ¹ãĤ±ãĥ¼ãĥ«ãģ«ãģĻãĤĭãģĭãĢĤ": 8067, "Ġç¸¦è»¸ãĥ©ãĥĻãĥ«ãĢĤ": 8068, "Ġç¾¤": 8069, "Ġç¾¤ãģĶãģ¨ãģ®": 8070, "Ġçİĩ": 8071, "ĠçİĩãĢĤ": 8072, "ĠçĶŁæĪĲ": 8073, "ĠçĶŁæĪĲãģĻãĤĭ": 8074, "Ġçķ": 8075, "ĠçķªçĽ®": 8076, "ĠçĽ": 8077, "ĠçĽ®æ¨Ļ": 8078, "ĠçĽ¸": 8079, "ĠçŃī": 8080, "Ġè¡Į": 8081, "Ġè¡ĮåĪĹ": 8082, "Ġè¨Ģèªŀãĥ¢ãĥĩãĥ«ãģ®": 8083, "Ġè¨ĵç·´": 8084, "Ġè¨ĺäºĭ": 8085, "Ġèªŀå½Ļ": 8086, "Ġèªŀå½ĻãĤµãĤ¤ãĤº": 8087, "Ġèªŀå½ĻãĤµãĤ¤ãĤº(vocabulary": 8088, "Ġè»¸": 8089, "Ġèĩª": 8090, "Ġèĩªå·±æ³¨æĦı": 8091, "Ġèĩªèº«ãģ®": 8092, "Ġèĩªèº«ãģ®åĪĨåī²çµĲæŀľ": 8093, "Ġéĩįãģ¿": 8094, "Ġéĩįãģ¿ãĥ»åĲĦãĤµãĥĸãĥ¬ãĤ¤ãĥ¤ãĥ¼åĩºåĬĽãģ«éģ©çĶ¨ãģĻãĤĭ": 8095, "Ġéĩįãģ¿æ¸Ľè¡°": 8096, "Ġéĩįãģ¿æ¸Ľè¡°ä¿Ĥæķ°": 8097, "ĠéĿŀåŁĭãĤģè¾¼ãģ¿ãĥĳãĥ©ãĥ¡ãĥ¼ãĤ¿æķ°": 8098, "ĠĠ": 8099, "ĠĠ#": 8100, "ĠĠ:": 8101, "ĠĠĠ:": 8102, "ĠĠĠĠ": 8103, "ĠĠĠĠ:": 8104, "ĠĠĠĠĠ:": 8105, "ĠĠĠĠĠĠĠ:": 8106, "ĠĠĠĠĠĠĠĠ": 8107, "ĠĠĠĠĠĠĠĠĠ:": 8108, "ĠĠĠĠĠĠĠĠĠĠĠ:": 8109, "ĠĠĠĠĠĠĠĠĠĠĠĠĠ:": 8110, "ġ": 8111, "Ģ": 8112, "Ģ¤": 8113, "Ģ§": 8114, "ĢæīĢ": 8115, "Ģĭ": 8116, "ĢĻ": 8117, "ĢĻè£ľ": 8118, "ģ": 8119, "ģ´": 8120, "Ĥ": 8121, "Ĥ¹": 8122, "ĤåĬł": 8123, "Ĥç": 8124, "Ĥç³»åĪĹ": 8125, "ĤçĤ¹": 8126, "ĥ": 8127, "ĥ³å®ļ": 8128, "ĥçķĮ": 8129, "ĥħ": 8130, "ĥħåł±": 8131, "Ħ": 8132, "Ħå½±": 8133, "Ħı": 8134, "ħ": 8135, "ħ§": 8136, "Ĩ": 8137, "Ĩħ": 8138, "ĩ": 8139, "ĩãĤ£ãĥ³ãĤ°": 8140, "Ī": 8141, "ī": 8142, "Ĭ": 8143, "ĭ": 8144, "ĭ¾éħį": 8145, "ĭãĤģ": 8146, "Į": 8147, "į": 8148, "įå¤±": 8149, "įĺ": 8150, "İ": 8151, "İåĽŀ": 8152, "İç¤º": 8153, "İĩ": 8154, "ı": 8155, "ı¾": 8156, "Ĳ": 8157, "ĲĨ": 8158, "Ĳįå¤±": 8159, "ĳ": 8160, "ĳ¼ãģ³åĩºãģĹ": 8161, "Ĵ": 8162, "ĴĮ": 8163, "ĵ": 8164, "Ķ": 8165, "ķ": 8166, "ķæķ°": 8167, "ĸ": 8168, "ĸĦ": 8169, "Ĺ": 8170, "ĺ": 8171, "ĺãĥĥãĥī": 8172, "ĺİç¤º": 8173, "Ļ": 8174, "Ļæºĸ": 8175, "ĻĤçĤ¹": 8176, "ļ": 8177, "ļĦ": 8178, "Ľ": 8179, "ľ": 8180, "ľãĥ«": 8181, "Ŀ": 8182, "ĿåŃĺ": 8183, "ŀ": 8184, "ŀæĢ§": 8185, "Ł": 8186, "ł": 8187, "ł´": 8188, "ł´åĲĪ": 8189, "Ń": 8190, "Ńī": 8191 } }