Tim Leiber
Upload tokenizer
4e1be82 verified
raw
history blame
535 Bytes
{
"'": 43,
"[PAD]": 46,
"[UNK]": 45,
"a": 0,
"b": 1,
"c": 2,
"d": 3,
"e": 4,
"f": 5,
"g": 6,
"h": 7,
"i": 8,
"j": 9,
"k": 10,
"l": 11,
"m": 12,
"n": 13,
"o": 14,
"p": 15,
"q": 16,
"r": 17,
"s": 18,
"t": 19,
"u": 20,
"v": 21,
"w": 22,
"x": 23,
"y": 24,
"z": 25,
"|": 44,
"à": 26,
"â": 27,
"ä": 28,
"æ": 41,
"ç": 29,
"è": 31,
"é": 30,
"ê": 32,
"ë": 33,
"î": 34,
"ï": 35,
"ô": 36,
"ù": 38,
"û": 39,
"ü": 40,
"ÿ": 42,
"œ": 37
}