Upload tokenizer
3535146
|
{ |
|
"tha": { |
|
"!": 1, |
|
"\"": 2, |
|
"'": 3, |
|
",": 4, |
|
"-": 5, |
|
".": 6, |
|
":": 7, |
|
";": 8, |
|
"?": 9, |
|
"A": 10, |
|
"B": 11, |
|
"C": 12, |
|
"F": 13, |
|
"J": 14, |
|
"K": 15, |
|
"L": 16, |
|
"M": 17, |
|
"X": 18, |
|
"[PAD]": 117, |
|
"[UNK]": 116, |
|
"_": 19, |
|
"a": 20, |
|
"b": 21, |
|
"c": 22, |
|
"e": 23, |
|
"f": 24, |
|
"g": 25, |
|
"h": 26, |
|
"i": 27, |
|
"k": 28, |
|
"l": 29, |
|
"m": 30, |
|
"n": 31, |
|
"o": 32, |
|
"p": 33, |
|
"r": 34, |
|
"s": 35, |
|
"t": 36, |
|
"u": 37, |
|
"w": 38, |
|
"x": 39, |
|
"y": 40, |
|
"|": 0, |
|
"~": 42, |
|
"ก": 43, |
|
"ข": 44, |
|
"ฃ": 45, |
|
"ค": 46, |
|
"ฆ": 47, |
|
"ง": 48, |
|
"จ": 49, |
|
"ฉ": 50, |
|
"ช": 51, |
|
"ซ": 52, |
|
"ฌ": 53, |
|
"ญ": 54, |
|
"ฎ": 55, |
|
"ฏ": 56, |
|
"ฐ": 57, |
|
"ฑ": 58, |
|
"ฒ": 59, |
|
"ณ": 60, |
|
"ด": 61, |
|
"ต": 62, |
|
"ถ": 63, |
|
"ท": 64, |
|
"ธ": 65, |
|
"น": 66, |
|
"บ": 67, |
|
"ป": 68, |
|
"ผ": 69, |
|
"ฝ": 70, |
|
"พ": 71, |
|
"ฟ": 72, |
|
"ภ": 73, |
|
"ม": 74, |
|
"ย": 75, |
|
"ร": 76, |
|
"ฤ": 77, |
|
"ล": 78, |
|
"ว": 79, |
|
"ศ": 80, |
|
"ษ": 81, |
|
"ส": 82, |
|
"ห": 83, |
|
"ฬ": 84, |
|
"อ": 85, |
|
"ฮ": 86, |
|
"ฯ": 87, |
|
"ะ": 88, |
|
"ั": 89, |
|
"า": 90, |
|
"ำ": 91, |
|
"ิ": 92, |
|
"ี": 93, |
|
"ึ": 94, |
|
"ื": 95, |
|
"ุ": 96, |
|
"ู": 97, |
|
"เ": 98, |
|
"แ": 99, |
|
"โ": 100, |
|
"ใ": 101, |
|
"ไ": 102, |
|
"ๅ": 103, |
|
"ๆ": 104, |
|
"็": 105, |
|
"่": 106, |
|
"้": 107, |
|
"๊": 108, |
|
"๋": 109, |
|
"์": 110, |
|
"ํ": 111, |
|
"—": 112, |
|
"‘": 113, |
|
"’": 114, |
|
"“": 115, |
|
"”": 116 |
|
} |
|
} |
|
|
