cdactvm's picture
Upload tokenizer
a838d04 verified
{
"/": 1,
"0": 2,
"1": 3,
"2": 4,
"3": 5,
"6": 6,
"8": 7,
"9": 8,
"I": 9,
"U": 10,
"[PAD]": 83,
"[UNK]": 82,
"_": 11,
"|": 0,
"ಂ": 12,
"ಃ": 13,
"ಅ": 14,
"ಆ": 15,
"ಇ": 16,
"ಈ": 17,
"ಉ": 18,
"ಊ": 19,
"ಋ": 20,
"ಎ": 21,
"ಏ": 22,
"ಐ": 23,
"ಒ": 24,
"ಓ": 25,
"ಔ": 26,
"ಕ": 27,
"ಖ": 28,
"ಗ": 29,
"ಘ": 30,
"ಙ": 31,
"ಚ": 32,
"ಛ": 33,
"ಜ": 34,
"ಝ": 35,
"ಞ": 36,
"ಟ": 37,
"ಠ": 38,
"ಡ": 39,
"ಢ": 40,
"ಣ": 41,
"ತ": 42,
"ಥ": 43,
"ದ": 44,
"ಧ": 45,
"ನ": 46,
"ಪ": 47,
"ಫ": 48,
"ಬ": 49,
"ಭ": 50,
"ಮ": 51,
"ಯ": 52,
"ರ": 53,
"ಱ": 54,
"ಲ": 55,
"ಳ": 56,
"ವ": 57,
"ಶ": 58,
"ಷ": 59,
"ಸ": 60,
"ಹ": 61,
"ಾ": 62,
"ಿ": 63,
"ೀ": 64,
"ು": 65,
"ೂ": 66,
"ೃ": 67,
"ೆ": 68,
"ೇ": 69,
"ೈ": 70,
"ೊ": 71,
"ೋ": 72,
"ೌ": 73,
"್": 74,
"ೕ": 75,
"ೖ": 76,
"೦": 77,
"೨": 78,
"೪": 79,
"​": 80,
"‍": 81
}