wav2vec2-indic-malvi / vocab.json
Anujgr8's picture
Upload tokenizer
eb384f3 verified
{
"!": 1,
"\"": 2,
",": 3,
"-": 4,
".": 5,
"0": 6,
"1": 7,
"2": 8,
"5": 9,
"?": 10,
"D": 11,
"[PAD]": 90,
"[UNK]": 89,
"a": 12,
"b": 13,
"c": 14,
"d": 15,
"e": 16,
"f": 17,
"h": 18,
"i": 19,
"j": 20,
"k": 21,
"l": 22,
"m": 23,
"n": 24,
"o": 25,
"p": 26,
"r": 27,
"s": 28,
"t": 29,
"v": 30,
"z": 31,
"|": 0,
"ँ": 32,
"ं": 33,
"ः": 34,
"अ": 35,
"आ": 36,
"इ": 37,
"ई": 38,
"उ": 39,
"ऊ": 40,
"ए": 41,
"ऐ": 42,
"ओ": 43,
"औ": 44,
"क": 45,
"ख": 46,
"ग": 47,
"घ": 48,
"च": 49,
"छ": 50,
"ज": 51,
"झ": 52,
"ट": 53,
"ठ": 54,
"ड": 55,
"ढ": 56,
"ण": 57,
"त": 58,
"थ": 59,
"द": 60,
"ध": 61,
"न": 62,
"प": 63,
"फ": 64,
"ब": 65,
"भ": 66,
"म": 67,
"य": 68,
"र": 69,
"ल": 70,
"व": 71,
"श": 72,
"ष": 73,
"स": 74,
"ह": 75,
"़": 76,
"ा": 77,
"ि": 78,
"ी": 79,
"ु": 80,
"ू": 81,
"े": 82,
"ै": 83,
"ॉ": 84,
"ो": 85,
"ौ": 86,
"्": 87,
"।": 88
}