excalibur12's picture
Upload tokenizer
52135b2 verified
raw
history blame
637 Bytes
{
"[PAD]": 53,
"[UNK]": 52,
"aa": 0,
"ae": 1,
"ah": 2,
"ao": 3,
"aw": 4,
"ax": 5,
"ay": 6,
"b": 7,
"bcl": 8,
"ch": 9,
"d": 10,
"dcl": 11,
"dh": 12,
"dx": 13,
"eh": 14,
"el": 15,
"epi": 16,
"er": 17,
"ey": 18,
"f": 19,
"g": 20,
"gcl": 21,
"h#": 22,
"hh": 23,
"ih": 24,
"iy": 25,
"jh": 26,
"k": 27,
"kcl": 28,
"l": 29,
"m": 30,
"n": 31,
"ng": 32,
"ow": 33,
"oy": 34,
"p": 35,
"pau": 36,
"pcl": 37,
"q": 38,
"r": 39,
"s": 40,
"sh": 41,
"t": 42,
"tcl": 43,
"th": 44,
"uh": 45,
"uw": 46,
"v": 47,
"w": 48,
"y": 49,
"z": 50,
"|": 51
}