w2v-bert-2.0-kbd-v2 / vocab.json
anzorq's picture
Upload tokenizer
c9fadb1 verified
raw
history blame
560 Bytes
{
"[PAD]": 46,
"[UNK]": 45,
"h": 35,
"j": 8,
"q": 18,
"|": 0,
"ɕ": 42,
"ɣ": 6,
"ɬ": 20,
"ʐ": 12,
"а": 1,
"б": 3,
"в": 4,
"г": 5,
"д": 7,
"е": 10,
"ж": 11,
"з": 13,
"и": 14,
"й": 15,
"к": 16,
"л": 19,
"м": 22,
"н": 23,
"о": 24,
"п": 25,
"р": 27,
"с": 28,
"т": 29,
"у": 31,
"ф": 32,
"х": 34,
"ц": 37,
"ч": 39,
"ш": 40,
"щ": 41,
"ы": 43,
"э": 2,
"қ": 17,
"ҭ": 30,
"ҳ": 36,
"ҵ": 38,
"ӏ": 44,
"ӡ": 9,
"ԥ": 26,
"ԯ": 21,
"ჶ": 33
}