pheme / ckpt /unique_text_tokens.k2symbols
taras-sereda's picture
speech tokenizer, requirements
3980644
raw
history blame
563 Bytes
<eps> 0
! 1
" 2
( 3
) 4
, 5
. 6
: 7
; 8
? 9
_ 10
aɪ 11
aɪə 12
aɪɚ 13
aɪʊ 14
aɪʊɹ 15
aʊ 16
b 17
d 18
19
e 20
enus 21
es 22
23
f 24
fr 25
h 26
i 27
iə 28
iː 29
j 30
k 31
l 32
m 33
n 34
35
36
37
oːɹ 38
p 39
r 40
s 41
t 42
43
44
v 45
w 46
x 47
z 48
æ 49
ç 50
ð 51
ø 52
ŋ 53
ɐ 54
ɑ 55
ɑː 56
ɑːɹ 57
ɔ 58
ɔɪ 59
ɔː 60
ɔːɹ 61
ə 62
əl 63
ɚ 64
ɛ 65
ɛɹ 66
ɛː 67
ɜː 68
ɡ 69
ɡʲ 70
ɣ 71
ɪ 72
ɪɹ 73
ɫ 74
ɬ 75
ɲ 76
ɹ 77
ɾ 78
ʃ 79
ʊ 80
ʊɹ 81
ʌ 82
ʒ 83
ʔ 84
̃ 85
̩ 86
θ 87
88
89