where to get the dictionary of vosk-model-en-us-0.22-lgraph about vosk-api HOT 8 OPEN

It is inside the compilation package

https://alphacephei.com/vosk/models/vosk-model-en-us-0.22-compile.zip

from vosk-api.

ok got it, thanks for your quick and accurate reply!

from vosk-api.

Hi, sorry to bother you, I find some phone which is hard to read, for example
electromagnetic electromagnetic @_B l_I E_I k_I t_I r_I oU_I m_I {_I g_I n_I E_I 4_I I_I k_E
electromagnetic electromagnetic @_B l_I E_I k_I t_I r_I oU_I m_I {_I g_I n_I E_I t_I I_I k_E
electromagnetic electromagnetic I_B l_I E_I k_I 4_I r_I oU_I m_I {_I g_I n_I E_I 4_I I_I k_E
electromagnetic electromagnetic I_B l_I E_I k_I t_I r_I oU_I m_I {_I g_I n_I E_I t_I I_I k_E

what's 4_I ?
And the word "electromagnetic" have 4 kinds of phone arrangements, I want to calculate GOP(goodness of pronunciation), how to decide the only one phone arrangement?

from vosk-api.

what's 4_I ?

4 is a sampa phone something like "ch".

_I is for word internal, you are probably looking inside intermediate lexicon instead of original one.

I want to calculate GOP(goodness of pronunciation), how to decide the only one phone arrangement?

run alignment

from vosk-api.

thanks for your reply, it help me a lot

from vosk-api.

Hi dear author, why the vosk use different phone system with Kaldi?
for example, "@", "{", "4" in vosk model, but doesn't in Kaldi model,
Any idea to map from one to the other?

from vosk-api.

That particular model uses different phoneset unfortunately. You can still map it easily, it is a simple mapping. Other models like gigaspeech use standard cmu dictionary

from vosk-api.

Thanks for you quick and accurate reply, I know gigaspeech model right know and it is good enough to me, I don't plan to map phoneset! thanks!

from vosk-api.