Suppose I want to adapt sphinx for Indian English speakers. Is it possible? Would changes in the HMM and LM directories be sufficient?... I mean reworking the HMM coeffs.
I have a large amount of audio data hand.
Please guide me in this matter. And if source requires changes kindly tell.
Thanks.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
We will be releasing the Trainer, along with a series of scripts, to do acoustic model adaptation. From there, we would also need a lexicon and a language model; with those, it
would be quite doable. One could start from American English models and adapt them, or start from flat distributions if we have enough training data for the target language/dialect.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Anonymous
-
2000-11-24
do you know when the trainer is likely to be released?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Suppose I want to adapt sphinx for Indian English speakers. Is it possible? Would changes in the HMM and LM directories be sufficient?... I mean reworking the HMM coeffs.
I have a large amount of audio data hand.
Please guide me in this matter. And if source requires changes kindly tell.
Thanks.
We will be releasing the Trainer, along with a series of scripts, to do acoustic model adaptation. From there, we would also need a lexicon and a language model; with those, it
would be quite doable. One could start from American English models and adapt them, or start from flat distributions if we have enough training data for the target language/dialect.
do you know when the trainer is likely to be released?
By the end of the year. We've been having good progress on moving it out of our local environment and successfully completing full training jobs.