I am thankful to you for your kind reply to my earlier post. I have one more
doubt. Is it possible to do incremental training using Sphinxtrain. By saying
this what I mean is
I have already trained my system with 20 speakers data. Now I have got 20 more
speakers data. Is it possible to refine the parameters of earlier models with
new 20 speakers data only used for training or Do I need to train 40 speakers
data again. If it is possilble which parameters I have to specify in sphinx
train.
I think I am asking question in right way. But sorry, If I not.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If it is possilble which parameters I have to specify in sphinx train.
It is possible, but there are no parameters to do that. You need to write a
new script to extend existing model using
the SphinxTrain binaries (bw, norm, etc).
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Sir,
I am thankful to you for your kind reply to my earlier post. I have one more
doubt. Is it possible to do incremental training using Sphinxtrain. By saying
this what I mean is
I have already trained my system with 20 speakers data. Now I have got 20 more
speakers data. Is it possible to refine the parameters of earlier models with
new 20 speakers data only used for training or Do I need to train 40 speakers
data again. If it is possilble which parameters I have to specify in sphinx
train.
I think I am asking question in right way. But sorry, If I not.
It is possible, but there are no parameters to do that. You need to write a
new script to extend existing model using
the SphinxTrain binaries (bw, norm, etc).