We would like to train your french model for usage within an IVR. Therefore we have to resample the training audio data to 8kHz and re-train the model. Can we have access to the model data (audios, prompts, dict, lm) required to train it?
Many thanks,
Ronny Reiprich
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It is not going to be very well suited for IVR though, desktop speech is quite different from telephone speech. I recommend you to look for real telephone database, there are some commercial offerings. Or you can collect such database yourself.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I already tried the Voxforge model, but it gives 20.000 Warnings of words that aren't in the dictionary. We're talking about 634 unique words, some are simply missing, most of them are combined words like c'est, j'ai, l'objet etc.
Do you know of any specific commercial offering?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
We would like to train your french model for usage within an IVR. Therefore we have to resample the training audio data to 8kHz and re-train the model. Can we have access to the model data (audios, prompts, dict, lm) required to train it?
Many thanks,
Ronny Reiprich
You can download French data on Voxforge
http://www.voxforge.org/fr/Downloads
It is not going to be very well suited for IVR though, desktop speech is quite different from telephone speech. I recommend you to look for real telephone database, there are some commercial offerings. Or you can collect such database yourself.
I already tried the Voxforge model, but it gives 20.000 Warnings of words that aren't in the dictionary. We're talking about 634 unique words, some are simply missing, most of them are combined words like c'est, j'ai, l'objet etc.
Do you know of any specific commercial offering?