Hi,
I am trying to decode proper nouns like names.
I have a audio with sentence as ‘Get me a bottle of vine Albert’.
Everything decodes except the name ‘Albert’
I need help to achieve decoding of such names.
Say, if i have a list of such names, that I use mostly.
1 . Should the acoustic model be retrained for such things.
2. The adaptation of the language model for names is sufficient.
If language model is sufficient, how should I add such list to the language model training.
Through interpolation and more weightage(lambda) to the names list?
Thanks,
Sandeep