102 resultados para instance-dependent
Modeling pronunciation variation using context-dependent weighting and B/S refined acoustic modeling