text-independent speaker identification

the source dir contain two model , gmm-256 and gmm-512。these tow model trained with 1000 pereons. only work for silence environment。