GPU를 이용한 대용량 음성인식기의 병렬 구현

김정석, 성원용 · 대한전자공학회 학술대회 · 2010

We have implemented a parallel large vocabulary continuous speech recognition (LVCSR) engine with a Graphics Processing Unit (GPU). A dynamic workload distribution method is employed in the emission probability computation for good load balancing. In the Viterbi search, we adopted Weighted Finite State Transducers (WFSTs) as a recognition network. We achieved the speed-up of x10.99 compared with a highly optimized sequential implementation. The recognition engine runs about 10 times faster than the real-time requirement without sacrificing accuracy.

Cited by 1 works.

Authors