Apr 9, 2018 · We describe initial work on an extension of the Kaldi toolkit that supports weighted finite-state transducer (WFST) decoding on Graphics Processing Units (GPUs ...

Abstract. We describe initial work on an extension of the Kaldi toolkit that supports weighted finite-state transducer (WFST) decoding.

We describe initial work on an extension of the Kaldi toolkit that supports weighted finite-state transducer (WFST) decoding on Graphics Processing Units (GPUs) ...

We present an optimized weighted finite-state transducer (WFST) decoder capable of online streaming and offline batch processing of audio using Graphics ...

GPU-Accelerated Viterbi Exact Lattice Decoder for Batched Online and Offline Speech Recognition · A GPU-based WFST Decoder with Exact Lattice Generation.

"A GPU-based WFST Decoder with Exact Lattice Generation", Zhehuai Chen ... "Generating exact lattices in the WFST framework", D. Povey, M. Hannemann et ...

Feb 13, 2019 · Our paper also expands an efficient parallel Viterbi beam decoding algorithm to decrease the speech recognition Real Time Factor (RTF) value.

The proposed algorithm yields up to a 240x speedup over single core CPU decoding, and up to 40x faster decoding than the current state-of-the-art GPU ...

The GPU decoder significantly speeds up both token passing and lattice processing, which take up most of the decoding time in CPU. Meanwhile, the "other" part ...

3 Introduction It is a general-purpose offline decoder, which does not impose special requirements on the form of AM or LM, and works for all GPU ...