libsio

A runtime library for Speech Input (stt) & Output (tts)

Speech To Text

  • unified CTC and WFST decoding via beam search
  • online(streaming) decoding
  • lattice-free on the fly rescoring with arbitrary language models, e.g.:

     CTC + external LM
     WFST - (lookahead LM) + (big LM)
     E2E - (estimated internal LM) + (external LM)
     flexible contextual biasing
    
  • streaming + lattice free -> low latency
  • modular design with potentials to deploy models from various speech toolkits such as Kaldi, K2, Espnet, SpeechBrain, WeNet.

Text To Speech

Long term plan, won’t happen soon.

Non Goal

  • GPU support

Zen

  • Keep it simple. All softwares are fated to fight against complexity.
  • Keep it small. Codes are debts, less code means less interests to pay.
  • Keep it explicit. Hidden behaviors behind non-obvious codes are evil.
  • Keep it local. Things should be organized locally for easy understanding.
  • Keep it concrete. Unnecessary abstractions bring cognitive costs, abstract only when it is absolutely necessary.

License

TBD

GitHub

View Github