The model files required to use Tacspeak are included in, and can be downloaded from, Tacspeak releases jwebmeister/tacspeak.
Just download the .zip of the model, extract the kaldi_model folder (and files) into the folder containing Tacspeak.exe (or cli.py).
Alternatively, models can be downloaded from my fork of Kaldi Active Grammar releases jwebmeister/kaldi-active-grammar, or the original Kaldi Active Grammar releases daanzu/kaldi-active-grammar. These alternative models are not guaranteed to work correctly (or as accurately) with Tacspeak.
Details of the originating projects source below:
- Name: kaldi-active-grammar
- Version: 3.1.0
- Summary: Kaldi speech recognition with grammars that can be set active/inactive dynamically at decode-time
- License: AGPL-3.0
- Home-page: https://github.com/daanzu/kaldi-active-grammar