Open andy8025 opened 2 years ago
Hello! You can follow the instructions at https://github.com/stanford-oval/genienlp and https://github.com/stanford-oval/genie-cloud to deploy the Semantic Parser on your own HW. The semantic parser can be downloaded from https://almond-static.stanford.edu/models/en-US/latest.tar.xz
As for STT/TTS, you will need to make changes in genie-cloud to replace the API calls to Microsoft with a different implementation that runs locally.
During a recent video demo of Genie https://youtu.be/6ZMXE5PXPqU?t=9694 the two team members stated that the stand-alone version of Genie sends TTS/STT data to Microsoft, and the Semantic Parser runs on remote (Stanford?) servers. Can someone provide instructions on how to build Genie on a workstation (with a GPU) so that it does not send data to any remote computers?