Project-FreeSpeech

Documentation on creating custom programs for Tape Measure's (TM) Core IO. This project aims to get three AI models running locally on TM's hardware. One, a live speech-to-text model based on OpenAI's whisper. Two, a large language model (multiple currently being considered). Three, a text-to-speech model based on Tortoise TTS, giving TM a more realistic and personal touch.

Basics of Docker & Example Program

Utilizing Docker this chapter shows how I got custom WAV files to be played on a mounted speaker.

Accessing GPU In Docker Containers

Speech-To-Text Utilizing Whisper

Large Language Models

Text-To-Speech Utilizing Tortoise TTS

Common Files & Cross Model Communication