By Lucas HAL-9000 Chatbot is a small one-day project connecting a local LLM with a voice synthesized version of the wonderful voice of Douglas Rain, sampled from William Campbell's wonderful model trained here https://huggingface.co/campwill/HAL-9000-Piper-TTS/tree/main
- simpleaudio
- wave
- gpt4all
- piper_tts (And a laptop with all the specs)
Run main.py. If you (likely) don't have the model installed, it will take 4 GBs to install a small LLM model. Then run and hope it won't break!
Occassionally the voice/wav conversion fails. I'm not sure how to optimize or make sure it never does, so it will return a simple warning.
Segmentation faults occur a lot... this is more of a proof of concept project. Feel free to grab and mantain it