they should make the zesty tiktok voice a public model so I can run it locally and have him say all the words for me

they should make a utau voice bank of him so he can sing

@sierra have qwen3 tts replicate it

@sierra there might be a better voice cloning tts model i haven’t looked at them in a bit

@karebu no earthly idea how to do any of that but like maybe. can claude vibecode an interface for me or something

@sierra this project makes it somewhat easy https://github.com/ServeurpersoCom/qwentts.cpp basically you point it to voices folder where you have reference audio in a .wav file and a .txt under the same name containing a transcription of whatever is being said and it serves an openai tts api which you can use with anything that supports that standard

@sierra if you’re comfortable using docker/podman its dead easy

docker run --rm --device /dev/dri --network host \
                                               -v ./models:/models:ro \
                                               -v ./voices:/voices:ro \
                                               -e MODEL_PATH=/models/qwen-talker-1.7b-base-Q8_0.gguf \
                                               -e CODEC_PATH=/models/qwen-tokenizer-12hz-Q8_0.gguf \
                                               ghcr.io/serveurpersocom/qwentts.cpp:vulkan

model files here https://huggingface.co/Serveurperso/Qwen3-TTS-GGUF/tree/main

@karebu @sierra will i ever see u post about something that isnt qwen

@fiore @sierra i post other things…..

@karebu @sierra u fell off

@fiore @sierra well… uhm… you fell off more

@karebu @sierra womp womp

@fiore @sierra 🖕🖕🖕🖕

@karebu @sierra wanna make out now sorry i didnt mean to say that

@fiore @sierra thats haram….

@karebu @sierra sorry