Hallucinator

[NVIDIA ONLY] Upload an audio clip and let the AI autocomplete the rest. Powered by hertz from Standard Intelligence.

See this thread from Standard Intelligence to learn what this model is capable of: https://x.com/si_pbc/status/1853184307063660723

https://github.com/user-attachments/assets/13fe1e65-eafc-44bf-8d49-ed6ad7a0aaef

Credits

This project is a slight modification of the official hertz-dev project. Changes made:

1. app.py: added a gradio web ui for audio autocompletion
2. transformer.py: Use SDPBackend.EFFICIENT_ATTENTION instead of SDPBackend.FLASH_ATTENTION because installing flash attention takes too long and we might be dead by the time it finishes installing.

Install

1. One Click Install

You can install it with one click on https://pinokio.computer

2. Manual Install

Install pytorch

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

instal dependencies

pip install -r requirements.txt

Run

python app.py