Hugging Face Space · source ready

Gemma 4 12B
Heretic GGUF

A text chat demo built for the exact i1-Q4_K_M GGUF, with llama.cpp serving the model and Gradio providing the chat interface.

Compute access required

The chat is staged, not running

This account cannot currently create a compute Space. The Dockerfile and app code are in this Space's repository. When compute access is available, switching its SDK from Static to Docker will build the live chat.

View exact model View demo code

Run this quant with llama.cpp

On a computer with sufficient RAM and a recent llama.cpp release:

llama serve -hf mradermacher/gemma-4-12b-heretic-abliterated-i1-GGUF:Q4_K_M