gemma4-heretical

Gemma4-heretical provides a streamlined setup for deploying an uncensored Gemma 4 model using Ollama, with enhanced capabilities through community-built abliteration techniques that significantly reduce model response refusals while maintaining low quality loss. Key features include easy integration of the model with Ollama, support for various quantization configurations for optimal resource management, and the option for native Apple Silicon inference. Additionally, it offers a DIY abliteration script for users looking to customize the model further.