>>106487068
You are stilll stuck using shitter quantization models, no one and absolutely no one would ever buy an AMD gpu for any LLM related task they are that bad