3 ms·use llama.cpp with cudaby slim 3mo agouse llama.cpp with cudaTheCycoONE 3mo agoThe problem may be that it's a 7800XT which handles memory contention by freezing.