neigsendoig on Nostr: llama.cpp, customized build, CUDA with your GPU. Model is HauhauCS Qwen 3.5 4B ...

llama.cpp, customized build, CUDA with your GPU.

Model is HauhauCS Qwen 3.5 4B Uncensored Aggressive on HuggingFace for basic chat capabilities. You need to upgrade your compute hardware in order to run good models.