2 ms·

Thanks for sharing. Pi5 + cheap AMD GPU = convenient modest LLM api server? ...if you find the right magic rocm incantations I guess Double thanks for the 3rd

Thanks for sharing. Pi5 + cheap AMD GPU = convenient modest LLM api server? ...if you find the right magic rocm incantations I guess Double thanks for the 3rd party mac mini SSD tip - eagerly awaiting delivery!

geerlingguy 2y ago
llama.cpp runs great with Vulkan, so no ROCm magic required!