3 ms·
In case you’ve been living in a cave, Nvidia is the defacto standard for LLM compute.
by nodesocket 1y ago
In case you’ve been living in a cave, Nvidia is the defacto standard for LLM compute.
- jakogut 1y agoLlama.cpp supports Vulkan, which is supported by all GPU vendors that care about standards and interoperability. The default should be open and portable APIs, not needlessly furthering a hegemony that is detrimental to us all.