3 ms·Way worse performance than llama.cpp based LLM deploymentsby villgax 3y agoWay worse performance than llama.cpp based LLM deployments