Yet another reminder why you should not use Ollama

[dcreater]

Georgi's relevant comment: https://github.com/ggml-org/llama.cpp/pull/19324#issuecommen...

[kermatt]

Can someone add some context as to what that diff is showing?

[dcreater]

and use the original llama.cpp directly. Its infinitely more easy to setup and use now

[tmtvl]

Setting up ollama is 2 steps:

1. yay -S ollama

2. sytemctl enable --now ollama

How is llama.cpp infinitely more easy to set up?