2 ms·
Do you recommend Ollama or bare llama.cpp?
by drnick1 4mo ago
Do you recommend Ollama or bare llama.cpp?
- shironnnn_ 4mo agoif on MacOS I recommend llm-mlx which currently renders tokens 10%-15% faster than llama.cpp.
- jboss10 4mo agollama.cpp It's faster and more open source. Ollama has some mixed history. I use llama-swap to emulate the Ollama experience.