Post

Interesting. Ollama is moving away from primarily using the llama.cpp project as a turnkey solution for model support, to building its own engine that leverages the underlying GGML library that llama.cpp uses. ollama.com/blog/multimo...