A pull request has been merged into llama.cpp to provide support for Granite models. This adds the IBM model family to the ggml-based inference runtime.
HOW THIS AFFECTS YOU
●
builderYou can now run Granite models locally using the llama.cpp inference engine.