* update llama.cpp submodule to `5921b8f089d3b7bda86aac5a66825df6a6c10603` * add patch
Attach a file by drag & drop or click to upload
