llama.cpp

cb871fa0 - gguf : do not support passing existing ggml_context to gguf_init