llama.cpp

d09770ca - ggml-alloc : fix list of allocated tensors with GGML_ALLOCATOR_DEBUG (#9573)