llama.cpp

models : replace ggml_tensor** out-params with output structs

#31142

Merged