llama.cpp

model-conversion : add token ids to prompt token output [no ci]

#17863

Merged