llama.cpp

Add tokenizer.ggml.pre to gguf-new-metadata.py

#7627

Merged