llama.cpp

eb57fee5 - gguf-py : Add tokenizer.ggml.pre to gguf-new-metadata.py (#7627)