llama.cpp

Fix convert-lora-to-ggml.py

#2738

Merged