A time-saving script for any Llama CPP/LoRA workflow: it handles merging the LoRA into the base model, converting it to GGUF format, and applying post-training quantization.

• Clone this repository: git clone https://github.com/georgepullen/Merge-LoRA-Into-GGUF.git .
• Clone the llama.cpp repository: git clone https://github.com/ggerganov/llama.cpp.git .
• Install requirements for both repositories: pip install -r requirements.txt, pip install -r merge_lora_requirements.txt

• -l Path to LoRA (HF Repo ID or Local Path)
• -b Path to base model (HF Repo ID or Local Path)
• -f Format (F32, F16, q6_k, q4_0, etc.)