KBaba7/llama.cpp

1# export-lora2 3Apply LORA adapters to base model and export the resulting model.4 5```6usage: llama-export-lora [options]7 8options:9  -m,    --model                  model path from which to load base model (default '')10         --lora FNAME             path to LoRA adapter  (can be repeated to use multiple adapters)11         --lora-scaled FNAME S    path to LoRA adapter with user defined scaling S  (can be repeated to use multiple adapters)12  -t,    --threads N              number of threads to use during computation (default: 4)13  -o,    --output FNAME           output file (default: 'ggml-lora-merged-f16.gguf')14```15 16For example:17 18```bash19./bin/llama-export-lora \20    -m open-llama-3b-v2.gguf \21    -o open-llama-3b-v2-english2tokipona-chat.gguf \22    --lora lora-open-llama-3b-v2-english2tokipona-chat-LATEST.gguf23```24 25Multiple LORA adapters can be applied by passing multiple `--lora FNAME` or `--lora-scaled FNAME S` command line parameters:26 27```bash28./bin/llama-export-lora \29    -m your_base_model.gguf \30    -o your_merged_model.gguf \31    --lora-scaled lora_task_A.gguf 0.5 \32    --lora-scaled lora_task_B.gguf 0.533```34