1、lora微调模型、融合导出模型
css
git clone https://github.com/hiyouga/LLaMA-Factory.git
cd LLaMA-Factory
pip install -e .
llamafactory-cli webui
将融合模型转移到另一台主机
2、安装llamacpp
css
git clone https://github.com/ggerganov/llama.cpp.git
source activate
conda create -n llamacpp
conda activate llamacpp
pip install -r llama.cpp/requirements.txt
3、使用llamacpp将模型转换成gguf
css
python llama.cpp/convert_hf_to_gguf.py /root/autodl-tmp/llmout --outtype f16 --verbose --outfile Qwen2.5-0.5B-Instruct-gguf.gguf
4、安装open-webui
todo~~