报错显示为：
Traceback (most recent call last): 

  File "/dgsoc/llf/LLaMA-Factory/src/llamafactory/launcher.py", line 23, in <module> 

    launch() 

  File "/dgsoc/llf/LLaMA-Factory/src/llamafactory/launcher.py", line 19, in launch 

    run_exp() 

  File "/dgsoc/llf/LLaMA-Factory/src/llamafactory/train/tuner.py", line 50, in run_exp 

    run_sft(model_args, data_args, training_args, finetuning_args, generating_args, callbacks) 

  File "/dgsoc/llf/LLaMA-Factory/src/llamafactory/train/sft/workflow.py", line 48, in run_sft 

    model = load_model(tokenizer, model_args, finetuning_args, training_args.do_train) 

  File "/dgsoc/llf/LLaMA-Factory/src/llamafactory/model/loader.py", line 160, in load_model 

    model = load_class.from_pretrained(**init_kwargs) 

  File "/workspace/installs/conda/envs/llm-dev/lib/python3.9/site-packages/transformers/models/auto/auto_factory.py", line 563, in from_pretrained 

    return model_class.from_pretrained( 

  File "/workspace/installs/conda/envs/llm-dev/lib/python3.9/site-packages/transformers/modeling_utils.py", line 3202, in from_pretrained 

    hf_quantizer.validate_environment( 

  File "/workspace/installs/conda/envs/llm-dev/lib/python3.9/site-packages/transformers/quantizers/quantizer_bnb_8bit.py", line 62, in validate_environment 

    raise RuntimeError("No GPU found. A GPU is needed for quantization.") 

如果不适用量化功能则是可以正常微调的。