运行命令⬇️

python main_sft.py  --model_name_or_path "../transformers-code/03-PEFT/21-lora/bloom-1b4-zh"
  --dataset_name "medalpaca/medical_meadow_medical_flashcards"
  --local_data_dir "alpaca-gpt4"
  --dataset_sample 20000  --fed_alg "fedavg"
  --num_clients 20
  --sample_clients 2
  --max_steps 10
  --num_rounds 200
  --batch_size 16
  --gradient_accumulation_steps 1
  --seq_length 512
  --peft_lora_r 32
  --peft_lora_alpha 64
  --use_peft
  --load_in_8bit
  --output_dir "./output"
  --template "alpaca"