llma2微调的3种方式

  • 本地:https://github.com/oobabooga/text-generation-webui
    • Download基座模型(huggingface路径)
    • 准备数据集,放到datasets中,注意格式format
  • colab:https://colab.research.google.com/drive/12dVqXZMIVxGI0uutU6HG9RWbWPXL3vts?usp=sharing (模型和数据集都在huggingface)
  • lamini,封装的库,调用比较简单

Leave a Reply

Your email address will not be published. Required fields are marked *