完整跑通Learning to Compress Prompts with Gist Tokens代码问题记录

论文链接：https://arxiv.org/abs/2304.08467
1、首先下载llama-7b模型，flant5没下载
下载地址：https://huggingface.co/jayelm/llama-7b-gist-1/tree/main
2、按照说明安装环境
pip install -r requirements.txt
3、问题记录
**

问题一：

**
pydantic.errors.PydanticUserError: If you use @root_validator with pre=False (the default) you MUST specify skip_on_failure=True. Note that @root_validator is deprecated and should be replaced with @model_validator.

原因：pydantic版本不符，需要降级
解决方法：
（1）查看自己的pydantic，我的版本为2.4.2

pip show pydantic
1

（2）卸载当前pydantic

pip uninstall pydantic
1

（3）安装新的pydantic

 pip install pydantic==1.10.9
1

ok,问题解决

问题二
OSError: Can’t load the configuration of ’ --instruction’. If you were trying to load it from ‘https://huggingface.co/models’, make sure you don’t have a local directory with the same name. Otherwise, make sure ’ --instruction’ is the correct path to a directory containing a config.json file

原因：我是一个傻杯
解决方法：将–model_name_or_path jayelm/llama-7b-gist-1里的jayelm/llama-7b-gist-1换成你自己下载的模型的地址，地址一定要对啊？我搞了快一天了，发现地址放错路径了，快要吐血了。

python -m src.compress --model_name_or_path jayelm/llama-7b-gist-1 --base_llama_path llama-7b \
    --instruction "Name the top cities in France that should not be missed. Include the best aspects of each place as well."
1
2

接下来就是考验服务器内存的时候了