生成式AI常见系统问题

人工智能 Alan 3周前 (06-06) 131次浏览 0个评论 扫描二维码

很多年没使用Windows了,这次组装了一个台式机,主要是为了能使用Nvidia的4090显卡,虽然可以做双系统,但简单使用下来发现WSL配合VScode使用也不错,先看看有哪些坑吧。

PyTorch安装

1. CategoryInfo : SecurityError: (:) [],PSSecurityException

2.Miniconda下载:https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda/

设置镜像源:

3.推荐版本为Python 3.10,最新的Python版本会出现一些难以debug的错误,比如stable-diffusion-webui即使安装了Rust也会报ERROR: Failed building wheel for tokenizers

4.指定Python版本

5.默认wsl --install会将系统安装到C盘,将WSL迁移到D盘:

6.查看Nvidia显卡实时数据watch -n 1 nvidia-smi

生成式AI常见系统问题

7.Setting pad_token_id to eos_token_id:50256 for open-end generation.

将model修改为正在使用的模型,如gpt2

8.401 Client Error: Unauthorized for url: https://huggingface.co/api/repos/create

9.PackageNotFoundError: No package metadata was found for bitsandbyte

10.Make sure to have access to it at https://huggingface.co/mistralai/Mistral-7B-v0.1

需要在相应页面同意使用协议并在命令行登录

11.The load_in_4bit and load_in_8bit arguments are deprecated and will be removed in the future versions. Please, pass a BitsAndBytesConfig object in quantization_config argument instead.

12.WSL局域网访问

此时在本机上可通过局域网IP 访问,但其它主机依然无法使用,需要进入Windows Defender 防火墙,点击高级设置>入站规则>新建规则,按照向导添加允许访问的端口。

 

喜欢 (0)
[]
分享 (0)
发表我的评论
取消评论

表情 贴图 加粗 删除线 居中 斜体 签到

Hi,您需要填写昵称和邮箱!

  • 昵称 (必填)
  • 邮箱 (必填)
  • 网址