【AIGC】HF-Mirror 使用说明（MacOS 版）

今天这篇文章主要是记录 HF-Mirror 的使用过程，官网上提供了 Linux 和 Windows 的使用方式。本人作为 MacOS 用户也将自己的使用办法记录一下，希望能够帮助到其他人。

HF-Mirror 是什么？

HF-Mirror 是大神 padeoe 开源的 huggingface 镜像站，网站属于公益性质，它帮助国内 AI 开发者能够快速、稳定地下载 huggingface 上模型和数据集。网站地址：https://hf-mirror.com/

怎么使用呢？

在下载模型前，需要对系统终端进行一些配置。

安装依赖

(base) yuanzhenhui@MacBook-Pro python % pip install -U huggingface_hub
Requirement already satisfied: huggingface_hub in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (0.15.1)
Collecting huggingface_hub
  Obtaining dependency information for huggingface_hub from https://files.pythonhosted.org/packages/3d/c8/c3342c97848896df5d78d18abd94c558e457a4f02feec99a79989d8c30e0/huggingface_hub-0.21.2-py3-none-any.whl.metadata
  Downloading huggingface_hub-0.21.2-py3-none-any.whl.metadata (13 kB)
Requirement already satisfied: filelock in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from huggingface_hub) (3.9.0)
Collecting fsspec>=2023.5.0 (from huggingface_hub)
  Obtaining dependency information for fsspec>=2023.5.0 from https://files.pythonhosted.org/packages/ad/30/2281c062222dc39328843bd1ddd30ff3005ef8e30b2fd09c4d2792766061/fsspec-2024.2.0-py3-none-any.whl.metadata
  Downloading fsspec-2024.2.0-py3-none-any.whl.metadata (6.8 kB)
Requirement already satisfied: requests in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from huggingface_hub) (2.31.0)
Requirement already satisfied: tqdm>=4.42.1 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from huggingface_hub) (4.65.0)
Requirement already satisfied: pyyaml>=5.1 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from huggingface_hub) (6.0)
Requirement already satisfied: typing-extensions>=3.7.4.3 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from huggingface_hub) (4.7.1)
Requirement already satisfied: packaging>=20.9 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from huggingface_hub) (23.1)
Requirement already satisfied: charset-normalizer<4,>=2 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from requests->huggingface_hub) (2.0.4)
Requirement already satisfied: idna<4,>=2.5 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from requests->huggingface_hub) (3.4)
Requirement already satisfied: urllib3<3,>=1.21.1 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from requests->huggingface_hub) (1.26.16)
Requirement already satisfied: certifi>=2017.4.17 in /Users/yuanzhenhui/anaconda3/lib/python3.11/site-packages (from requests->huggingface_hub) (2023.11.17)
Downloading huggingface_hub-0.21.2-py3-none-any.whl (346 kB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 346.2/346.2 kB 644.8 kB/s eta 0:00:00
Downloading fsspec-2024.2.0-py3-none-any.whl (170 kB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 170.9/170.9 kB 2.7 MB/s eta 0:00:00
Installing collected packages: fsspec, huggingface_hub
  Attempting uninstall: fsspec
    Found existing installation: fsspec 2023.4.0
    Uninstalling fsspec-2023.4.0:
      Successfully uninstalled fsspec-2023.4.0
  Attempting uninstall: huggingface_hub
    Found existing installation: huggingface-hub 0.15.1
    Uninstalling huggingface-hub-0.15.1:
      Successfully uninstalled huggingface-hub-0.15.1
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
s3fs 2023.4.0 requires fsspec==2023.4.0, but you have fsspec 2024.2.0 which is incompatible.
Successfully installed fsspec-2024.2.0 huggingface_hub-0.21.2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

设置环境变量

(base) yuanzhenhui@MacBook-Pro python % cd
(base) yuanzhenhui@MacBook-Pro python % vim .bash_profile

# 在 bash_profile 文件中随便找一行将下面的命令记录一下

export HF_ENDPOINT=https://hf-mirror.com
1
2
3
4
5
6

如果你止步于此，那么在下载时就会跳转到如下页面：

点击“前往官方入口”按钮会跳转到 Huggingface 官网，在这里你需要生成 access token （按照提示跟着走就可以了…），如下图：

使用 token 下载

拷贝后就可以通过以下命令进行下载，以 google/gemma-7b 为例：

(base) yuanzhenhui@MacBook-Pro python % huggingface-cli download \
--token hf_xxxxxxxxxxxxxxxxxxx \
--resume-download google/gemma-7b \
--local-dir gemma
1
2
3
4

至于下载后怎么用就留待后面的文章进行描述吧。