aarch64 arm64 部署 stable diffusion webui 笔记 【2】继续安装其他依赖 gfpgan

news/2024/7/10 21:13:22 标签: stable diffusion, pytorch, 笔记

接上篇

aarch64 arm64 部署 stable diffusion webui 笔记 【1】准备 venv 安装pytorch 验证cuda_hkNaruto的博客-CSDN博客

编辑requirements_versions.txt,注释掉torch

启动webui.sh

(venv) [root@ceph3 stable-diffusion-webui]# useradd yeqiang
useradd:警告:此主目录已经存在。
不从 skel 目录里向其中复制任何文件。
(venv) [root@ceph3 stable-diffusion-webui]# chown yeqiang:yeqiang . -R
(venv) [root@ceph3 stable-diffusion-webui]# su yeqiang
(venv) [yeqiang@ceph3 stable-diffusion-webui]$ ./webui.sh 

故障,通过修改源码,添加打印发现强制验证了torch,torchvision

继续。。

切换用户重新配置环境变量

[yeqiang@ceph3 root]$ cd /home/yeqiang/stable-diffusion-webui/

[yeqiang@ceph3 stable-diffusion-webui]$ source venv/bin/activate
(venv) [yeqiang@ceph3 stable-diffusion-webui]$ export LD_LIBRARY_PATH=/usr/local/lib64:/usr/local/lib:/usr/lib64:/usr/lib
 

手动安装torchvision

问题:pip会自动依赖安装torch==2.0.1,这个版本是CPU版本,Ctrl+C中断,选择手动编译安装torchvision

源码:

https://github.com/pytorch/vision/archive/refs/tags/v0.15.2.tar.gz

编译

[root@ceph3 ~]# su  yeqiang
[yeqiang@ceph3 root]$ cd ~
[yeqiang@ceph3 ~]$ cd stable-diffusion-webui/
[yeqiang@ceph3 stable-diffusion-webui]$ source venv/bin/activate

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ cd ~
(venv) [yeqiang@ceph3 ~]$ tar -xf vision-0.15.2.tar.gz 
(venv) [yeqiang@ceph3 ~]$ cd vision-0.15.2/

(venv) [yeqiang@ceph3 vision-0.15.2]$ export LD_LIBRARY_PATH=/usr/local/lib64:/usr/local/lib:/usr/lib64:/usr/lib

(venv) [yeqiang@ceph3 vision-0.15.2]$export PATH=/usr/local/cuda/bin:$PATH
(venv) [yeqiang@ceph3 vision-0.15.2]$ python3 setup.py build

(venv) [yeqiang@ceph3 vision-0.15.2]$ python3 setup.py install

网络故障处理(自己的环境问题,可不关注)

(venv) [yeqiang@ceph3 vision-0.15.2]$ python3 -m pip install Pillow==10.0.0 -i https://pypi.tuna.tsinghua.edu.cn/simple

(venv) [yeqiang@ceph3 vision-0.15.2]$ python3 -m pip install requests==2.31.0 -i https://pypi.tuna.tsinghua.edu.cn/simple

(venv) [yeqiang@ceph3 vision-0.15.2]$ python3 -m pip install numpy==1.26.0b1 -i https://pypi.tuna.tsinghua.edu.cn/simple

当前状态

启动webui.sh,开始自动安装,观察

(venv) [yeqiang@ceph3 vision-0.15.2]$ cd /home/yeqiang/stable-diffusion-webui/
(venv) [yeqiang@ceph3 stable-diffusion-webui]$ ./webui.sh 

离线安装(网络问题,可不关注)

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ python3 -m pip install /home/yeqiang/stable-diffusion-webui-libs/*.whl

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ python3 -m pip install file:///home/yeqiang/stable-diffusion-webui-libs/filterpy-1.4.5.zip

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ python3 -m pip install file:///home/yeqiang/stable-diffusion-webui-libs/basicsr-1.4.2.tar.gz

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ python3 -m pip install file:///home/yeqiang/stable-diffusion-webui-libs/filterpy-1.4.5.zip
Processing /home/yeqiang/stable-diffusion-webui-libs/filterpy-1.4.5.zip
  Preparing metadata (setup.py) ... done
Requirement already satisfied: numpy in ./venv/lib/python3.10/site-packages (from filterpy==1.4.5) (1.26.0b1)
Requirement already satisfied: scipy in ./venv/lib/python3.10/site-packages (from filterpy==1.4.5) (1.11.2)
Requirement already satisfied: matplotlib in ./venv/lib/python3.10/site-packages (from filterpy==1.4.5) (3.7.2)
Requirement already satisfied: packaging>=20.0 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (23.1)
Requirement already satisfied: fonttools>=4.22.0 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (4.42.1)
Requirement already satisfied: contourpy>=1.0.1 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (1.1.0)
Requirement already satisfied: python-dateutil>=2.7 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (2.8.2)
Requirement already satisfied: pyparsing<3.1,>=2.3.1 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (3.0.9)
Requirement already satisfied: cycler>=0.10 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (0.11.0)
Requirement already satisfied: kiwisolver>=1.0.1 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (1.4.5)
Requirement already satisfied: pillow>=6.2.0 in ./venv/lib/python3.10/site-packages (from matplotlib->filterpy==1.4.5) (10.0.0)
Requirement already satisfied: six>=1.5 in ./venv/lib/python3.10/site-packages (from python-dateutil>=2.7->matplotlib->filterpy==1.4.5) (1.16.0)
Building wheels for collected packages: filterpy
  Building wheel for filterpy (setup.py) ... done
  Created wheel for filterpy: filename=filterpy-1.4.5-py3-none-any.whl size=110334 sha256=ebe74d23306277b389d61d6a6450c026877f94faba646830e7bb6b4eced72562
  Stored in directory: /home/yeqiang/.cache/pip/wheels/33/87/3a/5541add2fe3261f47dc6b60b56b6c5356f09625c13e05234e8
Successfully built filterpy
Installing collected packages: filterpy
Successfully installed filterpy-1.4.5

[notice] A new release of pip is available: 23.0.1 -> 23.2.1
[notice] To update, run: pip install --upgrade pip

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ python3 -m pip install /home/yeqiang/stable-diffusion-webui-libs/gfpan-libs/*

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ python3 -m pip install file:///home/yeqiang/stable-diffusion-webui-libs/8d2447a2d918f8eba5a4a01463fd48e45126a379.zip

当前进度

(venv) [yeqiang@ceph3 stable-diffusion-webui]$ pip list
Package                 Version
----------------------- ------------------
absl-py                 1.4.0
addict                  2.4.0
basicsr                 1.4.2
cachetools              5.3.1
certifi                 2023.7.22
charset-normalizer      3.2.0
contourpy               1.1.0
cycler                  0.11.0
Cython                  3.0.2
facexlib                0.3.0
filelock                3.12.3
filterpy                1.4.5
fonttools               4.42.1
future                  0.18.3
gfpgan                  1.3.5
google-auth             2.22.0
google-auth-oauthlib    1.0.0
grpcio                  1.57.0
idna                    3.4
imageio                 2.31.3
importlib-metadata      6.8.0
Jinja2                  3.1.2
kiwisolver              1.4.5
lazy_loader             0.3
llvmlite                0.40.1
lmdb                    1.4.1
Markdown                3.4.4
MarkupSafe              2.1.3
matplotlib              3.7.2
mpmath                  1.3.0
networkx                3.1
numba                   0.57.1
numpy                   1.24.4
oauthlib                3.2.2
opencv-python           4.8.0.76
packaging               23.1
Pillow                  10.0.0
pip                     23.0.1
platformdirs            3.10.0
protobuf                4.23.4
pyasn1                  0.5.0
pyasn1-modules          0.3.0
pyparsing               3.0.9
python-dateutil         2.8.2
PyWavelets              1.4.1
PyYAML                  6.0.1
requests                2.31.0
requests-oauthlib       1.3.1
rsa                     4.9
scikit-image            0.21.0
scipy                   1.11.2
setuptools              65.5.0
six                     1.16.0
sympy                   1.12
tb-nightly              2.15.0a20230902
tensorboard-data-server 0.7.1
tifffile                2023.8.30
tomli                   2.0.1
torch                   2.0.0a0+gite9ebda2
torchvision             0.15.2a0
tqdm                    4.66.1
typing_extensions       4.7.1
urllib3                 1.26.16
Werkzeug                2.3.7
wheel                   0.41.2
yapf                    0.40.1
zipp                    3.16.2

[notice] A new release of pip is available: 23.0.1 -> 23.2.1
[notice] To update, run: pip install --upgrade pip
 


http://www.niftyadmin.cn/n/5006762.html

相关文章

【计算机网络】 子网划分

文章目录 IP地址分类子网掩码网关广播地址非默认子网掩码子网划分常见问题 IP地址分类 学会十进制和二进制的相互转换可以很快速的有规律的记住 子网掩码 又叫网络掩码&#xff0c;地址掩码&#xff0c;子网络遮罩&#xff0c;就是说把子网络遮起来&#xff0c;不让外界窥探到…

SpringMVC综合案例

目录 一、SpringMVC常用注解 二、传递参数 2.1 基础类型String 2.2 复杂类型 2.3 RequestParam 2.4 PathVariable 2.5 RequestBody 2.6 RequestHeader 2.7 请求方法 三、返回值 3.1 void 3.2 String 3.3 StringModel 3.4 ModelAndView 四、页面跳转 4.1 转发 4…

基于Matlab实现多个图像去噪案例(附上源码+数据集)

图像去噪是图像处理中的一项重要任务&#xff0c;旨在减少或消除图像中的噪声&#xff0c;以提高图像的质量和清晰度。在本文中&#xff0c;我们将介绍如何使用Matlab实现图像去噪。 实现步骤 首先&#xff0c;我们需要了解图像噪声的类型。常见的图像噪声类型包括高斯噪声、…

肖sir__设计测试用例方法之正交表08_(黑盒测试)

设计测试用例方法之正交 一、正交表定义 正交试验设计法&#xff0c;是从大量的试验点中挑选出适量的、有代表性的点&#xff0c;应用依据迦罗瓦理论导出的“正交表”&#xff0c;合理的安排试验的一种科学的试验设计方法。 二、 正交常用的术语 指标&#xff1a;通常把判断试验…

《向量数据库指南》——AI原生向量数据库Milvus Cloud 2.3新功能

New Feature Upsert 功能 支持用户通过 upsert 接口更新或插入数据。已知限制,自增 id 不支持 upsert;upsert 是内部实现是 delete + insert所以性能上会有一定损耗,如果明确知道是写入数据的场景请继续使用 insert。 Range Search 功能 支持用户通过输入参数指定 search 的…

聊聊mysql jdbc的prepareStatement

序 本文主要研究一下mysql jdbc的prepareStatement prepareStatement java/sql/Connection.java /*** Creates a <code>PreparedStatement</code> object for sending* parameterized SQL statements to the database.* <P>* A SQL statement with or wit…

国产信创服务器如何进行安全可靠的文件传输?

信创&#xff0c;即信息技术应用创新&#xff0c;2018年以来&#xff0c;受“华为、中兴事件”影响&#xff0c;国家将信创产业纳入国家战略&#xff0c;并提出了“28n”发展体系。从产业链角度&#xff0c;信创产业生态体系较为庞大&#xff0c;主要包括基础硬件、基础软件、应…

如何实现自动化测试?

一、首先我们要清楚自动化测试的分类 以实现方式可分为UI自动化和接口自动化。UI自动化可用selenium等工具实现&#xff0c;接口自动化可用使用RobotFramework和Jmeter等工具实现&#xff0c;Jmeter也可做性能自动化&#xff0c;压力测试。 二、平时自动化测试怎么做 1. UI和…