OS、GPU等版本确认
Ubuntu 22.04.3 LTS + NVIDIA GeForce RTX 3090
更新阿里云加速源 [可选]
Ubuntu 22.04.3 LTS jammy
将加速源更新至 /etc/apt/sources.list
1 2 3 4 5 6 7 8 9 10 echo 'deb http://mirrors.aliyun.com/ubuntu/ jammy main restricted universe multiverse deb-src http://mirrors.aliyun.com/ubuntu/ jammy main restricted universe multiverse deb http://mirrors.aliyun.com/ubuntu/ jammy-security main restricted universe multiverse deb-src http://mirrors.aliyun.com/ubuntu/ jammy-security main restricted universe multiverse deb http://mirrors.aliyun.com/ubuntu/ jammy-updates main restricted universe multiverse deb-src http://mirrors.aliyun.com/ubuntu/ jammy-updates main restricted universe multiverse deb http://mirrors.aliyun.com/ubuntu/ jammy-proposed main restricted universe multiverse deb-src http://mirrors.aliyun.com/ubuntu/ jammy-proposed main restricted universe multiverse deb http://mirrors.aliyun.com/ubuntu/ jammy-backports main restricted universe multiverse deb-src http://mirrors.aliyun.com/ubuntu/ jammy-backports main restricted universe multiverse' >/etc/apt/sources.list
获取OS版本 1 2 3 4 5 6 7 $ lsb_release -a No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 22.04.3 LTS Release: 22.04 Codename: jammy
查看显卡型号
系统安装了 NVIDIA GeForce RTX 3090 显卡，推荐安装的驱动程序是 nvidia-driver-535 。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 $ sudo ubuntu-drivers devices == /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0 == modalias : pci:v000010DEd00002204sv00007377sd0000120Bbc03sc00i00 vendor : NVIDIA Corporation model : GA102 [GeForce RTX 3090] driver : nvidia-driver-535-server - distro non-free driver : nvidia-driver-525-server - distro non-free driver : nvidia-driver-535-open - distro non-free driver : nvidia-driver-525 - distro non-free driver : nvidia-driver-525-open - distro non-free driver : nvidia-driver-470-server - distro non-free driver : nvidia-driver-470 - distro non-free driver : nvidia-driver-535-server-open - distro non-free driver : nvidia-driver-535 - distro non-free recommended driver : xserver-xorg-video-nouveau - distro free builtin
安装驱动和CUDA
因SD WebUI的webui.sh中指定的xFormers版本需要PyTorch 2.0.1+cu118，需要确保正确安装CUDA v11.8
通过 CUDA Toolkit 11.8 Downloads | NVIDIA Developer 检索发现 CUDA v11.8 配对的驱动最低为v520，且页面有对应的安装包地址
卸载错误的驱动（可选） 1 2 sudo apt-get remove --purge nvidia* sudo apt autoremove
卸载错误的CUDA（可选） 1 sudo apt-get --purge remove "*cublas*" "cuda*"
若有卸载操作，执行以下安装前建议reboot
安装正确的驱动 如果接受 ubuntu-drivers 给到的驱动版本建议不低于v520，可以直接使用 ubuntu-drivers 进行安装
1 $ sudo ubuntu-drivers autoinstall
或者，使用命令有选择地安装所需的驱动程序，如
1 $ sudo apt install nvidia-driver-535
安装正确的CUDA 1 2 wget wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda_11.8.0_520.61.05_linux.run sudo sh cuda_11.8.0_520.61.05_linux.run
完成安装后在~/.bashrc里加入环境变量：
1 2 export PATH="/usr/local/cuda-12.1/bin:$PATH" export LD_LIBRARY_PATH="/usr/local/cuda-12.1/lib64:$LD_LIBRARY_PATH"
异常处理 仍有其他方式安装的CUDA，nvcc -V仍有输出，如：
1 2 3 4 5 nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2022 NVIDIA Corporation Built on Wed_Jun__8_16:49:14_PDT_2022 Cuda compilation tools, release 11.7, V11.7.99 Build cuda_11.7.r11.7/compiler.31442593_0
可按以下方式继续尝试清理CUDA：
1 2 3 4 5 6 7 8 9 10 11 12 # 进入cuda的安装目录 cd /usr/local/cuda/bin # 执行cuda自带的卸载程序 sudo ./cuda-uninstaller # 清理目录残余 # sudo rm -fr /usr/local/cuda-* # 清理安装包残余 # 检索安装包 sudo dpkg -l |grep cuda # 清理后重启 sudo reboot
CUDA包附带的显卡驱动不匹配：
如sh cuda_x.x.x_y.y.y_linux.run时，提示Installation failed. See log at /var/log/cuda-installer.log for details.，而cuda-installer.log明确提示[ERROR]: Install of driver component failed.
因为funfile中带的驱动不兼容当前显卡，建议按sudo ubuntu-drivers devices给出的清单，选择大于y.y.y版本的驱动指定安装，如：
1 sudo apt install nvidia-driver-535
重启系统 安装完成后，重新启动系统即可完成。
检查驱动和CUDA 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 # 检查显卡驱动版本 # NVIDIA Driver v535.146.02 $ nvidia-smi +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 535.146.02 Driver Version: 535.146.02 CUDA Version: 12.2 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 NVIDIA GeForce RTX 3090 Off | 00000000:01:00.0 Off | N/A | | 0% 32C P8 18W / 350W | 26MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ +---------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=======================================================================================| | 0 N/A N/A 1257 G /usr/lib/xorg/Xorg 9MiB | | 0 N/A N/A 1414 G /usr/bin/gnome-shell 8MiB | +---------------------------------------------------------------------------------------+ # 检查CUDA版本 # CUDA v12.1 $ nvcc -V nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2023 NVIDIA Corporation Built on Tue_Feb__7_19:32:13_PST_2023 Cuda compilation tools, release 12.1, V12.1.66 Build cuda_12.1.r12.1/compiler.32415258_0
安装SD WebUI 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 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 # ENV # GPU::GeForce RTX 3090 # OS::Ubuntu Server 22.04 LTS 64位 # update ubuntu package 安装依赖包 sudo apt-get update sudo apt-get upgrade sudo apt-get install linux-headers-$(uname -r) sudo apt-get build-dep gcc sudo apt-get install build-essential gcc --version # install conda 安装 conda # latest：https://docs.conda.io/projects/conda/en/stable/user-guide/install/linux.html wget https://repo.anaconda.com/archive/Anaconda3-2023.09-0-Linux-x86_64.sh bash Anaconda3-2023.09-0-Linux-x86_64.sh conda # 回车&&yes bash # 升级最新conda conda update conda # 升级最新anaconda conda update anaconda # create env 创建环境 conda create -n sd python=3.10 conda activate sd # show env 查看环境 conda info -e # delete env 删除环境 # conda remoave -n sd --all # download sd 下载 webui # git clone https://ghproxy.com/https://github.com/AUTOMATIC1111/stable-diffusion-webui git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui cd stable-diffusion-webui # accelerate pytorch torch 下载过慢可使用腾讯源 # pip install -i https://mirrors.cloud.tencent.com/pypi/simple torch # 阿里：http://mirrors.aliyun.com/pypi/simple/ # 豆瓣：http://pypi.douban.com/simple/ # 清华：https://pypi.tuna.tsinghua.edu.cn/simple/ # 中科大：https://pypi.mirrors.ustc.edu.cn/simple/ # 中科大：http://pypi.mirrors.ustc.edu.cn/simple/ # install depend 安装依赖 # torchvision xformers相关的问题较多，先清理 pip uninstall torch torchvision xformers torchaudio -y sudo apt-get install libgoogle-perftools4 libtcmalloc-minimal4 -y # pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121 # pip install xformers # # # # # # # # bash webui.sh --listen --xformers --port 7860 --enable-insecure-extension-access
异常处理 Style database not found: /data/AI/stable-diffusion-webui/styles.csv styles.csv 下载后放入提示位置即可。
ImportError: Using SOCKS proxy 如果bash webui.sh运行中出现错误提示，证明当前session环境变量设置了代理，在运行webui.sh前先运行unset https_proxy http_proxy all_proxy
1 ImportError: Using SOCKS proxy, but the 'socksio' package is not installed. Make sure to install httpx using `pip install httpx[socks]`.
Cannot locate TCMalloc (improves CPU memory usage) 如果出现红字错误提醒：Cannot locate TCMalloc (improves CPU memory usage)，是因为系统缺少 libgoogle-perftools4 和 libtcmalloc-minimal4 这两个库，它们可以提高 CPU 的内存使用效率。可以通过以下命令安装它们：
1 sudo apt-get install libgoogle-perftools4 libtcmalloc-minimal4 -y
若出现以下提示，在启动命令添加--xformers参数，以强制使用xformers模块：bash webui.sh --xformers
1 2 3 4 Launching Web UI with arguments: no module 'xformers'. Processing without... no module 'xformers'. Processing without... No module 'xformers'. Proceeding without it.
如果监听端口需要改为0.0.0.0，在启动命令添加--listen参数，将监听改为0.0.0.0：bash webui.sh --listen
安装双语插件 由于SD WebUI的扩充功能发展太快，翻译可能跟不上，建议另外安装双语对照 sd-webui-bilingual-localization ，同时显示中文和英文的文本，这样看教学时就不会找无按钮了。
双语扩充功能安装方法：在Extensions页面按Install from URL，填入https://github.com/journey-ad/sd-webui-bilingual-localization，再按Install。
在激活双语扩充功能前，要到Settings → User interface → Localization设为None再重启WebUI才会生效。
AssertionError: extension access disabled because of command line flags 如果安装插件时提示：AssertionError: extension access disabled because of command line flags，以在启动 WebUI 的时候添加一个参数 --enable-insecure-extension-access，或者在 webui-user.sh 的 export COMMANDLINE_ARGS 参数中增加以下参数 --enable-insecure-extension-access。
常用 model 主模型 Stable-diffusion SD1.5 [基础模型，生成各种风格的图像] 1 wget https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt -P models/Stable-diffusion/
pastel-Mix [生成柔和色彩的二次元风格的图像] 1 wget https://civitai.com/api/download/models/6297 -O models/Stable-diffusion/pastelMixStylizedAnime_pastelMixPrunedFP16.safetensors
MIX-Pro-V3 [生成写实风格的图像，适合人物、风景、动物等] 1 wget https://civitai.com/api/download/models/8511 -O models/Stable-diffusion/mixProV3_v3.safetensors
chilloutmix [生成清新风格的图像，适合人物、风景、动物等] 1 wget https://civitai.com/api/download/models/11745 -O models/Stable-diffusion/chilloutmix_NiPrunedFp32Fix.safetensors
Cetus-Mix [生成海洋生物和水下风景的图像] 1 wget https://civitai.com/api/download/models/18496 -O models/Stable-diffusion/cetusMix_cetusVersion3.safetensors
MeinaMix [生成美女风格的图像，适合人物、服装、化妆等] 1 wget https://civitai.com/api/download/models/16925 -O models/Stable-diffusion/meinamix_meinaV7.safetensors
Counterfeit-V2.5 [生成仿真风格的图像，适合人物、风景、动物等] 1 wget https://civitai.com/api/download/models/7425 -O models/Stable-diffusion/CounterfeitV25_25.safetensors
ligneClaireAnime-V1 [生成线条清晰的二次元风格的图像] 1 wget https://civitai.com/api/download/models/4279 -O models/Stable-diffusion/ligneClaireAnime_v1.safetensors
AnythingAndEverything [生成各种风格的图像，是一个通用的模型] 1 wget https://civitai.com/api/download/models/8365 -O models/Stable-diffusion/anythingAndEverything_anythingAndEverythingVer.safetensors
VAE模型 滤镜和微调使用 Pastel-Waifu-Diffusion [增加柔和色彩的效果] 1 wget https://civitai.com/api/download/models/6297?type=VAE -O models/VAE/pastel-waifu-diffusion.vae.pt
kl-f8-anime2 [增加二次元风格的效果] 1 wget https://civitai.com/api/download/models/8511?type=VAE -O models/VAE/kl-f8-anime2.ckpt
Lora模型 微调使用 moxin [固定墨心风格的效果] 1 wget https://civitai.com/api/download/models/14856 -O models/Lora/Moxin_10.safetensors
fashion-gir [固定时尚女孩风格的效果] 1 wget https://civitai.com/api/download/models/18070 -O models/Lora/fashionGirl_v47.safetensors
line [固定线条风格的效果] 1 wget https://civitai.com/api/download/models/19075 -O models/Lora/animeLineartStyle_v20Offset.safetensors
eyes [固定眼睛风格的效果] 1 2 wget https://civitai.com/api/download/models/6433 -O models/Lora/eyeLora_eyesV10.safetensors wget https://civitai.com/api/download/models/19859 -O models/Lora/animeTarotCardArtStyleLora_v20Offset.safetensors
控制模型 control_canny [增加边缘检测的效果] 1 wget https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_canny-fp16.safetensors -P extensions/sd-webui-controlnet/models/
control_depth [增加深度检测的效果] 1 wget https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_depth-fp16.safetensors -P extensions/sd-webui-controlnet/models/
control_openpose [增加人体姿态检测的效果] 1 wget https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_openpose-fp16.safetensors -P extensions/sd-webui-controlnet/models/
control_hed [增加全局边缘检测的效果] 1 wget https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_hed-fp16.safetensors -P extensions/sd-webui-controlnet/models/
control_mlsd [增加多尺度线条检测的效果] 1 wget https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_mlsd-fp16.safetensors -P extensions/sd-webui-controlnet/models/