Deploy Stable Diffusion on JD Cloud GPU: A Complete Step‑by‑Step Guide

1. Create GPU Instance

The JD Cloud GPU instance (e.g., Tesla P40 24G) provides the compute power needed for Stable Diffusion.

Recommended configuration: 24 GB GPU memory, 12 CPU cores, 48 GB RAM.

1.2 Configure Security Group

Create a security group and open ports 7860, 7861, 8080, and 8888.

2. Environment Installation

2.1 Install GPU Driver

Find the appropriate driver version (e.g., 510) on NVIDIA’s website and install it.

# Install driver version 510
apt install nvidia-driver-510
# Verify installation
nvidia-smi

2.2 Install CUDA

Match the CUDA version to the driver (CUDA 11.6 for driver 510) and install it.

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin
sudo mv cuda-ubuntu2004.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/11.6.2/local_installers/cuda-repo-ubuntu2004-11-6-local_11.6.2-510.47.03-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu2004-11-6-local_11.6.2-510.47.03-1_amd64.deb
sudo apt-key add /var/cuda-repo-ubuntu2004-11-6-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda

2.3 Install Python 3.10

apt install software-properties-common
add-apt-repository ppa:deadsnakes/ppa
apt update
apt install python3.10
python3.10 --version

Configure pip to use a domestic mirror to avoid timeouts:

[global]
index-url = https://pypi.tuna.tsinghua.edu.cn/simple
[install]
trusted-host = https://pypi.tuna.tsinghua.edu.cn

2.4 Install Anaconda

wget https://repo.anaconda.com/archive/Anaconda3-2023.03-1-Linux-x86_64.sh
bash Anaconda3-2023.03-1-Linux-x86_64.sh
conda create -n python3.10.9 python==3.10.9
conda activate python3.10.9

2.5 Install PyTorch

# Using conda (recommended)
conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.6 -c pytorch -c nvidia
# Or using pip
pip install torch==1.13.1+cu116 torchvision==0.14.1+cu116 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu116

3. Deploy Stable Diffusion WebUI

3.1 Clone Repository

conda activate python3.10.9
git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

3.2 Install Dependencies

cd stable-diffusion-webui
pip install -r requirements_versions.txt
pip install -r requirements.txt

3.3 Launch WebUI

python launch.py --listen --enable-insecure-extension-access

The first run downloads common models; if a model fails to download, retrieve it manually from HuggingFace and place it in stable-diffusion-webui/models/Stable-diffusion.

3.4 Access UI

Open http://<IP>:7860 in a browser. For public access, set authentication:

python launch.py --listen --enable-insecure-extension-access --gradio-auth username:password

4. Recommended Plugins & Tools

4.1 LoRA (Additional Networks)

Install via Extensions → Install from URL with the repository

https://ghproxy.com/https://github.com/kohya-ss/sd-webui-additional-networks.git

, then set the LoRA folder path in Settings → Additional Networks .

4.2 ControlNet

Install via Extensions → Install from URL with

https://ghproxy.com/https://github.com/Mikubill/sd-webui-controlnet.git

. Download required model files from HuggingFace if automatic download fails.

4.3 Jupyter Notebook

jupyter notebook --allow-root --NotebookApp.token='your_token'

Access the notebook at http://<IP>:8888.

4.4 Kohya_ss (Model Training)

Install Docker and NVIDIA Container Toolkit, then clone the repository and adjust ports (map 7860 of Kohya to host 7861 to avoid conflict with WebUI).

sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit-base
nvidia-ctk --version
git clone https://github.com/bmaltais/kohya_ss.git
cd kohya_ss
# Edit docker-compose.yaml to map 0.0.0.0:7861:7860
docker compose build
docker compose run --service-ports kohya-ss-gui

Download missing model files from HuggingFace if needed.

5. Summary

After completing the installation and plugin configuration, Stable Diffusion runs efficiently on a JD Cloud GPU instance, providing a powerful, free, and extensible platform for AI‑generated artwork.