Deploy your own stable-diffusion using Tendong Cloud
1 Create project:
1. Enter the Trend Cloud user workbench and select: Current Space . Please ensure that the current space is the space automatically generated by the system during registration.
a. It is not a space automatically generated by the system, and there is no free computing power.
2. Click: Quick Create, select Create Project, and create a new project.
3. Fill in the relevant project information, and select the image and data set as follows.
4. After the project information is complete, click OK and a pop-up box will appear to select not to upload yet.
After selecting the image and data set, click "Create"
Click " Don't upload yet "
2 Initialize the development environment instance
Then find "Development" -> "Initialize Development Environment Instance" on the far right
Make your selection as shown below:
During the running initialization, wait for about 5-10 minutes. When the web terminal and JupyterLāb on the right are no longer gray, it indicates that the tool injection is successful. At this point,
you can use tools to tune the model on this development environment. See the next step for details.
3 Deployment model
After initialization, click "Web Terminal" on the right side as shown below
Just enter the following 4 series of commands in sequence. Note that after each command is entered, you must wait for it to finish running before entering the next command.
1. Unzip the code and model
tar xf /gemini/data-1/stable-diffusion-webui.tar -C /gemini/code/
2. Unzip the configuration file to the hidden directory /root/.cache
tar xf /gemini/data-1/cache.tar -C /root/
(The file location may change, please delete this line if an error is reported)
3. Copy the frpc intranet penetration file (note that there are two lines -> two instructions)
cp /gemini/data-1/frpc_linux_amd64 /root/miniconda3/lib/python3.10/site-packages/gradio/frpc_linux_amd64_v0.2
(The file location may change, please delete this line if an error is reported)
chmod +x /root/miniconda3/lib/python3.10/site-packages/gradio/frpc_linux_amd64_v0.2
Preprocessing operations on some instructions :
4. Run the project: (note there are two lines –> two instructions)
Enter the program directory:
cd /gemini/code/stable-diffusion-webui
Run the launch.py program:
python launch.py --deepdanbooru --share --theme dark --xformers --listen --gradio-auth qdy:123456
出现问题:
ERROR:Could not find a version that satisfies the requirement lpips (from versions:none)
ERROR:No matching distribution found for lpips
Change the image source: You can use the image source of task 1: Then run the python launch.py command
pip config set global.index-url https://mirrors.ustc.edu.cn/pypi/web/simple
When the following two URLs appear in the command window, the deployment is successful. You can copy the external access URL on the right to access wbui:
When accessing this website, a login page will appear. The account number is qdy and the password is 123456.
4 Model testing
Then load and enter:
使用测试:
(masterpiece),(best quality),(Realistic photos),a cute girl,war a cap,headphones,summer,sunset,mountain road,flowers,nice weather,healing sense,detailed,half-length shot,anime style
You can check the generation progress in the background:
Of course, you can also send your port information to others, or send it to your mobile phone and open it in a browser, and it can be used in the same way. You can also see it from the cloud server background.
Note: After the experience, be sure to stop the operation. The operation of Xiangri must be stopped. The project must be stopped. Otherwise, GPU resources will be consumed all the time
.