原版的Stable Diffusion Web UI实在是太复杂了,功能太多,安装中问题过多,很容易就爆了
这是一个简化版的Stable Diffusion WebUI+API,帮助你更好的在本地部署,使用UI、修改或调用接口。
-
删减了大量冗余代码,仅提供Linux的支持。
-
API功能全部打开,无需启动时输入参数。
你只需安装好Linux操作系统,conda环境,cuda,NV驱动等深度学习基础环境。
接下来创建并进入新的conda环境(py39+)
git clone https://github.com/uni-openai/stable-diffusion-simple.git
pip3 install -r requirements.txt
python3 ./launch.py
全部接口参考这个网址:https://documenter.getpostman.com/view/9347507/2s93saZYEN
- [POST]
/sdapi/v1/txt2img
{
"enable_hr": false,
"denoising_strength": 0,
"firstphase_width": 0,
"firstphase_height": 0,
"hr_scale": 2,
"hr_upscaler": "None",
"hr_second_pass_steps": 0,
"hr_resize_x": 0,
"hr_resize_y": 0,
"prompt": "a pretty girl",
"styles": [],
"seed": -1,
"subseed": -1,
"subseed_strength": 0,
"seed_resize_from_h": -1,
"seed_resize_from_w": -1,
"sampler_name": "DPM++ SDE Karras",
"batch_size": 1,
"n_iter": 1,
"steps": 50,
"cfg_scale": 7,
"width": 512,
"height": 512,
"restore_faces": false,
"tiling": false,
"do_not_save_samples": false,
"do_not_save_grid": false,
"negative_prompt": "",
"eta": 0,
"s_min_uncond": 0,
"s_churn": 0,
"s_tmax": 0,
"s_tmin": 0,
"s_noise": 1,
"override_settings": {},
"override_settings_restore_afterwards": true,
"script_args": [],
"sampler_index": "DPM++ SDE Karras",
"script_name": "",
"send_images": true,
"save_images": false,
"alwayson_scripts": {}
}
说明:
- 修改checkpoint model需要在options POST接口进行