sdxl_turbo_diffusers_api

A image generation service api, developed with diffusers + flask

Get Started

1. You can load the model locally through the environment variable declaration, if you do not declare it, it will be downloaded from huggingface.co

export BaseModelPath="your/model/path"
export PidiNetPath="your/model/path"
export AdapterSketchPath="your/model/path"
export LoraPath="your/model/path"

2. Start the service through flask

pip install -r requirements.txt

python -m flask run

Api doc

just like stable diffusion webui api, but only:

/sdapi/v1/txt2img

Recommended parameters:

{
  "prompt": "stylized by Daria Petrilli, Raw digital photo, 3D Rendering, Colorful Bird, it is Verdant, concept art, Stormy weather, Dark Academia, Light and shadow plays, film grain, Nikon d3300, 80mm, glitter texture, most beautiful artwork in the world, highly detailed, extremely content happy smile",
  "steps": 5,
  "cfg_scale": 1.5,
  "width": 1344,
  "height": 768
}

with sketch:

{
  "prompt": "stylized by Daria Petrilli, Raw digital photo, 3D Rendering, Colorful Bird, it is Verdant, concept art, Stormy weather, Dark Academia, Light and shadow plays, film grain, Nikon d3300, 80mm, glitter texture, most beautiful artwork in the world, highly detailed, extremely content happy smile",
  "steps": 5,
  "cfg_scale": 4,
  "width": 1344,
  "height": 768
}

• /sdapi/v1/progress
• /api/v1/text2img/controlnet/reference