Tencent’s open source SD workflow saving plug-in LightDiffusionFlow

Tencent announced the open source SD workflow preservation plug-in LightDiffusionFlow. This is an open source plug-in based on the AI ​​painting open source platform StableDiffusion webUI (SD for short). It can help you save all workflow data with one click. Next time you use it, just drag in the Flow file and you can quickly reproduce the entire workflow. As convenient as using Photoshop's PSD mockup files.

Function

  • Save and reproduce workflow

Save or reproduce the SD painting workflow with one click, including parameter settings for models, prompt words, mat images, and other third-party plug-ins.

  • Support reading PNG Info

Compatible with reading the PNG Info built into the image and automatically restoring it to the UI interface, and supporting parsing and restoring ControlNet settings.

  • Automatically search for models

When importing a Flow file, it will automatically search for relevant models in Civitai, match the corresponding model and provide a download link.

  • Support third-party plug-ins

Theoretically, it supports saving parameters of all third-party plug-ins.

  • SD Beginner Friendly

For SD beginners, some reliable workflow files are provided, which can be loaded with one click and support custom parameter adjustment.

value

  • For SD beginners

Through Flow files, you can quickly get started with SD, which greatly reduces the threshold for learning and using SD, and quickly improves AI painting capabilities.

  • For SD advanced people

Through LightDiffusionFlow, you can quickly save high-quality workflows, quickly reuse them with one click, reduce tedious operating costs, and better spread AI painting capabilities.

  • For enterprise teams

Through LightDiffusionFlow, you can establish a series of reusable AI painting workflows based on industry and business attributes, quickly build the team's AI painting capabilities, and achieve cost reduction and efficiency improvement.

Guess you like

Origin www.oschina.net/news/258751