Lots of features since the previous version:
-
Full support for Custom Models (UI selection)
-
Choose JPEG or PNG for output format
-
Don't reload the model when switching between img2img and txt2img
-
Reduced RAM memory usage for txt2img
-
Task Queue
-
Negative Prompts
-
Thumbnails for the image modifiers, written by @haka and @manny
-
Specify Multiple Prompts - choose a text file, or enter one prompt per line
-
Early version of Prompt Matrix - Separate your prompt with the | character to explore variations quickly. E.g. girl holding a rose | illustration | cinematic lighting creates four task combinations automatically: girl holding a rose, girl holding a rose, illustration, girl holding a rose, cinematic lighting and girl holding a rose, illustration, cinematic lighting
-
Use curly brackets in prompts to try different words E.g. man riding a {horse,motorcycle} results in man riding a horse and man riding a motorcycle being created automatically.
-
New Image Buttons: Make Similar Images, Draw another 25 steps, Upscale and Fix Faces - you can run these after an image has been generated. For e.g. you can now upscale or fix your images after they have been generated! (thanks @madrang)
-
Use our project simultaneously across multiple browser tabs/PCs/tablets/phones without errors (written by @madrang)
-
Choose between 7 themes for the UI, Mobile-friendly UI, Cleaner styling of image settings (written by @Bilbo's Last Clean Doily)
-
Write your own custom plugins for the UI - you can now write custom buttons for the UI in a javascript file, and put your_filename.plugin.js file inside the plugins/ui folder. You can make custom buttons for the image right now.
-
Auto-save image settings across browser restarts - (written by @Bilbo's Last Clean Doily)
-
Aspect Ratio preserved in the in-painting editor
-
Custom Image Modifiers
-
Custom themes (thanks @Bilbo's Last Clean Doily)
-
Use micromamba to install git/conda (if required), instead of bundling it in the installer