• Fixed character counter to match the currently selected model (@gdbing)
• Added button to view models in Finder (@alexey-detr)
• Updated project website url

Mochi Diffusion is looking for a new project maintainer. If you are interested in adopting, feel free to send me a DM on Discord @godly-devotion.

• Added ability to resize Inspector width
• Removed unnecessary Apply button from Settings window
• Fixed animation glitch when images are added or removed from the gallery
• Fixed gallery occasionally incorrectly sorting by Oldest First
• Fixed missing localization for initial model loading message
• Updated generation preview UI and show by default
• Updated system requirements to macOS 14
  • macOS 14 and later is required to use split-einsum v2 & SDXL models
• General performance improvements

• Fixed using slider for countries that use decimal commas (@haiodo)

• Added option to send notifications when images are ready (@mangoes-dev)
• Added ability to change slider control values by keyboard input (@gdbing)
• Changed Quick Look shortcut to spacebar (like Finder)
• Changed scheduler timestep to Karras for SDXL models
• Changed minimum step option to 1 (@amikot)
• Updated translations

• Added auto refresh of Image, Model, & ControlNet list (@gdbing)
• Added ability to queue images (@gdbing)
• Added ability to drag and drop gallery images out (@gdbing)
• Added ability to set the Starting Image by dragging and dropping an image (@gdbing)
• Added automatic resizing of input images based on output image size (@gdbing)
• Added Hungarian translation (Janos Hunyadi)
• Changed ControlNet list to only show those compatible with the currently selected model (@gdbing)

• Added time remaining info for currently generating image (@hoseins77)
• Changed starting image strength slider range (@gdbing)
• Changed magnifier button in Settings to show folder selection dialog (@vzsg)
• Updated translations

• Fixed problem selecting models with chunked Unets (@vzsg)
• Fixed conflicting keyboard shortcuts between text inputs and gallery (@hoseins77)
• Changed gallery selection keyboard shortcuts to no longer require command key

• Fixed the image gallery background's round things on macOS Sonoma (@vzsg)

• Added support for Stable Diffusion XL refiner (@ZachNagengast)
  • Model must have UnetRefiner.mlmodelc file
• Added Vietnamese translation (Toàn Hoàng Đức)
• Fixed noisy generated image (@ZachNagengast)
• Fixed Xcode build on macOS 14 (@ZachNagengast)

• Fixed setting ControlNet image
• Fixed restoring last gallery sort option
• Added support for Stable Diffusion XL models (requires macOS 14 beta) (@ZachNagengast)
• Added setting to show image preview during generation (@hoseins77)
• Sorted ControlNet list (@jrittvo)
• Updated link to HuggingFace Core ML Community

• Fixed ControlNet
• Fixed missing Dutch, Polish, & Ukrainian translations

Note: To use inpainting, select an inpainting ControlNet model and provide a mask image with transparent pixels.

• Fixed app crash when changing settings
• Added Norwegian Bokmal translation (Espen Bye)

Note: To use inpainting, select an inpainting ControlNet model and provide a mask image with transparent pixels.

• Fixed app crash when changing settings
• Added Norwegian Bokmal translation (Espen Bye)

Note: To use inpainting, select an inpainting ControlNet model and provide a mask image with transparent pixels.

• Added support for inpainting with ControlNet (@vzsg)
• Added option to sort images by date (@hoseins77)
• Added symlink support for model directory (@surjikal)
• Changed starting image strength value range (@jrittvo)
• Fixed potential crash when clearing ControlNet image (@vzsg)

Note: To use inpainting, select an inpainting ControlNet model and provide a mask image with transparent pixels.

• Added ControlNet (@stuartjmoore)
• Changed starting image strength value range (@jrittvo)
• Removed support for Intel Macs

Note: Previous Stable Diffusion models will need to be reconverted to support ControlNet. Several Core ML ControlNet models have been converted here by @jrittvo. See the wiki to find out how to convert the models.

• Added animation when converting image to high resolution (@jinhongw)
• Changed starting image implementation to follow model image size (@vzsg)
• Changed description too long message to use accent color
• Improved scheduler speed
• Updated translations

Special thanks to the following for supporting this project❣️
orange-wedge, RuralRob, vacekj, julien-c, BirdSesame, li775176364, Da-mi-en, monks1975, & various translators

• Added option to auto select ML Compute Unit (@vzsg)
• Added support for restoring jpeg files (@vzsg)
• Changed default model & image folder directory to user's home directory
• Changed import behavior to copy images
• Updated translations

Special thanks to the following for supporting me and making this app possible❣️
orange-wedge, RuralRob, vacekj, julien-c, BirdSesame, Da-mi-en, monks1975, & anonymous donors

• Added option to select a starting image (commonly known as Image2Image)
  • Model must have VAEEncoder.mlmodelc file
  • Starting image and model must be 512x512 in size
• Added Dutch translation (Richard Venneman)
• Added setting to change default save image type
• Added support for importing and saving HEIC images
  • HEIC images are so small and efficient that it only uses about 5% of the file size of upscaled PNG images 🤯
• Added link to project translation website (Help > Contribute Translation)
• Updated sidebar UI
• Removed focus from text fields if image is clicked to avoid accidentally changing text
• General performance improvements

Has Mochi Diffusion been useful? Support this project on GitHub or Liberapay❣️

Special thanks to the following for supporting me and making this app possible 🎉
raisingfightingspirit, serovar, orange-wedge, RuralRob, vacekj, julien-c, BirdSesame, Da-mi-en, monks1975, & anonymous donors

🧋

Don't you just hate it when you close the app and realize you forgot to save your images?

Soon this will only be something pepperidge farm remembers.

This release is entitled Never Gonna Let You Down, as you're never gonna lose your images again 😎

• Added option to autosave & restore images
• Added option to hide or show the Info panel
• Changed app behavior to quit when window is closed
• Images that are removed are now moved to Trash
• Fixed Copy image menu option
• Fixed Copy Options to Sidebar menu option
• Fixed incorrect image date on import
• Fixed removing an unselected image causing the selection to unexpectedly change (@vzsg)

Has Mochi Diffusion been useful? Support this project on GitHub or Liberapay.

Special thanks to the following for supporting me and making this app possible 🎉
raisingfightingspirit, serovar, orange-wedge, RuralRob, vacekj, julien-c, Quick-Eyed-Sky, BirdSesame, Da-mi-en, monks1975, & anonymous donors

• Added Ukrainian translation (Pavlo Pavlov)
• Added Copy to clipboard option to image context menu (@vzsg)
• Changed Number of Images option to slider
• Changed image selection to move to beginning if end is reached and vice versa (@vzsg)
• Fixed & improved search (@vzsg)
• Fixed copying Include in Image option to sidebar in Info panel

Can you speak another language? Visit the project page on Crowdin!

Special thanks to the following for supporting me and making this app possible 🎉
raisingfightingspirit, serovar, orange-wedge, RuralRob, vacekj, julien-c, Quick-Eyed-Sky, BirdSesame, Da-mi-en, monks1975, angeenes, & anonymous donors