StabilityMatrix

Added

• Added SVD Shared Model & Output Folders for Forge (fixes #580)

Changed

• Improved error message when logging in with a Lykos account fails due to incorrect email or password
• Model Browser & Workflow Browser now auto-load when first navigating to those pages
• Removed update confirmation dialog, instead showing the new version in the update button tooltip

Fixed

• Fixed package launch not working when environment variable SETUPTOOLS_USE_DISTUTILS is set due to conflict with a default environment variable. User environment variables will now correctly override any default environment variables.
• Fixed "No refresh token found" error when failing to login with Lykos account in some cases
• Fixed blank entries appearing in the Categories dropdown on the Checkpoints page
• Fixed crash when loading invalid connected model info files
• Fixed #585 - Crash when drag & drop source and destination are the same
• Fixed #584 - --launch-package argument not working
• Fixed #581 - Inference teaching tip showing more often than it should
• Fixed #578 - "python setup.py egg_info did not run successfully" failure when installing Auto1111 or SDWebUI Forge
• Fixed #574 - local images not showing on macOS or Linux

Added

• Added Reference-Only mode for Inference ControlNet, used for guiding the sampler with an image without a pretrained model. Part of the latent and attention layers will be connected to the reference image, similar to Image to Image or Inpainting.
• Inference ControlNet module now supports over 42 preprocessors, a new button next to the preprocessors dropdown allows previewing the output of the selected preprocessor on the image.
• Added resolution selection for Inference ControlNet module, this controls preprocessor resolution too.
• Added Layer Diffuse sampler addon to Inference, allows generating foreground with transparency with SD1.5 and SDXL.
• Added support for deep links from the new Stability Matrix Chrome extension, supporting model downloads from civitai.com
• Added OpenArt.AI workflow browser for ComfyUI workflows
• Added more metadata to the image dialog info flyout
• Added Output Sharing toggle in Advanced Options during install flow

Changed

• Revamped the Packages page to enable running multiple packages at the same time
• Changed the Outputs Page to use a TreeView for the directory selection instead of a dropdown selector
• Model download location selector now searches all subfolders
• Inference Primary Sampler Addons (i.e. ControlNet, FreeU) are now inherited by Hires Fix Samplers, this can be overriden from the Hires Fix module's settings menu by disabling the "Inherit Primary Sampler Addons" option.
• Revisited the way images are loaded on the outputs page, with improvements to loading speed & not freezing the UI while loading
• Updated translations for French, Spanish, and Turkish
• Changed to a new image control for pages with many images
• (Internal) Updated to Avalonia 11.0.10

Fixed

• Fixed #559 - "Unable to load bitmap from provided data" error in Checkpoints page
• Fixed #522 - Incorrect output directory path for latest Auto1111
• Fixed #529 - OneTrainer requesting input during update
• Fixed Civitai model browser error when sorting by Installed with more than 100 installed models
• Fixed CLIP Install errors due to setuptools distutils conflict, added default environment variable setting SETUPTOOLS_USE_DISTUTILS=stdlib
• Fixed progress bars not displaying properly during package installs & updates
• Fixed ComfyUI extension updates not running install.py / updating requirements.txt
• Improved performance when deleting many images from the Outputs page
• Fixed ComfyUI torch downgrading to 2.1.2 when updating
• Fixed Inference HiresFix module "Inherit Primary Sampler Addons" setting not effectively disabling when unchecked
• Fixed model download location options for VAEs in the CivitAI Model Browser

Removed

• Removed the main Launch page, as it is no longer needed with the new Packages page

Changed

• Removed Symlink option for InvokeAI to prevent InvokeAI from moving models into its own directories (will be replaced with a Config option in a future update)

Fixed

• Fixed images not appearing in Civitai Model Browser when "Show NSFW" was disabled
• Fixed #556 - "Could not find entry point for InvokeAI" error

Changed

• Due to changes with the CivitAI API, you can no longer select a specific page in the CivitAI Model Browser
• Due to the above API changes, new pages are now loaded via "infinite scrolling"

Fixed

• Fixed models not being removed from the installed models cache when deleting them from the Checkpoints page
• Fixed model download location options for VAEs in the CivitAI Model Browser
• Fixed One-Click install progress dialog not disappearing after completion
• Fixed ComfyUI with Inference pop-up during one-click install appearing below the visible scroll area
• Fixed no packages being available for one-click install on PCs without a GPU
• Fixed missing ratings on some models in the CivitAI Model Browser
• Fixed missing favorite count in the CivitAI Model Browser
• Fixed recommended models not showing all SDXL models

Added

• Fixed #498 Added "Pony" category to CivitAI Model Browser

Changed

• Changed package deletion warning dialog to require additional confirmation

Fixed

• Fixed #502 - missing launch options for Forge
• Fixed #500 - missing output images in Forge when using output sharing
• Fixed #490 - mpmath has no attribute 'rational' error on macOS
• Fixed #510 - kohya_ss packages with v23.0.x failing to install due to missing 'packaging' dependency
• Fixed incorrect progress text when deleting a checkpoint from the Checkpoints page
• Fixed incorrect icon colors on macOS

Added

• Added new package: StableSwarmUI by Stability AI
• Added new package: Stable Diffusion WebUI Forge by lllyasviel
• Added extension management for SD.Next and Stable Diffusion WebUI-UX
• Added the ability to choose where CivitAI model downloads are saved
• Added --launch-package argument to launch a specific package on startup, using display name or package ID (i.e. --launch-package "Stable Diffusion WebUI Forge" or --launch-package c0b3ecc5-9664-4be9-952d-a10b3dcaee14)
• Added more Base Model search options to the CivitAI Model Browser
• Added Stable Cascade to the HuggingFace Model Browser

Inference

• Added Inference Prompt Styles, with Prompt Expansion model support (i.e. Fooocus V2)
• Added option to load a .yaml config file next to the model with the same name. Can be used with VPred and other models that require a config file.
• Added copy image support on linux and macOS for Inference outputs viewer menu

Changed

• Updated translations for German, Spanish, French, Japanese, Portuguese, and Turkish
• (Internal) Updated to Avalonia 11.0.9

Fixed

• Fixed StableSwarmUI not installing properly on macOS
• Fixed #464 - error when installing InvokeAI on macOS
• Fixed #335 Update hanging indefinitely after git step for Auto1111 and SDWebUI Forge
• Fixed Inference output viewer menu "Copy" not copying image
• Fixed image viewer dialog arrow key navigation not working
• Fixed CivitAI login prompt not showing when downloading models that require CivitAI logins
• Fixed unknown model types not showing on checkpoints page (thanks Jerry!)
• Improved error handling for Inference Select Image hash calculation in case file is being written to while being read

Fixed

• Fixed user tokens read error causing failed downloads
• Failed downloads will now log error messages
• Fixed #458 - Save Intermediate Image not working
• Fixed #453 - Update Fooocus --output-directory argument to --output-path

v2.8.2

Added

• Added missing GFPGAN link to Automatic1111 packages

Fixed

• Fixed Inference Image to Image Denoise setting becoming hidden after changing schedulers
• Fixed Inference ControlNet models showing as downloadable even when they are already installed
• Fixed Inference Sampler Addon conditioning not applying (i.e. ControlNet)
• Fixed extension modification dialog not showing any progress messages

Fixed

• Fixed model links not working in RuinedFooocus for new installations
• Fixed incorrect nodejs download link on Linux (thanks to slogonomo for the fix)
• Fixed failing InvokeAI install on macOS due to missing nodejs
• Increased timeout on Recommended Models call to prevent potential timeout errors on slow connections
• Fixed SynchronizationLockException when saving settings
• Improved error messages with process output for 7z extraction errors
• Fixed missing tkinter dependency for OneTrainer on Windows
• Fixed auto-update on macOS not starting new version from an issue in starting .app bundles with arguments
• Fixed #436 - Crash on invalid json files during checkpoint indexing

v2.8.0 has arrived with many new features 🎉

macOS Support (Apple Silicon)

Inference - Image to Video

https://github.com/LykosAI/StabilityMatrix/assets/13956642/37303e2d-858a-49ed-8616-53b805ff8829

Inference - Enhanced model selection with metadata and thumbnails

Package Extensions Manager (ComfyUI, Automatic 1111)

Configurable native Desktop notifications

Added

• Added Image to Video project type
• Added CLIP Skip setting to inference, toggleable from the model settings button
• Added image and model details in model selection boxes
• Added new package: OneTrainer
• Added native desktop push notifications for some events (i.e. Downloads, Package installs, Inference generation)
  • Currently available on Windows and Linux, macOS support is pending
• Added Package Extensions (Plugins) management - accessible from the Packages' 3-dot menu. Currently supports ComfyUI and Automatic1111.
• Added new launch argument options for Fooocus
• Added "Config" Shared Model Folder option for Fooocus
• Added Recommended Models dialog after one-click installer
• Added "Copy Details" button to Unexpected Error dialog
• Added German language option, thanks to Mario da Graca for the translation
• Added Portuguese language options, thanks to nextosai for the translation
• Added base model filter to Checkpoints page
• Added "Compatible Images" category when selecting images for Inference projects
• Added "Find in Model Browser" option to the right-click menu on the Checkpoints page
• Added --use-directml launch argument for SDWebUI DirectML fork
• Added release builds for macOS (Apple Silicon)
• Added ComfyUI launch argument configs: Cross Attention Method, Force Floating Point Precision, VAE Precision
• Added Delete button to the CivitAI Model Browser details dialog
• Added "Copy Link to Clipboard" for connected models in the Checkpoints page
• Added support for webp files to the Output Browser
• Added "Send to Image to Image" and "Send to Image to Video" options to the context menu

Changed

• New package installation flow
• Changed one-click installer to match the new package installation style
• Automatic1111 packages will now use PyTorch v2.1.2. Upgrade will occur during the next package update or upon fresh installation.
• Search box on Checkpoints page now searches tags and trigger words
• Changed the Close button on the package install dialog to "Hide"
  • Functionality remains the same, just a name change
• Updated translations for the following languages:
  • Spanish
  • French
  • Japanese
  • Turkish
• Inference file name patterns with directory separator characters will now have the subdirectories created automatically
• Changed how settings file is written to disk to reduce potential data loss risk
• (Internal) Updated to Avalonia 11.0.7

Fixed

• Fixed error when ControlNet module image paths are not found, even if the module is disabled
• Fixed error when finding metadata for archived models
• Fixed error when extensions folder is missing
• Fixed crash when model was not selected in Inference
• Fixed Fooocus Config shared folder mode overwriting unknown config keys
• Fixed potential SD.Next update issues by moving to shared update process
• Fixed crash on startup when Outputs page failed to load categories properly
• Fixed image gallery arrow key navigation requiring clicking before responding
• Fixed crash when loading extensions list with no internet connection
• Fixed crash when invalid launch arguments are passed
• Fixed missing up/downgrade buttons on the Python Packages dialog when the version was not semver compatible

Fixed

• Fixed InvokeAI v3.6.0 "detail": "Not Found" error when opening the UI

Changed

• Python Packages install dialog now allows entering multiple arguments or option flags

Fixed

• Fixed InvokeAI Package dependency versions (#395)

Added

• Added --use-directml launch argument for SDWebUI DirectML fork

Changed

• Model Browser downloads will no longer be disabled if the free drive space is unavailable
• Default Linux installation folder changed to prevent issues with hidden folders
• Changed default Period to "AllTime" in the Model Browser

Fixed

• Fixed error where Environment Variables were not editable
• Fixed SDTurboScheduler's missing denoise parameter

Added

• Added SDXL Turbo and Stable Video Diffusion to the Hugging Face tab

Changed

• ControlNet model selector will now show the parent directory of a model when relevant

Fixed

• Fixed Python Packages dialog crash due to pip commands including warnings
• Fixed Base Model downloads from the Hugging Face tab downloading to the wrong folder
• Fixed InvokeAI ! [rejected] v3.4.0post2 -> v3.4.0post2 (would clobber existing tag) error on updating to the latest version
• Fixed settings not saving in some scenarios, such as when the settings.json file existed but was empty

Fixed

• Fixed Python Packages manager crash when pip list returns warnings in json
• Fixed slowdown when loading PNGs with large amounts of metadata
• Fixed crash when scanning directories for missing metadata

Changed

• Improved low disk space handling

Fixed

• Fixed denoise strength in Inference Text to Image
• Fixed PathTooLongException for IPAdapter folders when using ComfyUI in Symlink mode
• Fixed configs and symlinks not being cleaned up when switched to the opposite mode
• Fixed model indexing stopping when encountering paths longer than 1021 bytes in length
• Fixed repeated nested folders being created in Models/ControlNet when using ComfyUI in Symlink mode. Existing folders will be repaired to their original structure on launch.

Added

• Added missing IPAdapter and CLIP Vision folder links for ComfyUI

Fixed

• Fixed UnicodeDecodeError when using extra_model_paths.yaml in ComfyUI on certain locales
• Fixed SDXL CLIP Vision model directory name conflict
• Fixed #334 - Win32Exception if Settings are opened

Changed

• Changed Symlink shared folder link targets for Automatic1111 and ComfyUI. From ControlNet -> models/controlnet to ControlNet -> models/controlnet/ControlNet and T2IAdapter -> models/controlnet/T2IAdapter.
• Changed FreeU defaults to match recommended SD1.5 defaults
• Changed default denoise strength from 1.0 to 0.7

Fixed

• Fixed ControlNet / T2IAdapter shared folder links for Automatic1111 conflicting with each other
• Fixed URIScheme registration errors on Linux
• Fixed RuinedFooocus missing output folder on startup
• Fixed incorrect Fooocus VRAM launch arguments

Added

• Added Turkish UI language option, thanks to Progesor for the translation

Fixed

• Fixed Inference Image to Image projects missing denoise strength setting

Added

General

• New package: RuinedFooocus
• Added an X button to all search fields to instantly clear them (Esc key also works)
• Added System Information section to Settings

Inference

• Added Image to Image project type
• Added Modular custom steps
  • Use the plus button to add new steps (Hires Fix, Upscaler, and Save Image are currently available), and the edit button to enable removing or dragging steps to reorder them. This enables multi-pass Hires Fix, mixing different upscalers, and saving intermediate images at any point in the pipeline.
• Added Sampler addons
  • Addons usually affect guidance like ControlNet, T2I, FreeU, and other addons to come. They apply to the individual sampler, so you can mix and match different ControlNets for Base and Hires Fix, or use the current output from a previous sampler as ControlNet guidance image for HighRes passes.
• Added SD Turbo Scheduler
• Added display names for new samplers ("Heun++ 2", "DDPM", "LCM")
• Added Ctrl+Enter as a shortcut for the Generate Image button

Accounts Settings Subpage

• Lykos Account sign-up and login - currently for Patreon OAuth connections but GitHub requests caching and settings sync are planned
• Supporters can now connect your Patreon accounts, then head to the Updates page to choose to receive auto-updates from the Dev or Preview channels
• CivitAI Account login with API key - enables downloading models from the Browser page that require CivitAI logins, more integrations like liking and commenting are also planned

Updates Settings Subpage

• Toggle auto-update notifications and manually check for updates
• Choose between Stable, Preview, and Dev update channels

Inference Settings Subpage

• Moved Inference settings to subpage
• Updated with more localized labels

Outputs Page

• Added Refresh button to update gallery from file system changes

Checkpoints Page

• Added the ability to drag & drop checkpoints between different folders
• Added "Copy Trigger Words" option to the three-dots menu on the Checkpoints page (when data is available)
• Added trigger words on checkpoint card and tooltip
• Added "Find Connected Metadata" options for root-level and file-level scans
• Added "Update Existing Metadata" button

Model Browser

• Added Hugging Face tab to the Model Browser
• Added additional base model filter options for CivitAI ("SD 1.5 LCM", "SDXL 1.0 LCM", "SDXL Turbo", "Other")
• Added the ability to type in a specific page number in the CivitAI Model Browser
• Right clicking anywhere on the model card will open the same menu as the three-dots button
• New model downloads will save trigger words in metadata, if available
• Model author username and avatar display, with clickable link to their profile

Changed

General

• Model Browser page has been redesigned, featuring more information like rating and download counts
• Model Browser navigation has improved animations on hover and compact number formatting
• Updated Outputs Page button and menu layout
• Rearranged Add Package dialog slightly to accommodate longer package list
• Folder-level "Find Connected Metadata" now scans the selected folder and its subfolders
• Model Browser now split into "CivitAI" and "Hugging Face" tabs

Inference

• Selected images (i.e. Image2Image, Upscale, ControlNet) will now save their source paths saved and restored on load. If the image is moved or deleted, the selection will show as missing and can be reselected
• Project files (.smproj) have been updated to v3, existing projects will be upgraded on load and will no longer be compatible with older versions of Stability Matrix

Fixed

• Fixed Outputs page reverting back to Shared Output Folder every time the page is reloaded
• Potentially fixed updates sometimes clearing settings or launching in the wrong directory
• Improved startup time and window load time after exiting dialogs
• Fixed control character decoding that caused some progress bars to show as \u2588
• Fixed Python rich package's progress bars not showing in console
• Optimized ProgressRing animation bindings to reduce CPU usage
• Improved safety checks in custom control rendering to reduce potential graphical artifacts
• Improved console rendering safety with cursor line increment clamping, as potential fix for #111
• Fixed #290 - Model browser crash due to text trimming certain unicode characters
• Fixed crash when loading an empty settings file
• Improve Settings save and load performance with .NET 8 Source Generating Serialization
• Fixed ApplicationException during database shutdown
• InvokeAI model links for T2I/IpAdapters now point to the correct folders
• Added extra checks to help prevent settings resetting in certain scenarios
• Fixed Refiner model enabled state not saving to Inference project files
• Fixed NullReference error labels when clearing the Inference batch size settings, now shows improved message with minimum and maximum value constraints