v2.14.0 Preview 2

Howdy Patrons!

We've got another exciting update for you with v2.14.0 Preview 2, packed full of fantastic new features and important fixes!

We're thrilled to introduce Accelerated Model Discovery (Beta), exclusively for our Insider+ supporters. To enable it, click on the ⚡ icon in the Civitai Browser and click Enable. This new feature supercharges your browsing experience with dramatically faster and smoother access to online model repositories. Additionally, the new Prompt Amplifier is here to enrich your creative process! Just click the magic wand 🪄 in the prompt editor, pick your desired vibe with the 'Feel' selector, and see your prompts burst into vivid, detailed generations.

We've also added the popular Stable Diffusion WebUI Forge - Classic package, Undo/Redo options to text editor context menus, and support for HiDream in Inference. Plus, we've streamlined the UI by merging Inference GGUF workflows directly into the UNet loader and introduced toggles for filtering the extra network selectors.

On the backend, we've optimized performance for Checkpoint Manager and squashed several pesky bugs, including crashes related to model discovery and image caching issues.

Thank you all for your amazing support—stay tuned for more awesomeness coming your way!🌟

Added

• Added new package - Stable Diffusion WebUI Forge - Classic

• Added Accelerated Model Discovery (Beta) (⚡ icon in Civitai Browser) for Insider+ supporters. Utilizes an optimized connection for dramatically faster, more responsive browsing of online model repositories.

• Added Undo/Redo commands to text editor context menus

• Added Prompt Amplifier to Inference - click the magic wand 🪄 in the prompt editor to expand and enrich your ideas. Tailor the vibe with the ‘Feel’ selector and watch as your generations come to life with extra detail, coherence, and flair!

• Added support for HiDream in Inference - see ComfyUI Examples for more details

• Added toggle for filtering Inference Extra Networks by base model

Changed

• Updated install for kohya_ss to support RTX 5000-series GPUs

• Merged Inference GGUF workflows into the UNet model loader option (no longer need to choose GGUF separately)

• Updated some date strings to take into account the user's locale

• Fixed some crashes when using Accelerated Model Discovery

• Performance optimizations for Checkpoint Manager (progress indicators now fully uses Compiled Bindings)

Fixed

• Fixed Inference ControlNet Preprocessors using incorrect resolution and increased maximum of smallest dimension to 16384

• Fixed Triton/Sage install option showing for incompatible GPUs

• Fixed errors from invalid pip specifiers in requirements files

• Fixed package images sometimes showing as blank due to concurrent image caching. Requests to same image resources are now de-duplicated

• Fixed Inference Extra Networks card not allowing for more than one model at a time

• Reduced memory usage from ShowDisabledTooltipExtension

Special Thanks

Big shout-out to our Visionary-tier patrons: Waterclouds, Corey T, bluepopsicle, Bob S, and Akiro_Senkai! Your steadfast support keeps Stability Matrix moving forward, and we couldn’t do it without you. 🚀 Thank you!