v6.13.0.rc2
📦 invokeaiView on GitHub →
✨ 18 features🐛 1 fixes🔧 5 symbols
Summary
InvokeAI v6.13.0 RC2 introduces extensive support for new locally-hosted models (Qwen, Anima) and externally-hosted providers (OpenAI, Gemini). This release also brings significant improvements to the Model Manager, Canvas, Workflow Editor, and adds LLM-powered prompt manipulation features.
Migration Steps
- To use LLM-powered prompt expansion, install a causal LM model (e.g., Qwen/Qwen2.5-1.5B-Instruct) via the Model Manager (HuggingFace tab).
- To use custom workflows on Canvas images, create a workflow with an image input field in the Form Builder and a Canvas Output node, then right-click the canvas and select 'Run Workflow'.
- To use the workflow connector, right-click on the canvas and select 'Add Connector'.
✨ New Features
- Added support for remotely hosted models from external providers like OpenAI and Google Gemini.
- Added support for locally-hosted Qwen Image (txt2img and edit versions) and Anima models.
- Introduced Qwen Image Lightning and Qwen Image Edit Lightning LoRAs for fast rendering.
- Enabled image editing with Qwen Image Edit 2.5 using up to three reference images and declarative statements.
- Improved Model Manager with rewritten installation queue display, progress bars, bulk reidentify action, and multi-way sorting.
- Added LLM-powered prompt expansion (using local TextLLM) and Image to Prompt features accessible via new buttons in the prompt textfield.
- Added automatic creation of 'virtual' image boards that group images by characteristics (initially by date).
- Canvas improvements: ability to hide/show preview stage thumbnails and addition of a Lasso Tool (freehand and polygon modes) for masking/selection.
- Enabled applying custom workflows directly to a Canvas image via right-click menu.
- Introduced a Custom Node Manager tab for installing and managing node packs (groups of related nodes).
- Nodes are now organized by function in the node picker.
- Added the workflow connector, a new edge type for flexible connection of multiple input/output fields.
- New generation nodes for running external models.
- New denoising nodes for Qwen Image and Anima local models.
- Added a node for decoding invisible watermarks in images.
- The Collect node now supports chaining for deterministic iteration.
- Added an 'If' mode to workflows for conditional branching.
- In Multiuser Mode, users can create private, shared, and public image boards, and private/shared workflows.
🐛 Bug Fixes
- Corrected many small bugs and performance bottlenecks in image generation.