Release: https://github.com/SlavomirDurej/layershop/releases/tag/layershop-v0.20.1-preview.1

Release: https://github.com/Comfy-Org/ComfyUI/releases/tag/v0.44.0

Abstract

Instruction-driven image editing has reached native 4K, yet how such editors are trained remains undisclosed, and pretraining one from scratch is beyond most teams. The task itself compounds the difficulty: a single end-to-end editor cannot make precise parametric adjustments, match specialized experts, or interpret the vague instructions users actually write. We argue that 4K editing is a system problem rather than a single-model problem, and present BluePixel, which couples a native-4K editing model with an agent harness; the complete training recipe is disclosed in this paper. The model makes 4K training practical: it turns a 1K-pretrained editor into a native-4K one through progressive VAE adaptation that keeps 4K training as affordable as 1K, absorbs a long tail of vertical editing skills through continual learning, and refines the aesthetic-fidelity trade-off with reinforcement learning. Distillation via source-anchored target projection (SATP) then compresses 50 denoising steps to 4, completing a 4K edit in 8.7 s on a single Kunlunxin P800 XPU while preserving task success rate and background consistency. The harness, BluePixel Agent, makes editing controllable: it decomposes user instructions into specialist and general operations, verifies every result with the same reward model that guided training, and retries within bounded attempts. On VINS-4KEval++ (3,305 native-4K cases), BluePixel ranks second among seven models, behind only GPT Image 2 in non-edited-region consistency.

Paper: https://arxiv.org/abs/2610.19882

Code: https://github.com/vivo/BluePixel

Model: https://huggingface.co/openvivo/BluePixel-Turbo

Project Page: https://tech.vivo.com/BluePixel/

Release: https://github.com/KadoBOT/ComfyUI-H3-Continuity-Studio/releases/tag/v0.1.0

Abstract

The hair of stylized characters in games and avatars is built from hair cards: artist-placed ribbons whose geometry is the visible hair, animated by spring bones. Existing image-based methods produce realistic hair or a single fused mesh, while CHARM, the only generator of stylized hair cards, decodes a hairstyle autoregressively as a long sequence of control points. Observing that a card hairstyle is an unordered set of artist-defined primitives, we propose HairDeck, which generates it as one: each card is a single token, and a set diffusion transformer denoises all cards together. From a single image, HairDeck generates the cards and the spring-bone chains that drive them, yielding an animatable asset that exports to ready-to-use downstream formats. On CHARM's own test set, it lowers Chamfer distance by 44% and card-orientation error by more than half while running 33 times faster, and from single images it outperforms CHARM, image-to-3D pipelines and strand-based methods. Project page: this https URL

Paper: https://arxiv.org/abs/2610.40599

Code: https://github.com/hyz317/HairDeck

models: https://huggingface.co/hyz317/HairDeck

Project Page: https://hyzcluster.github.io/HairDeck/

Abstract

Existing sketch generation methods often rely on image guidance and are largely limited to individual objects or simplified scenes, leaving text-guided generation of multi-concept scene sketches underexplored. We present SketchDirector, an agentic framework that mimics the human sketching workflow for automatic multi-concept scene sketching. Given a scene-level text prompt, SketchDirector first employs a large language model (LLM) to decompose the scene into multiple draft assets, where each is represented by a set of Bézier curves. Then, an agentic reasoning loop iteratively plans and executes their spatial arrangement, guided by self-evolving memory. Finally, a detail refinement module leverages diffusion priors to enhance sketch quality while preserving geometric structure. To support future research and applications in multi-concept scene sketch generation, we further introduce the MCSketch8K dataset, comprising editable vector assets, compositional layouts, and object-level manipulation annotations. Extensive experiments show that SketchDirector outperforms state-of-the-art baselines in semantic and compositional quality while effectively preserving overall contour organization during refinement. The project page is: https://yiki77.github.io/SketchDirector

Paper: https://yiki77.github.io/SketchDirector/_AAAI_27__SketchDirector.pdf

Code: https://github.com/GVCLab/SketchDirector/tree/main

Project Page: https://yiki77.github.io/SketchDirector/

SD.Next Release 2026-10-29

What's new in this release?

• New base model: Bria FIBO 2.0 and improvements to MOSS-Image and Ming-Image workflows
• Major improvements to Prompt Enhance functionality to make it more versatile and cover usage in new structured/json based prompts
• New playground feature Story Mode
• Expanded native LoRA support with compatibility for multiple model types
• Updates to AI Coding instructions and skills: try your favorite model/tool on SD.Next, you might be surprised how quickly it adapts and performs
• Codebase cleanup: removed deprecated options that were left as legacy remnants
• And many more smaller improvements and fixes...

[!IMPORTANT] Python 3.10 is end-of-life as of October 2026!

See changelog below for full details...

Home | ChangeLog | Docs | Discord | Sponsor

Abstract

Generating images beyond the training resolution of Diffusion Transformers (DiTs) remains challenging, as the number of tokens grows rapidly with resolution, full attention becomes expensive, and outputs often suffer from structural distortions and degraded detail. We identify uniform full attention as the common source of both problems. Attending to all tokens at every step causes quadratic cost and exposes the model to more tokens and relative distances beyond the training range, pushing the model out of distribution. We observe that attention's effective receptive field narrows during denoising, from global in early steps that form structure to local in later steps that refine details. We propose AReA, a training-free attention mechanism whose receptive field evolves from coarse to fine, following the denoising progression. In early steps, AReA applies cross-scale coarse attention over downsampled keys and values, capturing global context, while in later steps it applies local attention within progressively shrinking windows. Without additional parameters or training, experiments show that AReA consistently outperforms state-of-the-art training-free baselines across target resolutions and different models, while also being faster and more efficient.

Paper: https://arxiv.org/abs/2610.17227

Code: https://github.com/KooroshRH/AReA

Project Page: https://rajabi2001.github.io/area/

• Add fibo 2.0 edit turbo openvino support by @rupeshs in #391

• Add edit image prompt presets

• Fixed garbage images from FLUX.3 Image Klein(OpenVlNO) on iGPU devices

For more details : https://github.com/rupeshs/fastsdcpu#fibo2-openvino

Full Changelog: v1.0.0-beta.610...v1.0.0-beta.620