The crew at Xiaohongshu has launched FireRed-Picture-Edit 1.1, an upgraded image-editing mannequin designed to enhance id consistency, multi-element composition, portrait magnificence enhancing, and font-style referencing.
In line with the corporate, the mannequin introduces what it describes as state-of-the-art id preservation expertise within the open-source ecosystem, enabling edited pictures to take care of clear recognition of the unique topic even after advanced modifications.
The mannequin can mix greater than 10 totally different visible components inside a single picture and contains an agent-based clever cropping and compositing function that mechanically arranges visible elements.
For portrait enhancing, FireRed-Picture-Edit 1.1 supplies dozens of make-up and sweetness types, permitting customers to carry out professional-grade retouching and create personalized appears to be like resembling brightening yellow or olive pores and skin tones.
The system additionally options an open LoRA coaching ecosystem, enabling creators to develop and apply customized visible types. As well as, the mannequin contains an optimized sampler designed to maximise GPU effectivity when dealing with equivalent duties, picture sizes, and enter volumes.
With 30GB of GPU reminiscence, FireRed-Picture-Edit 1.1 can generate edited pictures end-to-end in about 4.5 seconds. The mannequin helps cross-platform deployment, provides native node compatibility with ComfyUI, and helps the GGUF light-weight format, serving to speed up coaching and inference workflows.
Supply: IT House
Elevate your perspective with NextTech Information, the place innovation meets perception.
Uncover the most recent breakthroughs, get unique updates, and join with a world community of future-focused thinkers.
Unlock tomorrow’s tendencies at this time: learn extra, subscribe to our e-newsletter, and develop into a part of the NextTech neighborhood at NextTech-news.com
