Computer Vision

Black Forest Labs has launched FLUX.2, its second technology picture technology and enhancing system. FLUX.2 targets actual world artistic workflows equivalent to advertising and marketing belongings, product images, design layouts, and sophisticated infographics, with enhancing assist as much as 4 megapixels and powerful management over format, logos, and typography. FLUX.2 product household and FLUX.2 [dev]…

Even sturdy ‘long-context’ AI fashions fail badly once they should monitor objects and counts over lengthy, messy video streams, so the subsequent aggressive edge will come from fashions that predict what comes subsequent and selectively keep in mind solely shocking, necessary occasions, not from simply shopping for extra compute and greater context home windows. A…

In this tutorial, we discover superior laptop imaginative and prescient strategies utilizing TorchVision’s v2 transforms, fashionable augmentation methods, and highly effective coaching enhancements. We stroll by means of the method of constructing an augmentation pipeline, making use of MixUp and CutMix, designing a contemporary CNN with consideration, and implementing a sturdy coaching loop. By operating…

A staff of researchers from Meta Reality Labs and Carnegie Mellon University has launched MapAnything, an end-to-end transformer structure that straight regresses factored metric 3D scene geometry from photographs and elective sensor inputs. Released beneath Apache 2.0 with full coaching and benchmarking code, MapAnything advances past specialist pipelines by supporting over 12 distinct 3D imaginative…

In the domain of multimodal AI, instruction-based image editing models are transforming how users interact with visual content. Just released in August 2025 by Alibaba’s Qwen Team, Qwen-Image-Edit builds on the 20B-parameter Qwen-Image foundation to deliver advanced editing capabilities. This model excels in semantic editing (e.g., style transfer and novel view synthesis) and appearance editing…