Gal Metzer

Excited to share our work accepted to #SIGGRAPH2026 ! Video generation models struggle with something few talk about: their transformations don't evolve smoothly. You get long boring stretches... then a sudden semantic jump where everything "catches up" at once.

🔊 Audio on for this one! Code + checkpoint for ScenA, our multi-speaker audio scene generation model, are out! Reference voices + a prompt → a full conversational scene, prompt alone decides who speaks where. github.com/finmickey/scena arxiv.org/abs/2606.19325

ID-LoRA was accepted to ECCV 2026! @eccvconf Check out our project page if you're interested in how we generate videos with custom reference voice and visual identity @AviadDahan @NKraicer @RGiryes @liorwolf id-lora.github.io

Generating images with AI is fun 🤓 Finding the image? Not so much… The first result is rarely the one. Or the second. Or the third. 🥵 So we built a way to generate diverse possibilities along meaningful semantic axes 🧭 and organize them into a gallery you can effortlessly browse 🧚🏻 Introducing our #ECCV2026 paper: Semantic Browsing: Controllable Diversity for Image Generation ✨ 1/5

FUN to share that Sync-LoRA has been accepted to #ECCV2026! 🤗✨🎬 the field moves fast, but the core seems to remain: 1️⃣ IC-LoRA enables diverse edits 2️⃣ high-quality data is key (even if it's a small amount!) Huge thanks to @arash_mham @OPatashnik @DanielCohenOr1

Very excited to be presenting our highlight paper “Deep Feature Deformation Weights” at #CVPR2026! TLDR: Image features make surprisingly good deformation weights. Our feature distillation is the best in the biz. ExHall A&F 555 Poster Session 4 Website: threedle.github.io/dfd/

Playing around some more with Aleph 2.0. All edits here are guided from a single anchor editing frame, even across longer multi-shot videos. Excited about the jump to 1080p and more precise, localized editing. Can’t wait to share more results soon!

Even today, with powerful image editing models, making fine-grained structural changes to 3D shapes remains a major challenge. In our new #SIGGRAPH2026 paper, Prox-E, we use primitive-based abstraction to leverage VLMs for precise, reasoning-based 3D editing! 👇

Paper: arxiv.org/abs/2604.22554 Project: sagipolaczek.github.io/semantic-progr… Code: github.com/SagiPolaczek/s… HF: huggingface.co/papers/2604.22… w/ @PolaczekSagi, @arash_mham, @RGiryes & @DanielCohenOr1

Linearization is just one option. The SPF lets you target *any* pacing function. Exponential acceleration? Deceleration? Slow-motion semantic buildup? You control the semantic velocity of the output.

Excited to share our work accepted to #SIGGRAPH2026 ! Video generation models struggle with something few talk about: their transformations don't evolve smoothly. You get long boring stretches... then a sudden semantic jump where everything "catches up" at once.

1/7 When rewards conflict, what should RL post-training of diffusion models optimize? In visual generation, objectives are often in tension: Prompt adherence can conflict with source preservation. Photorealism can conflict with stylization. In our new paper, ParetoSlider, we introduce a multi-objective RL framework that trains a single diffusion model for continuous control over competing reward objectives 🧵

Video models as Physics simulators. 🌍🎥 [1/] In our latest work, WinDiNet, we finetuned a pre-trained video model into a differentiable physics engine. 1000x faster than traditional CFD solvers. Project page: rbischof.github.io/windinet_web/ Abs: arxiv.org/abs/2603.21210

Creative work often starts before we can describe what we're looking for. What role can generative models play at this stage? 🌱Our new work, Inspiration Seeds, reveals hidden visual connections between images, creating a purely visual exploration space. 🔗kfirgoldberg.github.io/InspirationSee…

presenting: Strait-of-Hormuz - the tower defense game I took Claude Code for a spin and vibe-coded this game during Passover and between the rocket sirens. your goal is to get the oil tankers OUT of the strait, while IRGC tries to STOP you. try it here: strait-of-hormuz.io

Many styles are not just textures or colors — they reshape geometry. Abstraction in Style (AiS) introduces an abstraction stage before stylization, enabling extreme abstraction-driven styles previously out of reach. arxiv.org/pdf/2603.29924

Modern T2I DiTs are incredibly powerful, but have a serious diversity problem. We introduce a surprisingly simple and efficient inference-time fix (+2s for Flux-dev, +1s for SD3.5-Turbo). Excited to share our SIGGRAPH 2026 (conditional) paper: “On-the-fly Repulsion in the Contextual Space for Rich Diversity in Diffusion Transformers”