camenduru

We've updated RconViaGen-v0.5 based on TRELLIS.2, supporting the generation of high-resolution meshes and PBR materials from multi-view images, and have also released the training code for ReconViaGen! Code：github.com/GAP-LAB-CUHK-S… Huggingface demo：huggingface.co/spaces/Stable-…

We're reintroducing and open-sourcing project "See-through". Given a single anime illustration, it automatically decomposes the character into fully-inpainted semantic layers with depth ordering. One image in, layered PSD out. (1/n) Repo: github.com/shitagaki-lab/…

🚨 We're very happy to introduce TRIBE v2: a foundation model of the brain's responses to sight, sound & language. 📄 Paper: ai.meta.com/research/publi… ▶️ Demo: aidemos.atmeta.com/tribev2/ 💻 Code: github.com/facebookresear… 🤗 Model: huggingface.co/facebook/tribe…

🪄 Introducing daVinci-MagiHuman: The Performance-Level Audio-Video Generative Foundation Model Proudly open-sourced and jointly developed by SII GAIR Lab & Sand.ai, it sets a new standard for multimodal AI. ⏳ 1/6

We have seen many works unlock the power of pretrained models for images and videos🏞️. But what about human motion🕺💃? Can we leverage a pretrained motion prior for a wide range of downstream tasks? Yes!! UMO is a simple yet effective framework that, for the first time, unlocks the priors of a motion foundation model (i.e., HY-Motion) for 10+ tasks, including editing, reaction generation, stylization, trajectory control, obstacle avoidance, keyframe infilling, and more. Amazing work! @xiaoyan_cong and @kunkun0w0. 🏠Webpage: oliver-cong02.github.io/UMO.github.io/ 📄 Paper: arxiv.org/abs/2603.15975 With the growing number of tools for transferring SMPL motion to humanoids, we hope it could also become a source of skills for humanoid robot learning. #Graphics #Motion #Animation #AIGC #GenerativeAI #Vision #3DV #Robotics #Robot #Humanoid #Learning #GenAI #Animation

Excited to share Track4World, feedforward 3D tracking of all pixels in the world-centric coordinate system. Code has been released, and welcome to try it! Homepage: jiah-cloud.github.io/Track4World.gi… Code: github.com/TencentARC/Tra… Paper: arxiv.org/abs/2603.02573

Ever tried inpainting an object into a scene with #AI, but details got lost? 🥴 Meet HiFi-Inpaint (#CVPR2026)! 🎉 High-fidelity detail preservation for reference-based inpainting — texts, logos, textures, all intact. No more blur in your ad images! 👇 correr-zhou.github.io/HiFi-Inpaint/

🚀 𝙈𝙖𝙩𝘼𝙣𝙮𝙤𝙣𝙚 2 is accepted to #CVPR2026! 🔥A stronger version of 𝙈𝙖𝙩𝘼𝙣𝙮𝙤𝙣𝙚 with finer details and enhanced robustness🔥 -🏡: pq-yang.github.io/projects/MatAn… -📜: arxiv.org/abs/2512.11782 -👩🏻‍💻: github.com/pq-yang/MatAny…

LTX-2.3 is a major upgrade. It’s a production-ready multimodal engine - designed to be built on. Here’s what’s new 🧵 1/7

🤯Real-time video generation just got HUGE. Introducing Helios: A 14B parameter model running at 19.5 FPS on a single H100. Cheaper, Faster, and Stronger than 1.3B models, generating minute-long videos without the usual tricks (even no quantization). github.com/PKU-YuanGroup/…

🚀 Introducing the Qwen 3.5 Small Model Series Qwen3.5-0.8B · Qwen3.5-2B · Qwen3.5-4B · Qwen3.5-9B ✨ More intelligence, less compute. These small models are built on the same Qwen3.5 foundation — native multimodal, improved architecture, scaled RL: • 0.8B / 2B → tiny, fast, great for edge device • 4B → a surprisingly strong multimodal base for lightweight agents • 9B → compact, but already closing the gap with much larger models And yes — we’re also releasing the Base models as well. We hope this better supports research, experimentation, and real-world industrial innovation. Hugging Face: huggingface.co/collections/Qw… ModelScope: modelscope.cn/collections/Qw…

I learned a lot while studying github.com/antirez/iris.c thanks to @antirez ❤ Here are my study notes written in pure C 🐣 github.com/camenduru/klei…

Bytedance just dropped a paper that might change how AI thinks. Literally. They figured out why LLMs fail at long reasoning — and framed it as chemistry. The discovery: Chain-of-thought isn't just words. It's molecular structure. Three bond types: • Deep reasoning = covalent bonds (strong, unbreakable) • Self-reflection = hydrogen bonds (flexible, context-aware) • Exploration = van der Waals (weak, ever-present) Why most AI "thinking" sucks: Everyone's been imitating keywords — "wait," "let me check" — without building the actual bonds. It's like copying the shape of a protein without the atomic forces holding it together. Bytedance proved: structure emerges from training, not prompting. The fix: Mole-Syn Their method doesn't just generate text. It synthesizes stable thought molecules. Results: better reasoning, more stable RL training. Bytedance is treating AI reasoning like organic chemistry — and it works. Paper: arxiv.org/abs/2601.06002

Introducing FireRed-Image-Edit-1.0 from FireRedTeam! 🚀 It’s officially the new SOTA for general image editing. ✅ Better than Closed-Source: Outperforms Nano-Banana & Seedream4.0 on GEdit benchmarks. ✅ Native Evolution: Built from T2I foundations, not just a "patch" on existing models. ✅ Style Mastery: Scored a record-breaking 4.97/5.0 in style transfer. ✅ High-Fidelity Text: Keeps original font styles perfectly. ✅ Virtual Try-on: Native support for multi-image joint editing. ✅ Bilingual: Native support for both English & Chinese prompts. Apache 2.0 license. Local deployment ready. 🤖 Model: modelscope.cn/models/FireRed… 🎠 Demo: modelscope.cn/studios/FireRe… 🛠️ Github: github.com/FireRedTeam/Fi…

Just shipped FastFlux2 Realtime Editor. A fully open-source real-time editing studio in your browser. Webcam → FLUX.2-klein-4B → Single 4090 @ 5 FPS, H100 @ 10+ FPS. Repo: github.com/Owen718/flux-s…

MiniMax-M2.5 is now open source. Trained with reinforcement learning across hundreds of thousands of complex real-world environments, it delivers SOTA performance in coding, agentic tool use, search, and office workflows. Hugging Face: huggingface.co/MiniMaxAI/Mini… GitHub: github.com/MiniMax-AI/Min… Coding Plan: platform.minimax.io/subscribe/codi… Intelligence with Everyone

🤦 @github @jaredpalmer 🚫 github.com/lucidrains 🆕 gitlab.com/lucidrains 🆕 codeberg.org/lucidrains

Excited to share our work: ArcFlow, a 2-Step Text-to-Image Generation Framework via High-Precision Non-Linear Flow Distillation. Code: github.com/pnotp/ArcFlow It ensures high-quality alignment with teacher, delivering 40× speedup and 4× faster convergence with <5% parameters.

Free training for Klein is live on ModelScope—supporting both image and edit models! 🚀modelscope.ai/civision/model…

@AnTh2107 yep reddit.com/r/comfyui/comm…

@camenduru Can i use my own Voice on Songs?

🎵 ACE-Step v1.5 ile Türkçe müzik yapabiliyormuşuzzz 😛 🪽 Video model: Imagine (480p 6s) 🧬 github.com/ace-step/ACE-S… 🎮 github.com/fspecii/ace-st… 🎮 acemusic.ai