🎙 Apache-2.0 🔥 OmniAvatar: Efficient Audio-Driven Avatar Video Generation with Adaptive Body Animation Thanks to Qijun Gan ❤ Ruizi Yang ❤ Jianke Zhu ❤ Shaofei Xue ❤ Steven Hoi ❤ 🌐page: omni-avatar.github.io 🧬code: github.com/Omni-Avatar/Om… 📄paper:

一张照片+一段音频就能做出全身能动的 AI 虚拟人。 浙江大学和阿里巴巴夸克技术团队合作的「OmniAvatar」模型，感觉很强啊！

SOTA in GUI grounding is here, and it's open source! 🔥

传说中的6B文生图开源模型Z-Image已经发布了，原来是阿里的。作为一个可以在消费级显卡运行的小尺寸模型，功能强大的有点超乎想象，测试下来非常出色，感觉像小号的🍌或seedream。 📷照片级真实感 📒精准的双语文本渲染 💡提示增强和推理 🎨指令图像编辑 modelscope.cn/models/Tongyi-…

Z-Image, a new image generation / editing model from Alibaba. - Ultra fast generation sub-second on H800 - Fits within 16G consumer device - Rich high frequency details - Base (non-distilled version) is also available - 6B, apache 2 license - Demo available on Hugging Face

Introducing Z-Image: a powerful, highly efficient 6B-parameter image generation model with three specialized variants! 🎨 🚀 Z-Image-Turbo Distilled for speed & quality Just 8 NFEs — matches or beats top competitors ⚡️Sub-second latency on H800 GPUs Runs smoothly on 16GB VRAM

阿里巴巴发布全新图像模型：Z-Image 看来下效果真的很不错啊 1/3 的参数 达到了与国际商业模型几乎相同的视觉质量 模型参数：6B 显存占用：16GB 以下即可运行 生成速度：8步采样即可出高清图 支持设备：消费级显卡即可流畅运行 能理解复杂指令并且会推理、会思考

1/ 10 We are pleased to introduce Z-Image, an efficient 6-billion-parameter foundation model for image generation. Through systematic optimization, it proves that top-tier performance is achievable without relying on enormous model sizes, delivering strong results in

🔥 Z Image Turbo is now available on 🔥 Free Google Colab (T4) 🔥 Thanks to Tongyi Lab ❤ 🍨 Decoupled DMD: CFG Augmentation as The Spear, Distribution Matching as The Shield 🔥 Jupyter Notebook 🥳 Thanks to Dongyang Liu ❤ Peng Gao ❤ David Liu ❤ Ruoyi Du ❤ Zhen Li ❤

Open source models are crazy. Alibaba just dropped a new image model called Z-Image. It's super fast, only 6B parameters and excels at producing images with photography-level realism. Here’s what it can do:

阿里发布了一个新图像模型 Z-Image 支持图像编辑，只有 6B 大小，非常适合本地跑 支持最高2K图像的生成，以及中文的文本渲染能力也不错，也有一定的世界知识 感觉非常适合本地部署一个来玩啊，这是拿刚才的提示词抛的Q版天气海报。

Z-Image Turbo in ComfyUI: Realism at Lightning Speed - Highly efficient 6B image model (distilled) - Production-level photorealistic 6 seconds / native 2K output on Comfy Cloud: links.comfy.org/zimgtw Examples 🧵

If you’re running Z-Image-Turbo, you need to slot this into your workflow. renderartist just dropped "Technically Color Z" for Z-Image-Turbo, and the saturation is gorgeous. It nails that lush, dramatic lighting from classic cinema without frying the image.

A week later, Alibaba's cloud service team has released Z-image Turbo Fun Controlnet Union again!

THANK YOU to the community for your support🙏 We’re so glad that you love Z-image 💕🎉