由深圳的香港中文大学、ByteDance、清华大学的研究者等开发的AI生成3D模型「Hi3DGen: High-fidelity 3D Geometry Generation from Images via Normal Bridging」的演示已公开！该项目开源，代码也预计不久后会在Github上公开。

Hi3DGen: High-fidelity 3D Geometry Generation from Images via Normal Bridging

Hunyuan3D-2 and TRELLIS are incredible,
But the 3D Open Source Jounery is not over yet,
We have one more thing…#3DModeling #AI #3D #GenerativeAI pic.twitter.com/DfYXciQp39

— Chongjie(CJ) Ye (@ychngji6) March 16, 2025

Excited to announce Hi👋3DGen v0.1 – my first 3D generation project! Released at https://t.co/7w3yMJAG3V. Huge thanks to the teams of @DeemosTech, @VastAIResearch, and @BytedanceTalk for their guidance. We sincerely believe open collaboration drives AI innovation forward! https://t.co/hmIBxBIP41 pic.twitter.com/5a6vybGDLQ

— Chongjie(CJ) Ye (@ychngji6) March 28, 2025

精度正不断提升呢，今年会进化到什么程度呢？
目前可以在Hugging Face上试用。Github页面暂时只公开了位置。听说到4月10日之前代码也会公开？请一定去看看。

链接

实时演示 | Hugging Face

Github