This is my adaptation of the amazing Hunyuan3D-2.0 project by Tencent. After a journey through dependency hell, CUDA complications, and virtual environment adventures, I've managed to get the 3D shape generation working using the Hunyuan3D-DiT model
The original implementation by Hunyuan has many more functions, which you can check out on Their Website. I was only interested in the geometry/3D rendering part of their tool. This is a much "simpler" version, which can only generate bare meshes without textures
The GPU Situation: I'm rocking an NVIDIA T600 Laptop GPU with a whopping 4GB of VRAM. The README cheerfully suggests 6GB minimum for shape generation, but we're out here living dangerously with --low_vram_mode and the smaller Hunyuan3D-2mini model. It's like trying to fit a luxury sedan in a compact parking spot - technically possible with enough determination and the right flags.
My laptop almost ran out of storage 🤣
-
PyTorch Installation:
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu130
-
Basic requirements:
pip install -r requirements.txt pip install -e . -
Launch the app:
python gradio_app.py --model_path tencent/Hunyuan3D-2mini --subfolder hunyuan3d-dit-v2-mini --texgen_model_path tencent/Hunyuan3D-2 --low_vram_mode
- Install Visual Studio Build Tools
- Install CUDA Toolkit 13.0
- Compile the C++ extensions for texture generation
- Upgrade GPU (in my dreams 💭)
All credit goes to the original Tencent Hunyuan3D Team for this incredible work!
If you found this repository helpful, please cite our reports:
(Hunyuan Reports👇🏻)
@misc{lai2025hunyuan3d25highfidelity3d,
title={Hunyuan3D 2.5: Towards High-Fidelity 3D Assets Generation with Ultimate Details},
author={Tencent Hunyuan3D Team},
year={2025},
eprint={2506.16504},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2506.16504},
}
@misc{hunyuan3d22025tencent,
title={Hunyuan3D 2.0: Scaling Diffusion Models for High Resolution Textured 3D Assets Generation},
author={Tencent Hunyuan3D Team},
year={2025},
eprint={2501.12202},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
@misc{yang2024hunyuan3d,
title={Hunyuan3D 1.0: A Unified Framework for Text-to-3D and Image-to-3D Generation},
author={Tencent Hunyuan3D Team},
year={2024},
eprint={2411.02293},
archivePrefix={arXiv},
primaryClass={cs.CV}
}Thanks for the contributions of community members, here we have these great extensions of Hunyuan3D 2.0: