3D-Adapter

Official PyTorch implementation of the papers:

3D-Adapter: Geometry-Consistent Multi-View Diffusion for High-Quality 3D Generation

Hansheng Chen¹, Bokui Shen², Yulin Liu^3,4, Ruoxi Shi³, Linqi Zhou², Connor Z. Lin², Jiayuan Gu³, Hao Su^3,4, Gordon Wetzstein¹, Leonidas Guibas¹

¹Stanford University, ²Apparate Labs, ³UCSD, ⁴Hillbot

[Project page] [🤗Demo] [Paper]

Generic 3D Diffusion Adapter Using Controlled Multi-View Editing

Hansheng Chen¹, Ruoxi Shi², Yulin Liu², Bokui Shen³, Jiayuan Gu², Gordon Wetzstein¹, Hao Su², Leonidas Guibas¹

¹Stanford University, ²UCSD, ³Apparate Labs

[Project page] [🤗Demo] [Paper]

https://github.com/user-attachments/assets/6cba3a92-04fe-46ee-88ca-e6dfe5443c36

Todos

• [ ] Release GRM-based 3D-Adapters (unfortunately, we cannot release these models before the official release of GRM)

Installation

The code has been tested in the environment described as follows:

• Linux (tested on Ubuntu 20 and above)
• CUDA Toolkit 11.8 and above
• PyTorch 2.1 and above
• FFmpeg, x264 (optional, for exporting videos)

Other dependencies can be installed via pip install -r requirements.txt.

An example of installation commands is shown below (you may change the CUDA version yourself):

# Export the PATH of CUDA toolkit
export PATH=/usr/local/cuda-12.1/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-12.1/lib64:$LD_LIBRARY_PATH

# Create conda environment
conda create -y -n mvedit python=3.10 numpy=1.26 ninja
conda activate mvedit

# Install FFmpeg (optional)
conda install -c conda-forge ffmpeg x264

# Install PyTorch
conda install pytorch==2.1.2 torchvision==0.16.2 pytorch-cuda=12.1 -c pytorch -c nvidia

# Clone this repo and install other dependencies
git clone https://github.com/Lakonik/MVEdit && cd MVEdit
pip install -r requirements.txt

This codebase also works on Windows systems if the environment is configured correctly. Certain packages (e.g., tiny-cuda-nn) may require adjustments for installation on Windows.

Inference

We recommend using the Gradio Web UI and its APIs for inference. A GPU with at least 24GB of VRAM is required to run the Web UI.

Web UI

Run the following command to start the Web UI:

python app.py --unload-models

The Web UI will be available at http://localhost:7860. If you add the --share flag, a temporary public URL will be generated for you to share the Web UI with others.

All models will be automatically loaded on demand. The first run will take a very long time to download the models. Check your network connection to GitHub, Google Drive and Hugging Face if the download fails.

To view other options, run:

python app.py -h

API

After starting the Web UI, the API docs will be available at http://localhost:7860/?view=api.

Alternatively, you can access the APIs provided by the official demo (https://mvedit.hanshengchen.com/?view=api), including APIs for GRM Adapters.

The API docs are automatically generated by Gradio, and the data types and default values may be incorrect. Please use the default values in the Web UI as a reference.

Please refer to our examples for API usage with python.

Training

Optimization-based 3D-Adapters (a.k.a. MVEdit adapters) adopt only off-the-shelf models and require no further training.

The training code for GRM-based 3D-Adapters will be released after the official release of GRM.

Acknowledgements

This codebase is built upon the following repositories: - Base library modified from SSDNeRF - NeRF renderer and DMTet modified from Stable-DreamFusion - Gaussian Splatting renderer modified from 3DGS and Differential Gaussian Rasterization - Mesh I/O modified from DreamGaussian - GRM for Gaussian reconstruction - Zero123++ for image-to-3D initialization - IP-Adapter for extra conditioning - TRACER for background removal - LoFTR for pose estimation in image-to-3D - Omnidata for normal prediction in image-to-3D - Image Packer for mesh preprocessing

Citation

@misc{3dadapter2024,
    title={3D-Adapter: Geometry-Consistent Multi-View Diffusion for High-Quality 3D Generation},
    author={Hansheng Chen and Bokui Shen and Yulin Liu and Ruoxi Shi and Linqi Zhou and Connor Z. Lin and Jiayuan Gu and Hao Su and Gordon Wetzstein and Leonidas Guibas},
    year={2024},
    eprint={2410.18974},
    archivePrefix={arXiv},
    primaryClass={cs.CV},
    url={https://arxiv.org/abs/2410.18974},
}

@misc{mvedit2024,
    title={Generic 3D Diffusion Adapter Using Controlled Multi-View Editing},
    author={Hansheng Chen and Ruoxi Shi and Yulin Liu and Bokui Shen and Jiayuan Gu and Gordon Wetzstein and Hao Su and Leonidas Guibas},
    year={2024},
    eprint={2403.12032},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}