Skip to content

Code for Lagrangian Hashes for Compressed Neural Fields Representation

License

Notifications You must be signed in to change notification settings

theialab/lagrangian_hashes

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 Cannot retrieve latest commit at this time.

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Official code for Lagrangian Hashes

[Project Page] [Arxiv]

Cloning the Repository

The repository contains submodules, thus please check it out with

# clone the repo with submodules.
git clone --recursive https://github.com/theialab/lagrangian_hashes.git

Installation

Dependence: Please install Pytorch first. This code is tested with Pytorch-2.1.0 with CUDA 12.1

Then please install nerfacc and lagrangian hashes code by running the setup.py file

python -m pip install .
cd laghash
python -m pip install -e .
cd ..

Then to install other libraries(including tcnn), use the requirements.txt

pip install -r requirements.txt

Experiments

Before running the example scripts, please check which dataset is needed, and download the dataset first. You could use dataset.data_root to specify the path or modify the config yaml.

NeRF-Synthetic (blender) dataset

python examples/train_laghash_nerf_occ.py --config-name "synthetic_occ.yaml" dataset.scene="chair"

Tanks and Temples (masked) dataset

python examples/train_laghash_nerf_occ.py --config-name "tanks_temples_occ.yaml" dataset.scene="Family"

Citation

@inproceedings{govindarajan2024laghashes,
  title     = {Lagrangian Hashing for Compressed Neural Field Representations},
  author    = {Shrisudhan Govindarajan, Zeno Sambugaro, Ahan Shabhanov, Towaki Takikawa, Weiwei Sun, Daniel Rebain, Nicola Conci, Kwang Moo  Yi, Andrea Tagliasacchi},
  booktitle = {ECCV},
  year      = {2024},
}

About

Code for Lagrangian Hashes for Compressed Neural Fields Representation

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published