Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/AlgoHunt/VQRF
Official implementation of our CVPR 2023 paper "Compressing Volumetric Radiance Fields to 1 MB"
https://github.com/AlgoHunt/VQRF
Last synced: 4 months ago
JSON representation
Official implementation of our CVPR 2023 paper "Compressing Volumetric Radiance Fields to 1 MB"
- Host: GitHub
- URL: https://github.com/AlgoHunt/VQRF
- Owner: AlgoHunt
- Created: 2022-11-14T01:23:42.000Z (over 2 years ago)
- Default Branch: master
- Last Pushed: 2023-07-24T09:19:36.000Z (over 1 year ago)
- Last Synced: 2024-08-01T13:29:14.767Z (7 months ago)
- Language: Python
- Homepage:
- Size: 581 KB
- Stars: 315
- Watchers: 11
- Forks: 19
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Compressing Volumetric Radiance Fields to 1 MB (CVPR2023)
Lingzhi Li*, Zhen Shen*, Zhongshu Wang, Li Shen, Liefeng Bo
Alibaba Group
:+1:**Update**: We Released [VQ-TensoRF](https://github.com/Spark001/VQ-TensoRF) :+1:
Citation:
```
@inproceedings{li2023compressing,
title={Compressing volumetric radiance fields to 1 mb},
author={Li, Lingzhi and Shen, Zhen and Wang, Zhongshu and Shen, Li and Bo, Liefeng},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={4222--4231},
year={2023}
}
```
**Note**: This repository only contain VQ-DVGO.
## Setup
- Download datasets:
[NeRF](https://drive.google.com/drive/folders/128yBriW1IG_3NJ5Rp7APSTZsJqdJdfc1)
[NSVF](https://dl.fbaipublicfiles.com/nsvf/dataset/Synthetic_NSVF.zip), [T&T (masked)](https://dl.fbaipublicfiles.com/nsvf/dataset/TanksAndTemple.zip)
- Install required libraries
```
pip install -r requirements.txt
cd lib/cuda
python setup.py install
```
Please install the correct version of [Pytorch](https://pytorch.org/) and [torch_scatter](https://github.com/rusty1s/pytorch_scatter) for your machine.
## Directory structure for the datasets
```
data
├── nerf_synthetic # Link: https://drive.google.com/drive/folders/128yBriW1IG_3NJ5Rp7APSTZsJqdJdfc1
│ └── [chair|drums|ficus|hotdog|lego|materials|mic|ship]
│ ├── [train|val|test]
│ │ └── r_*.png
│ └── transforms_[train|val|test].json
│
├── Synthetic_NSVF # Link: https://dl.fbaipublicfiles.com/nsvf/dataset/Synthetic_NSVF.zip
│ └── [Bike|Lifestyle|Palace|Robot|Spaceship|Steamtrain|Toad|Wineholder]
│ ├── intrinsics.txt
│ ├── rgb
│ │ └── [0_train|1_val|2_test]_*.png
│ └── pose
│ └── [0_train|1_val|2_test]_*.txt
│
│
└── TanksAndTemple # Link: https://dl.fbaipublicfiles.com/nsvf/dataset/TanksAndTemple.zip
└── [Barn|Caterpillar|Family|Ignatius|Truck]
├── intrinsics.txt
├── rgb
│ └── [0|1|2]_*.png
└── pose
└── [0|1|2]_*.txt
```
## Training
```
# for nerf_synthetic datasets
python autotask_final.py -g "0 1 2 3 4 5 6 7" --configname syn_4096code
# for Synthetic_NSVF datasets
python autotask_final.py -g "0 1 2 3 4 5 6 7" --configname nsvf_4096code --dataset nsvf
# for TanksAndTemple datasets
python autotask_final.py -g "0 1 2 3 4 5 6 7" --configname tnt_4096code --dataset tnt
```
Fixing TanksandTemples dataset following this [issue](https://github.com/sunset1995/DirectVoxGO/issues/4)
Set `-g` option according to the availible gpu on your machine.
## Testing
```
# for nerf_synthetic datasets
python autotask_eval_only.py -g "0 1 2 3 4 5 6 7" --configname syn_4096code
# for Synthetic_NSVF datasets
python autotask_eval_only.py -g "0 1 2 3 4 5 6 7" --configname nsvf_4096code --dataset nsvf
# for TanksAndTemple datasets
python autotask_eval_only.py -g "0 1 2 3 4 5 6 7" --configname tnt_4096code --dataset tnt
```