metadata
license: mit
Weight Space Representation Learning on Diverse NeRF Architectures (ICLR 2026)
[arXiv | project page]
Main dataset structure:
.
└── nerf/
└── shapenet/
├── hash/ # compressed into _hash.tar
│ └── class_id/
│ └── nerf_id/
│ ├── train/
│ │ └── *.png # object views used to train the NeRF
│ ├── grid.pth # occupancy grid parameters
│ ├── nerf_weights.pth # NeRF parameters
│ └── transforms_train.json # camera poses
├── mlp/ # compressed into _mlp.tar
│ └── class_id/
│ └── nerf_id/
│ ├── train/
│ │ └── *.png
│ ├── grid.pth
│ ├── nerf_weights.pth
│ └── transforms_train.json
├── triplane/ # compressed into _triplane.tar
│ └── class_id/
│ └── nerf_id/
│ ├── train/
│ │ └── *.png
│ ├── grid.pth
│ ├── nerf_weights.pth
│ └── transforms_train.json
├── test.txt # test split
├── train.txt # training split
└── val.txt # validation split
The official code repository will be available soon. In the meantime, here are some links to the (unpolished) code used to train the NeRFs contained in the dataset:
If you are interested in running this code, follow this README to install the required libraries.
If you find our work useful, please cite us:
@inproceedings{ballerini2026weight,
title = {Weight Space Representation Learning on Diverse {NeRF} Architectures},
author = {Ballerini, Francesco and Zama Ramirez, Pierluigi and Di Stefano, Luigi and Salti, Samuele},
booktitle = {The Fourteenth International Conference on Learning Representations},
year = {2026}