---
license: mit
pretty_name: "WorldParticle: Unified World Simulation of Lagrangian Particle Dynamics via Transformer"
viewer: false
---

# WorldParticle: Unified World Simulation of Lagrangian Particle Dynamics via Transformer

This repository provides datasets and model checkpoints for the paper
**WorldParticle: Unified World Simulation of Lagrangian Particle Dynamics via
Transformer**.

- **Paper:** [WorldParticle: Unified World Simulation of Lagrangian Particle Dynamics via Transformer](https://arxiv.org/abs/2605.15305)
- **Code:** [WorldParticle/WorldParticle](https://github.com/WorldParticle/WorldParticle)

## Overview

WorldParticle is a unified learning-based particle simulator for diverse physical
systems. The repository includes data and checkpoints for fluid dynamics, cloth
simulation, and inverse-design experiments.

## Repository Structure

```text
fancy_fluid/
  dataset/
  checkpoint/
cloth_on_ball/
  dataset/
  checkpoint/
inverse_design/
  dataset/
  checkpoint/
deep_lagrange_dataset/
  checkpoint/
```

## Dataset Layout

Each `dataset` directory contains:

```text
dataset/
  train.txt
  validation.txt
  <sample directories>/
```

`train.txt` and `validation.txt` list the sample identifiers used for the two
splits. The numbered sample directories contain the actual simulation data, such
as `.npz` files and task-specific metadata.

Dataset Viewer is disabled because the split text files contain sample identifiers
rather than standalone text samples. Download the repository to access the full
simulation data.

## Checkpoints

Each available `checkpoint` directory contains:

```text
checkpoint/
  config.yaml
  final.ckpt
```

The checkpoint configuration should be used together with the corresponding
model weights. Some components may contain only the files currently available for
that experiment.

## Citation

```bibtex
@article{wang2026worldparticle,
  title   = {WorldParticle: Unified World Simulation of Lagrangian Particle Dynamics via Transformer},
  author  = {Wang, Caoliwen and Guo, Minghao and Chen, Siyuan and Zhang, Heng and Wang, Mengdi and Ni, Xingyu and Sun, Hanson and Wang, Kunyi and Pan, Zherong and Wu, Kui and Liu, Lingjie and Yang, Yin and Jiang, Chenfanfu and Komura, Taku and Matusik, Wojciech and Chen, Peter Yichen},
  journal = {arXiv preprint arXiv:2605.15305},
  year    = {2026}
}
```