---
license: mit
pipeline_tag: text-to-3d
---

# Garment Particles: A 2D–3D Symmetric Garment Representation for Generation and Editing

Official checkpoints for **Garment Particles**, a framework for garment design spanning intuitive creation from high-level intent (text, image, sketch) to complex low-level editing across 2D sewing patterns and 3D draped geometry.

- **Paper**: [Garment Particles: A 2D--3D Symmetric Garment Representation for Generation and Editing](https://huggingface.co/papers/2605.26391)
- **Project Page**: [https://garment-particles.github.io](https://garment-particles.github.io)
- **Code**: [https://github.com/garment-particles/GarmentParticles](https://github.com/garment-particles/GarmentParticles)

## Overview
Garment Particles uses a 5D point-cloud representation to jointly encode 2D sewing patterns and 3D geometry. This representation enables Garment Particles Flow (GPF), a rectified flow framework that supports intuitive generation from high-level inputs (text, images, sketches) and various editing operations on 2D sewing patterns and 3D geometries.

## Installation
To use these checkpoints, clone the [official repository](https://github.com/garment-particles/GarmentParticles) and install the dependencies:

```bash
conda create -n interact_garment python=3.10
conda activate interact_garment
pip install torch torchvision
pip install -r src/requirements.txt
export PYTHONPATH=$PWD/src:$PYTHONPATH
```

## Sample Usage (Inference)

Download the checkpoints into `src/checkpoints/` and run the inference script from the `src` directory.

### Text / Unconditional Generation
```bash
torchrun --standalone --nproc_per_node=1 inference/infer_twostage.py \
  eval.sample_per_batch=1 eval.n_samples=0 eval.evaluate=False \
  train.exp_name=uncond_samples sample.num_sampling_steps=100 \
  gpf_ckpt=null \
  dataset.front_only=True dataset.use_all_captions=True \
  dataset.img_drop_prob=1 dataset.text_drop_prob=1 \
  model.use_qknorm=True \
  edge_model.use_qknorm=True \
  edge_model_ckpt=checkpoints/edge \
  model=sparse_lightningdit_v3_xl1_w_text_fsdp2 \
  pgf_weight_init=checkpoints/pgf_text \
  --config-name sparselightningdit_xl_garment_particle_inference
```

### Image-Conditioned Generation
```bash
torchrun --standalone --nproc_per_node=1 inference/infer_twostage.py \
  eval.sample_per_batch=1 eval.n_samples=0 eval.evaluate=False \
  train.exp_name=img_cond_samples sample.num_sampling_steps=100 \
  gpf_ckpt=null \
  dataset.front_only=True dataset.use_all_captions=True \
  dataset.img_drop_prob=0 dataset.text_drop_prob=1 \
  model.use_qknorm=True model.use_rope=False model.in_channels=6 model.freeze_everything=False \
  edge_model.use_qknorm=True \
  edge_model_ckpt=checkpoints/edge \
  model=sparse_lightningdit_v3_xl1_w_img_text_v2 \
  pgf_weight_init=checkpoints/pgf_image \
  --config-name sparselightningdit_xl_garment_particle_inference
```

## Citation
```bibtex
@inproceedings{garmentparticles2026,
  title={Garment Particles: A 2D--3D Symmetric Garment Representation for Generation and Editing},
  author={George Nakayama and others},
  booktitle={SIGGRAPH Conference Papers},
  year={2026}
}
```