Modelling complex vector drawings with stroke-clouds

ICLR 2024 Poster Paper


Paper · Data · Models

Introduction

This is the offical repository for "Modelling complex vector drawings with stroke-clouds". The data set and pre-trained models can be downloaded from the links given. Instructions for inference and training both the srm and lsg are given.

Environment

lxml - 4.9

lightning - 2.0.9

pytorch-lightning - 2.0.3

torch - 2.2.0

Training

SRM:

1. Add your wandb key or change the logger in srm_train.py.
2. Download the training data.
3. Run srm_train.py

LSG:

1. Generate latent codes with srm_test.py
2. Add your wandb key or change the logger in lsg_train.py.
3. Run lsg_train.py

Inference

1. Download the models and run lsg_test.py
2. Or run lsg_test.py with your own trained models.

License

Distributed under the MIT License. See LICENSE.txt for more information.

Contact

Alexander Ashcroft - [email protected]