vis4d 1.0.0

Vis4D: Python package for Visual 4D scene understanding

A modular library for 4D scene understanding

Quickstart

You can checkout our documentation.

You can use the template here to start your own project with Vis4D.

Installation

Installation is as easy as

python3 -m pip install vis4d

For more detailed information, check out our installation guide

Basic CLI usage

• To train a model, e.g. Faster-RCNN on COCO

# vis4d.engine
vis4d fit --config vis4d/zoo/faster_rcnn/faster_rcnn_coco.py --gpus 1

• To test a model

# vis4d.engine
vis4d test --config vis4d/zoo/faster_rcnn/faster_rcnn_coco.py --gpus 1

DDP

Training

• Local machine / SLURM interactivate job (job-name=bash)

vis4d fit --config <config-file> --gpus <num-gpus>

• SLURM

srun vis4d fit --config <config-file> --gpus <num-gpus>

Testing

• Local machine / SLURM interactivate job (job-name=bash)

vis4d test --config <config-file> --gpus <num-gpus>

• SLURM

srun vis4d test --config <config-file> --gpus <num-gpus>

Acknowledgement

Vis4D is a group effort by our team at ETH Zurich. Yung-Hsu Yang built the current version and will be the main maintainer of the codebase.

Vis4D was originally written by Tobias Fischer during the first three years of his Ph.D. at ETH Zurich, Thomas E. Huang helped contribute many models, Tao Sun implemented the ViT models and designed the evaluation pipeline, andRené Zurbrügg designed the config system.

Contributors

Project Leads

• Yung-Hsu Yang*
• Tobias Fischer*

Core Contributors

• Thomas E. Huang*
• Tao Sun
• René Zurbrügg

Advisors

• Fisher Yu

* denotes equal contribution

We are open to contributions and suggestions, feel free to reach out to us.

Check out our contribution guidelines for this project

Community Contributors

Citation

If you find Vis4D is useful for your research, please consider citing the following BibTeX entry.

@misc{vis4d2024,
  author = {Yang, Yung-Hsu and Fischer, Tobias and Huang, Thomas E. and Zurbr{\"u}gg, Ren{\'e} and Sun, Tao and Yu, Fisher},
  title = {Vis4{D}},
  howpublished = {\url{https://github.com/SysCV/vis4d}},
  year = {2024}
}