License

This work is published under a Creative Commons Attribution-ShareAlike 4.0 International License. By downloading SHIFT, you agree to follow this license.

If you use our dataset or codes in a research project, we kindly ask you to cite our paper as:

@InProceedings{shift2022,
    author    = {Sun, Tao and Segu, Mattia and Postels, Janis and Wang, Yuxuan and Van Gool, Luc and Schiele, Bernt and Tombari, Federico and Yu, Fisher},
    title     = {{SHIFT:} A Synthetic Driving Dataset for Continuous Multi-Task Domain Adaptation},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2022},
    pages     = {21371-21382}
}

Download

Download by script

We recommend to download SHIFT using the Python download script. You can select the subset of views, data group, splits and framerates of the data to download. A usage example is shown below. You can find the abbreviation for views and data groups in the following tables.

python download.py --view  "[front, front_left]" \    # list of view abbreviation to download
                   --group "[img, semseg]" \          # list of data group abbreviation to download 
                   --split "[train, val, test]" \     # list of splits to download 
                   --framerate "[images, videos]" \   # chooses the desired frame rate (images=1fps, videos=10fps)
                   --shift "discrete" \               # type of domain shifts: discrete, continuous/1x, continuous/10x, continuous/100x
                   dataset_root                       # path where to store the downloaded data

In our Github repo, you can find useful tools for data loading and data conversion!

Download manually

You can also access our file server here directly.