Fine-grained annotations for the ImageNet validation set

This is the data collected for our paper "From ImageNet to Image Classification: Contextualizing Progress on Benchmarks" (preprint, blog).

Parsing the annotations

Our annotations are available as pandas dataframes in data/annotations_{contains,classify}_task.pkl. These dataframes included both the raw data collected (after quality control), as well as the aggregate quantities we computed for our analysis. The rest of the files in data contain auxiliary information.

The easiest way to navigate these files is by running the jupyter notebook (basic_data_loading.ipynb) which loads all files, providing an explanation for each field.

Citation

@inproceedings{tsipras2020imagenet,
    title={From ImageNet to Image Classification: Contextualizing Progress on Benchmarks},
    author={Dimitris Tsipras and Shibani Santurkar and Logan Engstrom and Andrew Ilyas and Aleksander Madry},
    booktitle={ArXiv preprint arXiv:2005.11295},
    year={2020}
}