{"url":"/dataset/imagenet-32","name":"ImageNet-32","full_name":null,"description_markdown":"Imagenet32 is a huge dataset made up of small images called the down-sampled version of Imagenet. Imagenet32 is composed of 1,281,167 training data and 50,000 test data with 1,000 labels.\r\n\r\nSource: [Self-supervised Knowledge Distillation Using Singular Value Decomposition](https://arxiv.org/abs/1807.06819)\r\nImage Source: [https://arxiv.org/pdf/1707.08819v3.pdf](https://arxiv.org/pdf/1707.08819v3.pdf)","description_withheld":null,"homepage":"https://github.com/PatrykChrabaszcz/Imagenet32_Scripts","introduced_date":"2017-01-01","introduced_date_note":null,"introduced_by":{"paper":"/paper/a-downsampled-variant-of-imagenet-as-an","title":"A Downsampled Variant of ImageNet as an Alternative to the CIFAR datasets","first_author":"Patryk Chrabaszcz","url":null},"license":{"name":"Unknown","url":null},"modalities":[{"name":"Images","url":"/datasets/modality/images"}],"tasks":[{"name":"Image Classification","url":"/task/image-classification","datasets_with_task":"/datasets/task/image-classification"},{"name":"Image Compression","url":"/task/image-compression","datasets_with_task":"/datasets/task/image-compression"},{"name":"Learning with coarse labels","url":"/task/learning-with-coarse-labels","datasets_with_task":"/datasets/task/learning-with-coarse-labels"},{"name":"Sparse Learning","url":"/task/sparse-learning","datasets_with_task":"/datasets/task/sparse-learning"}],"languages":[],"variants":["ImageNet32","ImageNet-32"],"data_loaders":[{"repo":"https://github.com/tensorflow/datasets","url":"https://www.tensorflow.org/datasets/catalog/imagenet_resized","frameworks":["tf","jax"]},{"repo":"https://github.com/PatrykChrabaszcz/Imagenet32_Scripts","url":"https://github.com/PatrykChrabaszcz/Imagenet32_Scripts","frameworks":[]}],"num_papers_in_archive":112,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/image-compression-on-imagenet32","task":"Image Compression","dataset_variant":"ImageNet32","rows":5,"metrics":["bpsp"],"first_row_in_archive_order":{"model":"iFlow","paper":"/paper/iflow-numerically-invertible-flows-for","metrics":{"bpsp":"3.88"},"code_links":[]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/learning-with-coarse-labels-on-imagenet32","task":"Learning with coarse labels","dataset_variant":"ImageNet32","rows":2,"metrics":["Recall@1","Recall@2","Recall@5","Recall@10"],"first_row_in_archive_order":{"model":"MaskCon","paper":"/paper/maskcon-masked-contrastive-learning-for","metrics":{"Recall@1":"19.08","Recall@10":"47.96","Recall@2":"26.21","Recall@5":"38.17 "},"code_links":[{"title":"MrChenFeng/MaskCon_CVPR2023","url":"https://github.com/MrChenFeng/MaskCon_CVPR2023"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/image-classification-on-imagenet-32","task":"Image Classification","dataset_variant":"ImageNet-32","rows":1,"metrics":["Top 1 Error"],"first_row_in_archive_order":{"model":"WRN (N=28, k=10)","paper":"/paper/a-downsampled-variant-of-imagenet-as-an","metrics":{"Top 1 Error":"40.96"},"code_links":[{"title":"BayesWatch/cinic-10","url":"https://github.com/BayesWatch/cinic-10"},{"title":"PatrykChrabaszcz/Imagenet32_Scripts","url":"https://github.com/PatrykChrabaszcz/Imagenet32_Scripts"},{"title":"ZilinGao/GM-SOP","url":"https://github.com/ZilinGao/GM-SOP"},{"title":"Prev/downsampled-imagenet-path-fixer","url":"https://github.com/Prev/downsampled-imagenet-path-fixer"},{"title":"attaullah/Pretraining-WideResNet","url":"https://github.com/attaullah/Pretraining-WideResNet"},{"title":"curryandsun/AIOL","url":"https://github.com/curryandsun/AIOL"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/sparse-learning-on-imagenet32-1","task":"Sparse Learning","dataset_variant":"ImageNet32","rows":1,"metrics":["Sparsity"],"first_row_in_archive_order":{"model":"Resnet18","paper":"/paper/adaptive-neural-connections-for-sparsity","metrics":{"Sparsity":"93.63"},"code_links":[]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/maskcon-masked-contrastive-learning-for","title":"MaskCon: Masked Contrastive Learning for Coarse-Labelled Dataset","date":"2023-03-22","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":0,"samples_unverified":2,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/iflow-numerically-invertible-flows-for","title":"iFlow: Numerically Invertible Flows for Efficient Lossless Compression via a Uniform Coder","date":"2021-11-01","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/grafit-learning-fine-grained-image","title":"Grafit: Learning fine-grained image representations with coarse labels","date":"2020-11-25","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/adaptive-neural-connections-for-sparsity","title":"Adaptive Neural Connections for Sparsity Learning","date":"2020-03-05","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/practical-full-resolution-learned-lossless","title":"Practical Full Resolution Learned Lossless Image Compression","date":"2018-11-30","rows_on_this_dataset":3,"code_links":4,"syntology":null},{"paper":"/paper/a-downsampled-variant-of-imagenet-as-an","title":"A Downsampled Variant of ImageNet as an Alternative to the CIFAR datasets","date":"2017-07-27","rows_on_this_dataset":1,"code_links":6,"syntology":null},{"paper":"/paper/parallel-multiscale-autoregressive-density","title":"Parallel Multiscale Autoregressive Density Estimation","date":"2017-03-10","rows_on_this_dataset":1,"code_links":0,"syntology":null}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":1,"samples_harvested":2,"samples_ran":0,"samples_unverified":2,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":1,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}