{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/a-downsampled-variant-of-imagenet-as-an","title":"A Downsampled Variant of ImageNet as an Alternative to the CIFAR datasets","arxiv_id":"1707.08819","date":"2017-07-27","proceeding":null,"authors":["Patryk Chrabaszcz","Ilya Loshchilov","Frank Hutter"],"abstract":"The original ImageNet dataset is a popular large-scale benchmark for training\nDeep Neural Networks. Since the cost of performing experiments (e.g, algorithm\ndesign, architecture search, and hyperparameter tuning) on the original dataset\nmight be prohibitive, we propose to consider a downsampled version of ImageNet.\nIn contrast to the CIFAR datasets and earlier downsampled versions of ImageNet,\nour proposed ImageNet32$\\times$32 (and its variants ImageNet64$\\times$64 and\nImageNet16$\\times$16) contains exactly the same number of classes and images as\nImageNet, with the only difference that the images are downsampled to\n32$\\times$32 pixels per image (64$\\times$64 and 16$\\times$16 pixels for the\nvariants, respectively). Experiments on these downsampled variants are\ndramatically faster than on the original ImageNet and the characteristics of\nthe downsampled datasets with respect to optimal hyperparameters appear to\nremain similar. The proposed datasets and scripts to reproduce our results are\navailable at http://image-net.org/download-images and\nhttps://github.com/PatrykChrabaszcz/Imagenet32_Scripts","url_abs":"http://arxiv.org/abs/1707.08819v3","url_pdf":"http://arxiv.org/pdf/1707.08819v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"a-downsampled-variant-of-imagenet-as-an","repo_url":"https://github.com/PatrykChrabaszcz/Imagenet32_Scripts","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"a-downsampled-variant-of-imagenet-as-an","repo_url":"https://github.com/BayesWatch/cinic-10","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"a-downsampled-variant-of-imagenet-as-an","repo_url":"https://github.com/Prev/downsampled-imagenet-path-fixer","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-downsampled-variant-of-imagenet-as-an","repo_url":"https://github.com/ZilinGao/GM-SOP","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"a-downsampled-variant-of-imagenet-as-an","repo_url":"https://github.com/attaullah/Pretraining-WideResNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"a-downsampled-variant-of-imagenet-as-an","repo_url":"https://github.com/curryandsun/AIOL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"architecture-search","task_name":"Neural Architecture Search"}],"methods":[],"datasets_introduced":[{"slug":"imagenet-32","name":"ImageNet-32","full_name":""},{"slug":"imagenet-64","name":"ImageNet-64","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-imagenet-32","task":"Image Classification","dataset":"ImageNet-32","model":"WRN (N=28, k=10)","rank_in_archive_order":1,"of":1,"metrics":{"Top 1 Error":"40.96"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet-64","task":"Image Classification","dataset":"ImageNet-64","model":"WRN (N=36, k=5)","rank_in_archive_order":1,"of":1,"metrics":{"Top 1 Error":"32,34%"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1707.08819","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}