{"url":"/sota/data-augmentation-on-imagenet","task":{"name":"Data Augmentation","url":"/task/data-augmentation","note":null},"dataset":{"name":"ImageNet","url":"/dataset/imagenet"},"category":"Computer Vision","categories":["Computer Vision","Methodology","Natural Language Processing"],"category_note":null,"description":"Data augmentation involves techniques used for increasing the amount of data, based on different modifications, to expand the amount of examples in the original dataset. Data augmentation not only helps to grow the dataset but it also increases the diversity of the dataset. When training machine learning models, data augmentation acts as a regularizer and helps to avoid overfitting. \r\n\r\nData augmentation techniques have been found useful in domains like NLP and computer vision. In computer vision, transformations like cropping, flipping, and rotation are used. In NLP, data augmentation techniques can include swapping, deletion, random insertion, among others. \r\n\r\nFurther readings:\r\n\r\n- [A Survey of Data Augmentation Approaches for NLP](https://paperswithcode.com/paper/a-survey-of-data-augmentation-approaches-for)\r\n- [A survey on Image Data Augmentation for Deep Learning](https://journalofbigdata.springeropen.com/articles/10.1186/s40537-019-0197-0)\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [Albumentations](https://github.com/albumentations-team/albumentations) )</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["Accuracy (%)"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Accuracy (%)":"higher"}},"counts":{"rows":17,"rows_with_code":17,"rows_with_paper_page":17,"rows_dated":17,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"DeiT-B (+MixPro)","metrics":{"Accuracy (%)":"82.9"},"uses_additional_data":false,"paper_date":"2023-04-24","paper":"/paper/mixpro-data-augmentation-with-maskmix-and","paper_url":"https://arxiv.org/abs/2304.12043v2","paper_title":"MixPro: Data Augmentation with MaskMix and Progressive Attention Labeling for Vision Transformer","code":"https://github.com/fistyee/mixpro","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":8,"n_samples":18,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"ResNet-200 (DeepAA)","metrics":{"Accuracy (%)":"81.32"},"uses_additional_data":false,"paper_date":"2022-03-11","paper":"/paper/deep-autoaugment-1","paper_url":"https://arxiv.org/abs/2203.06172v2","paper_title":"Deep AutoAugment","code":"https://github.com/msu-mlsys-lab/deepaa","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"DeiT-S (+MixPro)","metrics":{"Accuracy (%)":"81.3"},"uses_additional_data":false,"paper_date":"2023-04-24","paper":"/paper/mixpro-data-augmentation-with-maskmix-and","paper_url":"https://arxiv.org/abs/2304.12043v2","paper_title":"MixPro: Data Augmentation with MaskMix and Progressive Attention Labeling for Vision Transformer","code":"https://github.com/fistyee/mixpro","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":8,"n_samples":18,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"ResNet-200 (Fast AA)","metrics":{"Accuracy (%)":"80.6"},"uses_additional_data":false,"paper_date":"2019-05-01","paper":"/paper/fast-autoaugment","paper_url":"https://arxiv.org/abs/1905.00397v2","paper_title":"Fast AutoAugment","code":"https://github.com/kakaobrain/fast-autoaugment","n_code_links":11,"syntology":{"n_ran":22,"n_unverified":18,"n_samples":40,"n_pointer_only_licence":3}},{"rank_in_archive_order":5,"model":"ResNet-200 (UA)","metrics":{"Accuracy (%)":"80.4"},"uses_additional_data":false,"paper_date":"2020-03-31","paper":"/paper/uniformaugment-a-search-free-probabilistic","paper_url":"https://arxiv.org/abs/2003.14348v1","paper_title":"UniformAugment: A Search-free Probabilistic Data Augmentation Approach","code":"https://github.com/tgilewicz/uniformaugment","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"ResNet-200 (AA)","metrics":{"Accuracy (%)":"80.0"},"uses_additional_data":false,"paper_date":"2018-05-24","paper":"/paper/autoaugment-learning-augmentation-policies","paper_url":"http://arxiv.org/abs/1805.09501v3","paper_title":"AutoAugment: Learning Augmentation Policies from Data","code":"https://github.com/tensorflow/models/tree/master/research/autoaugment","n_code_links":33,"syntology":{"n_ran":6,"n_unverified":37,"n_samples":43,"n_pointer_only_licence":2}},{"rank_in_archive_order":7,"model":"ResNet-50 (DeepAA)","metrics":{"Accuracy (%)":"78.30"},"uses_additional_data":false,"paper_date":"2022-03-11","paper":"/paper/deep-autoaugment-1","paper_url":"https://arxiv.org/abs/2203.06172v2","paper_title":"Deep AutoAugment","code":"https://github.com/msu-mlsys-lab/deepaa","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"ResNet-50 (TA wide)","metrics":{"Accuracy (%)":"78.07"},"uses_additional_data":false,"paper_date":"2021-03-18","paper":"/paper/trivialaugment-tuning-free-yet-state-of-the","paper_url":"https://arxiv.org/abs/2103.10158v2","paper_title":"TrivialAugment: Tuning-free Yet State-of-the-Art Data Augmentation","code":"https://github.com/pytorch/vision","n_code_links":2,"syntology":null},{"rank_in_archive_order":9,"model":"ResNet-50 (LoRot-E)","metrics":{"Accuracy (%)":"77.72"},"uses_additional_data":false,"paper_date":"2022-07-20","paper":"/paper/tailoring-self-supervision-for-supervised","paper_url":"https://arxiv.org/abs/2207.10023v1","paper_title":"Tailoring Self-Supervision for Supervised Learning","code":"https://github.com/wjun0830/localizable-rotation","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"ResNet-50 (LoRot-I)","metrics":{"Accuracy (%)":"77.71"},"uses_additional_data":false,"paper_date":"2022-07-20","paper":"/paper/tailoring-self-supervision-for-supervised","paper_url":"https://arxiv.org/abs/2207.10023v1","paper_title":"Tailoring Self-Supervision for Supervised Learning","code":"https://github.com/wjun0830/localizable-rotation","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"ResNet-50 (UA)","metrics":{"Accuracy (%)":"77.63"},"uses_additional_data":false,"paper_date":"2020-03-31","paper":"/paper/uniformaugment-a-search-free-probabilistic","paper_url":"https://arxiv.org/abs/2003.14348v1","paper_title":"UniformAugment: A Search-free Probabilistic Data Augmentation Approach","code":"https://github.com/tgilewicz/uniformaugment","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"ResNet-50 (AA)","metrics":{"Accuracy (%)":"77.6"},"uses_additional_data":false,"paper_date":"2018-05-24","paper":"/paper/autoaugment-learning-augmentation-policies","paper_url":"http://arxiv.org/abs/1805.09501v3","paper_title":"AutoAugment: Learning Augmentation Policies from Data","code":"https://github.com/tensorflow/models/tree/master/research/autoaugment","n_code_links":33,"syntology":{"n_ran":6,"n_unverified":37,"n_samples":43,"n_pointer_only_licence":2}},{"rank_in_archive_order":13,"model":"ResNet-50 (Fast AA)","metrics":{"Accuracy (%)":"77.6"},"uses_additional_data":false,"paper_date":"2019-05-01","paper":"/paper/fast-autoaugment","paper_url":"https://arxiv.org/abs/1905.00397v2","paper_title":"Fast AutoAugment","code":"https://github.com/kakaobrain/fast-autoaugment","n_code_links":11,"syntology":{"n_ran":22,"n_unverified":18,"n_samples":40,"n_pointer_only_licence":3}},{"rank_in_archive_order":14,"model":"ResNet-50 (RA)","metrics":{"Accuracy (%)":"77.6"},"uses_additional_data":false,"paper_date":"2019-09-30","paper":"/paper/randaugment-practical-data-augmentation-with","paper_url":"https://arxiv.org/abs/1909.13719v2","paper_title":"RandAugment: Practical automated data augmentation with a reduced search space","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":19,"syntology":{"n_ran":58,"n_unverified":7,"n_samples":65,"n_pointer_only_licence":17}},{"rank_in_archive_order":15,"model":"ResNet-50 (DADA)","metrics":{"Accuracy (%)":"77.5"},"uses_additional_data":false,"paper_date":"2020-03-08","paper":"/paper/dada-differentiable-automatic-data","paper_url":"https://arxiv.org/abs/2003.03780v3","paper_title":"DADA: Differentiable Automatic Data Augmentation","code":"https://github.com/VDIGPKU/DADA","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":4,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":16,"model":"ResNet-50 (Faster AA)","metrics":{"Accuracy (%)":"76.5"},"uses_additional_data":false,"paper_date":"2019-11-16","paper":"/paper/faster-autoaugment-learning-augmentation","paper_url":"https://arxiv.org/abs/1911.06987v1","paper_title":"Faster AutoAugment: Learning Augmentation Strategies using Backpropagation","code":"https://github.com/moskomule/dda","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":9,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"DeiT-T (+MixPro)","metrics":{"Accuracy (%)":"73.8"},"uses_additional_data":false,"paper_date":"2023-04-24","paper":"/paper/mixpro-data-augmentation-with-maskmix-and","paper_url":"https://arxiv.org/abs/2304.12043v2","paper_title":"MixPro: Data Augmentation with MaskMix and Progressive Attention Labeling for Vision Transformer","code":"https://github.com/fistyee/mixpro","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":8,"n_samples":18,"n_pointer_only_licence":0}}],"since_archive":{"present":false,"note":"No Syntology-extracted rows are published in this build."},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":12,"rows_with_any_sample_ran":9,"distinct_papers_with_graph_line":7,"distinct_papers_with_any_sample_ran":5,"samples_over_distinct_papers":{"n_ran":101,"n_unverified":84,"n_samples":185,"n_pointer_only_licence":22,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":149,"n_unverified":156,"n_samples":305,"n_pointer_only_licence":27,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}