{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/are-transformers-more-robust-than-cnns","title":"Are Transformers More Robust Than CNNs?","arxiv_id":"2111.05464","date":"2021-11-10","proceeding":"NeurIPS 2021 12","authors":["Yutong Bai","Jieru Mei","Alan Yuille","Cihang Xie"],"abstract":"Transformer emerges as a powerful tool for visual recognition. In addition to demonstrating competitive performance on a broad range of visual benchmarks, recent works also argue that Transformers are much more robust than Convolutions Neural Networks (CNNs). Nonetheless, surprisingly, we find these conclusions are drawn from unfair experimental settings, where Transformers and CNNs are compared at different scales and are applied with distinct training frameworks. In this paper, we aim to provide the first fair & in-depth comparisons between Transformers and CNNs, focusing on robustness evaluations. With our unified training setup, we first challenge the previous belief that Transformers outshine CNNs when measuring adversarial robustness. More surprisingly, we find CNNs can easily be as robust as Transformers on defending against adversarial attacks, if they properly adopt Transformers' training recipes. While regarding generalization on out-of-distribution samples, we show pre-training on (external) large-scale datasets is not a fundamental request for enabling Transformers to achieve better performance than CNNs. Moreover, our ablations suggest such stronger generalization is largely benefited by the Transformer's self-attention-like architectures per se, rather than by other training setups. We hope this work can help the community better understand and benchmark the robustness of Transformers and CNNs. The code and models are publicly available at https://github.com/ytongbai/ViTs-vs-CNNs.","url_abs":"https://arxiv.org/abs/2111.05464v1","url_pdf":"https://arxiv.org/pdf/2111.05464v1.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":"are-transformers-more-robust-than-cnns","repo_url":"https://github.com/ytongbai/ViTs-vs-CNNs","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"adversarial-robustness","task_name":"Adversarial Robustness"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/adversarial-robustness-on-imagenet","task":"Adversarial Robustness","dataset":"ImageNet","model":"ResNet-50 (SGD, Cosine)","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy":"77.4"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet","task":"Adversarial Robustness","dataset":"ImageNet","model":"ResNet-50 (SGD, Step)","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy":"76.9"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet","task":"Adversarial Robustness","dataset":"ImageNet","model":"DeiT-S (AdamW, Cosine)","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy":"76.8"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet","task":"Adversarial Robustness","dataset":"ImageNet","model":"ResNet-50 (AdamW, Cosine)","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy":"76.4"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-a","task":"Adversarial Robustness","dataset":"ImageNet-A","model":"DeiT-S (AdamW, Cosine)","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy":"12.2"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-a","task":"Adversarial Robustness","dataset":"ImageNet-A","model":"ResNet-50 (SGD, Cosine)","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy":"3.3"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-a","task":"Adversarial Robustness","dataset":"ImageNet-A","model":"ResNet-50 (SGD, Step)","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy":"3.2"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-a","task":"Adversarial Robustness","dataset":"ImageNet-A","model":"ResNet-50 (AdamW, Cosine)","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy":"3.1"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-c","task":"Adversarial Robustness","dataset":"ImageNet-C","model":"DeiT-S (AdamW, Cosine)","rank_in_archive_order":1,"of":4,"metrics":{"mean Corruption Error (mCE)":"48.0"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-c","task":"Adversarial Robustness","dataset":"ImageNet-C","model":"ResNet-50 (SGD, Cosine)","rank_in_archive_order":2,"of":4,"metrics":{"mean Corruption Error (mCE)":"56.9"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-c","task":"Adversarial Robustness","dataset":"ImageNet-C","model":"ResNet-50 (SGD, Step)","rank_in_archive_order":3,"of":4,"metrics":{"mean Corruption Error (mCE)":"57.9"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-imagenet-c","task":"Adversarial Robustness","dataset":"ImageNet-C","model":"ResNet-50 (AdamW, Cosine)","rank_in_archive_order":4,"of":4,"metrics":{"mean Corruption Error (mCE)":"59.3"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-stylized-imagenet","task":"Adversarial Robustness","dataset":"Stylized ImageNet","model":"DeiT-S (AdamW, Cosine)","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy":"13.0"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-stylized-imagenet","task":"Adversarial Robustness","dataset":"Stylized ImageNet","model":"ResNet-50 (SGD, Cosine)","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy":"8.4"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-stylized-imagenet","task":"Adversarial Robustness","dataset":"Stylized ImageNet","model":"ResNet-50 (SGD, Step)","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy":"8.3"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-stylized-imagenet","task":"Adversarial Robustness","dataset":"Stylized ImageNet","model":"ResNet-50 (AdamW, Cosine)","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy":"8.1"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2111.05464","atlas_url":"https://app.syntology.ai/?focus=2111.05464","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2111.05464"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2111.05464\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2111.05464"},"call":"get_concepts_for_paper(arxiv_id=\"2111.05464\")"}],"by_repo":[{"repo":"ytongbai/ViTs-vs-CNNs","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"6536d33aba17996c946e4f9865514cffe4518696226fcff65306f1bd14b9cf42","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}