{"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/neural-architecture-transfer","title":"Neural Architecture Transfer","arxiv_id":"2005.05859","date":"2020-05-12","proceeding":null,"authors":["Zhichao Lu","Gautam Sreekumar","Erik Goodman","Wolfgang Banzhaf","Kalyanmoy Deb","Vishnu Naresh Boddeti"],"abstract":"Neural architecture search (NAS) has emerged as a promising avenue for automatically designing task-specific neural networks. Existing NAS approaches require one complete search for each deployment specification of hardware or objective. This is a computationally impractical endeavor given the potentially large number of application scenarios. In this paper, we propose Neural Architecture Transfer (NAT) to overcome this limitation. NAT is designed to efficiently generate task-specific custom models that are competitive under multiple conflicting objectives. To realize this goal we learn task-specific supernets from which specialized subnets can be sampled without any additional training. The key to our approach is an integrated online transfer learning and many-objective evolutionary search procedure. A pre-trained supernet is iteratively adapted while simultaneously searching for task-specific subnets. We demonstrate the efficacy of NAT on 11 benchmark image classification tasks ranging from large-scale multi-class to small-scale fine-grained datasets. In all cases, including ImageNet, NATNets improve upon the state-of-the-art under mobile settings ($\\leq$ 600M Multiply-Adds). Surprisingly, small-scale fine-grained datasets benefit the most from NAT. At the same time, the architecture search and transfer is orders of magnitude more efficient than existing NAS methods. Overall, the experimental evaluation indicates that, across diverse image classification tasks and computational objectives, NAT is an appreciably more effective alternative to conventional transfer learning of fine-tuning weights of an existing network architecture learned on standard datasets. Code is available at https://github.com/human-analysis/neural-architecture-transfer","url_abs":"https://arxiv.org/abs/2005.05859v2","url_pdf":"https://arxiv.org/pdf/2005.05859v2.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":"neural-architecture-transfer","repo_url":"https://github.com/human-analysis/neural-architecture-transfer","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"neural-architecture-transfer","repo_url":"https://github.com/awesomelemon/encas","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"fine-grained-image-classification","task_name":"Fine-Grained Image Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"architecture-search","task_name":"Neural Architecture Search"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/fine-grained-image-classification-on-fgvc","task":"Fine-Grained Image Classification","dataset":"FGVC Aircraft","model":"NAT-M4","rank_in_archive_order":47,"of":57,"metrics":{"Accuracy":"90.8%","FLOPS":"581M","PARAMS":"5.3M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-fgvc","task":"Fine-Grained Image Classification","dataset":"FGVC Aircraft","model":"NAT-M3","rank_in_archive_order":48,"of":57,"metrics":{"Accuracy":"90.1%","FLOPS":"388M","PARAMS":"5.1M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-fgvc","task":"Fine-Grained Image Classification","dataset":"FGVC Aircraft","model":"NAT-M2","rank_in_archive_order":50,"of":57,"metrics":{"Accuracy":"89.0%","FLOPS":"235M","PARAMS":"3.4M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-fgvc","task":"Fine-Grained Image Classification","dataset":"FGVC Aircraft","model":"NAT-M1","rank_in_archive_order":52,"of":57,"metrics":{"Accuracy":"87.0%","FLOPS":"175M","PARAMS":"3.2M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-food-101","task":"Fine-Grained Image Classification","dataset":"Food-101","model":"NAT-M4","rank_in_archive_order":11,"of":15,"metrics":{"Accuracy":"89.4","FLOPS":"361M","PARAMS":"4.5M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-food-101","task":"Fine-Grained Image Classification","dataset":"Food-101","model":"NAT-M3","rank_in_archive_order":12,"of":15,"metrics":{"Accuracy":"89.0","FLOPS":"299M","PARAMS":"3.9M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-food-101","task":"Fine-Grained Image Classification","dataset":"Food-101","model":"NAT-M2","rank_in_archive_order":13,"of":15,"metrics":{"Accuracy":"88.5","FLOPS":"266M","PARAMS":"4.1M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-food-101","task":"Fine-Grained Image Classification","dataset":"Food-101","model":"NAT-M1","rank_in_archive_order":14,"of":15,"metrics":{"Accuracy":"87.4","FLOPS":"198M","PARAMS":"3.1M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford","task":"Fine-Grained Image Classification","dataset":"Oxford 102 Flowers","model":"NAT-M4","rank_in_archive_order":12,"of":25,"metrics":{"Accuracy":"98.3%","FLOPS":"400M","PARAMS":"4.2M"},"uses_additional_data":true},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford","task":"Fine-Grained Image Classification","dataset":"Oxford 102 Flowers","model":"NAT-M3","rank_in_archive_order":14,"of":25,"metrics":{"Accuracy":"98.1%","FLOPS":"250M","PARAMS":"3.7M"},"uses_additional_data":true},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford","task":"Fine-Grained Image Classification","dataset":"Oxford 102 Flowers","model":"NAT-M2","rank_in_archive_order":18,"of":25,"metrics":{"Accuracy":"97.9%","FLOPS":"195M","PARAMS":"3.4M"},"uses_additional_data":true},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford","task":"Fine-Grained Image Classification","dataset":"Oxford 102 Flowers","model":"NAT-M1","rank_in_archive_order":25,"of":25,"metrics":{"FLOPS":"152M","PARAMS":"3.3M"},"uses_additional_data":true},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford-1","task":"Fine-Grained Image Classification","dataset":"Oxford-IIIT Pet Dataset","model":"NAT-M1","rank_in_archive_order":15,"of":15,"metrics":{"FLOPS":"160M","PARAMS":"4.0M"},"uses_additional_data":true},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford-2","task":"Fine-Grained Image Classification","dataset":"Oxford-IIIT Pets","model":"NAT-M4","rank_in_archive_order":7,"of":19,"metrics":{"Accuracy":"94.3","FLOPS":"744M","PARAMS":"8.5M","Top-1 Error Rate":"5.7%"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford-2","task":"Fine-Grained Image Classification","dataset":"Oxford-IIIT Pets","model":"NAT-M3","rank_in_archive_order":8,"of":19,"metrics":{"Accuracy":"94.1","FLOPS":"471M","PARAMS":"5.7M","Top-1 Error Rate":"5.9%"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-oxford-2","task":"Fine-Grained Image Classification","dataset":"Oxford-IIIT Pets","model":"NAT-M2","rank_in_archive_order":9,"of":19,"metrics":{"Accuracy":"93.5","FLOPS":"306M","PARAMS":"5.5M","Top-1 Error Rate":"6.5%"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-stanford","task":"Fine-Grained Image Classification","dataset":"Stanford Cars","model":"NAT-M4","rank_in_archive_order":68,"of":83,"metrics":{"Accuracy":"92.9%","FLOPS":"369M","PARAMS":"3.7M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-stanford","task":"Fine-Grained Image Classification","dataset":"Stanford Cars","model":"NAT-M3","rank_in_archive_order":72,"of":83,"metrics":{"Accuracy":"92.6%","FLOPS":"289M","PARAMS":"3.5M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-stanford","task":"Fine-Grained Image Classification","dataset":"Stanford Cars","model":"NAT-M2","rank_in_archive_order":75,"of":83,"metrics":{"Accuracy":"92.2%","FLOPS":"222M","PARAMS":"2.7M"},"uses_additional_data":false},{"leaderboard":"/sota/fine-grained-image-classification-on-stanford","task":"Fine-Grained Image Classification","dataset":"Stanford Cars","model":"NAT-M1","rank_in_archive_order":77,"of":83,"metrics":{"Accuracy":"90.9%","FLOPS":"165M","PARAMS":"2.4M"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-10","task":"Image Classification","dataset":"CIFAR-10","model":"NAT-M4","rank_in_archive_order":41,"of":265,"metrics":{"Parameters":"6.9M","Percentage correct":"98.4","Top-1 Accuracy":"98.4"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-10","task":"Image Classification","dataset":"CIFAR-10","model":"NAT-M3","rank_in_archive_order":49,"of":265,"metrics":{"Parameters":"6.2M","Percentage correct":"98.2","Top-1 Accuracy":"98.2"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-10","task":"Image Classification","dataset":"CIFAR-10","model":"NAT-M2","rank_in_archive_order":63,"of":265,"metrics":{"Parameters":"4.6M","Percentage correct":"97.9","Top-1 Accuracy":"97.9"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-10","task":"Image Classification","dataset":"CIFAR-10","model":"NAT-M1","rank_in_archive_order":85,"of":265,"metrics":{"Parameters":"4.3M","Percentage correct":"97.4","Top-1 Accuracy":"97.4"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-100","task":"Image Classification","dataset":"CIFAR-100","model":"NAT-M4","rank_in_archive_order":39,"of":211,"metrics":{"PARAMS":"9.0M","Percentage correct":"88.3"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-100","task":"Image Classification","dataset":"CIFAR-100","model":"NAT-M3","rank_in_archive_order":41,"of":211,"metrics":{"PARAMS":"7.8M","Percentage correct":"87.7"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-100","task":"Image Classification","dataset":"CIFAR-100","model":"NAT-M2","rank_in_archive_order":43,"of":211,"metrics":{"PARAMS":"6.4M","Percentage correct":"87.5"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cifar-100","task":"Image Classification","dataset":"CIFAR-100","model":"NAT-M1","rank_in_archive_order":57,"of":211,"metrics":{"PARAMS":"3.8M","Percentage correct":"86.0"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-cinic-10","task":"Image Classification","dataset":"CINIC-10","model":"NAT-M3","rank_in_archive_order":3,"of":9,"metrics":{"Accuracy":"94.3","FLOPS":"501M","PARAMS":"8.1M"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cinic-10","task":"Image Classification","dataset":"CINIC-10","model":"NAT-M2","rank_in_archive_order":4,"of":9,"metrics":{"Accuracy":"94.1","FLOPS":"411M","PARAMS":"6.2M"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cinic-10","task":"Image Classification","dataset":"CINIC-10","model":"NAT-M1","rank_in_archive_order":5,"of":9,"metrics":{"Accuracy":"93.4","FLOPS":"317M","PARAMS":"4.6M"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-flowers-102","task":"Image Classification","dataset":"Flowers-102","model":"NAT-M4","rank_in_archive_order":26,"of":52,"metrics":{"Accuracy":"98.3%","FLOPS":"400M","PARAMS":"4.2M"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-flowers-102","task":"Image Classification","dataset":"Flowers-102","model":"NAT-M3","rank_in_archive_order":29,"of":52,"metrics":{"Accuracy":"98.1%","FLOPS":"250M","PARAMS":"3.7M"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-flowers-102","task":"Image Classification","dataset":"Flowers-102","model":"NAT-M2","rank_in_archive_order":33,"of":52,"metrics":{"Accuracy":"97.9%","FLOPS":"195M","PARAMS":"3.4M"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-flowers-102","task":"Image Classification","dataset":"Flowers-102","model":"NAT-M1","rank_in_archive_order":52,"of":52,"metrics":{"FLOPS":"152M","PARAMS":"3.3M"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"NAT-M4","rank_in_archive_order":697,"of":1060,"metrics":{"Number of params":"9.1M","Top 1 Accuracy":"80.5%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-stl-10","task":"Image Classification","dataset":"STL-10","model":"NAT-M4","rank_in_archive_order":10,"of":117,"metrics":{"FLOPS":"573M","PARAMS":"7.5M","Percentage correct":"97.9"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-stl-10","task":"Image Classification","dataset":"STL-10","model":"NAT-M3","rank_in_archive_order":11,"of":117,"metrics":{"FLOPS":"436M","PARAMS":"7.5M","Percentage correct":"97.8"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-stl-10","task":"Image Classification","dataset":"STL-10","model":"NAT-M2","rank_in_archive_order":13,"of":117,"metrics":{"FLOPS":"303M","PARAMS":"5.1M","Percentage correct":"97.2"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-stl-10","task":"Image Classification","dataset":"STL-10","model":"NAT-M1","rank_in_archive_order":15,"of":117,"metrics":{"FLOPS":"240M","PARAMS":"4.4M","Percentage correct":"96.7"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-10","task":"Neural Architecture Search","dataset":"CIFAR-10","model":"NAT-M4","rank_in_archive_order":1,"of":41,"metrics":{"FLOPS":"468M","Parameters":"6.9M","Search Time (GPU days)":"1.0","Top-1 Error Rate":"1.6%"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-10","task":"Neural Architecture Search","dataset":"CIFAR-10","model":"NAT-M3","rank_in_archive_order":4,"of":41,"metrics":{"FLOPS":"392M","Parameters":"6.2M","Search Time (GPU days)":"1.0","Top-1 Error Rate":"1.8%"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-10","task":"Neural Architecture Search","dataset":"CIFAR-10","model":"NAT-M2","rank_in_archive_order":8,"of":41,"metrics":{"FLOPS":"291M","Parameters":"4.6M","Search Time (GPU days)":"1.0","Top-1 Error Rate":"2.1%"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-10","task":"Neural Architecture Search","dataset":"CIFAR-10","model":"NAT-M1","rank_in_archive_order":28,"of":41,"metrics":{"FLOPS":"232M","Parameters":"4.3M","Search Time (GPU days)":"1.0","Top-1 Error Rate":"2.6%"},"uses_additional_data":false},{"leaderboard":"/sota/architecture-search-on-cifar-10-image","task":"Neural Architecture Search","dataset":"CIFAR-10 Image Classification","model":"NAT-M4","rank_in_archive_order":1,"of":19,"metrics":{"FLOPS":"468M","Params":"6.9M","Percentage error":"1.6"},"uses_additional_data":true},{"leaderboard":"/sota/architecture-search-on-cifar-10-image","task":"Neural Architecture Search","dataset":"CIFAR-10 Image Classification","model":"NAT-M3","rank_in_archive_order":2,"of":19,"metrics":{"FLOPS":"392M","Params":"6.2M","Percentage error":"1.8"},"uses_additional_data":true},{"leaderboard":"/sota/architecture-search-on-cifar-10-image","task":"Neural Architecture Search","dataset":"CIFAR-10 Image Classification","model":"NAT-M2","rank_in_archive_order":7,"of":19,"metrics":{"FLOPS":"291M","Params":"4.6M","Percentage error":"2.1"},"uses_additional_data":true},{"leaderboard":"/sota/architecture-search-on-cifar-10-image","task":"Neural Architecture Search","dataset":"CIFAR-10 Image Classification","model":"NAT-M1","rank_in_archive_order":15,"of":19,"metrics":{"FLOPS":"232M","Params":"4.3M","Percentage error":"2.6"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-100-1","task":"Neural Architecture Search","dataset":"CIFAR-100","model":"NAT-M4","rank_in_archive_order":2,"of":13,"metrics":{"FLOPS":"796M","PARAMS":"9.0M","Percentage Error":"11.7"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-100-1","task":"Neural Architecture Search","dataset":"CIFAR-100","model":"NAT-M3","rank_in_archive_order":3,"of":13,"metrics":{"FLOPS":"492M","PARAMS":"7.8M","Percentage Error":"12.3"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-100-1","task":"Neural Architecture Search","dataset":"CIFAR-100","model":"NAT-M2","rank_in_archive_order":4,"of":13,"metrics":{"FLOPS":"398M","PARAMS":"6.4M","Percentage Error":"12.5"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cifar-100-1","task":"Neural Architecture Search","dataset":"CIFAR-100","model":"NAT-M1","rank_in_archive_order":6,"of":13,"metrics":{"FLOPS":"261M","PARAMS":"3.8M","Percentage Error":"14.0"},"uses_additional_data":true},{"leaderboard":"/sota/neural-architecture-search-on-cinic-10","task":"Neural Architecture Search","dataset":"CINIC-10","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"94.8","FLOPS":"710M","PARAMS":"9.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-cinic-10","task":"Neural Architecture Search","dataset":"CINIC-10","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"94.3","FLOPS":"501M","PARAMS":"8.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-cinic-10","task":"Neural Architecture Search","dataset":"CINIC-10","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"94.1","FLOPS":"411M","PARAMS":"6.2M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-cinic-10","task":"Neural Architecture Search","dataset":"CINIC-10","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"93.4","FLOPS":"317M","PARAMS":"4.6M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-dtd","task":"Neural Architecture Search","dataset":"DTD","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"79.1","FLOPS":"560M","PARAMS":"6.3M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-dtd","task":"Neural Architecture Search","dataset":"DTD","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"78.4","FLOPS":"347M","PARAMS":"4.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-dtd","task":"Neural Architecture Search","dataset":"DTD","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"77.6","FLOPS":"297M","PARAMS":"4.0M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-dtd","task":"Neural Architecture Search","dataset":"DTD","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"76.1","FLOPS":"136M","PARAMS":"2.2M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-fgvc-aircraft","task":"Neural Architecture Search","dataset":"FGVC Aircraft","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"90.8","FLOPS":"581M","PARAMS":"5.3M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-fgvc-aircraft","task":"Neural Architecture Search","dataset":"FGVC Aircraft","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"90.1","FLOPS":"388M","PARAMS":"5.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-fgvc-aircraft","task":"Neural Architecture Search","dataset":"FGVC Aircraft","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"89.0","FLOPS":"235M","PARAMS":"3.4M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-fgvc-aircraft","task":"Neural Architecture Search","dataset":"FGVC Aircraft","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"87.0","FLOPS":"175M","PARAMS":"3.2M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-food-101","task":"Neural Architecture Search","dataset":"Food-101","model":"NAT-M4","rank_in_archive_order":1,"of":5,"metrics":{"Accuracy (%)":"89.4","FLOPS":"361M","PARAMS":"4.5M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-food-101","task":"Neural Architecture Search","dataset":"Food-101","model":"NAT-M3","rank_in_archive_order":2,"of":5,"metrics":{"Accuracy (%)":"89.0","FLOPS":"299M","PARAMS":"3.9M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-food-101","task":"Neural Architecture Search","dataset":"Food-101","model":"NAT-M2","rank_in_archive_order":3,"of":5,"metrics":{"Accuracy (%)":"88.5","FLOPS":"266M","PARAMS":"4.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-food-101","task":"Neural Architecture Search","dataset":"Food-101","model":"NAT-M1","rank_in_archive_order":4,"of":5,"metrics":{"Accuracy (%)":"87.4","FLOPS":"198M","PARAMS":"3.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-imagenet","task":"Neural Architecture Search","dataset":"ImageNet","model":"NAT-M4","rank_in_archive_order":15,"of":135,"metrics":{"Accuracy":"80.5","MACs":"600M","Params":"9.1M","Top-1 Error Rate":"19.5"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-imagenet","task":"Neural Architecture Search","dataset":"ImageNet","model":"NAT-M3","rank_in_archive_order":24,"of":135,"metrics":{"Accuracy":"79.9","MACs":"490M","Params":"9.1M","Top-1 Error Rate":"20.1"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-imagenet","task":"Neural Architecture Search","dataset":"ImageNet","model":"NAT-M2","rank_in_archive_order":44,"of":135,"metrics":{"Accuracy":"78.6","MACs":"312M","Params":"7.7M","Top-1 Error Rate":"21.4"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-imagenet","task":"Neural Architecture Search","dataset":"ImageNet","model":"NAT-M1","rank_in_archive_order":62,"of":135,"metrics":{"Accuracy":"77.5","MACs":"225M","Params":"6.0M","Top-1 Error Rate":"22.5"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-102","task":"Neural Architecture Search","dataset":"Oxford 102 Flowers","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"98.3","FLOPS":"400M","PARAMS":"4.2M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-102","task":"Neural Architecture Search","dataset":"Oxford 102 Flowers","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"98.1","FLOPS":"250M","PARAMS":"3.7M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-102","task":"Neural Architecture Search","dataset":"Oxford 102 Flowers","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"97.9","FLOPS":"195M","PARAMS":"3.4M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-102","task":"Neural Architecture Search","dataset":"Oxford 102 Flowers","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"97.5","FLOPS":"152M","PARAMS":"3.3M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-iiit","task":"Neural Architecture Search","dataset":"Oxford-IIIT Pet Dataset","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"94.3","FLOPS":"744M","PARAMS":"8.5M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-iiit","task":"Neural Architecture Search","dataset":"Oxford-IIIT Pet Dataset","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"94.1","FLOPS":"471M","PARAMS":"5.7M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-iiit","task":"Neural Architecture Search","dataset":"Oxford-IIIT Pet Dataset","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"93.5","FLOPS":"306M","PARAMS":"5.5M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-oxford-iiit","task":"Neural Architecture Search","dataset":"Oxford-IIIT Pet Dataset","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"91.8","FLOPS":"160M","PARAMS":"4.0M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stl-10","task":"Neural Architecture Search","dataset":"STL-10","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"97.9","FLOPS":"573M","PARAMS":"7.5M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stl-10","task":"Neural Architecture Search","dataset":"STL-10","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"97.8","FLOPS":"436M","PARAMS":"7.5M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stl-10","task":"Neural Architecture Search","dataset":"STL-10","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"97.2","FLOPS":"303M","PARAMS":"5.1M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stl-10","task":"Neural Architecture Search","dataset":"STL-10","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"96.7","FLOPS":"240M","PARAMS":"4.4M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stanford-cars","task":"Neural Architecture Search","dataset":"Stanford Cars","model":"NAT-M4","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy (%)":"92.9","FLOPS":"369M","PARAMS":"3.7M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stanford-cars","task":"Neural Architecture Search","dataset":"Stanford Cars","model":"NAT-M3","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy (%)":"92.6","FLOPS":"289M","PARAMS":"3.5M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stanford-cars","task":"Neural Architecture Search","dataset":"Stanford Cars","model":"NAT-M2","rank_in_archive_order":3,"of":4,"metrics":{"Accuracy (%)":"92.2","FLOPS":"222M","PARAMS":"2.7M"},"uses_additional_data":false},{"leaderboard":"/sota/neural-architecture-search-on-stanford-cars","task":"Neural Architecture Search","dataset":"Stanford Cars","model":"NAT-M1","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy (%)":"90.0","FLOPS":"165M","PARAMS":"2.4M"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2005.05859","atlas_url":"https://app.syntology.ai/?focus=2005.05859","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2005.05859"},"calls":[{"tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2005.05859"},"returns":"the samples harvested from this paper, each with its run record; it does not verify the paper"},{"tool":"get_code","arguments_in":"samples[].code_sha256","returns":"one stored sample by its code_sha256; not shown on this site"}]},"developers":"https://syntology.ai/developers","held":true,"reason":"in_graph_layer","checked_against":null,"agent_calls":[{"tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2005.05859"},"call":"get_harvested_code_for_paper(arxiv_id=\"2005.05859\")"},{"tool":"get_code","arguments_in":{"code_sha256":"syntology.samples[].code_sha256"},"call":"get_code(code_sha256=\"…\")"},{"tool":"get_citation_path","arguments":{"paper_1":"2005.05859"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2005.05859\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2005.05859"},"call":"get_concepts_for_paper(arxiv_id=\"2005.05859\")"}],"by_repo":[{"repo":"human-analysis/neural-architecture-transfer","kind":"official","samples":1,"ran":0,"instrument":0,"unverified":1},{"repo":"awesomelemon/encas","kind":"listed","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","read_at":"2026-10-05T00:36:11+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","ran_record":{"url":"https://syntology.ai/api/ran/2005.05859.json","schema":"syntology.ran_record/1","counts_crosswalk":{"n_samples":"syntology.counts.lifted","n_ran":"syntology.counts.executed","n_ran_checked":"syntology.counts.checked","n_instrument":"syntology.counts.instrument_failures","n_unverified":"syntology.counts.no_recorded_run"},"note":"live record, measured at request time; may be newer than this page's graph read","badge":"https://syntology.ai/api/ran/2005.05859.svg"},"claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/awesomelemon/encas","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/human-analysis/neural-architecture-transfer","reach":null}],"summary":{"unverified":1},"n_samples":1,"n_ran":0,"n_constructed":0,"n_ran_checked":0,"n_instrument":0,"n_unverified":1,"n_instrument_is":"failures of Syntology's instrument, not of the code","every_run_is_an_instrument_failure":false,"key_notes":{"n_ran_checked":"legacy name, kept unchanged so existing readers do not break: it counts the samples that ran with no instrument failure (honoured, violated, and ran with no contract checked); it does not mean a contract was checked, and the pages print it as 'K with no instrument failure', not 'K checked'","n_constructed":"a sub-count of the samples that ran, never subtracted from them and never a failure: an executed sample whose run returned an instance of its own class (fixture_out_type equals the entry name): the run built an object and did not compute a result (Syntology's RAN record, counts.constructed)"},"by_repo_kind":{"official":{"samples":1,"ran":0,"n_ran_checked":0,"n_instrument":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":1,"inline_ok_is":"inline_ok is true only when both licence gates allow Syntology to serve this code's text: this paper's link (inline_ok_link, from the licence of this paper's copy) and the stored copy of the same code that Syntology serves, which may have been first harvested from another repository. withheld_clause is Syntology's own wording of which gate keeps the text a pointer and why, and is null when inline_ok is true. This site shows no code text either way; n_pointer_only_for_licence counts samples with inline_ok false (both gates), n_pointer_only_link_allows those of them whose inline_ok_link is true","n_pointer_only_link_allows":0,"code_page_min_papers":5,"code_page_is":"a page for the entry name across every paper's harvested code; the name is a string, not an identity","census_is":"Syntology's function-name census ran implementations with this entry name on one shared input per bucket (the positional rank, kind and dtype of each array argument) and counted the distinct output digests (sha256 of the output rounded to 6 decimals). implementations and papers count the compared samples; distinct_outputs counts digests across all buckets; n_buckets is how many shared inputs were used. It is not a statement that any two samples compute one thing, and the name is a string, not an identity.","samples":[{"code_sha256":"7028564474e3fefd","code_sha256_prefix":"7028564474e3fefd","entry":"drop_connect","repo":"human-analysis/neural-architecture-transfer","repo_kind":"official","path":"codebase/networks/natnet.py","file_url":"https://github.com/human-analysis/neural-architecture-transfer/blob/HEAD/codebase/networks/natnet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"code_page":"/code/drop-connect","census":{"url":"/census/drop-connect","implementations":9,"papers":7,"distinct_outputs":4,"n_buckets":2,"bucket_key":null,"bucket_index":null,"bucket_label":null},"census_bucket_fragment":null,"licence":"NONE","inline_ok":false,"inline_ok_link":false,"withheld_clause":"GitHub detected no licence in this paper's repository, human-analysis/neural-architecture-transfer when it was checked","mcp_get_code":{"code_sha256":"7028564474e3fefd"}}]},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"d6225f4d15d60eb4bc2362e94579443ad5cf3e161c690e78f6339a8f5d2c701a","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}