{"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/aggregating-nested-transformers","title":"Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding","arxiv_id":"2105.12723","date":"2021-05-26","proceeding":null,"authors":["Zizhao Zhang","Han Zhang","Long Zhao","Ting Chen","Sercan O. Arik","Tomas Pfister"],"abstract":"Hierarchical structures are popular in recent vision transformers, however, they require sophisticated designs and massive datasets to work well. In this paper, we explore the idea of nesting basic local transformers on non-overlapping image blocks and aggregating them in a hierarchical way. We find that the block aggregation function plays a critical role in enabling cross-block non-local information communication. This observation leads us to design a simplified architecture that requires minor code changes upon the original vision transformer. The benefits of the proposed judiciously-selected design are threefold: (1) NesT converges faster and requires much less training data to achieve good generalization on both ImageNet and small datasets like CIFAR; (2) when extending our key ideas to image generation, NesT leads to a strong decoder that is 8$\\times$ faster than previous transformer-based generators; and (3) we show that decoupling the feature learning and abstraction processes via this nested hierarchy in our design enables constructing a novel method (named GradCAT) for visually interpreting the learned model. Source code is available https://github.com/google-research/nested-transformer.","url_abs":"https://arxiv.org/abs/2105.12723v4","url_pdf":"https://arxiv.org/pdf/2105.12723v4.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":"aggregating-nested-transformers","repo_url":"https://github.com/google-research/nested-transformer","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"jax","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"aggregating-nested-transformers","repo_url":"https://github.com/ahmedelmahy/myownvit","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"aggregating-nested-transformers","repo_url":"https://github.com/conceptofmind/NesT-flax","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"jax","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"aggregating-nested-transformers","repo_url":"https://github.com/freder-chen/vitp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"aggregating-nested-transformers","repo_url":"https://github.com/rwightman/pytorch-image-models","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"aggregating-nested-transformers","repo_url":"https://github.com/ttt496/vit-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-generation","task_name":"Image Generation"}],"methods":[{"method_slug":"attention","method_name":"Attention"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"nest","method_name":"NesT"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"vision-transformer","method_name":"Vision Transformer"}],"datasets_introduced":[],"methods_introduced":[{"slug":"nest","name":"NesT","full_name":"NesT"}],"results":[{"leaderboard":"/sota/image-classification-on-cifar-10","task":"Image Classification","dataset":"CIFAR-10","model":"Transformer local-attention (NesT-B)","rank_in_archive_order":90,"of":265,"metrics":{"Percentage correct":"97.2"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-100","task":"Image Classification","dataset":"CIFAR-100","model":"Transformer local-attention (NesT-B)","rank_in_archive_order":103,"of":211,"metrics":{"Percentage correct":"82.56"},"uses_additional_data":true},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"Transformer local-attention (NesT-B)","rank_in_archive_order":390,"of":1060,"metrics":{"GFLOPs":"17.9","Number of params":"68M","Top 1 Accuracy":"83.8%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"Transformer local-attention (NesT-S)","rank_in_archive_order":441,"of":1060,"metrics":{"GFLOPs":"10.4","Number of params":"38M","Top 1 Accuracy":"83.3%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-imagenet","task":"Image Classification","dataset":"ImageNet","model":"Transformer local-attention (NesT-T)","rank_in_archive_order":635,"of":1060,"metrics":{"GFLOPs":"5.8","Number of params":"17M","Top 1 Accuracy":"81.5%"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2105.12723","atlas_url":"https://app.syntology.ai/?focus=2105.12723","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2105.12723"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","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/rwightman/pytorch-image-models","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/freder-chen/vitp","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ttt496/vit-pytorch","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/conceptofmind/NesT-flax","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/google-research/nested-transformer","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ahmedelmahy/myownvit","reach":null}],"summary":{"ran_honours":1,"ran":7,"ran_violates":1,"ran_fixture":1,"unverified":16},"by_repo_kind":{"official":{"samples":12,"ran":1,"repositories":1},"listed":{"samples":14,"ran":9,"repositories":2}},"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":0,"samples":[{"code_sha256_prefix":"bd3578806ecf5ac2","entry":"adjust_learning_rate","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/lr_sched.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/lr_sched.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"bd3578806ecf5ac2"}},{"code_sha256_prefix":"d83a787ef8a41d29","entry":"adjust_mask_ratio","repo":"freder-chen/vitp","repo_kind":"listed","path":"engine/scratch.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/engine/scratch.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"d83a787ef8a41d29"}},{"code_sha256_prefix":"3dc19396537db789","entry":"all_reduce_mean","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/misc.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/misc.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3dc19396537db789"}},{"code_sha256_prefix":"b86ebb0fa1909e0d","entry":"build_transform","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/datasets.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/datasets.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"b86ebb0fa1909e0d"}},{"code_sha256_prefix":"1f7c550e65142ef8","entry":"cast_tuple","repo":"conceptofmind/NesT-flax","repo_kind":"listed","path":"nest_flax.py/nest_flax.py","file_url":"https://github.com/conceptofmind/NesT-flax/blob/HEAD/nest_flax.py/nest_flax.py","link_basis":"harvester_set","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1f7c550e65142ef8"}},{"code_sha256_prefix":"202b055d01574c6b","entry":"generate_list","repo":"freder-chen/vitp","repo_kind":"listed","path":"models/swelling_vit.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/models/swelling_vit.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"202b055d01574c6b"}},{"code_sha256_prefix":"ffc38a27fd9393ba","entry":"get_hyper","repo":"google-research/nested-transformer","repo_kind":"official","path":"configs/resnet50_imagenet.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/configs/resnet50_imagenet.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"ffc38a27fd9393ba"}},{"code_sha256_prefix":"89edf5372214b57d","entry":"get_layer_id_for_vit","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/lr_decay.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/lr_decay.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"89edf5372214b57d"}},{"code_sha256_prefix":"8baf25cef41cba16","entry":"param_groups_lrd","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/lr_decay.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/lr_decay.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"8baf25cef41cba16"}},{"code_sha256_prefix":"191b2735955d18f4","entry":"resize_pos_embed","repo":"freder-chen/vitp","repo_kind":"listed","path":"models/swelling_vit.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/models/swelling_vit.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"191b2735955d18f4"}},{"code_sha256_prefix":"3523de76ecdc1efc","entry":"activation","repo":"google-research/nested-transformer","repo_kind":"official","path":"models/wide_resnet.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/models/wide_resnet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3523de76ecdc1efc"}},{"code_sha256_prefix":"08baacad242c29f2","entry":"apply_label_smoothing","repo":"google-research/nested-transformer","repo_kind":"official","path":"libml/losses.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/libml/losses.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"08baacad242c29f2"}},{"code_sha256_prefix":"d6d5db4678e1915c","entry":"apply_weights","repo":"google-research/nested-transformer","repo_kind":"official","path":"libml/losses.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/libml/losses.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"d6d5db4678e1915c"}},{"code_sha256_prefix":"11d1d6883016bc0a","entry":"build_dataset","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/datasets.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/datasets.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"11d1d6883016bc0a"}},{"code_sha256_prefix":"6d8a1f8a0697180f","entry":"checkpoint_filter_fn","repo":"freder-chen/vitp","repo_kind":"listed","path":"models/swelling_vit.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/models/swelling_vit.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6d8a1f8a0697180f"}},{"code_sha256_prefix":"796c65138aa563c7","entry":"create_model","repo":"google-research/nested-transformer","repo_kind":"official","path":"models/resnet_v1.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/models/resnet_v1.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"796c65138aa563c7"}},{"code_sha256_prefix":"13ce18ddacc0988a","entry":"create_model","repo":"google-research/nested-transformer","repo_kind":"official","path":"models/wide_resnet.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/models/wide_resnet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"13ce18ddacc0988a"}},{"code_sha256_prefix":"6f6b06fdb24a7472","entry":"dense_layer_init_fn","repo":"google-research/nested-transformer","repo_kind":"official","path":"models/wide_resnet.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/models/wide_resnet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6f6b06fdb24a7472"}},{"code_sha256_prefix":"12035a2f77d8016c","entry":"get_1d_sincos_pos_embed_from_grid","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/pos_embed.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/pos_embed.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"12035a2f77d8016c"}},{"code_sha256_prefix":"3185afc3e87293ed","entry":"get_2d_sincos_pos_embed","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/pos_embed.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/pos_embed.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3185afc3e87293ed"}},{"code_sha256_prefix":"f10004e059714d42","entry":"get_2d_sincos_pos_embed_from_grid","repo":"freder-chen/vitp","repo_kind":"listed","path":"util/pos_embed.py","file_url":"https://github.com/freder-chen/vitp/blob/HEAD/util/pos_embed.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f10004e059714d42"}},{"code_sha256_prefix":"d04f187058cb2a9c","entry":"random_apply","repo":"google-research/nested-transformer","repo_kind":"official","path":"augment/color_util.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/augment/color_util.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"d04f187058cb2a9c"}},{"code_sha256_prefix":"e1aaf77a8784526a","entry":"random_brightness","repo":"google-research/nested-transformer","repo_kind":"official","path":"augment/color_util.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/augment/color_util.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"e1aaf77a8784526a"}},{"code_sha256_prefix":"103f46a75894451d","entry":"register","repo":"google-research/nested-transformer","repo_kind":"official","path":"models/nest_net.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/models/nest_net.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"103f46a75894451d"}},{"code_sha256_prefix":"49c827442d2fb761","entry":"to_grayscale","repo":"google-research/nested-transformer","repo_kind":"official","path":"augment/color_util.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/augment/color_util.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"49c827442d2fb761"}},{"code_sha256_prefix":"a70cd69707274f4d","entry":"weighted_unnormalized_softmax_cross_entropy","repo":"google-research/nested-transformer","repo_kind":"official","path":"libml/losses.py","file_url":"https://github.com/google-research/nested-transformer/blob/HEAD/libml/losses.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"a70cd69707274f4d"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}