{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:S7Q2N5R44MQYICTLXNQ4UR7WXR","short_pith_number":"pith:S7Q2N5R4","schema_version":"1.0","canonical_sha256":"97e1a6f63ce321840a6bbb61ca47f6bc4f070e30110c09841eec7d85c55ad239","source":{"kind":"arxiv","id":"2101.03697","version":3},"attestation_state":"computed","paper":{"title":"RepVGG: Making VGG-style ConvNets Great Again","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Guiguang Ding, Jian Sun, Jungong Han, Ningning Ma, Xiangyu Zhang, Xiaohan Ding","submitted_at":"2021-01-11T04:46:11Z","abstract_excerpt":"We present a simple but powerful architecture of convolutional neural network, which has a VGG-like inference-time body composed of nothing but a stack of 3x3 convolution and ReLU, while the training-time model has a multi-branch topology. Such decoupling of the training-time and inference-time architecture is realized by a structural re-parameterization technique so that the model is named RepVGG. On ImageNet, RepVGG reaches over 80% top-1 accuracy, which is the first time for a plain model, to the best of our knowledge. On NVIDIA 1080Ti GPU, RepVGG models run 83% faster than ResNet-50 or 101"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2101.03697","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-01-11T04:46:11Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"f38cdcc3be4893b9f46165493a8d1a88a2ac77f1b84ee2a74a08b0cb72a33bc2","abstract_canon_sha256":"2586c9e1e7a47ceb0349954a2a0891bf34fbb4cf7de84a530441def23c16e2b6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:26:54.589452Z","signature_b64":"AJoCDZMAGqTMv7AalTsjdYINnlcrKgWFTaGzCsbFyzJ96N3jKKQQduatJuuqsW3QVgxRouMJo5m2jargG8IiDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"97e1a6f63ce321840a6bbb61ca47f6bc4f070e30110c09841eec7d85c55ad239","last_reissued_at":"2026-07-05T02:26:54.588906Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:26:54.588906Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RepVGG: Making VGG-style ConvNets Great Again","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Guiguang Ding, Jian Sun, Jungong Han, Ningning Ma, Xiangyu Zhang, Xiaohan Ding","submitted_at":"2021-01-11T04:46:11Z","abstract_excerpt":"We present a simple but powerful architecture of convolutional neural network, which has a VGG-like inference-time body composed of nothing but a stack of 3x3 convolution and ReLU, while the training-time model has a multi-branch topology. Such decoupling of the training-time and inference-time architecture is realized by a structural re-parameterization technique so that the model is named RepVGG. On ImageNet, RepVGG reaches over 80% top-1 accuracy, which is the first time for a plain model, to the best of our knowledge. On NVIDIA 1080Ti GPU, RepVGG models run 83% faster than ResNet-50 or 101"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.03697","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2101.03697/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2101.03697","created_at":"2026-07-05T02:26:54.588966+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.03697v3","created_at":"2026-07-05T02:26:54.588966+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.03697","created_at":"2026-07-05T02:26:54.588966+00:00"},{"alias_kind":"pith_short_12","alias_value":"S7Q2N5R44MQY","created_at":"2026-07-05T02:26:54.588966+00:00"},{"alias_kind":"pith_short_16","alias_value":"S7Q2N5R44MQYICTL","created_at":"2026-07-05T02:26:54.588966+00:00"},{"alias_kind":"pith_short_8","alias_value":"S7Q2N5R4","created_at":"2026-07-05T02:26:54.588966+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.14946","citing_title":"Reliable Evaluation of Attribution Maps in CNNs: A Perturbation-Based Approach","ref_index":58,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR","json":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR.json","graph_json":"https://pith.science/api/pith-number/S7Q2N5R44MQYICTLXNQ4UR7WXR/graph.json","events_json":"https://pith.science/api/pith-number/S7Q2N5R44MQYICTLXNQ4UR7WXR/events.json","paper":"https://pith.science/paper/S7Q2N5R4"},"agent_actions":{"view_html":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR","download_json":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR.json","view_paper":"https://pith.science/paper/S7Q2N5R4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.03697&json=true","fetch_graph":"https://pith.science/api/pith-number/S7Q2N5R44MQYICTLXNQ4UR7WXR/graph.json","fetch_events":"https://pith.science/api/pith-number/S7Q2N5R44MQYICTLXNQ4UR7WXR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR/action/storage_attestation","attest_author":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR/action/author_attestation","sign_citation":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR/action/citation_signature","submit_replication":"https://pith.science/pith/S7Q2N5R44MQYICTLXNQ4UR7WXR/action/replication_record"}},"created_at":"2026-07-05T02:26:54.588966+00:00","updated_at":"2026-07-05T02:26:54.588966+00:00"}