{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XHZXEQVU2OWDT2WLR75KRJGFJS","short_pith_number":"pith:XHZXEQVU","schema_version":"1.0","canonical_sha256":"b9f37242b4d3ac39eacb8ffaa8a4c54c91eec8bf23ad79c00af8c6707dc9a2df","source":{"kind":"arxiv","id":"2404.10518","version":2},"attestation_state":"computed","paper":{"title":"MobileNetV4 -- Universal Models for the Mobile Ecosystem","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Andrew Howard, Berkin Akin, Chas Leichner, Chengxi Ye, Colby Banbury, Danfeng Qin, Daniele Moro, Fan Yang, Manolis Delakis, Marco Fornoni, Shixin Luo, Tenghui Zhu, Vaibhav Aggarwal, Weijun Wang","submitted_at":"2024-04-16T12:41:25Z","abstract_excerpt":"We present the latest generation of MobileNets, known as MobileNetV4 (MNv4), featuring universally efficient architecture designs for mobile devices. At its core, we introduce the Universal Inverted Bottleneck (UIB) search block, a unified and flexible structure that merges Inverted Bottleneck (IB), ConvNext, Feed Forward Network (FFN), and a novel Extra Depthwise (ExtraDW) variant. Alongside UIB, we present Mobile MQA, an attention block tailored for mobile accelerators, delivering a significant 39% speedup. An optimized neural architecture search (NAS) recipe is also introduced which improve"},"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":"2404.10518","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-04-16T12:41:25Z","cross_cats_sorted":[],"title_canon_sha256":"911155c9f040664408378ef42b1e68a110da3ce8f309a237d693103eb3be266a","abstract_canon_sha256":"564e70a3afb38cef58abc2b30f16de7172037f93ff2da2f2752ee200d635a4c7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:13:01.263254Z","signature_b64":"BSb7nTceH8lOze5QzrbbZKmrwtCkj5Bfs3wklXx8Tjlv31JOepyG42dRARkJgd8QFmunwnFVpnoxTi3/9IhaDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b9f37242b4d3ac39eacb8ffaa8a4c54c91eec8bf23ad79c00af8c6707dc9a2df","last_reissued_at":"2026-07-05T09:13:01.262766Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:13:01.262766Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MobileNetV4 -- Universal Models for the Mobile Ecosystem","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Andrew Howard, Berkin Akin, Chas Leichner, Chengxi Ye, Colby Banbury, Danfeng Qin, Daniele Moro, Fan Yang, Manolis Delakis, Marco Fornoni, Shixin Luo, Tenghui Zhu, Vaibhav Aggarwal, Weijun Wang","submitted_at":"2024-04-16T12:41:25Z","abstract_excerpt":"We present the latest generation of MobileNets, known as MobileNetV4 (MNv4), featuring universally efficient architecture designs for mobile devices. At its core, we introduce the Universal Inverted Bottleneck (UIB) search block, a unified and flexible structure that merges Inverted Bottleneck (IB), ConvNext, Feed Forward Network (FFN), and a novel Extra Depthwise (ExtraDW) variant. Alongside UIB, we present Mobile MQA, an attention block tailored for mobile accelerators, delivering a significant 39% speedup. An optimized neural architecture search (NAS) recipe is also introduced which improve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.10518","kind":"arxiv","version":2},"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/2404.10518/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":"2404.10518","created_at":"2026-07-05T09:13:01.262824+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.10518v2","created_at":"2026-07-05T09:13:01.262824+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.10518","created_at":"2026-07-05T09:13:01.262824+00:00"},{"alias_kind":"pith_short_12","alias_value":"XHZXEQVU2OWD","created_at":"2026-07-05T09:13:01.262824+00:00"},{"alias_kind":"pith_short_16","alias_value":"XHZXEQVU2OWDT2WL","created_at":"2026-07-05T09:13:01.262824+00:00"},{"alias_kind":"pith_short_8","alias_value":"XHZXEQVU","created_at":"2026-07-05T09:13:01.262824+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19195","citing_title":"Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17260","citing_title":"LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLMs","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17260","citing_title":"LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLMs","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10229","citing_title":"VPD-100K: Towards Generalizable and Fine-grained Visual Privacy Protection","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS","json":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS.json","graph_json":"https://pith.science/api/pith-number/XHZXEQVU2OWDT2WLR75KRJGFJS/graph.json","events_json":"https://pith.science/api/pith-number/XHZXEQVU2OWDT2WLR75KRJGFJS/events.json","paper":"https://pith.science/paper/XHZXEQVU"},"agent_actions":{"view_html":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS","download_json":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS.json","view_paper":"https://pith.science/paper/XHZXEQVU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.10518&json=true","fetch_graph":"https://pith.science/api/pith-number/XHZXEQVU2OWDT2WLR75KRJGFJS/graph.json","fetch_events":"https://pith.science/api/pith-number/XHZXEQVU2OWDT2WLR75KRJGFJS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS/action/storage_attestation","attest_author":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS/action/author_attestation","sign_citation":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS/action/citation_signature","submit_replication":"https://pith.science/pith/XHZXEQVU2OWDT2WLR75KRJGFJS/action/replication_record"}},"created_at":"2026-07-05T09:13:01.262824+00:00","updated_at":"2026-07-05T09:13:01.262824+00:00"}