{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RLI34DUUBE3LTKDC4RZYVRW337","short_pith_number":"pith:RLI34DUU","schema_version":"1.0","canonical_sha256":"8ad1be0e940936b9a862e4738ac6dbdfcf40996872643958fff1354a0d0d076a","source":{"kind":"arxiv","id":"2407.12893","version":1},"attestation_state":"computed","paper":{"title":"Hybrid Dynamic Pruning: A Pathway to Efficient Transformer Inference","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Baker Mohammad, Ghada Alsuhli, Ghadeer Jaradat, Hani Saleh, Mahmoud Al-Qutayri, Mohammed Tolba, Thanos Stouraitis","submitted_at":"2024-07-17T11:15:16Z","abstract_excerpt":"In the world of deep learning, Transformer models have become very significant, leading to improvements in many areas from understanding language to recognizing images, covering a wide range of applications. Despite their success, the deployment of these models in real-time applications, particularly on edge devices, poses significant challenges due to their quadratic computational intensity and memory demands. To overcome these challenges we introduce a novel Hybrid Dynamic Pruning (HDP), an efficient algorithm-architecture co-design approach that accelerates transformers using head sparsity,"},"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":"2407.12893","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-17T11:15:16Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7174ae2a78d6e0ef03e7e596cc295e3345b5a477e5d3d8215209896d71bcbe68","abstract_canon_sha256":"e81a49d537aed7ffb49b798e17733d256dde287da7100cc91b82afaf95177318"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:45:13.240216Z","signature_b64":"IzI3ddeH5QTWNPTnWZmYULRT47pohbwMKYuS+P8KtZvAXyjfgYPf4gU03KbQKCtdoNn+CPdKY5QvxhixBW/5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8ad1be0e940936b9a862e4738ac6dbdfcf40996872643958fff1354a0d0d076a","last_reissued_at":"2026-07-05T08:45:13.239698Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:45:13.239698Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hybrid Dynamic Pruning: A Pathway to Efficient Transformer Inference","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Baker Mohammad, Ghada Alsuhli, Ghadeer Jaradat, Hani Saleh, Mahmoud Al-Qutayri, Mohammed Tolba, Thanos Stouraitis","submitted_at":"2024-07-17T11:15:16Z","abstract_excerpt":"In the world of deep learning, Transformer models have become very significant, leading to improvements in many areas from understanding language to recognizing images, covering a wide range of applications. Despite their success, the deployment of these models in real-time applications, particularly on edge devices, poses significant challenges due to their quadratic computational intensity and memory demands. To overcome these challenges we introduce a novel Hybrid Dynamic Pruning (HDP), an efficient algorithm-architecture co-design approach that accelerates transformers using head sparsity,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.12893","kind":"arxiv","version":1},"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/2407.12893/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":"2407.12893","created_at":"2026-07-05T08:45:13.239757+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.12893v1","created_at":"2026-07-05T08:45:13.239757+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.12893","created_at":"2026-07-05T08:45:13.239757+00:00"},{"alias_kind":"pith_short_12","alias_value":"RLI34DUUBE3L","created_at":"2026-07-05T08:45:13.239757+00:00"},{"alias_kind":"pith_short_16","alias_value":"RLI34DUUBE3LTKDC","created_at":"2026-07-05T08:45:13.239757+00:00"},{"alias_kind":"pith_short_8","alias_value":"RLI34DUU","created_at":"2026-07-05T08:45:13.239757+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.22583","citing_title":"Adaptive Head Budgeting for Efficient Multi-Head Attention","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337","json":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337.json","graph_json":"https://pith.science/api/pith-number/RLI34DUUBE3LTKDC4RZYVRW337/graph.json","events_json":"https://pith.science/api/pith-number/RLI34DUUBE3LTKDC4RZYVRW337/events.json","paper":"https://pith.science/paper/RLI34DUU"},"agent_actions":{"view_html":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337","download_json":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337.json","view_paper":"https://pith.science/paper/RLI34DUU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.12893&json=true","fetch_graph":"https://pith.science/api/pith-number/RLI34DUUBE3LTKDC4RZYVRW337/graph.json","fetch_events":"https://pith.science/api/pith-number/RLI34DUUBE3LTKDC4RZYVRW337/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337/action/storage_attestation","attest_author":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337/action/author_attestation","sign_citation":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337/action/citation_signature","submit_replication":"https://pith.science/pith/RLI34DUUBE3LTKDC4RZYVRW337/action/replication_record"}},"created_at":"2026-07-05T08:45:13.239757+00:00","updated_at":"2026-07-05T08:45:13.239757+00:00"}