{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4CC5WIBQZ5AZ4VWVZIEJBAHTSK","short_pith_number":"pith:4CC5WIBQ","schema_version":"1.0","canonical_sha256":"e085db2030cf419e56d5ca089080f392ad64be9d3a2896efd26d5c71df48f53e","source":{"kind":"arxiv","id":"2502.06155","version":2},"attestation_state":"computed","paper":{"title":"Efficient-vDiT: Efficient Video Diffusion Transformers With Attention Tile","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dacheng Li, Hangliang Ding, Hao Zhang, Ion Stoica, Peiyuan Zhang, Runlong Su, Zhijie Deng","submitted_at":"2025-02-10T05:00:56Z","abstract_excerpt":"Despite the promise of synthesizing high-fidelity videos, Diffusion Transformers (DiTs) with 3D full attention suffer from expensive inference due to the complexity of attention computation and numerous sampling steps. For example, the popular Open-Sora-Plan model consumes more than 9 minutes for generating a single video of 29 frames. This paper addresses the inefficiency issue from two aspects: 1) Prune the 3D full attention based on the redundancy within video data; We identify a prevalent tile-style repetitive pattern in the 3D attention maps for video data, and advocate a new family of sp"},"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":"2502.06155","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-02-10T05:00:56Z","cross_cats_sorted":[],"title_canon_sha256":"b8523b67386239610f84b9f5c688f540b82520df1387d54ec1d9c7b6c35417e7","abstract_canon_sha256":"df133597f5402f432e9d95adea92fb643d2cd269928913117d6ea314ea53a38a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:15:26.021569Z","signature_b64":"XIzbAJ7Wy+83p3l9661xcjKhwa37muAxTaxs4E3crqvEV1WCJBwXrVQcOfYGi0f/Jti6OBEUSjA2clF+qe59DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e085db2030cf419e56d5ca089080f392ad64be9d3a2896efd26d5c71df48f53e","last_reissued_at":"2026-07-05T10:15:26.021085Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:15:26.021085Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient-vDiT: Efficient Video Diffusion Transformers With Attention Tile","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dacheng Li, Hangliang Ding, Hao Zhang, Ion Stoica, Peiyuan Zhang, Runlong Su, Zhijie Deng","submitted_at":"2025-02-10T05:00:56Z","abstract_excerpt":"Despite the promise of synthesizing high-fidelity videos, Diffusion Transformers (DiTs) with 3D full attention suffer from expensive inference due to the complexity of attention computation and numerous sampling steps. For example, the popular Open-Sora-Plan model consumes more than 9 minutes for generating a single video of 29 frames. This paper addresses the inefficiency issue from two aspects: 1) Prune the 3D full attention based on the redundancy within video data; We identify a prevalent tile-style repetitive pattern in the 3D attention maps for video data, and advocate a new family of sp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.06155","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/2502.06155/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":"2502.06155","created_at":"2026-07-05T10:15:26.021139+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.06155v2","created_at":"2026-07-05T10:15:26.021139+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.06155","created_at":"2026-07-05T10:15:26.021139+00:00"},{"alias_kind":"pith_short_12","alias_value":"4CC5WIBQZ5AZ","created_at":"2026-07-05T10:15:26.021139+00:00"},{"alias_kind":"pith_short_16","alias_value":"4CC5WIBQZ5AZ4VWV","created_at":"2026-07-05T10:15:26.021139+00:00"},{"alias_kind":"pith_short_8","alias_value":"4CC5WIBQ","created_at":"2026-07-05T10:15:26.021139+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.21002","citing_title":"SURF: Signature-Retained Fast Video Generation","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2603.09721","citing_title":"FrameDiT: Diffusion Transformer with Matrix Attention for Efficient Video Generation","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15911","citing_title":"Efficient Video Diffusion Models: Advancements and Challenges","ref_index":253,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK","json":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK.json","graph_json":"https://pith.science/api/pith-number/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/graph.json","events_json":"https://pith.science/api/pith-number/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/events.json","paper":"https://pith.science/paper/4CC5WIBQ"},"agent_actions":{"view_html":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK","download_json":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK.json","view_paper":"https://pith.science/paper/4CC5WIBQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.06155&json=true","fetch_graph":"https://pith.science/api/pith-number/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/graph.json","fetch_events":"https://pith.science/api/pith-number/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/action/storage_attestation","attest_author":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/action/author_attestation","sign_citation":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/action/citation_signature","submit_replication":"https://pith.science/pith/4CC5WIBQZ5AZ4VWVZIEJBAHTSK/action/replication_record"}},"created_at":"2026-07-05T10:15:26.021139+00:00","updated_at":"2026-07-05T10:15:26.021139+00:00"}