{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4M4TCNHBHSEUB5UF6QSEF5ZCAR","short_pith_number":"pith:4M4TCNHB","schema_version":"1.0","canonical_sha256":"e3393134e13c8940f685f42442f722044a1b09e4470684377e48fd308bfca6c1","source":{"kind":"arxiv","id":"2605.23445","version":1},"attestation_state":"computed","paper":{"title":"DFSAttn: Dynamic Fine-grained Sparse Attention for Efficient Video Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jie Hu, Kun Yuan, Yutong He, Zixiang Gao","submitted_at":"2026-05-22T09:58:23Z","abstract_excerpt":"Diffusion transformers have achieved remarkable success in high-quality video generation, yet their reliance on spatiotemporal 3D full attention incurs prohibitive computational cost due to the quadratic complexity of attention. Block sparse attention is a common approach to mitigate this by focusing computation on important regions. However, attention maps in DiTs exhibit inherently dynamic and fine-grained sparsity, which causes existing block sparse attention methods to degrade significantly in quality, especially at high sparsity ratios. In this paper, we revisit block sparse attention and"},"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":"2605.23445","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-22T09:58:23Z","cross_cats_sorted":[],"title_canon_sha256":"86cfe2b688b7b8dc69d7a604dd1e56c50bffa3032d0119369853afc018226ce9","abstract_canon_sha256":"33f685a5c869087337234d2477870a1ea796149175ca1e07bb892fbefb3fe011"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:01:54.577218Z","signature_b64":"ph4xMtLbWt1/OuQ3BJZe3rJAnYbZt/7S+LZZWx/RCNExPVzw7rD7RS5K5dJFiYC4hDosnk2eZq06MdmjR+4wCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e3393134e13c8940f685f42442f722044a1b09e4470684377e48fd308bfca6c1","last_reissued_at":"2026-05-25T02:01:54.576627Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:01:54.576627Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DFSAttn: Dynamic Fine-grained Sparse Attention for Efficient Video Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jie Hu, Kun Yuan, Yutong He, Zixiang Gao","submitted_at":"2026-05-22T09:58:23Z","abstract_excerpt":"Diffusion transformers have achieved remarkable success in high-quality video generation, yet their reliance on spatiotemporal 3D full attention incurs prohibitive computational cost due to the quadratic complexity of attention. Block sparse attention is a common approach to mitigate this by focusing computation on important regions. However, attention maps in DiTs exhibit inherently dynamic and fine-grained sparsity, which causes existing block sparse attention methods to degrade significantly in quality, especially at high sparsity ratios. In this paper, we revisit block sparse attention and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23445","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/2605.23445/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":"2605.23445","created_at":"2026-05-25T02:01:54.576735+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.23445v1","created_at":"2026-05-25T02:01:54.576735+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23445","created_at":"2026-05-25T02:01:54.576735+00:00"},{"alias_kind":"pith_short_12","alias_value":"4M4TCNHBHSEU","created_at":"2026-05-25T02:01:54.576735+00:00"},{"alias_kind":"pith_short_16","alias_value":"4M4TCNHBHSEUB5UF","created_at":"2026-05-25T02:01:54.576735+00:00"},{"alias_kind":"pith_short_8","alias_value":"4M4TCNHB","created_at":"2026-05-25T02:01:54.576735+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR","json":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR.json","graph_json":"https://pith.science/api/pith-number/4M4TCNHBHSEUB5UF6QSEF5ZCAR/graph.json","events_json":"https://pith.science/api/pith-number/4M4TCNHBHSEUB5UF6QSEF5ZCAR/events.json","paper":"https://pith.science/paper/4M4TCNHB"},"agent_actions":{"view_html":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR","download_json":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR.json","view_paper":"https://pith.science/paper/4M4TCNHB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.23445&json=true","fetch_graph":"https://pith.science/api/pith-number/4M4TCNHBHSEUB5UF6QSEF5ZCAR/graph.json","fetch_events":"https://pith.science/api/pith-number/4M4TCNHBHSEUB5UF6QSEF5ZCAR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR/action/storage_attestation","attest_author":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR/action/author_attestation","sign_citation":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR/action/citation_signature","submit_replication":"https://pith.science/pith/4M4TCNHBHSEUB5UF6QSEF5ZCAR/action/replication_record"}},"created_at":"2026-05-25T02:01:54.576735+00:00","updated_at":"2026-05-25T02:01:54.576735+00:00"}