{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5S2ZWVWMV5BJRQGM56VYRAFMGZ","short_pith_number":"pith:5S2ZWVWM","schema_version":"1.0","canonical_sha256":"ecb59b56ccaf4298c0ccefab8880ac364b63d929a6800bc2208d06c10a6b7c72","source":{"kind":"arxiv","id":"2506.04648","version":2},"attestation_state":"computed","paper":{"title":"FPSAttention: Training-Aware FP8 and Sparsity Co-Design for Fast Video Diffusion","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Akide Liu, Bohan Zhuang, Fan Wang, Gholamreza Haffari, Jiahao He, Jiasheng Tang, Jichao Wu, Mingyang Yang, Weihua Chen, Xuehai Bai, Yizeng Han, Yuanjie Xing, Yuanyu He, Zeyu Zhang, Zhexin Li","submitted_at":"2025-06-05T05:30:30Z","abstract_excerpt":"Diffusion generative models have become the standard for producing high-quality, coherent video content, yet their slow inference speeds and high computational demands hinder practical deployment. Although both quantization and sparsity can independently accelerate inference while maintaining generation quality, naively combining these techniques in existing training-free approaches leads to significant performance degradation due to the lack of joint optimization. We introduce FPSAttention, a novel training-aware co-design of FP8 quantization and sparsity for video generation, with a focus on"},"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":"2506.04648","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-06-05T05:30:30Z","cross_cats_sorted":[],"title_canon_sha256":"3e62977e9d437c4cc77f9fe9102f91ebd338611f9a97c0022803043dcb2320d0","abstract_canon_sha256":"8c50685dcccc6f634a08be08a5f28e8c9a60655fd429f5d970ea42d1b637b36e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:00.520412Z","signature_b64":"Znn/RqQMgd8Qrld1Zdziyh0q9RipLdg3dcgBG1paswwPvUSFyTbRuwxTvQxb9WCgajlPiBtkoNcQ8OTNPvM3CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ecb59b56ccaf4298c0ccefab8880ac364b63d929a6800bc2208d06c10a6b7c72","last_reissued_at":"2026-07-05T11:17:00.519911Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:00.519911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FPSAttention: Training-Aware FP8 and Sparsity Co-Design for Fast Video Diffusion","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Akide Liu, Bohan Zhuang, Fan Wang, Gholamreza Haffari, Jiahao He, Jiasheng Tang, Jichao Wu, Mingyang Yang, Weihua Chen, Xuehai Bai, Yizeng Han, Yuanjie Xing, Yuanyu He, Zeyu Zhang, Zhexin Li","submitted_at":"2025-06-05T05:30:30Z","abstract_excerpt":"Diffusion generative models have become the standard for producing high-quality, coherent video content, yet their slow inference speeds and high computational demands hinder practical deployment. Although both quantization and sparsity can independently accelerate inference while maintaining generation quality, naively combining these techniques in existing training-free approaches leads to significant performance degradation due to the lack of joint optimization. We introduce FPSAttention, a novel training-aware co-design of FP8 quantization and sparsity for video generation, with a focus on"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.04648","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/2506.04648/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":"2506.04648","created_at":"2026-07-05T11:17:00.519968+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.04648v2","created_at":"2026-07-05T11:17:00.519968+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.04648","created_at":"2026-07-05T11:17:00.519968+00:00"},{"alias_kind":"pith_short_12","alias_value":"5S2ZWVWMV5BJ","created_at":"2026-07-05T11:17:00.519968+00:00"},{"alias_kind":"pith_short_16","alias_value":"5S2ZWVWMV5BJRQGM","created_at":"2026-07-05T11:17:00.519968+00:00"},{"alias_kind":"pith_short_8","alias_value":"5S2ZWVWM","created_at":"2026-07-05T11:17:00.519968+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.15911","citing_title":"Efficient Video Diffusion Models: Advancements and Challenges","ref_index":78,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ","json":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ.json","graph_json":"https://pith.science/api/pith-number/5S2ZWVWMV5BJRQGM56VYRAFMGZ/graph.json","events_json":"https://pith.science/api/pith-number/5S2ZWVWMV5BJRQGM56VYRAFMGZ/events.json","paper":"https://pith.science/paper/5S2ZWVWM"},"agent_actions":{"view_html":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ","download_json":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ.json","view_paper":"https://pith.science/paper/5S2ZWVWM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.04648&json=true","fetch_graph":"https://pith.science/api/pith-number/5S2ZWVWMV5BJRQGM56VYRAFMGZ/graph.json","fetch_events":"https://pith.science/api/pith-number/5S2ZWVWMV5BJRQGM56VYRAFMGZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ/action/storage_attestation","attest_author":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ/action/author_attestation","sign_citation":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ/action/citation_signature","submit_replication":"https://pith.science/pith/5S2ZWVWMV5BJRQGM56VYRAFMGZ/action/replication_record"}},"created_at":"2026-07-05T11:17:00.519968+00:00","updated_at":"2026-07-05T11:17:00.519968+00:00"}