{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:OYI5CF6QZW6GKAUIUZ4CTENKCQ","short_pith_number":"pith:OYI5CF6Q","schema_version":"1.0","canonical_sha256":"7611d117d0cdbc650288a6782991aa1400fc9584a2cab9af6a35aef16a60444d","source":{"kind":"arxiv","id":"2602.20497","version":3},"attestation_state":"computed","paper":{"title":"LESA: Learnable Stage-Aware Predictors for Diffusion Model Acceleration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chang Zou, Haowen Xu, Jiacheng Liu, Linfeng Zhang, Peiliang Cai, Xinyu Wang","submitted_at":"2026-02-24T02:53:28Z","abstract_excerpt":"Diffusion models have achieved remarkable success in image and video generation tasks. However, the high computational demands of Diffusion Transformers (DiTs) pose a significant challenge to their practical deployment. While feature caching is a promising acceleration strategy, existing methods based on simple reusing or training-free forecasting struggle to adapt to the complex, stage-dependent dynamics of the diffusion process, often resulting in quality degradation and failing to maintain consistency with the standard denoising process. To address this, we propose a LEarnable Stage-Aware ("},"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":"2602.20497","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-24T02:53:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"04c6c37961f0ab681d442528b6850e6443099e5c64934fbd115e7a03dc7ec4c9","abstract_canon_sha256":"398f489c467330c28d6ecaa424a508eb1eefb6c9dd7b5cf0be7228f119113451"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:37.297236Z","signature_b64":"6dIqryusBFZPM1JVcRRZCFdfpzoA6cB3r7slFX+9QlB2ilL0EYxNXvQTqWFKeg9EqRI5qfsbTK4m2JLlCsfvBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7611d117d0cdbc650288a6782991aa1400fc9584a2cab9af6a35aef16a60444d","last_reissued_at":"2026-05-28T01:04:37.296762Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:37.296762Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LESA: Learnable Stage-Aware Predictors for Diffusion Model Acceleration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chang Zou, Haowen Xu, Jiacheng Liu, Linfeng Zhang, Peiliang Cai, Xinyu Wang","submitted_at":"2026-02-24T02:53:28Z","abstract_excerpt":"Diffusion models have achieved remarkable success in image and video generation tasks. However, the high computational demands of Diffusion Transformers (DiTs) pose a significant challenge to their practical deployment. While feature caching is a promising acceleration strategy, existing methods based on simple reusing or training-free forecasting struggle to adapt to the complex, stage-dependent dynamics of the diffusion process, often resulting in quality degradation and failing to maintain consistency with the standard denoising process. To address this, we propose a LEarnable Stage-Aware ("},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.20497","kind":"arxiv","version":3},"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/2602.20497/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":"2602.20497","created_at":"2026-05-28T01:04:37.296819+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.20497v3","created_at":"2026-05-28T01:04:37.296819+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.20497","created_at":"2026-05-28T01:04:37.296819+00:00"},{"alias_kind":"pith_short_12","alias_value":"OYI5CF6QZW6G","created_at":"2026-05-28T01:04:37.296819+00:00"},{"alias_kind":"pith_short_16","alias_value":"OYI5CF6QZW6GKAUI","created_at":"2026-05-28T01:04:37.296819+00:00"},{"alias_kind":"pith_short_8","alias_value":"OYI5CF6Q","created_at":"2026-05-28T01:04:37.296819+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.18346","citing_title":"Focused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video Diffusion","ref_index":5,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ","json":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ.json","graph_json":"https://pith.science/api/pith-number/OYI5CF6QZW6GKAUIUZ4CTENKCQ/graph.json","events_json":"https://pith.science/api/pith-number/OYI5CF6QZW6GKAUIUZ4CTENKCQ/events.json","paper":"https://pith.science/paper/OYI5CF6Q"},"agent_actions":{"view_html":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ","download_json":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ.json","view_paper":"https://pith.science/paper/OYI5CF6Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.20497&json=true","fetch_graph":"https://pith.science/api/pith-number/OYI5CF6QZW6GKAUIUZ4CTENKCQ/graph.json","fetch_events":"https://pith.science/api/pith-number/OYI5CF6QZW6GKAUIUZ4CTENKCQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ/action/storage_attestation","attest_author":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ/action/author_attestation","sign_citation":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ/action/citation_signature","submit_replication":"https://pith.science/pith/OYI5CF6QZW6GKAUIUZ4CTENKCQ/action/replication_record"}},"created_at":"2026-05-28T01:04:37.296819+00:00","updated_at":"2026-05-28T01:04:37.296819+00:00"}