{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NTCTJ2K5RG5VDBTHZRN2C5KQDG","short_pith_number":"pith:NTCTJ2K5","schema_version":"1.0","canonical_sha256":"6cc534e95d89bb518667cc5ba1755019a10758db9768461c0604fe6c69693ebe","source":{"kind":"arxiv","id":"2404.13040","version":2},"attestation_state":"computed","paper":{"title":"Analysis of Classifier-Free Guidance Weight Schedulers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"David Picard, Marie-Paule Cani, Nefeli Andreou, Nicolas Dufour, Vicky Kalogeiton, Victoria Fernandez Abrevaya, Xi Wang","submitted_at":"2024-04-19T17:53:43Z","abstract_excerpt":"Classifier-Free Guidance (CFG) enhances the quality and condition adherence of text-to-image diffusion models. It operates by combining the conditional and unconditional predictions using a fixed weight. However, recent works vary the weights throughout the diffusion process, reporting superior results but without providing any rationale or analysis. By conducting comprehensive experiments, this paper provides insights into CFG weight schedulers. Our findings suggest that simple, monotonically increasing weight schedulers consistently lead to improved performances, requiring merely a single li"},"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":"2404.13040","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-04-19T17:53:43Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"78633acdab9caa4cab1ba3751103fcdc0ab6553571f55fed85512dcc9a926add","abstract_canon_sha256":"67517b87577f8785cf4299d74b37febc4ba14eedbaf39782d3081a1debd7b873"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:03:42.243508Z","signature_b64":"bGduSdYTLUm/5mzWm5q4ffytT7hDpfIMc7Tl+hAg6m2RQrthXjENrQQEhMD9M/iVV5lyuG/SoNQZGezAIT7CBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6cc534e95d89bb518667cc5ba1755019a10758db9768461c0604fe6c69693ebe","last_reissued_at":"2026-07-05T10:03:42.243064Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:03:42.243064Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Analysis of Classifier-Free Guidance Weight Schedulers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"David Picard, Marie-Paule Cani, Nefeli Andreou, Nicolas Dufour, Vicky Kalogeiton, Victoria Fernandez Abrevaya, Xi Wang","submitted_at":"2024-04-19T17:53:43Z","abstract_excerpt":"Classifier-Free Guidance (CFG) enhances the quality and condition adherence of text-to-image diffusion models. It operates by combining the conditional and unconditional predictions using a fixed weight. However, recent works vary the weights throughout the diffusion process, reporting superior results but without providing any rationale or analysis. By conducting comprehensive experiments, this paper provides insights into CFG weight schedulers. Our findings suggest that simple, monotonically increasing weight schedulers consistently lead to improved performances, requiring merely a single li"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.13040","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/2404.13040/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":"2404.13040","created_at":"2026-07-05T10:03:42.243128+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.13040v2","created_at":"2026-07-05T10:03:42.243128+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.13040","created_at":"2026-07-05T10:03:42.243128+00:00"},{"alias_kind":"pith_short_12","alias_value":"NTCTJ2K5RG5V","created_at":"2026-07-05T10:03:42.243128+00:00"},{"alias_kind":"pith_short_16","alias_value":"NTCTJ2K5RG5VDBTH","created_at":"2026-07-05T10:03:42.243128+00:00"},{"alias_kind":"pith_short_8","alias_value":"NTCTJ2K5","created_at":"2026-07-05T10:03:42.243128+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.08155","citing_title":"C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15661","citing_title":"VAGS: Velocity Adaptive Guidance Scale for Image Editing and Generation","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2603.08155","citing_title":"C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07701","citing_title":"Guidance Is Not a Hyperparameter: Learning Dynamic Control in Diffusion Language Models","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG","json":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG.json","graph_json":"https://pith.science/api/pith-number/NTCTJ2K5RG5VDBTHZRN2C5KQDG/graph.json","events_json":"https://pith.science/api/pith-number/NTCTJ2K5RG5VDBTHZRN2C5KQDG/events.json","paper":"https://pith.science/paper/NTCTJ2K5"},"agent_actions":{"view_html":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG","download_json":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG.json","view_paper":"https://pith.science/paper/NTCTJ2K5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.13040&json=true","fetch_graph":"https://pith.science/api/pith-number/NTCTJ2K5RG5VDBTHZRN2C5KQDG/graph.json","fetch_events":"https://pith.science/api/pith-number/NTCTJ2K5RG5VDBTHZRN2C5KQDG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG/action/storage_attestation","attest_author":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG/action/author_attestation","sign_citation":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG/action/citation_signature","submit_replication":"https://pith.science/pith/NTCTJ2K5RG5VDBTHZRN2C5KQDG/action/replication_record"}},"created_at":"2026-07-05T10:03:42.243128+00:00","updated_at":"2026-07-05T10:03:42.243128+00:00"}