{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:54LPPHTM626FAFRSWPXTW7E3SD","short_pith_number":"pith:54LPPHTM","schema_version":"1.0","canonical_sha256":"ef16f79e6cf6bc501632b3ef3b7c9b90ff04ff3450ce570ef1737cce4a8592e0","source":{"kind":"arxiv","id":"2404.05384","version":1},"attestation_state":"computed","paper":{"title":"Rethinking the Spatial Inconsistency in Classifier-Free Diffusion Guidance","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Dazhong Shen, Fu-Yun Wang, Guanglu Song, Yu Liu, Zeyue Xue","submitted_at":"2024-04-08T10:45:29Z","abstract_excerpt":"Classifier-Free Guidance (CFG) has been widely used in text-to-image diffusion models, where the CFG scale is introduced to control the strength of text guidance on the whole image space. However, we argue that a global CFG scale results in spatial inconsistency on varying semantic strengths and suboptimal image quality. To address this problem, we present a novel approach, Semantic-aware Classifier-Free Guidance (S-CFG), to customize the guidance degrees for different semantic units in text-to-image diffusion models. Specifically, we first design a training-free semantic segmentation method t"},"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":"2404.05384","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2024-04-08T10:45:29Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"108a0035cdfae7c2cba1a19b610c392357bd29841a4b1e77ce71026b2d57036e","abstract_canon_sha256":"c100cad004ad405bf8b5c6cae526c0d9dea3b77bd6dc7f3db9c4e6d1efb11468"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:05:30.026917Z","signature_b64":"ceqWhWmnwD92j23s7xqWFC5hDg54DjisQR9Ag06P83I8VDnjxcY8rWBGYJ69FBmwWbLLu32O+n7rLB4CUkPvDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ef16f79e6cf6bc501632b3ef3b7c9b90ff04ff3450ce570ef1737cce4a8592e0","last_reissued_at":"2026-07-05T08:05:30.026448Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:05:30.026448Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rethinking the Spatial Inconsistency in Classifier-Free Diffusion Guidance","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Dazhong Shen, Fu-Yun Wang, Guanglu Song, Yu Liu, Zeyue Xue","submitted_at":"2024-04-08T10:45:29Z","abstract_excerpt":"Classifier-Free Guidance (CFG) has been widely used in text-to-image diffusion models, where the CFG scale is introduced to control the strength of text guidance on the whole image space. However, we argue that a global CFG scale results in spatial inconsistency on varying semantic strengths and suboptimal image quality. To address this problem, we present a novel approach, Semantic-aware Classifier-Free Guidance (S-CFG), to customize the guidance degrees for different semantic units in text-to-image diffusion models. Specifically, we first design a training-free semantic segmentation method t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.05384","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/2404.05384/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":"2404.05384","created_at":"2026-07-05T08:05:30.026505+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.05384v1","created_at":"2026-07-05T08:05:30.026505+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.05384","created_at":"2026-07-05T08:05:30.026505+00:00"},{"alias_kind":"pith_short_12","alias_value":"54LPPHTM626F","created_at":"2026-07-05T08:05:30.026505+00:00"},{"alias_kind":"pith_short_16","alias_value":"54LPPHTM626FAFRS","created_at":"2026-07-05T08:05:30.026505+00:00"},{"alias_kind":"pith_short_8","alias_value":"54LPPHTM","created_at":"2026-07-05T08:05:30.026505+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.18010","citing_title":"Functionalization via Structure Completion and Motion Rectification","ref_index":151,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD","json":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD.json","graph_json":"https://pith.science/api/pith-number/54LPPHTM626FAFRSWPXTW7E3SD/graph.json","events_json":"https://pith.science/api/pith-number/54LPPHTM626FAFRSWPXTW7E3SD/events.json","paper":"https://pith.science/paper/54LPPHTM"},"agent_actions":{"view_html":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD","download_json":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD.json","view_paper":"https://pith.science/paper/54LPPHTM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.05384&json=true","fetch_graph":"https://pith.science/api/pith-number/54LPPHTM626FAFRSWPXTW7E3SD/graph.json","fetch_events":"https://pith.science/api/pith-number/54LPPHTM626FAFRSWPXTW7E3SD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD/action/storage_attestation","attest_author":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD/action/author_attestation","sign_citation":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD/action/citation_signature","submit_replication":"https://pith.science/pith/54LPPHTM626FAFRSWPXTW7E3SD/action/replication_record"}},"created_at":"2026-07-05T08:05:30.026505+00:00","updated_at":"2026-07-05T08:05:30.026505+00:00"}