{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:ADIZELKNMXEIG3WOCO4WWH2JVK","short_pith_number":"pith:ADIZELKN","schema_version":"1.0","canonical_sha256":"00d1922d4d65c8836ece13b96b1f49aa810c39bcbecd3d8edf3a4c55fb2c0b98","source":{"kind":"arxiv","id":"2309.11497","version":2},"attestation_state":"computed","paper":{"title":"FreeU: Free Lunch in Diffusion U-Net","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenyang Si, Yuming Jiang, Ziqi Huang, Ziwei Liu","submitted_at":"2023-09-20T17:56:18Z","abstract_excerpt":"In this paper, we uncover the untapped potential of diffusion U-Net, which serves as a \"free lunch\" that substantially improves the generation quality on the fly. We initially investigate the key contributions of the U-Net architecture to the denoising process and identify that its main backbone primarily contributes to denoising, whereas its skip connections mainly introduce high-frequency features into the decoder module, causing the network to overlook the backbone semantics. Capitalizing on this discovery, we propose a simple yet effective method-termed \"FreeU\" - that enhances generation q"},"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":"2309.11497","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-09-20T17:56:18Z","cross_cats_sorted":[],"title_canon_sha256":"f3f518624b5950d3c781dfe1c3aec064a788fd548be36da03cab9152adaf3df9","abstract_canon_sha256":"5c29a28bc0a08a74d296b4e8e80ef27b3fc7dfc0d9749d211fbdf6563843cbff"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:02:05.091019Z","signature_b64":"DdOom7X2jdTvqpwD10O4AGPq4RR7I9JydBOQV3FFJJHIHx6lgXh2xestPOQedxPUj9buchGfBT1WArGMoiNSBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"00d1922d4d65c8836ece13b96b1f49aa810c39bcbecd3d8edf3a4c55fb2c0b98","last_reissued_at":"2026-07-05T07:02:05.090595Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:02:05.090595Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FreeU: Free Lunch in Diffusion U-Net","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenyang Si, Yuming Jiang, Ziqi Huang, Ziwei Liu","submitted_at":"2023-09-20T17:56:18Z","abstract_excerpt":"In this paper, we uncover the untapped potential of diffusion U-Net, which serves as a \"free lunch\" that substantially improves the generation quality on the fly. We initially investigate the key contributions of the U-Net architecture to the denoising process and identify that its main backbone primarily contributes to denoising, whereas its skip connections mainly introduce high-frequency features into the decoder module, causing the network to overlook the backbone semantics. Capitalizing on this discovery, we propose a simple yet effective method-termed \"FreeU\" - that enhances generation q"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.11497","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/2309.11497/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":"2309.11497","created_at":"2026-07-05T07:02:05.090650+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.11497v2","created_at":"2026-07-05T07:02:05.090650+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.11497","created_at":"2026-07-05T07:02:05.090650+00:00"},{"alias_kind":"pith_short_12","alias_value":"ADIZELKNMXEI","created_at":"2026-07-05T07:02:05.090650+00:00"},{"alias_kind":"pith_short_16","alias_value":"ADIZELKNMXEIG3WO","created_at":"2026-07-05T07:02:05.090650+00:00"},{"alias_kind":"pith_short_8","alias_value":"ADIZELKN","created_at":"2026-07-05T07:02:05.090650+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2405.18716","citing_title":"SketchDeco: Training-Free Latent Composition for Precise Sketch Colourisation","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2407.12173","citing_title":"Beta Sampling is All You Need: Efficient Image Generation Strategy for Diffusion Models using Stepwise Spectral Analysis","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26503","citing_title":"Delta Score Matters! Spatial Adaptive Multi Guidance in Diffusion Models","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK","json":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK.json","graph_json":"https://pith.science/api/pith-number/ADIZELKNMXEIG3WOCO4WWH2JVK/graph.json","events_json":"https://pith.science/api/pith-number/ADIZELKNMXEIG3WOCO4WWH2JVK/events.json","paper":"https://pith.science/paper/ADIZELKN"},"agent_actions":{"view_html":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK","download_json":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK.json","view_paper":"https://pith.science/paper/ADIZELKN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.11497&json=true","fetch_graph":"https://pith.science/api/pith-number/ADIZELKNMXEIG3WOCO4WWH2JVK/graph.json","fetch_events":"https://pith.science/api/pith-number/ADIZELKNMXEIG3WOCO4WWH2JVK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK/action/storage_attestation","attest_author":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK/action/author_attestation","sign_citation":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK/action/citation_signature","submit_replication":"https://pith.science/pith/ADIZELKNMXEIG3WOCO4WWH2JVK/action/replication_record"}},"created_at":"2026-07-05T07:02:05.090650+00:00","updated_at":"2026-07-05T07:02:05.090650+00:00"}