{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:4LMZRYZFULW2QT7LIBJLYEBM3Q","short_pith_number":"pith:4LMZRYZF","canonical_record":{"source":{"id":"2402.13929","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-02-21T16:51:05Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"81c5c4e9f2261c876842f43d5b9bd8f72c619adc558aa52d9377023389e7de38","abstract_canon_sha256":"944b350b051a34f080e3d07ec97f5b20c4901bf62844bd2ddccf55539707df75"},"schema_version":"1.0"},"canonical_sha256":"e2d998e325a2eda84feb4052bc102cdc1c892ce6ec7d9a8b911010ba85775255","source":{"kind":"arxiv","id":"2402.13929","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.13929","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"arxiv_version","alias_value":"2402.13929v3","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.13929","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"pith_short_12","alias_value":"4LMZRYZFULW2","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"4LMZRYZFULW2QT7L","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"4LMZRYZF","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:4LMZRYZFULW2QT7LIBJLYEBM3Q","target":"record","payload":{"canonical_record":{"source":{"id":"2402.13929","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-02-21T16:51:05Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"81c5c4e9f2261c876842f43d5b9bd8f72c619adc558aa52d9377023389e7de38","abstract_canon_sha256":"944b350b051a34f080e3d07ec97f5b20c4901bf62844bd2ddccf55539707df75"},"schema_version":"1.0"},"canonical_sha256":"e2d998e325a2eda84feb4052bc102cdc1c892ce6ec7d9a8b911010ba85775255","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:15.038137Z","signature_b64":"NYZ00208QNTO9AF74zCZdscqwQ5rDE6vk2wM7bzjgtHwUGZ2IS2hBgQW5VYHzTu0hhS+lHaGPJfqXR7jmNEcDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e2d998e325a2eda84feb4052bc102cdc1c892ce6ec7d9a8b911010ba85775255","last_reissued_at":"2026-05-17T23:38:15.037380Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:15.037380Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.13929","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"15r35Ee8nzto6NvfeaDVd9uSETAfbPVWsJTTIFzdhD/px/mx3JyhCJ4uhNWAoAsfE8B9AHt7kZVDb98gitIzDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-18T17:40:04.192287Z"},"content_sha256":"92318ccd122bf771ef3e1ec55dfb73b93e23d1e67454a5a552e511b387ef3475","schema_version":"1.0","event_id":"sha256:92318ccd122bf771ef3e1ec55dfb73b93e23d1e67454a5a552e511b387ef3475"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:4LMZRYZFULW2QT7LIBJLYEBM3Q","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SDXL-Lightning: Progressive Adversarial Diffusion Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A distillation method combines progressive and adversarial training to enable one-step high-quality 1024-pixel image generation from SDXL.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Anran Wang, Shanchuan Lin, Xiao Yang","submitted_at":"2024-02-21T16:51:05Z","abstract_excerpt":"We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL. Our method combines progressive and adversarial distillation to achieve a balance between quality and mode coverage. In this paper, we discuss the theoretical analysis, discriminator design, model formulation, and training techniques. We open-source our distilled SDXL-Lightning models both as LoRA and full UNet weights."},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL by combining progressive and adversarial distillation to achieve a balance between quality and mode coverage.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that progressive adversarial training will maintain both perceptual quality and broad mode coverage without introducing artifacts or collapse, particularly when scaling the discriminator design and training schedule to the SDXL architecture.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SDXL-Lightning uses progressive adversarial distillation to reach new state-of-the-art quality in one-step and few-step 1024px text-to-image generation from the SDXL base model.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A distillation method combines progressive and adversarial training to enable one-step high-quality 1024-pixel image generation from SDXL.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1bf43eafdc66e096b0aecdb89acf30c469eeaedfc56e54e22a3bce1aa15d625d"},"source":{"id":"2402.13929","kind":"arxiv","version":3},"verdict":{"id":"d1b735b0-c26a-412a-bd33-abd343ff07c5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T05:04:24.605423Z","strongest_claim":"We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL by combining progressive and adversarial distillation to achieve a balance between quality and mode coverage.","one_line_summary":"SDXL-Lightning uses progressive adversarial distillation to reach new state-of-the-art quality in one-step and few-step 1024px text-to-image generation from the SDXL base model.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that progressive adversarial training will maintain both perceptual quality and broad mode coverage without introducing artifacts or collapse, particularly when scaling the discriminator design and training schedule to the SDXL architecture.","pith_extraction_headline":"A distillation method combines progressive and adversarial training to enable one-step high-quality 1024-pixel image generation from SDXL."},"references":{"count":82,"sample":[{"doi":"","year":null,"title":"https : / / civitai","work_id":"fcf55c25-ec16-434b-b454-6e682e87e8d0","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Stable video diffusion: Scaling latent video diffusion models to large datasets, 2023","work_id":"6ccf5d98-42d9-4662-a50a-59f3fbdbebee","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja Fidler, and Karsten Kreis","work_id":"92ea557a-eaa4-4e69-81ff-f2d447b0332d","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Coyo-700m: Image-text pair dataset","work_id":"33df4c3f-ffbc-4046-aa62-20cc70dafa4c","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Pixart-$\\alpha$: Fast training of diffusion transformer for photorealistic text-to-image syn- thesis","work_id":"456cbd06-948f-4cb3-b31c-2b425f535c25","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":82,"snapshot_sha256":"f79ea90efbc71693686e1e1aee9fe5afff7b06f8227fda6e7e0119be7445e278","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"d30b77f15db7e79e9236e8b506df5efa3795ef4a71e650d9b945b179ecf60651"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"d1b735b0-c26a-412a-bd33-abd343ff07c5"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"E1jB3LC3JfZtS3D3ybttwEH/DTsCSmR+s61sUB/xlSDTOHuJTPSba5JY5EYB9KPV72t0ZxAwfUmv8ZpaXjB7AA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-18T17:40:04.193263Z"},"content_sha256":"8a33e86e39ffc091ea68e4e9777482a848af7b8cfb5a7699e7720e3fe47908ae","schema_version":"1.0","event_id":"sha256:8a33e86e39ffc091ea68e4e9777482a848af7b8cfb5a7699e7720e3fe47908ae"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q/bundle.json","state_url":"https://pith.science/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-18T17:40:04Z","links":{"resolver":"https://pith.science/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q","bundle":"https://pith.science/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q/bundle.json","state":"https://pith.science/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4LMZRYZFULW2QT7LIBJLYEBM3Q/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:4LMZRYZFULW2QT7LIBJLYEBM3Q","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"944b350b051a34f080e3d07ec97f5b20c4901bf62844bd2ddccf55539707df75","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-02-21T16:51:05Z","title_canon_sha256":"81c5c4e9f2261c876842f43d5b9bd8f72c619adc558aa52d9377023389e7de38"},"schema_version":"1.0","source":{"id":"2402.13929","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.13929","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"arxiv_version","alias_value":"2402.13929v3","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.13929","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"pith_short_12","alias_value":"4LMZRYZFULW2","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"4LMZRYZFULW2QT7L","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"4LMZRYZF","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:8a33e86e39ffc091ea68e4e9777482a848af7b8cfb5a7699e7720e3fe47908ae","target":"graph","created_at":"2026-05-17T23:38:15Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL by combining progressive and adversarial distillation to achieve a balance between quality and mode coverage."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that progressive adversarial training will maintain both perceptual quality and broad mode coverage without introducing artifacts or collapse, particularly when scaling the discriminator design and training schedule to the SDXL architecture."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"SDXL-Lightning uses progressive adversarial distillation to reach new state-of-the-art quality in one-step and few-step 1024px text-to-image generation from the SDXL base model."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A distillation method combines progressive and adversarial training to enable one-step high-quality 1024-pixel image generation from SDXL."}],"snapshot_sha256":"1bf43eafdc66e096b0aecdb89acf30c469eeaedfc56e54e22a3bce1aa15d625d"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"d30b77f15db7e79e9236e8b506df5efa3795ef4a71e650d9b945b179ecf60651"},"paper":{"abstract_excerpt":"We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL. Our method combines progressive and adversarial distillation to achieve a balance between quality and mode coverage. In this paper, we discuss the theoretical analysis, discriminator design, model formulation, and training techniques. We open-source our distilled SDXL-Lightning models both as LoRA and full UNet weights.","authors_text":"Anran Wang, Shanchuan Lin, Xiao Yang","cross_cats":["cs.AI","cs.LG"],"headline":"A distillation method combines progressive and adversarial training to enable one-step high-quality 1024-pixel image generation from SDXL.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-02-21T16:51:05Z","title":"SDXL-Lightning: Progressive Adversarial Diffusion Distillation"},"references":{"count":82,"internal_anchors":1,"resolved_work":82,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"https : / / civitai","work_id":"fcf55c25-ec16-434b-b454-6e682e87e8d0","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Stable video diffusion: Scaling latent video diffusion models to large datasets, 2023","work_id":"6ccf5d98-42d9-4662-a50a-59f3fbdbebee","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja Fidler, and Karsten Kreis","work_id":"92ea557a-eaa4-4e69-81ff-f2d447b0332d","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Coyo-700m: Image-text pair dataset","work_id":"33df4c3f-ffbc-4046-aa62-20cc70dafa4c","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Pixart-$\\alpha$: Fast training of diffusion transformer for photorealistic text-to-image syn- thesis","work_id":"456cbd06-948f-4cb3-b31c-2b425f535c25","year":2024}],"snapshot_sha256":"f79ea90efbc71693686e1e1aee9fe5afff7b06f8227fda6e7e0119be7445e278"},"source":{"id":"2402.13929","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-17T05:04:24.605423Z","id":"d1b735b0-c26a-412a-bd33-abd343ff07c5","model_set":{"reader":"grok-4.3"},"one_line_summary":"SDXL-Lightning uses progressive adversarial distillation to reach new state-of-the-art quality in one-step and few-step 1024px text-to-image generation from the SDXL base model.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A distillation method combines progressive and adversarial training to enable one-step high-quality 1024-pixel image generation from SDXL.","strongest_claim":"We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL by combining progressive and adversarial distillation to achieve a balance between quality and mode coverage.","weakest_assumption":"The assumption that progressive adversarial training will maintain both perceptual quality and broad mode coverage without introducing artifacts or collapse, particularly when scaling the discriminator design and training schedule to the SDXL architecture."}},"verdict_id":"d1b735b0-c26a-412a-bd33-abd343ff07c5"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:92318ccd122bf771ef3e1ec55dfb73b93e23d1e67454a5a552e511b387ef3475","target":"record","created_at":"2026-05-17T23:38:15Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"944b350b051a34f080e3d07ec97f5b20c4901bf62844bd2ddccf55539707df75","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-02-21T16:51:05Z","title_canon_sha256":"81c5c4e9f2261c876842f43d5b9bd8f72c619adc558aa52d9377023389e7de38"},"schema_version":"1.0","source":{"id":"2402.13929","kind":"arxiv","version":3}},"canonical_sha256":"e2d998e325a2eda84feb4052bc102cdc1c892ce6ec7d9a8b911010ba85775255","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e2d998e325a2eda84feb4052bc102cdc1c892ce6ec7d9a8b911010ba85775255","first_computed_at":"2026-05-17T23:38:15.037380Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:15.037380Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NYZ00208QNTO9AF74zCZdscqwQ5rDE6vk2wM7bzjgtHwUGZ2IS2hBgQW5VYHzTu0hhS+lHaGPJfqXR7jmNEcDg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:15.038137Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.13929","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:92318ccd122bf771ef3e1ec55dfb73b93e23d1e67454a5a552e511b387ef3475","sha256:8a33e86e39ffc091ea68e4e9777482a848af7b8cfb5a7699e7720e3fe47908ae"],"state_sha256":"dc52ff1875ff3d2690e7483a50d945924c7aa96cec29ed1e1051be5c9fca36f5"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qVPqfSmyyAjkT9tiF6NCm9qXOO65THUm3iIbXFEHcdNIjNEgKH3X7ZbpGni5ML81D1crfXS83mFjJb3USMvWCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-18T17:40:04.196094Z","bundle_sha256":"405cff627b8dfb5daea79f9115bb183704e23b1f255f555325d93a28c2b15325"}}