{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:G5EM5IWWSC6GM3Y2VGWTGSDUYG","short_pith_number":"pith:G5EM5IWW","schema_version":"1.0","canonical_sha256":"3748cea2d690bc666f1aa9ad334874c18985b5121a921237e11d9e8e516f4391","source":{"kind":"arxiv","id":"2405.02730","version":3},"attestation_state":"computed","paper":{"title":"U-DiTs: Downsample Tokens in U-Shaped Diffusion Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chao Xu, Hanting Chen, Jie Hu, Yuchuan Tian, Yunhe Wang, Zhijun Tu","submitted_at":"2024-05-04T18:27:29Z","abstract_excerpt":"Diffusion Transformers (DiTs) introduce the transformer architecture to diffusion tasks for latent-space image generation. With an isotropic architecture that chains a series of transformer blocks, DiTs demonstrate competitive performance and good scalability; but meanwhile, the abandonment of U-Net by DiTs and their following improvements is worth rethinking. To this end, we conduct a simple toy experiment by comparing a U-Net architectured DiT with an isotropic one. It turns out that the U-Net architecture only gain a slight advantage amid the U-Net inductive bias, indicating potential redun"},"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":"2405.02730","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-04T18:27:29Z","cross_cats_sorted":[],"title_canon_sha256":"49640a21c97f3e99a4aff2140069cbd864de350e58601017b1c9f0fee3578f16","abstract_canon_sha256":"d8ba25a382883b7fa943dcc37f3b10df9695495ef5a1295ceb10c57e17f8f2bf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:28:30.583388Z","signature_b64":"AyYmy+1NAVGNMtxCxNE5w5J+JOUWCyFGX4LKwjnSX9KyXkIsDswmyJsz6xtb2sLonFyfQ2F8de8sTpbc2qbMCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3748cea2d690bc666f1aa9ad334874c18985b5121a921237e11d9e8e516f4391","last_reissued_at":"2026-07-05T09:28:30.582950Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:28:30.582950Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"U-DiTs: Downsample Tokens in U-Shaped Diffusion Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chao Xu, Hanting Chen, Jie Hu, Yuchuan Tian, Yunhe Wang, Zhijun Tu","submitted_at":"2024-05-04T18:27:29Z","abstract_excerpt":"Diffusion Transformers (DiTs) introduce the transformer architecture to diffusion tasks for latent-space image generation. With an isotropic architecture that chains a series of transformer blocks, DiTs demonstrate competitive performance and good scalability; but meanwhile, the abandonment of U-Net by DiTs and their following improvements is worth rethinking. To this end, we conduct a simple toy experiment by comparing a U-Net architectured DiT with an isotropic one. It turns out that the U-Net architecture only gain a slight advantage amid the U-Net inductive bias, indicating potential redun"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.02730","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/2405.02730/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":"2405.02730","created_at":"2026-07-05T09:28:30.583007+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.02730v3","created_at":"2026-07-05T09:28:30.583007+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.02730","created_at":"2026-07-05T09:28:30.583007+00:00"},{"alias_kind":"pith_short_12","alias_value":"G5EM5IWWSC6G","created_at":"2026-07-05T09:28:30.583007+00:00"},{"alias_kind":"pith_short_16","alias_value":"G5EM5IWWSC6GM3Y2","created_at":"2026-07-05T09:28:30.583007+00:00"},{"alias_kind":"pith_short_8","alias_value":"G5EM5IWW","created_at":"2026-07-05T09:28:30.583007+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.15684","citing_title":"ElasticDiT: Efficient Diffusion Transformers via Elastic Architecture and Sparse Attention for High-Resolution Image Generation on Mobile Devices","ref_index":70,"is_internal_anchor":false},{"citing_arxiv_id":"2410.10629","citing_title":"SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG","json":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG.json","graph_json":"https://pith.science/api/pith-number/G5EM5IWWSC6GM3Y2VGWTGSDUYG/graph.json","events_json":"https://pith.science/api/pith-number/G5EM5IWWSC6GM3Y2VGWTGSDUYG/events.json","paper":"https://pith.science/paper/G5EM5IWW"},"agent_actions":{"view_html":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG","download_json":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG.json","view_paper":"https://pith.science/paper/G5EM5IWW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.02730&json=true","fetch_graph":"https://pith.science/api/pith-number/G5EM5IWWSC6GM3Y2VGWTGSDUYG/graph.json","fetch_events":"https://pith.science/api/pith-number/G5EM5IWWSC6GM3Y2VGWTGSDUYG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG/action/storage_attestation","attest_author":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG/action/author_attestation","sign_citation":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG/action/citation_signature","submit_replication":"https://pith.science/pith/G5EM5IWWSC6GM3Y2VGWTGSDUYG/action/replication_record"}},"created_at":"2026-07-05T09:28:30.583007+00:00","updated_at":"2026-07-05T09:28:30.583007+00:00"}