{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RIYD23OMPN5SSDYGLFE7EALGVF","short_pith_number":"pith:RIYD23OM","schema_version":"1.0","canonical_sha256":"8a303d6dcc7b7b290f065949f20166a947502fafeb80edcc7063efbf8b33acad","source":{"kind":"arxiv","id":"2605.17923","version":1},"attestation_state":"computed","paper":{"title":"AdaptiveLoad: Towards Efficient Video Diffusion Transformer Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.DC","authors_text":"Haoran Sun, Jing Long, Junwu Xiong, Shuai Di, Wanting Xu, Wen Huang, Yongjian Guo, Yucheng Guo, Zhong Guan","submitted_at":"2026-05-18T06:30:31Z","abstract_excerpt":"In video generation models, particularly world models, training large-scale video diffusion Transformers (such as DiT and MMDiT) poses significant computational challenges due to the extreme variance in sequence lengths within mixed-mode datasets. Existing bucket-based data loading strategies typically rely on \"equal token length\" constraints. This approach fails to account for the quadratic complexity of self-attention mechanisms, leading to severe load imbalance and underutilization of GPU resources.\n  This paper proposes \\textit{AdaptiveLoad}, an integrated optimization framework consisting"},"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":"2605.17923","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DC","submitted_at":"2026-05-18T06:30:31Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"44d1ce3bb6368f09df92a4730b7c8639c35a41ab42853486c8763f9ad23a48e7","abstract_canon_sha256":"9eda7d22013bfc4cb2b604f2245c9e2ed91e2bbfba1665964e2f722331aea792"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:06.242633Z","signature_b64":"mg7whDUSljBg79gcRi5Z00JYPsg/s5DHlA1VaDEqkLHlgUqvEjL5jkg5/fmwoQL9jPxMUO8BPBunVwCsntl9Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8a303d6dcc7b7b290f065949f20166a947502fafeb80edcc7063efbf8b33acad","last_reissued_at":"2026-05-20T00:05:06.241816Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:06.241816Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AdaptiveLoad: Towards Efficient Video Diffusion Transformer Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.DC","authors_text":"Haoran Sun, Jing Long, Junwu Xiong, Shuai Di, Wanting Xu, Wen Huang, Yongjian Guo, Yucheng Guo, Zhong Guan","submitted_at":"2026-05-18T06:30:31Z","abstract_excerpt":"In video generation models, particularly world models, training large-scale video diffusion Transformers (such as DiT and MMDiT) poses significant computational challenges due to the extreme variance in sequence lengths within mixed-mode datasets. Existing bucket-based data loading strategies typically rely on \"equal token length\" constraints. This approach fails to account for the quadratic complexity of self-attention mechanisms, leading to severe load imbalance and underutilization of GPU resources.\n  This paper proposes \\textit{AdaptiveLoad}, an integrated optimization framework consisting"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.17923","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/2605.17923/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":"2605.17923","created_at":"2026-05-20T00:05:06.241943+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17923v1","created_at":"2026-05-20T00:05:06.241943+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17923","created_at":"2026-05-20T00:05:06.241943+00:00"},{"alias_kind":"pith_short_12","alias_value":"RIYD23OMPN5S","created_at":"2026-05-20T00:05:06.241943+00:00"},{"alias_kind":"pith_short_16","alias_value":"RIYD23OMPN5SSDYG","created_at":"2026-05-20T00:05:06.241943+00:00"},{"alias_kind":"pith_short_8","alias_value":"RIYD23OM","created_at":"2026-05-20T00:05:06.241943+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF","json":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF.json","graph_json":"https://pith.science/api/pith-number/RIYD23OMPN5SSDYGLFE7EALGVF/graph.json","events_json":"https://pith.science/api/pith-number/RIYD23OMPN5SSDYGLFE7EALGVF/events.json","paper":"https://pith.science/paper/RIYD23OM"},"agent_actions":{"view_html":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF","download_json":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF.json","view_paper":"https://pith.science/paper/RIYD23OM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17923&json=true","fetch_graph":"https://pith.science/api/pith-number/RIYD23OMPN5SSDYGLFE7EALGVF/graph.json","fetch_events":"https://pith.science/api/pith-number/RIYD23OMPN5SSDYGLFE7EALGVF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF/action/storage_attestation","attest_author":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF/action/author_attestation","sign_citation":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF/action/citation_signature","submit_replication":"https://pith.science/pith/RIYD23OMPN5SSDYGLFE7EALGVF/action/replication_record"}},"created_at":"2026-05-20T00:05:06.241943+00:00","updated_at":"2026-05-20T00:05:06.241943+00:00"}