{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:NMLUEL2BYWVZPNNTWRMXWB7OPH","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":"747a06210b9658b84611782597ee66a017d372239935246bd3b7b2ca72a2ab37","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T21:25:11Z","title_canon_sha256":"7f7f5b724ee0cc18e291ecb4b4340962e067bc3f1df973a9c15f47ec4da49e31"},"schema_version":"1.0","source":{"id":"2507.09029","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.09029","created_at":"2026-06-02T01:03:32Z"},{"alias_kind":"arxiv_version","alias_value":"2507.09029v5","created_at":"2026-06-02T01:03:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.09029","created_at":"2026-06-02T01:03:32Z"},{"alias_kind":"pith_short_12","alias_value":"NMLUEL2BYWVZ","created_at":"2026-06-02T01:03:32Z"},{"alias_kind":"pith_short_16","alias_value":"NMLUEL2BYWVZPNNT","created_at":"2026-06-02T01:03:32Z"},{"alias_kind":"pith_short_8","alias_value":"NMLUEL2B","created_at":"2026-06-02T01:03:32Z"}],"graph_snapshots":[{"event_id":"sha256:247838aad46470925b9aec4e3c8213a68926ef742ab0c318a273e9fba495b517","target":"graph","created_at":"2026-06-02T01:03:32Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2507.09029/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pre-training large neural networks at scale imposes heavy memory demands on accelerators and often requires costly communication. We introduce Subnetwork Data Parallelism (SDP), a distributed training framework that partitions a model into structured subnetworks trained across workers without exchanging activations. We study two complementary masking regimes: backward masking, which applies sparsity only in the backward step to retain unbiased gradients, and forward masking, which also removes parameters in the forward pass to deliver stronger efficiency gains while providing additional regula","authors_text":"Edouard Oyallon, Eugene Belilovsky, Pietro Cagnasso, Vaibhav Singh, Zafir Khalid","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T21:25:11Z","title":"Model Parallelism With Subnetwork Data Parallelism"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.09029","kind":"arxiv","version":5},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7f2cd7cfa2ec17efcf7b1fdb65d5c165eaa56930e303721b5c6c15cf2c9859dd","target":"record","created_at":"2026-06-02T01:03:32Z","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":"747a06210b9658b84611782597ee66a017d372239935246bd3b7b2ca72a2ab37","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-11T21:25:11Z","title_canon_sha256":"7f7f5b724ee0cc18e291ecb4b4340962e067bc3f1df973a9c15f47ec4da49e31"},"schema_version":"1.0","source":{"id":"2507.09029","kind":"arxiv","version":5}},"canonical_sha256":"6b17422f41c5ab97b5b3b4597b07ee79eed265ab7ea403b8ac80a99eaf355d08","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6b17422f41c5ab97b5b3b4597b07ee79eed265ab7ea403b8ac80a99eaf355d08","first_computed_at":"2026-06-02T01:03:32.558600Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T01:03:32.558600Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"bw8n6fzpjHu5bOuK5TRqiAbHPqQ4PNFqRckRJGOS43uaTmR8g5WOlSJqUMaPgCHlmBadN97nMMma27JzsII/DA==","signature_status":"signed_v1","signed_at":"2026-06-02T01:03:32.559068Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.09029","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7f2cd7cfa2ec17efcf7b1fdb65d5c165eaa56930e303721b5c6c15cf2c9859dd","sha256:247838aad46470925b9aec4e3c8213a68926ef742ab0c318a273e9fba495b517"],"state_sha256":"b17757e22b8e7059ca973d415545d45254c9e66a2aea3eb78eadf152b5683ae4"}