{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:U7SHXPZNZISKMFA4DTK7A7B6EU","short_pith_number":"pith:U7SHXPZN","canonical_record":{"source":{"id":"2507.08091","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-10T18:04:52Z","cross_cats_sorted":[],"title_canon_sha256":"8cdffc81b94e0cd95df95e5d78966e4e4d44e8ec5be4f01826df0ecded66e054","abstract_canon_sha256":"6873464763bb7b714119525471bfdc0d9b71154254e7b2ba98d718b60a65e6ca"},"schema_version":"1.0"},"canonical_sha256":"a7e47bbf2dca24a6141c1cd5f07c3e253b63d1d5c42c9dbef32b86f0c2ca4e56","source":{"kind":"arxiv","id":"2507.08091","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.08091","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"arxiv_version","alias_value":"2507.08091v1","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.08091","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"pith_short_12","alias_value":"U7SHXPZNZISK","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"pith_short_16","alias_value":"U7SHXPZNZISKMFA4","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"pith_short_8","alias_value":"U7SHXPZN","created_at":"2026-07-05T11:35:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:U7SHXPZNZISKMFA4DTK7A7B6EU","target":"record","payload":{"canonical_record":{"source":{"id":"2507.08091","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-10T18:04:52Z","cross_cats_sorted":[],"title_canon_sha256":"8cdffc81b94e0cd95df95e5d78966e4e4d44e8ec5be4f01826df0ecded66e054","abstract_canon_sha256":"6873464763bb7b714119525471bfdc0d9b71154254e7b2ba98d718b60a65e6ca"},"schema_version":"1.0"},"canonical_sha256":"a7e47bbf2dca24a6141c1cd5f07c3e253b63d1d5c42c9dbef32b86f0c2ca4e56","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:35:17.173459Z","signature_b64":"7Q9eYsDjAw9dN48hLi4M99I1gVa6WWVhrfA4yrDYYPrFe9xoku3IVByjiF9F3ns7Gf5ouzaCXS6lvOdjkmy5AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a7e47bbf2dca24a6141c1cd5f07c3e253b63d1d5c42c9dbef32b86f0c2ca4e56","last_reissued_at":"2026-07-05T11:35:17.172965Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:35:17.172965Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.08091","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:35:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5s4PaaUBUTdDhnHKOOrdJiJahK78LAMCW3+tmR1nHGMJ8wjxDQxgYelu4XvrhSvbBSowQvZxFUhhxWtdkAc8BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T23:58:41.144872Z"},"content_sha256":"1ba6df9e1bb7b48a00c2db64881235b32b1963b5a842daadf2849ea519dd82fb","schema_version":"1.0","event_id":"sha256:1ba6df9e1bb7b48a00c2db64881235b32b1963b5a842daadf2849ea519dd82fb"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:U7SHXPZNZISKMFA4DTK7A7B6EU","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Low-rank Momentum Factorization for Memory Efficient Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Mehrdad Mahdavi, Pouria Mahdavinia","submitted_at":"2025-07-10T18:04:52Z","abstract_excerpt":"Fine-tuning large foundation models presents significant memory challenges due to stateful optimizers like AdamW, often requiring several times more GPU memory than inference. While memory-efficient methods like parameter-efficient fine-tuning (e.g., LoRA) and optimizer state compression exist, recent approaches like GaLore bridge these by using low-rank gradient projections and subspace moment accumulation. However, such methods may struggle with fixed subspaces or computationally costly offline resampling (e.g., requiring full-matrix SVDs). We propose Momentum Factorized SGD (MoFaSGD), which"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.08091","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/2507.08091/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:35:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4yw0t/GxBiTkwF/f9VC8Am4pNHRvMUywnfUx0kKJ5lDIWadr+IFThYvEPIkeG1k74CpHUaA4brtlWNhDjcpHCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T23:58:41.145404Z"},"content_sha256":"000b29dac856a49780fa14f8c8bbeb9bf7b87117bb970946155d419c5161494e","schema_version":"1.0","event_id":"sha256:000b29dac856a49780fa14f8c8bbeb9bf7b87117bb970946155d419c5161494e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/U7SHXPZNZISKMFA4DTK7A7B6EU/bundle.json","state_url":"https://pith.science/pith/U7SHXPZNZISKMFA4DTK7A7B6EU/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/U7SHXPZNZISKMFA4DTK7A7B6EU/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":"2026-08-09T23:58:41Z","links":{"resolver":"https://pith.science/pith/U7SHXPZNZISKMFA4DTK7A7B6EU","bundle":"https://pith.science/pith/U7SHXPZNZISKMFA4DTK7A7B6EU/bundle.json","state":"https://pith.science/pith/U7SHXPZNZISKMFA4DTK7A7B6EU/state.json","well_known_bundle":"https://pith.science/.well-known/pith/U7SHXPZNZISKMFA4DTK7A7B6EU/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:U7SHXPZNZISKMFA4DTK7A7B6EU","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":"6873464763bb7b714119525471bfdc0d9b71154254e7b2ba98d718b60a65e6ca","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-10T18:04:52Z","title_canon_sha256":"8cdffc81b94e0cd95df95e5d78966e4e4d44e8ec5be4f01826df0ecded66e054"},"schema_version":"1.0","source":{"id":"2507.08091","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.08091","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"arxiv_version","alias_value":"2507.08091v1","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.08091","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"pith_short_12","alias_value":"U7SHXPZNZISK","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"pith_short_16","alias_value":"U7SHXPZNZISKMFA4","created_at":"2026-07-05T11:35:17Z"},{"alias_kind":"pith_short_8","alias_value":"U7SHXPZN","created_at":"2026-07-05T11:35:17Z"}],"graph_snapshots":[{"event_id":"sha256:000b29dac856a49780fa14f8c8bbeb9bf7b87117bb970946155d419c5161494e","target":"graph","created_at":"2026-07-05T11:35:17Z","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/2507.08091/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning large foundation models presents significant memory challenges due to stateful optimizers like AdamW, often requiring several times more GPU memory than inference. While memory-efficient methods like parameter-efficient fine-tuning (e.g., LoRA) and optimizer state compression exist, recent approaches like GaLore bridge these by using low-rank gradient projections and subspace moment accumulation. However, such methods may struggle with fixed subspaces or computationally costly offline resampling (e.g., requiring full-matrix SVDs). We propose Momentum Factorized SGD (MoFaSGD), which","authors_text":"Mehrdad Mahdavi, Pouria Mahdavinia","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-10T18:04:52Z","title":"Low-rank Momentum Factorization for Memory Efficient Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.08091","kind":"arxiv","version":1},"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":"sha256:1ba6df9e1bb7b48a00c2db64881235b32b1963b5a842daadf2849ea519dd82fb","target":"record","created_at":"2026-07-05T11:35:17Z","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":"6873464763bb7b714119525471bfdc0d9b71154254e7b2ba98d718b60a65e6ca","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-10T18:04:52Z","title_canon_sha256":"8cdffc81b94e0cd95df95e5d78966e4e4d44e8ec5be4f01826df0ecded66e054"},"schema_version":"1.0","source":{"id":"2507.08091","kind":"arxiv","version":1}},"canonical_sha256":"a7e47bbf2dca24a6141c1cd5f07c3e253b63d1d5c42c9dbef32b86f0c2ca4e56","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a7e47bbf2dca24a6141c1cd5f07c3e253b63d1d5c42c9dbef32b86f0c2ca4e56","first_computed_at":"2026-07-05T11:35:17.172965Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:35:17.172965Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7Q9eYsDjAw9dN48hLi4M99I1gVa6WWVhrfA4yrDYYPrFe9xoku3IVByjiF9F3ns7Gf5ouzaCXS6lvOdjkmy5AA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:35:17.173459Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.08091","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1ba6df9e1bb7b48a00c2db64881235b32b1963b5a842daadf2849ea519dd82fb","sha256:000b29dac856a49780fa14f8c8bbeb9bf7b87117bb970946155d419c5161494e"],"state_sha256":"f71ad492d63fffbed01bf66e8996dbc41fd87b9f39d0f76d63afe1b478331127"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jH6wc7Bx4KZrSVbbpQ+7deTOLo4v0Ex1wUmBawDmoFK3W4oN5GlUGM9m53JAJSp+zlp9hn6RKTXa7Hzy/h0rCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T23:58:41.150808Z","bundle_sha256":"cb03084cae491c0006a999a9e36cb34d679dc288f730955e67661dd60959d854"}}