{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IO3GNSYNZPCEFEBRH2JIDAUGC7","short_pith_number":"pith:IO3GNSYN","schema_version":"1.0","canonical_sha256":"43b666cb0dcbc44290313e9281828617d407c23196203526f02593d59d84ee12","source":{"kind":"arxiv","id":"2406.02214","version":2},"attestation_state":"computed","paper":{"title":"SLTrain: a sparse plus low-rank approach for parameter and memory efficient pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Akiko Takeda, Andi Han, Bamdev Mishra, Jiaxiang Li, Mingyi Hong, Pratik Jawanpuria, Wei Huang","submitted_at":"2024-06-04T11:14:21Z","abstract_excerpt":"Large language models (LLMs) have shown impressive capabilities across various tasks. However, training LLMs from scratch requires significant computational power and extensive memory capacity. Recent studies have explored low-rank structures on weights for efficient fine-tuning in terms of parameters and memory, either through low-rank adaptation or factorization. While effective for fine-tuning, low-rank structures are generally less suitable for pretraining because they restrict parameters to a low-dimensional subspace. In this work, we propose to parameterize the weights as a sum of low-ra"},"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":"2406.02214","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-04T11:14:21Z","cross_cats_sorted":[],"title_canon_sha256":"2aa91feeff8aa186be605e6576fbe446853ee1bf2e46ce2f74e9031b83f91808","abstract_canon_sha256":"813f8d77565d1965c107e2d4b1ad2c36126efb072b881c78218dedcfc1254cfa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:30:15.876433Z","signature_b64":"bp5xi9anB5ePt/9sdBnEWAe9eZWDWs2eFCQAen3Npyf6+H+VgZHGgxFfKLDFeqM6hYIuVr4Zo3pjarerwngBDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"43b666cb0dcbc44290313e9281828617d407c23196203526f02593d59d84ee12","last_reissued_at":"2026-07-05T09:30:15.875887Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:30:15.875887Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SLTrain: a sparse plus low-rank approach for parameter and memory efficient pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Akiko Takeda, Andi Han, Bamdev Mishra, Jiaxiang Li, Mingyi Hong, Pratik Jawanpuria, Wei Huang","submitted_at":"2024-06-04T11:14:21Z","abstract_excerpt":"Large language models (LLMs) have shown impressive capabilities across various tasks. However, training LLMs from scratch requires significant computational power and extensive memory capacity. Recent studies have explored low-rank structures on weights for efficient fine-tuning in terms of parameters and memory, either through low-rank adaptation or factorization. While effective for fine-tuning, low-rank structures are generally less suitable for pretraining because they restrict parameters to a low-dimensional subspace. In this work, we propose to parameterize the weights as a sum of low-ra"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.02214","kind":"arxiv","version":2},"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/2406.02214/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":"2406.02214","created_at":"2026-07-05T09:30:15.875957+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.02214v2","created_at":"2026-07-05T09:30:15.875957+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.02214","created_at":"2026-07-05T09:30:15.875957+00:00"},{"alias_kind":"pith_short_12","alias_value":"IO3GNSYNZPCE","created_at":"2026-07-05T09:30:15.875957+00:00"},{"alias_kind":"pith_short_16","alias_value":"IO3GNSYNZPCEFEBR","created_at":"2026-07-05T09:30:15.875957+00:00"},{"alias_kind":"pith_short_8","alias_value":"IO3GNSYN","created_at":"2026-07-05T09:30:15.875957+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.13652","citing_title":"Beyond Perplexity: A Geometric and Spectral Study of Low-Rank Pre-Training","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2509.18993","citing_title":"CR-Net: Scaling Parameter-Efficient Training with Cross-Layer Low-Rank Structure","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13652","citing_title":"Beyond Perplexity: A Geometric and Spectral Study of Low-Rank Pre-Training","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7","json":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7.json","graph_json":"https://pith.science/api/pith-number/IO3GNSYNZPCEFEBRH2JIDAUGC7/graph.json","events_json":"https://pith.science/api/pith-number/IO3GNSYNZPCEFEBRH2JIDAUGC7/events.json","paper":"https://pith.science/paper/IO3GNSYN"},"agent_actions":{"view_html":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7","download_json":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7.json","view_paper":"https://pith.science/paper/IO3GNSYN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.02214&json=true","fetch_graph":"https://pith.science/api/pith-number/IO3GNSYNZPCEFEBRH2JIDAUGC7/graph.json","fetch_events":"https://pith.science/api/pith-number/IO3GNSYNZPCEFEBRH2JIDAUGC7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7/action/storage_attestation","attest_author":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7/action/author_attestation","sign_citation":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7/action/citation_signature","submit_replication":"https://pith.science/pith/IO3GNSYNZPCEFEBRH2JIDAUGC7/action/replication_record"}},"created_at":"2026-07-05T09:30:15.875957+00:00","updated_at":"2026-07-05T09:30:15.875957+00:00"}