{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:MZPNCN2IDFTZETUXHPCHTTD3GL","short_pith_number":"pith:MZPNCN2I","schema_version":"1.0","canonical_sha256":"665ed137481967924e973bc479cc7b32f96e2d1667bba467314b6cdcf7a77295","source":{"kind":"arxiv","id":"2311.12023","version":4},"attestation_state":"computed","paper":{"title":"LQ-LoRA: Low-rank Plus Quantized Matrix Decomposition for Efficient Language Model Finetuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Eric P. Xing, Han Guo, Philip Greengard, Yoon Kim","submitted_at":"2023-11-20T18:57:41Z","abstract_excerpt":"We propose a simple approach for memory-efficient adaptation of pretrained language models. Our approach uses an iterative algorithm to decompose each pretrained matrix into a high-precision low-rank component and a memory-efficient quantized component. During finetuning, the quantized component remains fixed and only the low-rank component is updated. We present an integer linear programming formulation of the quantization component which enables dynamic configuration of quantization parameters (e.g., bit-width, block size) for each matrix given an overall target memory budget. We further exp"},"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":"2311.12023","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-20T18:57:41Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6ffd1f3284ff030aab5343398e523a0e72bc887ec13c18b4f965721c36d080ab","abstract_canon_sha256":"f66a4447744af110551dd09bcdb00797f35473e50ac7d2991276b20d14839278"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:59:24.658801Z","signature_b64":"eJlJQX5WMu7vQzEgVUPIosoWZLP1iv6f+xsh/JsiCFvqDz389ne09w5jNUWQ6CNdtIG72PXDTuIM3R9oBhJNDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"665ed137481967924e973bc479cc7b32f96e2d1667bba467314b6cdcf7a77295","last_reissued_at":"2026-07-05T08:59:24.658217Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:59:24.658217Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LQ-LoRA: Low-rank Plus Quantized Matrix Decomposition for Efficient Language Model Finetuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Eric P. Xing, Han Guo, Philip Greengard, Yoon Kim","submitted_at":"2023-11-20T18:57:41Z","abstract_excerpt":"We propose a simple approach for memory-efficient adaptation of pretrained language models. Our approach uses an iterative algorithm to decompose each pretrained matrix into a high-precision low-rank component and a memory-efficient quantized component. During finetuning, the quantized component remains fixed and only the low-rank component is updated. We present an integer linear programming formulation of the quantization component which enables dynamic configuration of quantization parameters (e.g., bit-width, block size) for each matrix given an overall target memory budget. We further exp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.12023","kind":"arxiv","version":4},"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/2311.12023/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":"2311.12023","created_at":"2026-07-05T08:59:24.658287+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.12023v4","created_at":"2026-07-05T08:59:24.658287+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.12023","created_at":"2026-07-05T08:59:24.658287+00:00"},{"alias_kind":"pith_short_12","alias_value":"MZPNCN2IDFTZ","created_at":"2026-07-05T08:59:24.658287+00:00"},{"alias_kind":"pith_short_16","alias_value":"MZPNCN2IDFTZETUX","created_at":"2026-07-05T08:59:24.658287+00:00"},{"alias_kind":"pith_short_8","alias_value":"MZPNCN2I","created_at":"2026-07-05T08:59:24.658287+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01412","citing_title":"GPTQ-intrinsic LoRA: A Near-optimal Algorithm for Low-precision Quantization with Low-rank Adaptation","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00494","citing_title":"ProjQ: Project-and-Quantize for Adapter-Aware LLM Compression","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2403.14608","citing_title":"Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey","ref_index":126,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05819","citing_title":"HCInfer: An Efficient Inference System via Error Compensation for Resource-Constrained Devices","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL","json":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL.json","graph_json":"https://pith.science/api/pith-number/MZPNCN2IDFTZETUXHPCHTTD3GL/graph.json","events_json":"https://pith.science/api/pith-number/MZPNCN2IDFTZETUXHPCHTTD3GL/events.json","paper":"https://pith.science/paper/MZPNCN2I"},"agent_actions":{"view_html":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL","download_json":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL.json","view_paper":"https://pith.science/paper/MZPNCN2I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.12023&json=true","fetch_graph":"https://pith.science/api/pith-number/MZPNCN2IDFTZETUXHPCHTTD3GL/graph.json","fetch_events":"https://pith.science/api/pith-number/MZPNCN2IDFTZETUXHPCHTTD3GL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL/action/storage_attestation","attest_author":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL/action/author_attestation","sign_citation":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL/action/citation_signature","submit_replication":"https://pith.science/pith/MZPNCN2IDFTZETUXHPCHTTD3GL/action/replication_record"}},"created_at":"2026-07-05T08:59:24.658287+00:00","updated_at":"2026-07-05T08:59:24.658287+00:00"}