{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WVXWXUP264WSGZNNC5W5PFSSTD","short_pith_number":"pith:WVXWXUP2","schema_version":"1.0","canonical_sha256":"b56f6bd1faf72d2365ad176dd7965298de037a69098dd7b3f450c3a12c629d03","source":{"kind":"arxiv","id":"2405.20973","version":2},"attestation_state":"computed","paper":{"title":"LCQ: Low-Rank Codebook based Quantization for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ming-Yang Li, Wen-Pu Cai, Wu-Jun Li","submitted_at":"2024-05-31T16:21:05Z","abstract_excerpt":"Large language models~(LLMs) have recently demonstrated promising performance in many tasks. However, the high storage and computational cost of LLMs has become a challenge for deploying LLMs. Weight quantization has been widely used for model compression, which can reduce both storage and computational cost. Most existing weight quantization methods for LLMs use a rank-one codebook for quantization, which results in substantial accuracy loss when the compression ratio is high. In this paper, we propose a novel weight quantization method, called low-rank codebook based quantization~(LCQ), for "},"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":"2405.20973","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-31T16:21:05Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"d5ad5892777a0d09bc56dcfb4197b48589720741680903494a55b088af4a84f8","abstract_canon_sha256":"f34969cec7fef9239c02c7cd57f9cf0ccab978bdc20dc071e88e4eb5b8b8e902"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:11:58.956751Z","signature_b64":"pCTB/OfuEdBJnVNum7wV5JBc83byVvUZEXUzpAWCplr2nDrQlsmbdx7jy8ispoAW0WXu3Uwe0Mn4uWrMQy1KAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b56f6bd1faf72d2365ad176dd7965298de037a69098dd7b3f450c3a12c629d03","last_reissued_at":"2026-07-05T10:11:58.956231Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:11:58.956231Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LCQ: Low-Rank Codebook based Quantization for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ming-Yang Li, Wen-Pu Cai, Wu-Jun Li","submitted_at":"2024-05-31T16:21:05Z","abstract_excerpt":"Large language models~(LLMs) have recently demonstrated promising performance in many tasks. However, the high storage and computational cost of LLMs has become a challenge for deploying LLMs. Weight quantization has been widely used for model compression, which can reduce both storage and computational cost. Most existing weight quantization methods for LLMs use a rank-one codebook for quantization, which results in substantial accuracy loss when the compression ratio is high. In this paper, we propose a novel weight quantization method, called low-rank codebook based quantization~(LCQ), for "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.20973","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/2405.20973/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":"2405.20973","created_at":"2026-07-05T10:11:58.956295+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.20973v2","created_at":"2026-07-05T10:11:58.956295+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.20973","created_at":"2026-07-05T10:11:58.956295+00:00"},{"alias_kind":"pith_short_12","alias_value":"WVXWXUP264WS","created_at":"2026-07-05T10:11:58.956295+00:00"},{"alias_kind":"pith_short_16","alias_value":"WVXWXUP264WSGZNN","created_at":"2026-07-05T10:11:58.956295+00:00"},{"alias_kind":"pith_short_8","alias_value":"WVXWXUP2","created_at":"2026-07-05T10:11:58.956295+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.12984","citing_title":"Tilus: A Tile-Level GPGPU Programming Language for Low-Precision Computation","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD","json":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD.json","graph_json":"https://pith.science/api/pith-number/WVXWXUP264WSGZNNC5W5PFSSTD/graph.json","events_json":"https://pith.science/api/pith-number/WVXWXUP264WSGZNNC5W5PFSSTD/events.json","paper":"https://pith.science/paper/WVXWXUP2"},"agent_actions":{"view_html":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD","download_json":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD.json","view_paper":"https://pith.science/paper/WVXWXUP2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.20973&json=true","fetch_graph":"https://pith.science/api/pith-number/WVXWXUP264WSGZNNC5W5PFSSTD/graph.json","fetch_events":"https://pith.science/api/pith-number/WVXWXUP264WSGZNNC5W5PFSSTD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD/action/storage_attestation","attest_author":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD/action/author_attestation","sign_citation":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD/action/citation_signature","submit_replication":"https://pith.science/pith/WVXWXUP264WSGZNNC5W5PFSSTD/action/replication_record"}},"created_at":"2026-07-05T10:11:58.956295+00:00","updated_at":"2026-07-05T10:11:58.956295+00:00"}