{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:AATBMNCAUWIHTIPX3UA32QHOEE","short_pith_number":"pith:AATBMNCA","schema_version":"1.0","canonical_sha256":"0026163440a59079a1f7dd01bd40ee21020384295b77f4ac10101680518e72ed","source":{"kind":"arxiv","id":"2403.06504","version":2},"attestation_state":"computed","paper":{"title":"LoHan: Low-Cost High-Performance Framework to Fine-Tune 100B Model on a Consumer GPU","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Binhang Yuan, Changyue Liao, Fei Wu, Jun Xie, Kaiqi Chen, Mo Sun, Zeke Wang, Zihan Yang","submitted_at":"2024-03-11T08:25:53Z","abstract_excerpt":"Nowadays, AI researchers become more and more interested in fine-tuning a pre-trained LLM, whose size has grown to up to over 100B parameters, for their downstream tasks. One approach to fine-tune such huge models is to aggregate device memory from many GPUs. However, this approach introduces prohibitive costs for most data scientists with a limited budget for high-end GPU servers. In this paper, we focus on LLM fine-tuning on a single consumer-grade GPU in a commodity server with limited main memory capacity, which is accessible to most AI researchers. In such a scenario, existing offloading-"},"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":"2403.06504","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.DC","submitted_at":"2024-03-11T08:25:53Z","cross_cats_sorted":[],"title_canon_sha256":"9acf218c78de9b5fec4fdf49a59adacb3ee8f9bb2f25a83244bfdf7f35252f7f","abstract_canon_sha256":"b1ceace6e9d95386e2305a419220222fba8b77ecbdf98fcf785be8552fa27fdd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:53:40.436787Z","signature_b64":"iH0fwkDA5cQq0Jab8WchqQQLxmFtx55h1r784fZuoDFPvUWvkkAZjc2GjXNLxmmwu9qSSNp1o0Ji2ZTT0ptyCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0026163440a59079a1f7dd01bd40ee21020384295b77f4ac10101680518e72ed","last_reissued_at":"2026-07-05T09:53:40.436255Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:53:40.436255Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LoHan: Low-Cost High-Performance Framework to Fine-Tune 100B Model on a Consumer GPU","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Binhang Yuan, Changyue Liao, Fei Wu, Jun Xie, Kaiqi Chen, Mo Sun, Zeke Wang, Zihan Yang","submitted_at":"2024-03-11T08:25:53Z","abstract_excerpt":"Nowadays, AI researchers become more and more interested in fine-tuning a pre-trained LLM, whose size has grown to up to over 100B parameters, for their downstream tasks. One approach to fine-tune such huge models is to aggregate device memory from many GPUs. However, this approach introduces prohibitive costs for most data scientists with a limited budget for high-end GPU servers. In this paper, we focus on LLM fine-tuning on a single consumer-grade GPU in a commodity server with limited main memory capacity, which is accessible to most AI researchers. In such a scenario, existing offloading-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.06504","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/2403.06504/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":"2403.06504","created_at":"2026-07-05T09:53:40.436323+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.06504v2","created_at":"2026-07-05T09:53:40.436323+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.06504","created_at":"2026-07-05T09:53:40.436323+00:00"},{"alias_kind":"pith_short_12","alias_value":"AATBMNCAUWIH","created_at":"2026-07-05T09:53:40.436323+00:00"},{"alias_kind":"pith_short_16","alias_value":"AATBMNCAUWIHTIPX","created_at":"2026-07-05T09:53:40.436323+00:00"},{"alias_kind":"pith_short_8","alias_value":"AATBMNCA","created_at":"2026-07-05T09:53:40.436323+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.16269","citing_title":"Train the Trainers -- An Agentic AI Framework for Peer-Based Mental Health Support in Battlefield Environments","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27085","citing_title":"Efficient Training on Multiple Consumer GPUs with RoundPipe","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05987","citing_title":"Flowr -- Scaling Up Retail Supply Chain Operations Through Agentic AI in Large Scale Supermarket Chains","ref_index":47,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE","json":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE.json","graph_json":"https://pith.science/api/pith-number/AATBMNCAUWIHTIPX3UA32QHOEE/graph.json","events_json":"https://pith.science/api/pith-number/AATBMNCAUWIHTIPX3UA32QHOEE/events.json","paper":"https://pith.science/paper/AATBMNCA"},"agent_actions":{"view_html":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE","download_json":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE.json","view_paper":"https://pith.science/paper/AATBMNCA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.06504&json=true","fetch_graph":"https://pith.science/api/pith-number/AATBMNCAUWIHTIPX3UA32QHOEE/graph.json","fetch_events":"https://pith.science/api/pith-number/AATBMNCAUWIHTIPX3UA32QHOEE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE/action/storage_attestation","attest_author":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE/action/author_attestation","sign_citation":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE/action/citation_signature","submit_replication":"https://pith.science/pith/AATBMNCAUWIHTIPX3UA32QHOEE/action/replication_record"}},"created_at":"2026-07-05T09:53:40.436323+00:00","updated_at":"2026-07-05T09:53:40.436323+00:00"}