{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:X66QFZET2DQRFF7ITXEQCK2XUU","short_pith_number":"pith:X66QFZET","schema_version":"1.0","canonical_sha256":"bfbd02e493d0e11297e89dc9012b57a52ff2d3219fbf86bc484e131c3fbd2d38","source":{"kind":"arxiv","id":"2210.11948","version":2},"attestation_state":"computed","paper":{"title":"lo-fi: distributed fine-tuning without communication","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ali Farhadi, Ari S. Morcos, Ludwig Schmidt, Michael Rabbat, Mitchell Wortsman, Shen Li, Suchin Gururangan","submitted_at":"2022-10-19T20:15:18Z","abstract_excerpt":"When fine-tuning large neural networks, it is common to use multiple nodes and to communicate gradients at each optimization step. By contrast, we investigate completely local fine-tuning, which we refer to as lo-fi. During lo-fi, each node is fine-tuned independently without any communication. Then, the weights are averaged across nodes at the conclusion of fine-tuning. When fine-tuning DeiT-base and DeiT-large on ImageNet, this procedure matches accuracy in-distribution and improves accuracy under distribution shift compared to the baseline, which observes the same amount of data but communi"},"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":"2210.11948","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T20:15:18Z","cross_cats_sorted":[],"title_canon_sha256":"fb8776a97e37e8ad2ea7c56faff3223e63d821e98283432399e5b4d86bf5ed18","abstract_canon_sha256":"82abca90cfc1d8451869f09fec5ce31f7ee61f873f5b1eb074bc49a8f61b6e25"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:16:37.009398Z","signature_b64":"1knolnQWJDrwt4MnuMWbh6+w0mIY2uqkiS2Wp9O4Uc9SblsRhw+XsQ6tB88b3FcFW7DV2PF7CqnR19Px/zjpDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bfbd02e493d0e11297e89dc9012b57a52ff2d3219fbf86bc484e131c3fbd2d38","last_reissued_at":"2026-07-05T05:16:37.008900Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:16:37.008900Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"lo-fi: distributed fine-tuning without communication","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ali Farhadi, Ari S. Morcos, Ludwig Schmidt, Michael Rabbat, Mitchell Wortsman, Shen Li, Suchin Gururangan","submitted_at":"2022-10-19T20:15:18Z","abstract_excerpt":"When fine-tuning large neural networks, it is common to use multiple nodes and to communicate gradients at each optimization step. By contrast, we investigate completely local fine-tuning, which we refer to as lo-fi. During lo-fi, each node is fine-tuned independently without any communication. Then, the weights are averaged across nodes at the conclusion of fine-tuning. When fine-tuning DeiT-base and DeiT-large on ImageNet, this procedure matches accuracy in-distribution and improves accuracy under distribution shift compared to the baseline, which observes the same amount of data but communi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.11948","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/2210.11948/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":"2210.11948","created_at":"2026-07-05T05:16:37.008961+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.11948v2","created_at":"2026-07-05T05:16:37.008961+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.11948","created_at":"2026-07-05T05:16:37.008961+00:00"},{"alias_kind":"pith_short_12","alias_value":"X66QFZET2DQR","created_at":"2026-07-05T05:16:37.008961+00:00"},{"alias_kind":"pith_short_16","alias_value":"X66QFZET2DQRFF7I","created_at":"2026-07-05T05:16:37.008961+00:00"},{"alias_kind":"pith_short_8","alias_value":"X66QFZET","created_at":"2026-07-05T05:16:37.008961+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2405.07987","citing_title":"The Platonic Representation Hypothesis","ref_index":132,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU","json":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU.json","graph_json":"https://pith.science/api/pith-number/X66QFZET2DQRFF7ITXEQCK2XUU/graph.json","events_json":"https://pith.science/api/pith-number/X66QFZET2DQRFF7ITXEQCK2XUU/events.json","paper":"https://pith.science/paper/X66QFZET"},"agent_actions":{"view_html":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU","download_json":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU.json","view_paper":"https://pith.science/paper/X66QFZET","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.11948&json=true","fetch_graph":"https://pith.science/api/pith-number/X66QFZET2DQRFF7ITXEQCK2XUU/graph.json","fetch_events":"https://pith.science/api/pith-number/X66QFZET2DQRFF7ITXEQCK2XUU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU/action/storage_attestation","attest_author":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU/action/author_attestation","sign_citation":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU/action/citation_signature","submit_replication":"https://pith.science/pith/X66QFZET2DQRFF7ITXEQCK2XUU/action/replication_record"}},"created_at":"2026-07-05T05:16:37.008961+00:00","updated_at":"2026-07-05T05:16:37.008961+00:00"}