{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:3TK7BFQHNO5RISDH3T5HVK6332","short_pith_number":"pith:3TK7BFQH","schema_version":"1.0","canonical_sha256":"dcd5f096076bbb144867dcfa7aabdbdead0cd79aa20f1b7430aa20ef55efec86","source":{"kind":"arxiv","id":"2002.11770","version":1},"attestation_state":"computed","paper":{"title":"Rethinking the Hyperparameters for Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.CV","authors_text":"Avinash Ravichandran, Hao Li, Hao Yang, Michael Lam, Pratik Chaudhari, Rahul Bhotika, Stefano Soatto","submitted_at":"2020-02-19T18:59:52Z","abstract_excerpt":"Fine-tuning from pre-trained ImageNet models has become the de-facto standard for various computer vision tasks. Current practices for fine-tuning typically involve selecting an ad-hoc choice of hyperparameters and keeping them fixed to values normally used for training from scratch. This paper re-examines several common practices of setting hyperparameters for fine-tuning. Our findings are based on extensive empirical evaluation for fine-tuning on various transfer learning benchmarks. (1) While prior works have thoroughly investigated learning rate and batch size, momentum for fine-tuning is "},"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":"2002.11770","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2020-02-19T18:59:52Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"35cf37c4b0c326843f33bb7c629ae624a50d54fb6d80189001f8181568a0611c","abstract_canon_sha256":"c6151f1acdd152613e55ec9305cd3b40cc2b7942a1db3a1fee5306d9e08ec887"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:44:08.776918Z","signature_b64":"rpV1+OxNlzHS8VgrC9fU4M6Rgz8bifp9TAfVOheAmuQx7cLY1tM/skAMHHkWfaIeJMdgk1obyGIoskcH5SY5Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dcd5f096076bbb144867dcfa7aabdbdead0cd79aa20f1b7430aa20ef55efec86","last_reissued_at":"2026-07-05T00:44:08.776458Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:44:08.776458Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rethinking the Hyperparameters for Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.CV","authors_text":"Avinash Ravichandran, Hao Li, Hao Yang, Michael Lam, Pratik Chaudhari, Rahul Bhotika, Stefano Soatto","submitted_at":"2020-02-19T18:59:52Z","abstract_excerpt":"Fine-tuning from pre-trained ImageNet models has become the de-facto standard for various computer vision tasks. Current practices for fine-tuning typically involve selecting an ad-hoc choice of hyperparameters and keeping them fixed to values normally used for training from scratch. This paper re-examines several common practices of setting hyperparameters for fine-tuning. Our findings are based on extensive empirical evaluation for fine-tuning on various transfer learning benchmarks. (1) While prior works have thoroughly investigated learning rate and batch size, momentum for fine-tuning is "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.11770","kind":"arxiv","version":1},"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/2002.11770/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":"2002.11770","created_at":"2026-07-05T00:44:08.776518+00:00"},{"alias_kind":"arxiv_version","alias_value":"2002.11770v1","created_at":"2026-07-05T00:44:08.776518+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.11770","created_at":"2026-07-05T00:44:08.776518+00:00"},{"alias_kind":"pith_short_12","alias_value":"3TK7BFQHNO5R","created_at":"2026-07-05T00:44:08.776518+00:00"},{"alias_kind":"pith_short_16","alias_value":"3TK7BFQHNO5RISDH","created_at":"2026-07-05T00:44:08.776518+00:00"},{"alias_kind":"pith_short_8","alias_value":"3TK7BFQH","created_at":"2026-07-05T00:44:08.776518+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.20009","citing_title":"Training Neural Networks with Optimal Double-Bayesian Learning","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332","json":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332.json","graph_json":"https://pith.science/api/pith-number/3TK7BFQHNO5RISDH3T5HVK6332/graph.json","events_json":"https://pith.science/api/pith-number/3TK7BFQHNO5RISDH3T5HVK6332/events.json","paper":"https://pith.science/paper/3TK7BFQH"},"agent_actions":{"view_html":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332","download_json":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332.json","view_paper":"https://pith.science/paper/3TK7BFQH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2002.11770&json=true","fetch_graph":"https://pith.science/api/pith-number/3TK7BFQHNO5RISDH3T5HVK6332/graph.json","fetch_events":"https://pith.science/api/pith-number/3TK7BFQHNO5RISDH3T5HVK6332/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332/action/storage_attestation","attest_author":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332/action/author_attestation","sign_citation":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332/action/citation_signature","submit_replication":"https://pith.science/pith/3TK7BFQHNO5RISDH3T5HVK6332/action/replication_record"}},"created_at":"2026-07-05T00:44:08.776518+00:00","updated_at":"2026-07-05T00:44:08.776518+00:00"}