{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:BJ2NLSZEZB3UFOYPNLZWPROJGU","short_pith_number":"pith:BJ2NLSZE","schema_version":"1.0","canonical_sha256":"0a74d5cb24c87742bb0f6af367c5c93517ef7defacf3580979c068463bf17ecf","source":{"kind":"arxiv","id":"2212.06138","version":1},"attestation_state":"computed","paper":{"title":"CLIP Itself is a Strong Fine-tuner: Achieving 85.7% and 88.0% Top-1 Accuracy with ViT-B and ViT-L on ImageNet","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Dong Chen, Dongdong Chen, Fang Wen, Jianmin Bao, Lu Yuan, Nenghai Yu, Shuyang Gu, Ting Zhang, Weiming Zhang, Xiaoyi Dong","submitted_at":"2022-12-12T18:59:59Z","abstract_excerpt":"Recent studies have shown that CLIP has achieved remarkable success in performing zero-shot inference while its fine-tuning performance is not satisfactory. In this paper, we identify that fine-tuning performance is significantly impacted by hyper-parameter choices. We examine various key hyper-parameters and empirically evaluate their impact in fine-tuning CLIP for classification tasks through a comprehensive study. We find that the fine-tuning performance of CLIP is substantially underestimated. Equipped with hyper-parameter refinement, we demonstrate CLIP itself is better or at least compet"},"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":"2212.06138","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-12-12T18:59:59Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"5993597c12b588fb43a4fce757d0dd5eff39d0af3d5cdd96a9bbdcbb0ff1b057","abstract_canon_sha256":"ca2cf0324f77ca9ad63b8231b395f4adb8f625e16d16f801de5be93091c98f97"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:24:23.252040Z","signature_b64":"rrEpjAKzNRfkwD6WrPLUv7KJnwc0IeOAbagaO6HwegXO4Kq26iqeV9s2P01eQbr5CGzQ/eH7RPCWlI5uGlYECw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a74d5cb24c87742bb0f6af367c5c93517ef7defacf3580979c068463bf17ecf","last_reissued_at":"2026-07-05T05:24:23.251569Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:24:23.251569Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CLIP Itself is a Strong Fine-tuner: Achieving 85.7% and 88.0% Top-1 Accuracy with ViT-B and ViT-L on ImageNet","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Dong Chen, Dongdong Chen, Fang Wen, Jianmin Bao, Lu Yuan, Nenghai Yu, Shuyang Gu, Ting Zhang, Weiming Zhang, Xiaoyi Dong","submitted_at":"2022-12-12T18:59:59Z","abstract_excerpt":"Recent studies have shown that CLIP has achieved remarkable success in performing zero-shot inference while its fine-tuning performance is not satisfactory. In this paper, we identify that fine-tuning performance is significantly impacted by hyper-parameter choices. We examine various key hyper-parameters and empirically evaluate their impact in fine-tuning CLIP for classification tasks through a comprehensive study. We find that the fine-tuning performance of CLIP is substantially underestimated. Equipped with hyper-parameter refinement, we demonstrate CLIP itself is better or at least compet"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.06138","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/2212.06138/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":"2212.06138","created_at":"2026-07-05T05:24:23.251628+00:00"},{"alias_kind":"arxiv_version","alias_value":"2212.06138v1","created_at":"2026-07-05T05:24:23.251628+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.06138","created_at":"2026-07-05T05:24:23.251628+00:00"},{"alias_kind":"pith_short_12","alias_value":"BJ2NLSZEZB3U","created_at":"2026-07-05T05:24:23.251628+00:00"},{"alias_kind":"pith_short_16","alias_value":"BJ2NLSZEZB3UFOYP","created_at":"2026-07-05T05:24:23.251628+00:00"},{"alias_kind":"pith_short_8","alias_value":"BJ2NLSZE","created_at":"2026-07-05T05:24:23.251628+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.18878","citing_title":"Prognostic Value of Lung Ultrasound Biomarkers for Readmission Risk in Congestive Heart Failure: A Pilot Data-Driven Analysis","ref_index":208,"is_internal_anchor":false},{"citing_arxiv_id":"2303.15343","citing_title":"Sigmoid Loss for Language Image Pre-Training","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU","json":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU.json","graph_json":"https://pith.science/api/pith-number/BJ2NLSZEZB3UFOYPNLZWPROJGU/graph.json","events_json":"https://pith.science/api/pith-number/BJ2NLSZEZB3UFOYPNLZWPROJGU/events.json","paper":"https://pith.science/paper/BJ2NLSZE"},"agent_actions":{"view_html":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU","download_json":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU.json","view_paper":"https://pith.science/paper/BJ2NLSZE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2212.06138&json=true","fetch_graph":"https://pith.science/api/pith-number/BJ2NLSZEZB3UFOYPNLZWPROJGU/graph.json","fetch_events":"https://pith.science/api/pith-number/BJ2NLSZEZB3UFOYPNLZWPROJGU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU/action/storage_attestation","attest_author":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU/action/author_attestation","sign_citation":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU/action/citation_signature","submit_replication":"https://pith.science/pith/BJ2NLSZEZB3UFOYPNLZWPROJGU/action/replication_record"}},"created_at":"2026-07-05T05:24:23.251628+00:00","updated_at":"2026-07-05T05:24:23.251628+00:00"}