{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:3EXDUGZSVUJD26NZH4R5POOEKU","short_pith_number":"pith:3EXDUGZS","schema_version":"1.0","canonical_sha256":"d92e3a1b32ad123d79b93f23d7b9c455029334bb0199b5ab4be7ed20b5de2b01","source":{"kind":"arxiv","id":"2111.07991","version":3},"attestation_state":"computed","paper":{"title":"LiT: Zero-Shot Transfer with Locked-image text Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander Kolesnikov, Andreas Steiner, Basil Mustafa, Daniel Keysers, Lucas Beyer, Xiaohua Zhai, Xiao Wang","submitted_at":"2021-11-15T18:53:48Z","abstract_excerpt":"This paper presents contrastive-tuning, a simple method employing contrastive training to align image and text models while still taking advantage of their pre-training. In our empirical study we find that locked pre-trained image models with unlocked text models work best. We call this instance of contrastive-tuning \"Locked-image Tuning\" (LiT), which just teaches a text model to read out good representations from a pre-trained image model for new tasks. A LiT model gains the capability of zero-shot transfer to new vision tasks, such as image classification or retrieval. The proposed LiT is wi"},"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":"2111.07991","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-11-15T18:53:48Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"efa505a980fa227f6dcc37285dd80502dd829eaa201a6888dfb529a81f5318ac","abstract_canon_sha256":"a7bb4468babb42232d04f64e38e16ab46bf43b95ef0783fc0fea5f0f4fa23cf4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:34:02.997209Z","signature_b64":"2mjw9mVL54nSpVk1XodTBvcCeNrMF1AXoDKJWspgcneE5MLD9sm65BYPDZdLWFAiUskkEsq19FcqKjepww+7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d92e3a1b32ad123d79b93f23d7b9c455029334bb0199b5ab4be7ed20b5de2b01","last_reissued_at":"2026-07-05T04:34:02.996662Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:34:02.996662Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LiT: Zero-Shot Transfer with Locked-image text Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander Kolesnikov, Andreas Steiner, Basil Mustafa, Daniel Keysers, Lucas Beyer, Xiaohua Zhai, Xiao Wang","submitted_at":"2021-11-15T18:53:48Z","abstract_excerpt":"This paper presents contrastive-tuning, a simple method employing contrastive training to align image and text models while still taking advantage of their pre-training. In our empirical study we find that locked pre-trained image models with unlocked text models work best. We call this instance of contrastive-tuning \"Locked-image Tuning\" (LiT), which just teaches a text model to read out good representations from a pre-trained image model for new tasks. A LiT model gains the capability of zero-shot transfer to new vision tasks, such as image classification or retrieval. The proposed LiT is wi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2111.07991","kind":"arxiv","version":3},"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/2111.07991/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":"2111.07991","created_at":"2026-07-05T04:34:02.996728+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.07991v3","created_at":"2026-07-05T04:34:02.996728+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.07991","created_at":"2026-07-05T04:34:02.996728+00:00"},{"alias_kind":"pith_short_12","alias_value":"3EXDUGZSVUJD","created_at":"2026-07-05T04:34:02.996728+00:00"},{"alias_kind":"pith_short_16","alias_value":"3EXDUGZSVUJD26NZ","created_at":"2026-07-05T04:34:02.996728+00:00"},{"alias_kind":"pith_short_8","alias_value":"3EXDUGZS","created_at":"2026-07-05T04:34:02.996728+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2204.00598","citing_title":"Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2205.01917","citing_title":"CoCa: Contrastive Captioners are Image-Text Foundation Models","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2210.08402","citing_title":"LAION-5B: An open large-scale dataset for training next generation image-text models","ref_index":98,"is_internal_anchor":false},{"citing_arxiv_id":"2204.14198","citing_title":"Flamingo: a Visual Language Model for Few-Shot Learning","ref_index":147,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU","json":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU.json","graph_json":"https://pith.science/api/pith-number/3EXDUGZSVUJD26NZH4R5POOEKU/graph.json","events_json":"https://pith.science/api/pith-number/3EXDUGZSVUJD26NZH4R5POOEKU/events.json","paper":"https://pith.science/paper/3EXDUGZS"},"agent_actions":{"view_html":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU","download_json":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU.json","view_paper":"https://pith.science/paper/3EXDUGZS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.07991&json=true","fetch_graph":"https://pith.science/api/pith-number/3EXDUGZSVUJD26NZH4R5POOEKU/graph.json","fetch_events":"https://pith.science/api/pith-number/3EXDUGZSVUJD26NZH4R5POOEKU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU/action/storage_attestation","attest_author":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU/action/author_attestation","sign_citation":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU/action/citation_signature","submit_replication":"https://pith.science/pith/3EXDUGZSVUJD26NZH4R5POOEKU/action/replication_record"}},"created_at":"2026-07-05T04:34:02.996728+00:00","updated_at":"2026-07-05T04:34:02.996728+00:00"}