{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3OLMVMZ33U7EWQIJQGR75CY35W","short_pith_number":"pith:3OLMVMZ3","schema_version":"1.0","canonical_sha256":"db96cab33bdd3e4b410981a3fe8b1beda61ca4ec4cd90dc1f28715c206370e57","source":{"kind":"arxiv","id":"2312.00674","version":1},"attestation_state":"computed","paper":{"title":"LightCLIP: Learning Multi-Level Interaction for Lightweight Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fanyi Du, Kai Han, Tianyu Guo, Wei He, Yehui Tang, Ying Nie, Yunhe Wang","submitted_at":"2023-12-01T15:54:55Z","abstract_excerpt":"Vision-language pre-training like CLIP has shown promising performance on various downstream tasks such as zero-shot image classification and image-text retrieval. Most of the existing CLIP-alike works usually adopt relatively large image encoders like ResNet50 and ViT, while the lightweight counterparts are rarely discussed. In this paper, we propose a multi-level interaction paradigm for training lightweight CLIP models. Firstly, to mitigate the problem that some image-text pairs are not strictly one-to-one correspondence, we improve the conventional global instance-level alignment objective"},"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":"2312.00674","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-12-01T15:54:55Z","cross_cats_sorted":[],"title_canon_sha256":"fff913fb4886e40e23b2f7f9d9d5d5650b9b5e68c3fb8e03ad72bfa1f30e4236","abstract_canon_sha256":"2fa7b5279f7ad84a15ae6a3cb506b48ab71941d92871d5f1ebf0cfc895b82879"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:19:12.920810Z","signature_b64":"sUKPbYAzUbSeDJCBOuMlmoiR74UHNKzdBlWpiS/uWbKtcLIIJDWWso5iiiIiOrc/UJNNov/jbW4PkE5Q0T67DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db96cab33bdd3e4b410981a3fe8b1beda61ca4ec4cd90dc1f28715c206370e57","last_reissued_at":"2026-07-05T07:19:12.920359Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:19:12.920359Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LightCLIP: Learning Multi-Level Interaction for Lightweight Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fanyi Du, Kai Han, Tianyu Guo, Wei He, Yehui Tang, Ying Nie, Yunhe Wang","submitted_at":"2023-12-01T15:54:55Z","abstract_excerpt":"Vision-language pre-training like CLIP has shown promising performance on various downstream tasks such as zero-shot image classification and image-text retrieval. Most of the existing CLIP-alike works usually adopt relatively large image encoders like ResNet50 and ViT, while the lightweight counterparts are rarely discussed. In this paper, we propose a multi-level interaction paradigm for training lightweight CLIP models. Firstly, to mitigate the problem that some image-text pairs are not strictly one-to-one correspondence, we improve the conventional global instance-level alignment objective"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.00674","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/2312.00674/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":"2312.00674","created_at":"2026-07-05T07:19:12.920444+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.00674v1","created_at":"2026-07-05T07:19:12.920444+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.00674","created_at":"2026-07-05T07:19:12.920444+00:00"},{"alias_kind":"pith_short_12","alias_value":"3OLMVMZ33U7E","created_at":"2026-07-05T07:19:12.920444+00:00"},{"alias_kind":"pith_short_16","alias_value":"3OLMVMZ33U7EWQIJ","created_at":"2026-07-05T07:19:12.920444+00:00"},{"alias_kind":"pith_short_8","alias_value":"3OLMVMZ3","created_at":"2026-07-05T07:19:12.920444+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W","json":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W.json","graph_json":"https://pith.science/api/pith-number/3OLMVMZ33U7EWQIJQGR75CY35W/graph.json","events_json":"https://pith.science/api/pith-number/3OLMVMZ33U7EWQIJQGR75CY35W/events.json","paper":"https://pith.science/paper/3OLMVMZ3"},"agent_actions":{"view_html":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W","download_json":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W.json","view_paper":"https://pith.science/paper/3OLMVMZ3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.00674&json=true","fetch_graph":"https://pith.science/api/pith-number/3OLMVMZ33U7EWQIJQGR75CY35W/graph.json","fetch_events":"https://pith.science/api/pith-number/3OLMVMZ33U7EWQIJQGR75CY35W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W/action/storage_attestation","attest_author":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W/action/author_attestation","sign_citation":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W/action/citation_signature","submit_replication":"https://pith.science/pith/3OLMVMZ33U7EWQIJQGR75CY35W/action/replication_record"}},"created_at":"2026-07-05T07:19:12.920444+00:00","updated_at":"2026-07-05T07:19:12.920444+00:00"}