{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NG3EV4WOLZ3ZPYVMTXFGB4PI5L","short_pith_number":"pith:NG3EV4WO","schema_version":"1.0","canonical_sha256":"69b64af2ce5e7797e2ac9dca60f1e8eae86cea74d20667fff86f5399225255e6","source":{"kind":"arxiv","id":"2412.04424","version":1},"attestation_state":"computed","paper":{"title":"Florence-VL: Enhancing Vision-Language Models with Generative Vision Encoder and Depth-Breadth Fusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bin Xiao, Dianqi Li, Haiping Wu, Jianfeng Gao, Jianwei Yang, Jiuhai Chen, Tianyi Zhou","submitted_at":"2024-12-05T18:50:39Z","abstract_excerpt":"We present Florence-VL, a new family of multimodal large language models (MLLMs) with enriched visual representations produced by Florence-2, a generative vision foundation model. Unlike the widely used CLIP-style vision transformer trained by contrastive learning, Florence-2 can capture different levels and aspects of visual features, which are more versatile to be adapted to diverse downstream tasks. We propose a novel feature-fusion architecture and an innovative training recipe that effectively integrates Florence-2's visual features into pretrained LLMs, such as Phi 3.5 and LLama 3. In pa"},"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":"2412.04424","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-05T18:50:39Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ef574480848cd70f311365bb6258e915ecc3d5f92987715c43af3b9a1e5397e3","abstract_canon_sha256":"b815f73a9b29458f507935aea8ca52779ee7397e0082c61887e7b5aad3a959a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:45:03.738512Z","signature_b64":"MU+/HyJmwdVtwq1gYyuN4j5aqOl7ZKtKJ/BzU6jYasEJfZaTgyzPwgty2o9EWUlXPVaJDcfiGRT8XI7wA+q7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69b64af2ce5e7797e2ac9dca60f1e8eae86cea74d20667fff86f5399225255e6","last_reissued_at":"2026-07-05T09:45:03.738046Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:45:03.738046Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Florence-VL: Enhancing Vision-Language Models with Generative Vision Encoder and Depth-Breadth Fusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bin Xiao, Dianqi Li, Haiping Wu, Jianfeng Gao, Jianwei Yang, Jiuhai Chen, Tianyi Zhou","submitted_at":"2024-12-05T18:50:39Z","abstract_excerpt":"We present Florence-VL, a new family of multimodal large language models (MLLMs) with enriched visual representations produced by Florence-2, a generative vision foundation model. Unlike the widely used CLIP-style vision transformer trained by contrastive learning, Florence-2 can capture different levels and aspects of visual features, which are more versatile to be adapted to diverse downstream tasks. We propose a novel feature-fusion architecture and an innovative training recipe that effectively integrates Florence-2's visual features into pretrained LLMs, such as Phi 3.5 and LLama 3. In pa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.04424","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/2412.04424/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":"2412.04424","created_at":"2026-07-05T09:45:03.738109+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.04424v1","created_at":"2026-07-05T09:45:03.738109+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.04424","created_at":"2026-07-05T09:45:03.738109+00:00"},{"alias_kind":"pith_short_12","alias_value":"NG3EV4WOLZ3Z","created_at":"2026-07-05T09:45:03.738109+00:00"},{"alias_kind":"pith_short_16","alias_value":"NG3EV4WOLZ3ZPYVM","created_at":"2026-07-05T09:45:03.738109+00:00"},{"alias_kind":"pith_short_8","alias_value":"NG3EV4WO","created_at":"2026-07-05T09:45:03.738109+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2504.09925","citing_title":"FLARE: Fully Integration of Vision-Language Representations for Deep Cross-Modal Understanding","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L","json":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L.json","graph_json":"https://pith.science/api/pith-number/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/graph.json","events_json":"https://pith.science/api/pith-number/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/events.json","paper":"https://pith.science/paper/NG3EV4WO"},"agent_actions":{"view_html":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L","download_json":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L.json","view_paper":"https://pith.science/paper/NG3EV4WO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.04424&json=true","fetch_graph":"https://pith.science/api/pith-number/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/graph.json","fetch_events":"https://pith.science/api/pith-number/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/action/storage_attestation","attest_author":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/action/author_attestation","sign_citation":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/action/citation_signature","submit_replication":"https://pith.science/pith/NG3EV4WOLZ3ZPYVMTXFGB4PI5L/action/replication_record"}},"created_at":"2026-07-05T09:45:03.738109+00:00","updated_at":"2026-07-05T09:45:03.738109+00:00"}