{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:INYRK5U66U2NYQOUBJVMSLAFTX","short_pith_number":"pith:INYRK5U6","schema_version":"1.0","canonical_sha256":"437115769ef534dc41d40a6ac92c059ddc3f20f7157f144b00b8078d30100e02","source":{"kind":"arxiv","id":"2412.05185","version":2},"attestation_state":"computed","paper":{"title":"LinVT: Empower Your Image-level Large Language Model to Understand Videos","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.MM"],"primary_cat":"cs.CV","authors_text":"Dengjie Li, Haoxian Tan, Lishuai Gao, Yingsen Zeng, Yujie Zhong, Zheng Zhao","submitted_at":"2024-12-06T17:04:42Z","abstract_excerpt":"Large Language Models (LLMs) have been widely used in various tasks, motivating us to develop an LLM-based assistant for videos. Instead of training from scratch, we propose a module to transform arbitrary well-trained image-based LLMs into video-LLMs (after being trained on video data). To better adapt image-LLMs for processing videos, we introduce two design principles: linear transformation to preserve the original visual-language alignment and representative information condensation from redundant video content. Guided by these principles, we propose a plug-and-play Linear Video Tokenizer("},"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":"2412.05185","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-06T17:04:42Z","cross_cats_sorted":["cs.LG","cs.MM"],"title_canon_sha256":"da5c277768807f719a47cc19c1c87c518ad1ad93c9fb1f0b72ebd0781ad36513","abstract_canon_sha256":"e7f062b7ab18c6dd8eae9d89755175c32e48e488de11afebe6a61070caab9963"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:47:37.579988Z","signature_b64":"k0V4ZhReY5cfks8AA400Vy7IbNQEk8xDskS0PIPWnEsGBop+SKtL+nlPHqfpCEuGLLxkpqjVlU9sEac9A8e6Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"437115769ef534dc41d40a6ac92c059ddc3f20f7157f144b00b8078d30100e02","last_reissued_at":"2026-07-05T09:47:37.579506Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:47:37.579506Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LinVT: Empower Your Image-level Large Language Model to Understand Videos","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.MM"],"primary_cat":"cs.CV","authors_text":"Dengjie Li, Haoxian Tan, Lishuai Gao, Yingsen Zeng, Yujie Zhong, Zheng Zhao","submitted_at":"2024-12-06T17:04:42Z","abstract_excerpt":"Large Language Models (LLMs) have been widely used in various tasks, motivating us to develop an LLM-based assistant for videos. Instead of training from scratch, we propose a module to transform arbitrary well-trained image-based LLMs into video-LLMs (after being trained on video data). To better adapt image-LLMs for processing videos, we introduce two design principles: linear transformation to preserve the original visual-language alignment and representative information condensation from redundant video content. Guided by these principles, we propose a plug-and-play Linear Video Tokenizer("},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.05185","kind":"arxiv","version":2},"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/2412.05185/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":"2412.05185","created_at":"2026-07-05T09:47:37.579565+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.05185v2","created_at":"2026-07-05T09:47:37.579565+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.05185","created_at":"2026-07-05T09:47:37.579565+00:00"},{"alias_kind":"pith_short_12","alias_value":"INYRK5U66U2N","created_at":"2026-07-05T09:47:37.579565+00:00"},{"alias_kind":"pith_short_16","alias_value":"INYRK5U66U2NYQOU","created_at":"2026-07-05T09:47:37.579565+00:00"},{"alias_kind":"pith_short_8","alias_value":"INYRK5U6","created_at":"2026-07-05T09:47:37.579565+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.06532","citing_title":"GOPAgen: Motion-Aware and Efficient Agentic Long-Video Understanding with Structural Memory and Hierarchical Reasoning","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29445","citing_title":"Bridging VideoQA and Video-Guided Agentic Tasks via Generalized Keyframe Extraction","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2602.20913","citing_title":"LongVideo-R1: Smart Navigation for Low-cost Long Video Understanding","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX","json":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX.json","graph_json":"https://pith.science/api/pith-number/INYRK5U66U2NYQOUBJVMSLAFTX/graph.json","events_json":"https://pith.science/api/pith-number/INYRK5U66U2NYQOUBJVMSLAFTX/events.json","paper":"https://pith.science/paper/INYRK5U6"},"agent_actions":{"view_html":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX","download_json":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX.json","view_paper":"https://pith.science/paper/INYRK5U6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.05185&json=true","fetch_graph":"https://pith.science/api/pith-number/INYRK5U66U2NYQOUBJVMSLAFTX/graph.json","fetch_events":"https://pith.science/api/pith-number/INYRK5U66U2NYQOUBJVMSLAFTX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX/action/storage_attestation","attest_author":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX/action/author_attestation","sign_citation":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX/action/citation_signature","submit_replication":"https://pith.science/pith/INYRK5U66U2NYQOUBJVMSLAFTX/action/replication_record"}},"created_at":"2026-07-05T09:47:37.579565+00:00","updated_at":"2026-07-05T09:47:37.579565+00:00"}