{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TMBRF6KZKHA5GU5VXTPCVVP7HL","short_pith_number":"pith:TMBRF6KZ","schema_version":"1.0","canonical_sha256":"9b0312f95951c1d353b5bcde2ad5ff3af5099b67d7a00511d3d3890cfd128f51","source":{"kind":"arxiv","id":"2507.12566","version":1},"attestation_state":"computed","paper":{"title":"Mono-InternVL-1.5: Towards Cheaper and Faster Monolithic Multimodal Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Changyao Tian, Gen Luo, Hao Li, Jifeng Dai, Weiyun Wang, Wenhai Wang, Wenhan Dou, Wenhao Li, Xizhou Zhu, Xue Yang, Yu Qiao, Zhaokai Wang","submitted_at":"2025-07-16T18:31:23Z","abstract_excerpt":"This paper focuses on monolithic Multimodal Large Language Models (MLLMs), which integrate visual encoding and language decoding into a single model. Existing structures and pre-training strategies for monolithic MLLMs often suffer from unstable optimization and catastrophic forgetting. To address these challenges, our key idea is to embed a new visual parameter space into a pre-trained LLM, enabling stable learning of visual knowledge from noisy data via delta tuning. Based on this principle, we first introduce Mono-InternVL, an advanced monolithic MLLM that incorporates a set of visual exper"},"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":"2507.12566","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-07-16T18:31:23Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"d81a309edf9a571b4c3f2abd5886ba6527ae67663a7c8f1c944bbc1aba628e48","abstract_canon_sha256":"5bc4a20fed870c3e0b635c353333fe69978eb02b76ed842006650b48e53ad853"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:38:34.687231Z","signature_b64":"83A22roSNYYc+KI50f1DbEBk81qibXar9SLA8s/26JiRqH9fV1O7c7IyYybEg2E8doTP5Jgw8xAVQ2XVe7d2AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9b0312f95951c1d353b5bcde2ad5ff3af5099b67d7a00511d3d3890cfd128f51","last_reissued_at":"2026-07-05T11:38:34.686533Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:38:34.686533Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mono-InternVL-1.5: Towards Cheaper and Faster Monolithic Multimodal Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Changyao Tian, Gen Luo, Hao Li, Jifeng Dai, Weiyun Wang, Wenhai Wang, Wenhan Dou, Wenhao Li, Xizhou Zhu, Xue Yang, Yu Qiao, Zhaokai Wang","submitted_at":"2025-07-16T18:31:23Z","abstract_excerpt":"This paper focuses on monolithic Multimodal Large Language Models (MLLMs), which integrate visual encoding and language decoding into a single model. Existing structures and pre-training strategies for monolithic MLLMs often suffer from unstable optimization and catastrophic forgetting. To address these challenges, our key idea is to embed a new visual parameter space into a pre-trained LLM, enabling stable learning of visual knowledge from noisy data via delta tuning. Based on this principle, we first introduce Mono-InternVL, an advanced monolithic MLLM that incorporates a set of visual exper"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.12566","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/2507.12566/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":"2507.12566","created_at":"2026-07-05T11:38:34.686634+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.12566v1","created_at":"2026-07-05T11:38:34.686634+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.12566","created_at":"2026-07-05T11:38:34.686634+00:00"},{"alias_kind":"pith_short_12","alias_value":"TMBRF6KZKHA5","created_at":"2026-07-05T11:38:34.686634+00:00"},{"alias_kind":"pith_short_16","alias_value":"TMBRF6KZKHA5GU5V","created_at":"2026-07-05T11:38:34.686634+00:00"},{"alias_kind":"pith_short_8","alias_value":"TMBRF6KZ","created_at":"2026-07-05T11:38:34.686634+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.12500","citing_title":"SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2508.18265","citing_title":"InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency","ref_index":79,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL","json":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL.json","graph_json":"https://pith.science/api/pith-number/TMBRF6KZKHA5GU5VXTPCVVP7HL/graph.json","events_json":"https://pith.science/api/pith-number/TMBRF6KZKHA5GU5VXTPCVVP7HL/events.json","paper":"https://pith.science/paper/TMBRF6KZ"},"agent_actions":{"view_html":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL","download_json":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL.json","view_paper":"https://pith.science/paper/TMBRF6KZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.12566&json=true","fetch_graph":"https://pith.science/api/pith-number/TMBRF6KZKHA5GU5VXTPCVVP7HL/graph.json","fetch_events":"https://pith.science/api/pith-number/TMBRF6KZKHA5GU5VXTPCVVP7HL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL/action/storage_attestation","attest_author":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL/action/author_attestation","sign_citation":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL/action/citation_signature","submit_replication":"https://pith.science/pith/TMBRF6KZKHA5GU5VXTPCVVP7HL/action/replication_record"}},"created_at":"2026-07-05T11:38:34.686634+00:00","updated_at":"2026-07-05T11:38:34.686634+00:00"}