{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4VDWM77NGHJTOYV5KKHMFY2HIG","short_pith_number":"pith:4VDWM77N","schema_version":"1.0","canonical_sha256":"e547667fed31d33762bd528ec2e34741af18073ab94082f5fa4e08b5ae1443e1","source":{"kind":"arxiv","id":"2410.16261","version":3},"attestation_state":"computed","paper":{"title":"Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Erfei Cui, Hao Tian, Jifeng Dai, Jinguo Zhu, Junjun He, Lewei Lu, Shenglong Ye, Tong Lu, Weiyun Wang, Wenhai Wang, Xizhou Zhu, Yiming Ren, Yu Qiao, Zhangwei Gao, Zhe Chen","submitted_at":"2024-10-21T17:58:20Z","abstract_excerpt":"Multimodal large language models (MLLMs) have demonstrated impressive performance in vision-language tasks across a broad spectrum of domains. However, the large model scale and associated high computational costs pose significant challenges for training and deploying MLLMs on consumer-grade GPUs or edge devices, thereby hindering their widespread application. In this work, we introduce Mini-InternVL, a series of MLLMs with parameters ranging from 1B to 4B, which achieves 90% of the performance with only 5% of the parameters. This significant improvement in efficiency and effectiveness makes o"},"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":"2410.16261","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-10-21T17:58:20Z","cross_cats_sorted":[],"title_canon_sha256":"384021ef382e534de4b50b887c771bf1a4bf910dd99e7c18196c369af8a46293","abstract_canon_sha256":"3d84875451bb17ca8fea8dda18a88d48e53145ac1e3994d30aff3f76aa89fd6d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:32:33.423401Z","signature_b64":"svPUHS+2GtGLYrwaCfUAgVdsLSWDjt0tVihbi++Kg/kZ1rdaBxWXuIoOKyOxa7tQR3cERGvI3/CcOPMZJJ1HAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e547667fed31d33762bd528ec2e34741af18073ab94082f5fa4e08b5ae1443e1","last_reissued_at":"2026-07-05T09:32:33.422898Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:32:33.422898Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Erfei Cui, Hao Tian, Jifeng Dai, Jinguo Zhu, Junjun He, Lewei Lu, Shenglong Ye, Tong Lu, Weiyun Wang, Wenhai Wang, Xizhou Zhu, Yiming Ren, Yu Qiao, Zhangwei Gao, Zhe Chen","submitted_at":"2024-10-21T17:58:20Z","abstract_excerpt":"Multimodal large language models (MLLMs) have demonstrated impressive performance in vision-language tasks across a broad spectrum of domains. However, the large model scale and associated high computational costs pose significant challenges for training and deploying MLLMs on consumer-grade GPUs or edge devices, thereby hindering their widespread application. In this work, we introduce Mini-InternVL, a series of MLLMs with parameters ranging from 1B to 4B, which achieves 90% of the performance with only 5% of the parameters. This significant improvement in efficiency and effectiveness makes o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.16261","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/2410.16261/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":"2410.16261","created_at":"2026-07-05T09:32:33.422958+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.16261v3","created_at":"2026-07-05T09:32:33.422958+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.16261","created_at":"2026-07-05T09:32:33.422958+00:00"},{"alias_kind":"pith_short_12","alias_value":"4VDWM77NGHJT","created_at":"2026-07-05T09:32:33.422958+00:00"},{"alias_kind":"pith_short_16","alias_value":"4VDWM77NGHJTOYV5","created_at":"2026-07-05T09:32:33.422958+00:00"},{"alias_kind":"pith_short_8","alias_value":"4VDWM77N","created_at":"2026-07-05T09:32:33.422958+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23539","citing_title":"LightSTAR: Efficient Visual Document Retrieval via Lightweight Selection with Vision-Adaptive Refinement","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18242","citing_title":"EventDrive: Event Cameras for Vision-Language Driving Intelligence","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18750","citing_title":"A Readiness-Driven Runtime for Pipeline-Parallel Training under Runtime Variability","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04857","citing_title":"The Blind Spot of Adaptation: Quantifying and Mitigating Forgetting in Fine-tuned Driving Models","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2504.10479","citing_title":"InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05271","citing_title":"Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling","ref_index":71,"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":37,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG","json":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG.json","graph_json":"https://pith.science/api/pith-number/4VDWM77NGHJTOYV5KKHMFY2HIG/graph.json","events_json":"https://pith.science/api/pith-number/4VDWM77NGHJTOYV5KKHMFY2HIG/events.json","paper":"https://pith.science/paper/4VDWM77N"},"agent_actions":{"view_html":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG","download_json":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG.json","view_paper":"https://pith.science/paper/4VDWM77N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.16261&json=true","fetch_graph":"https://pith.science/api/pith-number/4VDWM77NGHJTOYV5KKHMFY2HIG/graph.json","fetch_events":"https://pith.science/api/pith-number/4VDWM77NGHJTOYV5KKHMFY2HIG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG/action/storage_attestation","attest_author":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG/action/author_attestation","sign_citation":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG/action/citation_signature","submit_replication":"https://pith.science/pith/4VDWM77NGHJTOYV5KKHMFY2HIG/action/replication_record"}},"created_at":"2026-07-05T09:32:33.422958+00:00","updated_at":"2026-07-05T09:32:33.422958+00:00"}