{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:S5GUI7DOUCFILUJNROTPDEBMLR","short_pith_number":"pith:S5GUI7DO","schema_version":"1.0","canonical_sha256":"974d447c6ea08a85d12d8ba6f1902c5c6fd559a6e25d0ab95774e900f4c9094e","source":{"kind":"arxiv","id":"2502.07838","version":2},"attestation_state":"computed","paper":{"title":"NanoVLMs: How small can we go and still make coherent Vision Language Models?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Himanshu kumar, Mukund Agarwalla, Rajat Dandekar, Raj Dandekar, Sreedath Panat","submitted_at":"2025-02-11T02:31:45Z","abstract_excerpt":"Vision-Language Models (VLMs), such as GPT-4V and Llama 3.2 vision, have garnered significant research attention for their ability to leverage Large Language Models (LLMs) in multimodal tasks. However, their potential is constrained by inherent challenges, including proprietary restrictions, substantial computational demands, and limited accessibility. Smaller models, such as GIT and BLIP, exhibit marked limitations, often failing to generate coherent and consistent text beyond a few tokens, even with extensive training. This underscores a pivotal inquiry: how small can a VLM be and still prod"},"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":"2502.07838","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-11T02:31:45Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"96bac150d241d9bbc6adc667fbe6035764c95df52e98ef8d3f79da20cd58a925","abstract_canon_sha256":"300787e0fa78dcdfb668dff7a34d24447e833f8a41c11dc3b71f3db60744d3e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:20.861360Z","signature_b64":"tJp0eykUe8MYpDv4mjPHEfzc0EjJrEhSLgRdaUD4qHukgwghdkUylBvK4dp04Q3Ov+4yiSreAzhHm+IBypYpAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"974d447c6ea08a85d12d8ba6f1902c5c6fd559a6e25d0ab95774e900f4c9094e","last_reissued_at":"2026-07-05T10:16:20.860890Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:20.860890Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NanoVLMs: How small can we go and still make coherent Vision Language Models?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Himanshu kumar, Mukund Agarwalla, Rajat Dandekar, Raj Dandekar, Sreedath Panat","submitted_at":"2025-02-11T02:31:45Z","abstract_excerpt":"Vision-Language Models (VLMs), such as GPT-4V and Llama 3.2 vision, have garnered significant research attention for their ability to leverage Large Language Models (LLMs) in multimodal tasks. However, their potential is constrained by inherent challenges, including proprietary restrictions, substantial computational demands, and limited accessibility. Smaller models, such as GIT and BLIP, exhibit marked limitations, often failing to generate coherent and consistent text beyond a few tokens, even with extensive training. This underscores a pivotal inquiry: how small can a VLM be and still prod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.07838","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/2502.07838/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":"2502.07838","created_at":"2026-07-05T10:16:20.860949+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.07838v2","created_at":"2026-07-05T10:16:20.860949+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.07838","created_at":"2026-07-05T10:16:20.860949+00:00"},{"alias_kind":"pith_short_12","alias_value":"S5GUI7DOUCFI","created_at":"2026-07-05T10:16:20.860949+00:00"},{"alias_kind":"pith_short_16","alias_value":"S5GUI7DOUCFILUJN","created_at":"2026-07-05T10:16:20.860949+00:00"},{"alias_kind":"pith_short_8","alias_value":"S5GUI7DO","created_at":"2026-07-05T10:16:20.860949+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.11530","citing_title":"Beyond Attention Scores: SVD-Based Vision Token Pruning for Efficient Vision-Language Models","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11530","citing_title":"Beyond Attention Scores: SVD-Based Vision Token Pruning for Efficient Vision-Language Models","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR","json":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR.json","graph_json":"https://pith.science/api/pith-number/S5GUI7DOUCFILUJNROTPDEBMLR/graph.json","events_json":"https://pith.science/api/pith-number/S5GUI7DOUCFILUJNROTPDEBMLR/events.json","paper":"https://pith.science/paper/S5GUI7DO"},"agent_actions":{"view_html":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR","download_json":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR.json","view_paper":"https://pith.science/paper/S5GUI7DO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.07838&json=true","fetch_graph":"https://pith.science/api/pith-number/S5GUI7DOUCFILUJNROTPDEBMLR/graph.json","fetch_events":"https://pith.science/api/pith-number/S5GUI7DOUCFILUJNROTPDEBMLR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR/action/storage_attestation","attest_author":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR/action/author_attestation","sign_citation":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR/action/citation_signature","submit_replication":"https://pith.science/pith/S5GUI7DOUCFILUJNROTPDEBMLR/action/replication_record"}},"created_at":"2026-07-05T10:16:20.860949+00:00","updated_at":"2026-07-05T10:16:20.860949+00:00"}