{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PQTFLYAWN7NMGMVNFV5OQBQXF7","short_pith_number":"pith:PQTFLYAW","schema_version":"1.0","canonical_sha256":"7c2655e0166fdac332ad2d7ae806172fc30477d660f6e80a290242aeaf971f73","source":{"kind":"arxiv","id":"2508.11624","version":1},"attestation_state":"computed","paper":{"title":"LoRAtorio: An intrinsic approach to LoRA Skill Composition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ignas Budvytis, Niki Foteinopoulou, Stephan Liwicki","submitted_at":"2025-08-15T17:52:56Z","abstract_excerpt":"Low-Rank Adaptation (LoRA) has become a widely adopted technique in text-to-image diffusion models, enabling the personalisation of visual concepts such as characters, styles, and objects. However, existing approaches struggle to effectively compose multiple LoRA adapters, particularly in open-ended settings where the number and nature of required skills are not known in advance. In this work, we present LoRAtorio, a novel train-free framework for multi-LoRA composition that leverages intrinsic model behaviour. Our method is motivated by two key observations: (1) LoRA adapters trained on narro"},"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":"2508.11624","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-15T17:52:56Z","cross_cats_sorted":[],"title_canon_sha256":"fb0cba01237c2468d9662d0b65e1978cf94567b7de2f87a6a1d7a7b2461161ea","abstract_canon_sha256":"351dcbc2d40a339beb95c6b170516862676b25740f37ac17081958629f168e39"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:54:36.413163Z","signature_b64":"/oPKLEdjOvlPK8jqdkHKPxpCuJCmD6mCHiZ6JdDH46qn5yixNcmhqbdN1BnU7/0F04E/d7eJ9DfXMNdj4FWBCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7c2655e0166fdac332ad2d7ae806172fc30477d660f6e80a290242aeaf971f73","last_reissued_at":"2026-07-05T11:54:36.412730Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:54:36.412730Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LoRAtorio: An intrinsic approach to LoRA Skill Composition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ignas Budvytis, Niki Foteinopoulou, Stephan Liwicki","submitted_at":"2025-08-15T17:52:56Z","abstract_excerpt":"Low-Rank Adaptation (LoRA) has become a widely adopted technique in text-to-image diffusion models, enabling the personalisation of visual concepts such as characters, styles, and objects. However, existing approaches struggle to effectively compose multiple LoRA adapters, particularly in open-ended settings where the number and nature of required skills are not known in advance. In this work, we present LoRAtorio, a novel train-free framework for multi-LoRA composition that leverages intrinsic model behaviour. Our method is motivated by two key observations: (1) LoRA adapters trained on narro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.11624","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/2508.11624/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":"2508.11624","created_at":"2026-07-05T11:54:36.412789+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.11624v1","created_at":"2026-07-05T11:54:36.412789+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.11624","created_at":"2026-07-05T11:54:36.412789+00:00"},{"alias_kind":"pith_short_12","alias_value":"PQTFLYAWN7NM","created_at":"2026-07-05T11:54:36.412789+00:00"},{"alias_kind":"pith_short_16","alias_value":"PQTFLYAWN7NMGMVN","created_at":"2026-07-05T11:54:36.412789+00:00"},{"alias_kind":"pith_short_8","alias_value":"PQTFLYAW","created_at":"2026-07-05T11:54:36.412789+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03792","citing_title":"Training-Free Multi-Concept LoRA Composition with Prompt-Aware Weighting","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7","json":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7.json","graph_json":"https://pith.science/api/pith-number/PQTFLYAWN7NMGMVNFV5OQBQXF7/graph.json","events_json":"https://pith.science/api/pith-number/PQTFLYAWN7NMGMVNFV5OQBQXF7/events.json","paper":"https://pith.science/paper/PQTFLYAW"},"agent_actions":{"view_html":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7","download_json":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7.json","view_paper":"https://pith.science/paper/PQTFLYAW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.11624&json=true","fetch_graph":"https://pith.science/api/pith-number/PQTFLYAWN7NMGMVNFV5OQBQXF7/graph.json","fetch_events":"https://pith.science/api/pith-number/PQTFLYAWN7NMGMVNFV5OQBQXF7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7/action/storage_attestation","attest_author":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7/action/author_attestation","sign_citation":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7/action/citation_signature","submit_replication":"https://pith.science/pith/PQTFLYAWN7NMGMVNFV5OQBQXF7/action/replication_record"}},"created_at":"2026-07-05T11:54:36.412789+00:00","updated_at":"2026-07-05T11:54:36.412789+00:00"}