{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UI6BJPZMOR73OM7JXTJAFV5SBQ","short_pith_number":"pith:UI6BJPZM","schema_version":"1.0","canonical_sha256":"a23c14bf2c747fb733e9bcd202d7b20c0c290823a8c707941fdad99165237db6","source":{"kind":"arxiv","id":"2504.18590","version":1},"attestation_state":"computed","paper":{"title":"A multilevel approach to accelerate the training of Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","math.OC"],"primary_cat":"cs.LG","authors_text":"Arthur Lebeurrier (OCKHAM), Edgar Desainte-Mar\\'eville (OCKHAM), \\'Etienne Lasalle (OCKHAM), Guillaume Lauga (OCKHAM), Ma\\\"el Chaumette (OCKHAM)","submitted_at":"2025-04-24T08:23:50Z","abstract_excerpt":"In this article, we investigate the potential of multilevel approaches to accelerate the training of transformer architectures. Using an ordinary differential equation (ODE) interpretation of these architectures, we propose an appropriate way of varying the discretization of these ODE Transformers in order to accelerate the training. We validate our approach experimentally by a comparison with the standard training procedure."},"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":"2504.18590","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T08:23:50Z","cross_cats_sorted":["cs.AI","math.OC"],"title_canon_sha256":"3f855a5bdd4050c4df82caffe916ec55a2076690e6e1c3bbac5613bb7ebcfedf","abstract_canon_sha256":"e23a15c730df9e89c5d799749359f6a1ab40bdb0e8b28740d956b84dd7415674"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:54:17.573863Z","signature_b64":"FK0pwMIvHRN3p6j4mXb45haXu2yxEdgOaLzaqjYCif9Vr2ePKhad2xpw3F4CQRS2gZilZz2eN5ZlBli1/UjIBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a23c14bf2c747fb733e9bcd202d7b20c0c290823a8c707941fdad99165237db6","last_reissued_at":"2026-07-05T10:54:17.573364Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:54:17.573364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A multilevel approach to accelerate the training of Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","math.OC"],"primary_cat":"cs.LG","authors_text":"Arthur Lebeurrier (OCKHAM), Edgar Desainte-Mar\\'eville (OCKHAM), \\'Etienne Lasalle (OCKHAM), Guillaume Lauga (OCKHAM), Ma\\\"el Chaumette (OCKHAM)","submitted_at":"2025-04-24T08:23:50Z","abstract_excerpt":"In this article, we investigate the potential of multilevel approaches to accelerate the training of transformer architectures. Using an ordinary differential equation (ODE) interpretation of these architectures, we propose an appropriate way of varying the discretization of these ODE Transformers in order to accelerate the training. We validate our approach experimentally by a comparison with the standard training procedure."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.18590","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/2504.18590/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":"2504.18590","created_at":"2026-07-05T10:54:17.573429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.18590v1","created_at":"2026-07-05T10:54:17.573429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.18590","created_at":"2026-07-05T10:54:17.573429+00:00"},{"alias_kind":"pith_short_12","alias_value":"UI6BJPZMOR73","created_at":"2026-07-05T10:54:17.573429+00:00"},{"alias_kind":"pith_short_16","alias_value":"UI6BJPZMOR73OM7J","created_at":"2026-07-05T10:54:17.573429+00:00"},{"alias_kind":"pith_short_8","alias_value":"UI6BJPZM","created_at":"2026-07-05T10:54:17.573429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ","json":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ.json","graph_json":"https://pith.science/api/pith-number/UI6BJPZMOR73OM7JXTJAFV5SBQ/graph.json","events_json":"https://pith.science/api/pith-number/UI6BJPZMOR73OM7JXTJAFV5SBQ/events.json","paper":"https://pith.science/paper/UI6BJPZM"},"agent_actions":{"view_html":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ","download_json":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ.json","view_paper":"https://pith.science/paper/UI6BJPZM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.18590&json=true","fetch_graph":"https://pith.science/api/pith-number/UI6BJPZMOR73OM7JXTJAFV5SBQ/graph.json","fetch_events":"https://pith.science/api/pith-number/UI6BJPZMOR73OM7JXTJAFV5SBQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ/action/storage_attestation","attest_author":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ/action/author_attestation","sign_citation":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ/action/citation_signature","submit_replication":"https://pith.science/pith/UI6BJPZMOR73OM7JXTJAFV5SBQ/action/replication_record"}},"created_at":"2026-07-05T10:54:17.573429+00:00","updated_at":"2026-07-05T10:54:17.573429+00:00"}