{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:UN5V344BBGX35T5AS7XGVER4R4","short_pith_number":"pith:UN5V344B","schema_version":"1.0","canonical_sha256":"a37b5df38109afbecfa097ee6a923c8f0f234d52ddc56fb88f55a1954a8aa3c8","source":{"kind":"arxiv","id":"2401.00134","version":1},"attestation_state":"computed","paper":{"title":"Unicron: Economizing Self-Healing LLM Training at Scale","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Jingbo Xu, Jingren Zhou, Kun Qian, Tao He, Wenyuan Yu, Xue Li, Zhibin Wang","submitted_at":"2023-12-30T04:06:16Z","abstract_excerpt":"Training large-scale language models is increasingly critical in various domains, but it is hindered by frequent failures, leading to significant time and economic costs. Current failure recovery methods in cloud-based settings inadequately address the diverse and complex scenarios that arise, focusing narrowly on erasing downtime for individual tasks without considering the overall cost impact on a cluster. We introduce Unicron, a workload manager designed for efficient self-healing in large-scale language model training. Unicron optimizes the training process by minimizing failure-related co"},"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":"2401.00134","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2023-12-30T04:06:16Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"5ba39ad3df9bb5ca11cb48f89aa76e85a9230dc32b60240f29f17c6a9cb28cb0","abstract_canon_sha256":"6d868266ccf118cac3094b63fb4471aaa6ba5041d4a243be1e52aeec5869eed8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:30:24.476908Z","signature_b64":"c89/npQIjjtKF8xaxWkGRhFrSizsEYT5JJV0ewRuzcDnieJku+XuJzfjZ7aYWKQJ8sBaTYMx9+hsxatrwCskDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a37b5df38109afbecfa097ee6a923c8f0f234d52ddc56fb88f55a1954a8aa3c8","last_reissued_at":"2026-07-05T07:30:24.476434Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:30:24.476434Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unicron: Economizing Self-Healing LLM Training at Scale","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Jingbo Xu, Jingren Zhou, Kun Qian, Tao He, Wenyuan Yu, Xue Li, Zhibin Wang","submitted_at":"2023-12-30T04:06:16Z","abstract_excerpt":"Training large-scale language models is increasingly critical in various domains, but it is hindered by frequent failures, leading to significant time and economic costs. Current failure recovery methods in cloud-based settings inadequately address the diverse and complex scenarios that arise, focusing narrowly on erasing downtime for individual tasks without considering the overall cost impact on a cluster. We introduce Unicron, a workload manager designed for efficient self-healing in large-scale language model training. Unicron optimizes the training process by minimizing failure-related co"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.00134","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/2401.00134/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":"2401.00134","created_at":"2026-07-05T07:30:24.476491+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.00134v1","created_at":"2026-07-05T07:30:24.476491+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.00134","created_at":"2026-07-05T07:30:24.476491+00:00"},{"alias_kind":"pith_short_12","alias_value":"UN5V344BBGX3","created_at":"2026-07-05T07:30:24.476491+00:00"},{"alias_kind":"pith_short_16","alias_value":"UN5V344BBGX35T5A","created_at":"2026-07-05T07:30:24.476491+00:00"},{"alias_kind":"pith_short_8","alias_value":"UN5V344B","created_at":"2026-07-05T07:30:24.476491+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.09562","citing_title":"Cloud abstractions for AI workloads","ref_index":23,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4","json":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4.json","graph_json":"https://pith.science/api/pith-number/UN5V344BBGX35T5AS7XGVER4R4/graph.json","events_json":"https://pith.science/api/pith-number/UN5V344BBGX35T5AS7XGVER4R4/events.json","paper":"https://pith.science/paper/UN5V344B"},"agent_actions":{"view_html":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4","download_json":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4.json","view_paper":"https://pith.science/paper/UN5V344B","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.00134&json=true","fetch_graph":"https://pith.science/api/pith-number/UN5V344BBGX35T5AS7XGVER4R4/graph.json","fetch_events":"https://pith.science/api/pith-number/UN5V344BBGX35T5AS7XGVER4R4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4/action/storage_attestation","attest_author":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4/action/author_attestation","sign_citation":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4/action/citation_signature","submit_replication":"https://pith.science/pith/UN5V344BBGX35T5AS7XGVER4R4/action/replication_record"}},"created_at":"2026-07-05T07:30:24.476491+00:00","updated_at":"2026-07-05T07:30:24.476491+00:00"}