{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:KHABQ3PAAVQQ3H7YN2ITFV5B7P","short_pith_number":"pith:KHABQ3PA","schema_version":"1.0","canonical_sha256":"51c0186de005610d9ff86e9132d7a1fbd6a1010420bad3df7d025db75bab85dc","source":{"kind":"arxiv","id":"2212.14034","version":1},"attestation_state":"computed","paper":{"title":"Cramming: Training a Language Model on a Single GPU in One Day","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jonas Geiping, Tom Goldstein","submitted_at":"2022-12-28T18:59:28Z","abstract_excerpt":"Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners. While most in the community are asking how to push the limits of extreme computation, we ask the opposite question: How far can we get with a single GPU in just one day?\n  We investigate the downstream performance achievable with a transformer-based language model trained completely from scratch with masked language modeling for a single day on a single consumer GPU. Aside from re-an"},"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":"2212.14034","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-12-28T18:59:28Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"522f877dacb1d5aba8b60c61e11260840308655c748e8650f3e6d500e65bd7c9","abstract_canon_sha256":"63944e3af8509b61c229a3fe94967cb6b840d6b5a7a30a2cf3b5959fd401ddbb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:28:57.700577Z","signature_b64":"u1Mn9tA2O8MtdCcztrFM7ilkwtGc1UXD0c+bNgtAqFFrKEcQ3bon1L2g32546vbEqtYcJGfDVJ5dsojDOxnuAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"51c0186de005610d9ff86e9132d7a1fbd6a1010420bad3df7d025db75bab85dc","last_reissued_at":"2026-07-05T05:28:57.700119Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:28:57.700119Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Cramming: Training a Language Model on a Single GPU in One Day","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jonas Geiping, Tom Goldstein","submitted_at":"2022-12-28T18:59:28Z","abstract_excerpt":"Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners. While most in the community are asking how to push the limits of extreme computation, we ask the opposite question: How far can we get with a single GPU in just one day?\n  We investigate the downstream performance achievable with a transformer-based language model trained completely from scratch with masked language modeling for a single day on a single consumer GPU. Aside from re-an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.14034","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/2212.14034/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":"2212.14034","created_at":"2026-07-05T05:28:57.700170+00:00"},{"alias_kind":"arxiv_version","alias_value":"2212.14034v1","created_at":"2026-07-05T05:28:57.700170+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.14034","created_at":"2026-07-05T05:28:57.700170+00:00"},{"alias_kind":"pith_short_12","alias_value":"KHABQ3PAAVQQ","created_at":"2026-07-05T05:28:57.700170+00:00"},{"alias_kind":"pith_short_16","alias_value":"KHABQ3PAAVQQ3H7Y","created_at":"2026-07-05T05:28:57.700170+00:00"},{"alias_kind":"pith_short_8","alias_value":"KHABQ3PA","created_at":"2026-07-05T05:28:57.700170+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2311.04799","citing_title":"DA-Cramming: Enhancing Cost-Effective Language Model Pretraining with Dependency Agreement Integration","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P","json":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P.json","graph_json":"https://pith.science/api/pith-number/KHABQ3PAAVQQ3H7YN2ITFV5B7P/graph.json","events_json":"https://pith.science/api/pith-number/KHABQ3PAAVQQ3H7YN2ITFV5B7P/events.json","paper":"https://pith.science/paper/KHABQ3PA"},"agent_actions":{"view_html":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P","download_json":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P.json","view_paper":"https://pith.science/paper/KHABQ3PA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2212.14034&json=true","fetch_graph":"https://pith.science/api/pith-number/KHABQ3PAAVQQ3H7YN2ITFV5B7P/graph.json","fetch_events":"https://pith.science/api/pith-number/KHABQ3PAAVQQ3H7YN2ITFV5B7P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P/action/storage_attestation","attest_author":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P/action/author_attestation","sign_citation":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P/action/citation_signature","submit_replication":"https://pith.science/pith/KHABQ3PAAVQQ3H7YN2ITFV5B7P/action/replication_record"}},"created_at":"2026-07-05T05:28:57.700170+00:00","updated_at":"2026-07-05T05:28:57.700170+00:00"}