{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:I6DYAHCRPXIDWQD2YN7B6XO2OP","short_pith_number":"pith:I6DYAHCR","canonical_record":{"source":{"id":"2405.13698","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-22T14:43:02Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"60f5a6dba065aad38c2755b3ffa9a7b638871bd98da2ae4e67cbbb3fbbda9809","abstract_canon_sha256":"b198faf1cd92e511d937f9847a533da60d64c338a64f7d2ca6c8d42e4cc9eee9"},"schema_version":"1.0"},"canonical_sha256":"4787801c517dd03b407ac37e1f5dda73ff17dfd8f5338e24dfc28964e629a4e1","source":{"kind":"arxiv","id":"2405.13698","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.13698","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"arxiv_version","alias_value":"2405.13698v3","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.13698","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"pith_short_12","alias_value":"I6DYAHCRPXID","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"pith_short_16","alias_value":"I6DYAHCRPXIDWQD2","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"pith_short_8","alias_value":"I6DYAHCR","created_at":"2026-07-05T11:13:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:I6DYAHCRPXIDWQD2YN7B6XO2OP","target":"record","payload":{"canonical_record":{"source":{"id":"2405.13698","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-22T14:43:02Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"60f5a6dba065aad38c2755b3ffa9a7b638871bd98da2ae4e67cbbb3fbbda9809","abstract_canon_sha256":"b198faf1cd92e511d937f9847a533da60d64c338a64f7d2ca6c8d42e4cc9eee9"},"schema_version":"1.0"},"canonical_sha256":"4787801c517dd03b407ac37e1f5dda73ff17dfd8f5338e24dfc28964e629a4e1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:09.862334Z","signature_b64":"4vLZPjGd5aJr4Luiwusc3Gq1R49VkM8/8OIa/ZIU1rwnugVHByXZUKK0u+WopGkh5hxFKqXYL0gBPhwqVeF/DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4787801c517dd03b407ac37e1f5dda73ff17dfd8f5338e24dfc28964e629a4e1","last_reissued_at":"2026-07-05T11:13:09.861774Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:09.861774Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.13698","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:13:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"P7giTWmrK69NtpI0ArWdElFaDHa7z13TR+FQqLy2NpRpor2n54BS4Zt/sXTCB/yKHJIEORde1wTHtmJQmn81DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T19:44:06.610336Z"},"content_sha256":"551994cc63568edf710eff7f17347d0acb427b88312eb6d92597ebb8673a663d","schema_version":"1.0","event_id":"sha256:551994cc63568edf710eff7f17347d0acb427b88312eb6d92597ebb8673a663d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:I6DYAHCRPXIDWQD2YN7B6XO2OP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"How to set AdamW's weight decay as you scale model and dataset size","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Laurence Aitchison, Xi Wang","submitted_at":"2024-05-22T14:43:02Z","abstract_excerpt":"The scaling of the optimal AdamW weight decay hyperparameter with model and dataset size is critical as we seek to build larger models, but is poorly understood. We show that weights learned by AdamW can be understood as an exponential moving average (EMA) of recent updates. This gives critical insights for how to set the weight decay in AdamW, and how the weight decay should scale with model and dataset size. In particular, the key hyperparameter for an exponential moving average is the EMA timescale. Intuitively, the EMA timescale can be understood as the number of recent iterations the EMA "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.13698","kind":"arxiv","version":3},"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/2405.13698/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:13:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EbHm/q9Uyvgra6g0WY1ZeGihVabH0GiFFsk6AYQnoeZjeW29ErQkb4iW+pFdJecGPQ4LcS/gemIaDDSDIvdFBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T19:44:06.610845Z"},"content_sha256":"4bb086f04cc240e0a216244b5ffc2b14dcd3bf8413bc427211ca7ff2448d9cdb","schema_version":"1.0","event_id":"sha256:4bb086f04cc240e0a216244b5ffc2b14dcd3bf8413bc427211ca7ff2448d9cdb"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP/bundle.json","state_url":"https://pith.science/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-07T19:44:06Z","links":{"resolver":"https://pith.science/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP","bundle":"https://pith.science/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP/bundle.json","state":"https://pith.science/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/I6DYAHCRPXIDWQD2YN7B6XO2OP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:I6DYAHCRPXIDWQD2YN7B6XO2OP","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b198faf1cd92e511d937f9847a533da60d64c338a64f7d2ca6c8d42e4cc9eee9","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-22T14:43:02Z","title_canon_sha256":"60f5a6dba065aad38c2755b3ffa9a7b638871bd98da2ae4e67cbbb3fbbda9809"},"schema_version":"1.0","source":{"id":"2405.13698","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.13698","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"arxiv_version","alias_value":"2405.13698v3","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.13698","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"pith_short_12","alias_value":"I6DYAHCRPXID","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"pith_short_16","alias_value":"I6DYAHCRPXIDWQD2","created_at":"2026-07-05T11:13:09Z"},{"alias_kind":"pith_short_8","alias_value":"I6DYAHCR","created_at":"2026-07-05T11:13:09Z"}],"graph_snapshots":[{"event_id":"sha256:4bb086f04cc240e0a216244b5ffc2b14dcd3bf8413bc427211ca7ff2448d9cdb","target":"graph","created_at":"2026-07-05T11:13:09Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2405.13698/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The scaling of the optimal AdamW weight decay hyperparameter with model and dataset size is critical as we seek to build larger models, but is poorly understood. We show that weights learned by AdamW can be understood as an exponential moving average (EMA) of recent updates. This gives critical insights for how to set the weight decay in AdamW, and how the weight decay should scale with model and dataset size. In particular, the key hyperparameter for an exponential moving average is the EMA timescale. Intuitively, the EMA timescale can be understood as the number of recent iterations the EMA ","authors_text":"Laurence Aitchison, Xi Wang","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-22T14:43:02Z","title":"How to set AdamW's weight decay as you scale model and dataset size"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.13698","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:551994cc63568edf710eff7f17347d0acb427b88312eb6d92597ebb8673a663d","target":"record","created_at":"2026-07-05T11:13:09Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b198faf1cd92e511d937f9847a533da60d64c338a64f7d2ca6c8d42e4cc9eee9","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-22T14:43:02Z","title_canon_sha256":"60f5a6dba065aad38c2755b3ffa9a7b638871bd98da2ae4e67cbbb3fbbda9809"},"schema_version":"1.0","source":{"id":"2405.13698","kind":"arxiv","version":3}},"canonical_sha256":"4787801c517dd03b407ac37e1f5dda73ff17dfd8f5338e24dfc28964e629a4e1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4787801c517dd03b407ac37e1f5dda73ff17dfd8f5338e24dfc28964e629a4e1","first_computed_at":"2026-07-05T11:13:09.861774Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:13:09.861774Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"4vLZPjGd5aJr4Luiwusc3Gq1R49VkM8/8OIa/ZIU1rwnugVHByXZUKK0u+WopGkh5hxFKqXYL0gBPhwqVeF/DA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:13:09.862334Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.13698","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:551994cc63568edf710eff7f17347d0acb427b88312eb6d92597ebb8673a663d","sha256:4bb086f04cc240e0a216244b5ffc2b14dcd3bf8413bc427211ca7ff2448d9cdb"],"state_sha256":"88a71eabfd84f40e09cf8833e21c96a1b5f1ff51c6cbdae3b01bfbec5b269ca0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UyEkLVHzjlDGmd8ACgM31Li5/lpW7VZPUoQwpOPOCCO7/JFIGSpGhMKokgJy9wCFA6/6bDqsNS60XP9YV7GrCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T19:44:06.616217Z","bundle_sha256":"1d1feea538cad604da917a81211e0b7d1aa1e9c94a4061826866cc55e90f1c2b"}}