{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:DS6OM53CHG3PNB7PJ5OE3APKXQ","short_pith_number":"pith:DS6OM53C","canonical_record":{"source":{"id":"2405.18047","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-28T11:02:01Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"3f4e0ef7d26af89fac11b3d279a17a29510838834dbdebeb8b1981e92ad5a2d7","abstract_canon_sha256":"5affd8222391154adad8170ad579016463b4977076d57200df870dde1979571e"},"schema_version":"1.0"},"canonical_sha256":"1cbce6776239b6f687ef4f5c4d81eabc1d76ea9f13a6a2c82486a5f39726879f","source":{"kind":"arxiv","id":"2405.18047","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.18047","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"arxiv_version","alias_value":"2405.18047v1","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18047","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"pith_short_12","alias_value":"DS6OM53CHG3P","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"pith_short_16","alias_value":"DS6OM53CHG3PNB7P","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"pith_short_8","alias_value":"DS6OM53C","created_at":"2026-07-05T08:24:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:DS6OM53CHG3PNB7PJ5OE3APKXQ","target":"record","payload":{"canonical_record":{"source":{"id":"2405.18047","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-28T11:02:01Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"3f4e0ef7d26af89fac11b3d279a17a29510838834dbdebeb8b1981e92ad5a2d7","abstract_canon_sha256":"5affd8222391154adad8170ad579016463b4977076d57200df870dde1979571e"},"schema_version":"1.0"},"canonical_sha256":"1cbce6776239b6f687ef4f5c4d81eabc1d76ea9f13a6a2c82486a5f39726879f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:24:15.023378Z","signature_b64":"HIxfnvXcA1p6gcRlPYw53CYxDXJa6GJ+lKf18j9IMcyn2ojBmk5SENf1xStGJGHOroTQLbrm5OviqqcmoMlUBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1cbce6776239b6f687ef4f5c4d81eabc1d76ea9f13a6a2c82486a5f39726879f","last_reissued_at":"2026-07-05T08:24:15.022921Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:24:15.022921Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.18047","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:24:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cLSTo+9ROiXjMYaIgBr4sU8rObe/VlneLyaHYbBPbOxF9mLvPnnCu1qAjcXGvIXE4c3PXeD5Z2nWeiQ64HCFBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T15:28:32.922245Z"},"content_sha256":"ccf8261acf53a0d86cb4df4ab43769c832a3c7e160f22845ed3fc0674d5af9e1","schema_version":"1.0","event_id":"sha256:ccf8261acf53a0d86cb4df4ab43769c832a3c7e160f22845ed3fc0674d5af9e1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:DS6OM53CHG3PNB7PJ5OE3APKXQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"2BP: 2-Stage Backpropagation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Christopher Rae, James Richings, Joseph K. L. Lee","submitted_at":"2024-05-28T11:02:01Z","abstract_excerpt":"As Deep Neural Networks (DNNs) grow in size and complexity, they often exceed the memory capacity of a single accelerator, necessitating the sharding of model parameters across multiple accelerators. Pipeline parallelism is a commonly used sharding strategy for training large DNNs. However, current implementations of pipeline parallelism are being unintentionally bottlenecked by the automatic differentiation tools provided by ML frameworks. This paper introduces 2-stage backpropagation (2BP). By splitting the backward propagation step into two separate stages, we can reduce idle compute time. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18047","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/2405.18047/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":"2026-07-05T08:24:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LCXGsHXjqfNFXny5zgtS+eaRhALPMlkWrUKie3MeuL4VCDeYlBDtna50jEFI6bYsrdYojnUGfpk93nk5hW1kDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T15:28:32.923747Z"},"content_sha256":"10293675e989ea2b7c2f9a5f8be48e86e3cba6c5a5b89679f14c306e3d9e69ed","schema_version":"1.0","event_id":"sha256:10293675e989ea2b7c2f9a5f8be48e86e3cba6c5a5b89679f14c306e3d9e69ed"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ/bundle.json","state_url":"https://pith.science/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ/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":"2026-08-03T15:28:32Z","links":{"resolver":"https://pith.science/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ","bundle":"https://pith.science/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ/bundle.json","state":"https://pith.science/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DS6OM53CHG3PNB7PJ5OE3APKXQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:DS6OM53CHG3PNB7PJ5OE3APKXQ","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":"5affd8222391154adad8170ad579016463b4977076d57200df870dde1979571e","cross_cats_sorted":["cs.AI","cs.DC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-28T11:02:01Z","title_canon_sha256":"3f4e0ef7d26af89fac11b3d279a17a29510838834dbdebeb8b1981e92ad5a2d7"},"schema_version":"1.0","source":{"id":"2405.18047","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.18047","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"arxiv_version","alias_value":"2405.18047v1","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18047","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"pith_short_12","alias_value":"DS6OM53CHG3P","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"pith_short_16","alias_value":"DS6OM53CHG3PNB7P","created_at":"2026-07-05T08:24:15Z"},{"alias_kind":"pith_short_8","alias_value":"DS6OM53C","created_at":"2026-07-05T08:24:15Z"}],"graph_snapshots":[{"event_id":"sha256:10293675e989ea2b7c2f9a5f8be48e86e3cba6c5a5b89679f14c306e3d9e69ed","target":"graph","created_at":"2026-07-05T08:24:15Z","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/2405.18047/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As Deep Neural Networks (DNNs) grow in size and complexity, they often exceed the memory capacity of a single accelerator, necessitating the sharding of model parameters across multiple accelerators. Pipeline parallelism is a commonly used sharding strategy for training large DNNs. However, current implementations of pipeline parallelism are being unintentionally bottlenecked by the automatic differentiation tools provided by ML frameworks. This paper introduces 2-stage backpropagation (2BP). By splitting the backward propagation step into two separate stages, we can reduce idle compute time. ","authors_text":"Christopher Rae, James Richings, Joseph K. L. Lee","cross_cats":["cs.AI","cs.DC"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-28T11:02:01Z","title":"2BP: 2-Stage Backpropagation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18047","kind":"arxiv","version":1},"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:ccf8261acf53a0d86cb4df4ab43769c832a3c7e160f22845ed3fc0674d5af9e1","target":"record","created_at":"2026-07-05T08:24:15Z","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":"5affd8222391154adad8170ad579016463b4977076d57200df870dde1979571e","cross_cats_sorted":["cs.AI","cs.DC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-28T11:02:01Z","title_canon_sha256":"3f4e0ef7d26af89fac11b3d279a17a29510838834dbdebeb8b1981e92ad5a2d7"},"schema_version":"1.0","source":{"id":"2405.18047","kind":"arxiv","version":1}},"canonical_sha256":"1cbce6776239b6f687ef4f5c4d81eabc1d76ea9f13a6a2c82486a5f39726879f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1cbce6776239b6f687ef4f5c4d81eabc1d76ea9f13a6a2c82486a5f39726879f","first_computed_at":"2026-07-05T08:24:15.022921Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:24:15.022921Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HIxfnvXcA1p6gcRlPYw53CYxDXJa6GJ+lKf18j9IMcyn2ojBmk5SENf1xStGJGHOroTQLbrm5OviqqcmoMlUBg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:24:15.023378Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.18047","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ccf8261acf53a0d86cb4df4ab43769c832a3c7e160f22845ed3fc0674d5af9e1","sha256:10293675e989ea2b7c2f9a5f8be48e86e3cba6c5a5b89679f14c306e3d9e69ed"],"state_sha256":"f097137feb0138b1a896c233690901851ed5774cfb5d271d8805e855033673b3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"paf9725yVAyO4dJnOFB5R+x79kY3pCe3VYdYe8lfFLuFDCzbBTtNS4DwD4WlAISvtTgzAIKzyJmipQtuYytfDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T15:28:32.932628Z","bundle_sha256":"9ed5eb51bdf494426d0311490b19aadbb8536738ebbf9b8ae71a9ba18db0adda"}}