{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:MEAED32TIIUNQLVBK5JNU75GBL","short_pith_number":"pith:MEAED32T","canonical_record":{"source":{"id":"2404.01847","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T11:12:42Z","cross_cats_sorted":[],"title_canon_sha256":"dfec37736c4f46436e8acf20250af28f45f32a19fa648919181fbd817502f5d1","abstract_canon_sha256":"4d634a0538956f31d72d56a0d1f47ce5fd0c90448b244bdb3f35ddfd56a19d7e"},"schema_version":"1.0"},"canonical_sha256":"610041ef534228d82ea15752da7fa60adf46064a44dfd02b335d77462b61fde0","source":{"kind":"arxiv","id":"2404.01847","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.01847","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"arxiv_version","alias_value":"2404.01847v3","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.01847","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"pith_short_12","alias_value":"MEAED32TIIUN","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"pith_short_16","alias_value":"MEAED32TIIUNQLVB","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"pith_short_8","alias_value":"MEAED32T","created_at":"2026-07-05T09:26:25Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:MEAED32TIIUNQLVBK5JNU75GBL","target":"record","payload":{"canonical_record":{"source":{"id":"2404.01847","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T11:12:42Z","cross_cats_sorted":[],"title_canon_sha256":"dfec37736c4f46436e8acf20250af28f45f32a19fa648919181fbd817502f5d1","abstract_canon_sha256":"4d634a0538956f31d72d56a0d1f47ce5fd0c90448b244bdb3f35ddfd56a19d7e"},"schema_version":"1.0"},"canonical_sha256":"610041ef534228d82ea15752da7fa60adf46064a44dfd02b335d77462b61fde0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:26:25.368414Z","signature_b64":"zb/pGRQOZ0ss2iEuPHBBPDne/O0LBWamzF/0L+ccIqbwq3XxQeRcXl/3QPm8qljN60FzvOfFrXLUdNuDzNw7CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"610041ef534228d82ea15752da7fa60adf46064a44dfd02b335d77462b61fde0","last_reissued_at":"2026-07-05T09:26:25.367931Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:26:25.367931Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.01847","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":"2026-07-05T09:26:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fHW8zvLFQvXBM0RA6WhzBRcaqcfpCVbyPRMU4JQmt+tD5pptU8P8mIwFTWYJSour5TwOFOSj+g3yi5XC8SHQBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T02:40:33.907645Z"},"content_sha256":"fdd1c793cf9184b25413462dc84d77ec3649bbf8efe336daf3aa92746cc3b899","schema_version":"1.0","event_id":"sha256:fdd1c793cf9184b25413462dc84d77ec3649bbf8efe336daf3aa92746cc3b899"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:MEAED32TIIUNQLVBK5JNU75GBL","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Accelerating Transformer Pre-training with 2:4 Sparsity","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jianfei Chen, Jun Zhu, Kang Zhao, Weiyu Huang, Yuezhou Hu","submitted_at":"2024-04-02T11:12:42Z","abstract_excerpt":"Training large transformers is slow, but recent innovations on GPU architecture give us an advantage. NVIDIA Ampere GPUs can execute a fine-grained 2:4 sparse matrix multiplication twice as fast as its dense equivalent. In the light of this property, we comprehensively investigate the feasibility of accelerating feed-forward networks (FFNs) of transformers in pre-training. First, we define a ``flip rate'' to monitor the stability of a 2:4 training process. Utilizing this metric, we propose three techniques to preserve accuracy: to modify the sparse-refined straight-through estimator by applyin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.01847","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/2404.01847/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":"2026-07-05T09:26:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3CqALyXZ/S6HyhFcaRnDkiQSn39T6DTR+QIhSWGDLV0CG6lAQSNKUZF3lbMuJrA/cvszZG/LqOlljcIVVBf5AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T02:40:33.908471Z"},"content_sha256":"45b006e54d0b8217746f6f022d0a059cc2b6d9773b6853aecdae6902cedaf5a7","schema_version":"1.0","event_id":"sha256:45b006e54d0b8217746f6f022d0a059cc2b6d9773b6853aecdae6902cedaf5a7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MEAED32TIIUNQLVBK5JNU75GBL/bundle.json","state_url":"https://pith.science/pith/MEAED32TIIUNQLVBK5JNU75GBL/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MEAED32TIIUNQLVBK5JNU75GBL/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":"2026-08-06T02:40:33Z","links":{"resolver":"https://pith.science/pith/MEAED32TIIUNQLVBK5JNU75GBL","bundle":"https://pith.science/pith/MEAED32TIIUNQLVBK5JNU75GBL/bundle.json","state":"https://pith.science/pith/MEAED32TIIUNQLVBK5JNU75GBL/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MEAED32TIIUNQLVBK5JNU75GBL/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:MEAED32TIIUNQLVBK5JNU75GBL","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":"4d634a0538956f31d72d56a0d1f47ce5fd0c90448b244bdb3f35ddfd56a19d7e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T11:12:42Z","title_canon_sha256":"dfec37736c4f46436e8acf20250af28f45f32a19fa648919181fbd817502f5d1"},"schema_version":"1.0","source":{"id":"2404.01847","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.01847","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"arxiv_version","alias_value":"2404.01847v3","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.01847","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"pith_short_12","alias_value":"MEAED32TIIUN","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"pith_short_16","alias_value":"MEAED32TIIUNQLVB","created_at":"2026-07-05T09:26:25Z"},{"alias_kind":"pith_short_8","alias_value":"MEAED32T","created_at":"2026-07-05T09:26:25Z"}],"graph_snapshots":[{"event_id":"sha256:45b006e54d0b8217746f6f022d0a059cc2b6d9773b6853aecdae6902cedaf5a7","target":"graph","created_at":"2026-07-05T09:26:25Z","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/2404.01847/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training large transformers is slow, but recent innovations on GPU architecture give us an advantage. NVIDIA Ampere GPUs can execute a fine-grained 2:4 sparse matrix multiplication twice as fast as its dense equivalent. In the light of this property, we comprehensively investigate the feasibility of accelerating feed-forward networks (FFNs) of transformers in pre-training. First, we define a ``flip rate'' to monitor the stability of a 2:4 training process. Utilizing this metric, we propose three techniques to preserve accuracy: to modify the sparse-refined straight-through estimator by applyin","authors_text":"Jianfei Chen, Jun Zhu, Kang Zhao, Weiyu Huang, Yuezhou Hu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T11:12:42Z","title":"Accelerating Transformer Pre-training with 2:4 Sparsity"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.01847","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:fdd1c793cf9184b25413462dc84d77ec3649bbf8efe336daf3aa92746cc3b899","target":"record","created_at":"2026-07-05T09:26:25Z","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":"4d634a0538956f31d72d56a0d1f47ce5fd0c90448b244bdb3f35ddfd56a19d7e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T11:12:42Z","title_canon_sha256":"dfec37736c4f46436e8acf20250af28f45f32a19fa648919181fbd817502f5d1"},"schema_version":"1.0","source":{"id":"2404.01847","kind":"arxiv","version":3}},"canonical_sha256":"610041ef534228d82ea15752da7fa60adf46064a44dfd02b335d77462b61fde0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"610041ef534228d82ea15752da7fa60adf46064a44dfd02b335d77462b61fde0","first_computed_at":"2026-07-05T09:26:25.367931Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:26:25.367931Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zb/pGRQOZ0ss2iEuPHBBPDne/O0LBWamzF/0L+ccIqbwq3XxQeRcXl/3QPm8qljN60FzvOfFrXLUdNuDzNw7CA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:26:25.368414Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.01847","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fdd1c793cf9184b25413462dc84d77ec3649bbf8efe336daf3aa92746cc3b899","sha256:45b006e54d0b8217746f6f022d0a059cc2b6d9773b6853aecdae6902cedaf5a7"],"state_sha256":"5d014d38944bc1fa06862577d1c895ce503fe4cf450e1893fbd15464b1420d1e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8kS8ZoTD7eM3qZo1LL2GrdSg0I/Cx0J68LCgmB/4oRnPTdNxbQySUeoitpuy7clHxOSSs7lOFiOH427/fivjDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T02:40:33.914426Z","bundle_sha256":"31d3b43bbbdbee460fe9e86735d13240042017a752e69c29e750312876237aa6"}}