{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2019:SZPRQ3BLDY3FMJSBPEEB7O4XKO","short_pith_number":"pith:SZPRQ3BL","canonical_record":{"source":{"id":"1905.12334","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-29T11:25:09Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"acc3c9cb967e315105362111b7f25995c082f938f230db5f8cbd26fb3c50976c","abstract_canon_sha256":"bee589f83336b808f2495f122955ec6d18f788f31ee50ac4082f859ef257631a"},"schema_version":"1.0"},"canonical_sha256":"965f186c2b1e3656264179081fbb9753927fd129f59195fb44fdca9bfa016ca6","source":{"kind":"arxiv","id":"1905.12334","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1905.12334","created_at":"2026-05-17T23:44:44Z"},{"alias_kind":"arxiv_version","alias_value":"1905.12334v1","created_at":"2026-05-17T23:44:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.12334","created_at":"2026-05-17T23:44:44Z"},{"alias_kind":"pith_short_12","alias_value":"SZPRQ3BLDY3F","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_16","alias_value":"SZPRQ3BLDY3FMJSB","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_8","alias_value":"SZPRQ3BL","created_at":"2026-05-18T12:33:27Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2019:SZPRQ3BLDY3FMJSBPEEB7O4XKO","target":"record","payload":{"canonical_record":{"source":{"id":"1905.12334","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-29T11:25:09Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"acc3c9cb967e315105362111b7f25995c082f938f230db5f8cbd26fb3c50976c","abstract_canon_sha256":"bee589f83336b808f2495f122955ec6d18f788f31ee50ac4082f859ef257631a"},"schema_version":"1.0"},"canonical_sha256":"965f186c2b1e3656264179081fbb9753927fd129f59195fb44fdca9bfa016ca6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:44:44.614298Z","signature_b64":"/+Kd6Hl+NhfaRhDtiWQjA8shVa56/l6wWtB5fEgHZEVyjvmWZbgTNYFIn8nNyq28lLENBLrr9RmxnVUyJes1BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"965f186c2b1e3656264179081fbb9753927fd129f59195fb44fdca9bfa016ca6","last_reissued_at":"2026-05-17T23:44:44.613818Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:44:44.613818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1905.12334","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":"2026-05-17T23:44:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"p3xhCZBLz27/esQq0kwdPI6XXvp/O5WI83tUn6wJmDdyb0umQK5lbndUGn6o4+AL1sD/2BwWRZaKCcAo7dUVCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-22T09:12:44.405465Z"},"content_sha256":"5232e37288ebebe730880d8ba793381a11ec33752642a2075f804c31857b8d4b","schema_version":"1.0","event_id":"sha256:5232e37288ebebe730880d8ba793381a11ec33752642a2075f804c31857b8d4b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2019:SZPRQ3BLDY3FMJSBPEEB7O4XKO","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Mixed Precision Training With 8-bit Floating Point","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Bharat Kaul, Dipankar Das, Naveen Mellempudi, Sudarshan Srinivasan","submitted_at":"2019-05-29T11:25:09Z","abstract_excerpt":"Reduced precision computation for deep neural networks is one of the key areas addressing the widening compute gap driven by an exponential growth in model size. In recent years, deep learning training has largely migrated to 16-bit precision, with significant gains in performance and energy efficiency. However, attempts to train DNNs at 8-bit precision have met with significant challenges because of the higher precision and dynamic range requirements of back-propagation. In this paper, we propose a method to train deep neural networks using 8-bit floating point representation for weights, act"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.12334","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":""},"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":"2026-05-17T23:44:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rMImHXtf9CgTHw6o9bha8jrq4XlfC3rtDpnDHO0owexCPxAU3VcICJiDbM69+p6ccgLFEzwNa6QCIskvdjvpAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-22T09:12:44.406160Z"},"content_sha256":"0a87d22387aed7016c9bf8fe04eb8fe9dd4770e89bef0ec8b020ed723bcaafc6","schema_version":"1.0","event_id":"sha256:0a87d22387aed7016c9bf8fe04eb8fe9dd4770e89bef0ec8b020ed723bcaafc6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO/bundle.json","state_url":"https://pith.science/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO/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":"2026-05-22T09:12:44Z","links":{"resolver":"https://pith.science/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO","bundle":"https://pith.science/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO/bundle.json","state":"https://pith.science/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SZPRQ3BLDY3FMJSBPEEB7O4XKO/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2019:SZPRQ3BLDY3FMJSBPEEB7O4XKO","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":"bee589f83336b808f2495f122955ec6d18f788f31ee50ac4082f859ef257631a","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-29T11:25:09Z","title_canon_sha256":"acc3c9cb967e315105362111b7f25995c082f938f230db5f8cbd26fb3c50976c"},"schema_version":"1.0","source":{"id":"1905.12334","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1905.12334","created_at":"2026-05-17T23:44:44Z"},{"alias_kind":"arxiv_version","alias_value":"1905.12334v1","created_at":"2026-05-17T23:44:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.12334","created_at":"2026-05-17T23:44:44Z"},{"alias_kind":"pith_short_12","alias_value":"SZPRQ3BLDY3F","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_16","alias_value":"SZPRQ3BLDY3FMJSB","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_8","alias_value":"SZPRQ3BL","created_at":"2026-05-18T12:33:27Z"}],"graph_snapshots":[{"event_id":"sha256:0a87d22387aed7016c9bf8fe04eb8fe9dd4770e89bef0ec8b020ed723bcaafc6","target":"graph","created_at":"2026-05-17T23:44:44Z","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"},"paper":{"abstract_excerpt":"Reduced precision computation for deep neural networks is one of the key areas addressing the widening compute gap driven by an exponential growth in model size. In recent years, deep learning training has largely migrated to 16-bit precision, with significant gains in performance and energy efficiency. However, attempts to train DNNs at 8-bit precision have met with significant challenges because of the higher precision and dynamic range requirements of back-propagation. In this paper, we propose a method to train deep neural networks using 8-bit floating point representation for weights, act","authors_text":"Bharat Kaul, Dipankar Das, Naveen Mellempudi, Sudarshan Srinivasan","cross_cats":["stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-29T11:25:09Z","title":"Mixed Precision Training With 8-bit Floating Point"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.12334","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":"sha256:5232e37288ebebe730880d8ba793381a11ec33752642a2075f804c31857b8d4b","target":"record","created_at":"2026-05-17T23:44:44Z","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":"bee589f83336b808f2495f122955ec6d18f788f31ee50ac4082f859ef257631a","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-29T11:25:09Z","title_canon_sha256":"acc3c9cb967e315105362111b7f25995c082f938f230db5f8cbd26fb3c50976c"},"schema_version":"1.0","source":{"id":"1905.12334","kind":"arxiv","version":1}},"canonical_sha256":"965f186c2b1e3656264179081fbb9753927fd129f59195fb44fdca9bfa016ca6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"965f186c2b1e3656264179081fbb9753927fd129f59195fb44fdca9bfa016ca6","first_computed_at":"2026-05-17T23:44:44.613818Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:44:44.613818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/+Kd6Hl+NhfaRhDtiWQjA8shVa56/l6wWtB5fEgHZEVyjvmWZbgTNYFIn8nNyq28lLENBLrr9RmxnVUyJes1BQ==","signature_status":"signed_v1","signed_at":"2026-05-17T23:44:44.614298Z","signed_message":"canonical_sha256_bytes"},"source_id":"1905.12334","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5232e37288ebebe730880d8ba793381a11ec33752642a2075f804c31857b8d4b","sha256:0a87d22387aed7016c9bf8fe04eb8fe9dd4770e89bef0ec8b020ed723bcaafc6"],"state_sha256":"add60e677b3535dad28aac857d0f2c4f091628db07751a81b0c19328ebbba5fd"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h48RHOZDlFi/QyLLxdYpz2DFNJYLoBgykfQ+1FzPQbu23fPSzgYYSkzkTAaj1P/M4Jwjo3Hni6Hpy5tJhvY0Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-22T09:12:44.409403Z","bundle_sha256":"d13eacad4d7da35132a7b2e9cf931ae73d2528178552217b080a5c01a2542df9"}}