{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:C37Z7BMUVQS7EDN26RTWV5SG3Q","short_pith_number":"pith:C37Z7BMU","canonical_record":{"source":{"id":"2004.10568","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-04-22T13:44:28Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"944032695a7cefad18e745cca34d4a4dfbf620a24906393394f6ad3df094f2c4","abstract_canon_sha256":"5236fab203f270bacddb57e2a0725383c4c0f214fbb38c24a108896173f0325c"},"schema_version":"1.0"},"canonical_sha256":"16ff9f8594ac25f20dbaf4676af646dc245692c6006b18c777eaab8d03b7b45f","source":{"kind":"arxiv","id":"2004.10568","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2004.10568","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"arxiv_version","alias_value":"2004.10568v2","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2004.10568","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"pith_short_12","alias_value":"C37Z7BMUVQS7","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"pith_short_16","alias_value":"C37Z7BMUVQS7EDN2","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"pith_short_8","alias_value":"C37Z7BMU","created_at":"2026-07-05T01:14:48Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:C37Z7BMUVQS7EDN26RTWV5SG3Q","target":"record","payload":{"canonical_record":{"source":{"id":"2004.10568","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-04-22T13:44:28Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"944032695a7cefad18e745cca34d4a4dfbf620a24906393394f6ad3df094f2c4","abstract_canon_sha256":"5236fab203f270bacddb57e2a0725383c4c0f214fbb38c24a108896173f0325c"},"schema_version":"1.0"},"canonical_sha256":"16ff9f8594ac25f20dbaf4676af646dc245692c6006b18c777eaab8d03b7b45f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:14:48.611288Z","signature_b64":"fli0smbvXzY8w8uMLEyH23YSS5ATH65RlQv88olhhBpy1puyDJtPQAY3R2xAoxayqLFLQgdPimw2j0ydMdxqAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"16ff9f8594ac25f20dbaf4676af646dc245692c6006b18c777eaab8d03b7b45f","last_reissued_at":"2026-07-05T01:14:48.610871Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:14:48.610871Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2004.10568","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T01:14:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qIR2dIaoqBgNpQkk/kVjAXSP5TAPlHDgIuVf8xPFzTHRQgqQhDBoCP0tywwkByzFtJ5XEZ2UUvs+aummwxjcCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T20:49:46.413679Z"},"content_sha256":"331c4be82e32570b31b0a5d7e2a1a71385ca45410956599b2400b071d4b792e8","schema_version":"1.0","event_id":"sha256:331c4be82e32570b31b0a5d7e2a1a71385ca45410956599b2400b071d4b792e8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:C37Z7BMUVQS7EDN26RTWV5SG3Q","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Up or Down? Adaptive Rounding for Post-Training Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Christos Louizos, Markus Nagel, Mart van Baalen, Rana Ali Amjad, Tijmen Blankevoort","submitted_at":"2020-04-22T13:44:28Z","abstract_excerpt":"When quantizing neural networks, assigning each floating-point weight to its nearest fixed-point value is the predominant approach. We find that, perhaps surprisingly, this is not the best we can do. In this paper, we propose AdaRound, a better weight-rounding mechanism for post-training quantization that adapts to the data and the task loss. AdaRound is fast, does not require fine-tuning of the network, and only uses a small amount of unlabelled data. We start by theoretically analyzing the rounding problem for a pre-trained neural network. By approximating the task loss with a Taylor series "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2004.10568","kind":"arxiv","version":2},"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/2004.10568/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":"2026-07-05T01:14:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uZGCe4ThvvZqqRmEhre6d/G/GZJ55xfIVWnFyZEvupYCDoNefQOtO23PhejQsN27llxdN1/emvoGxbHEDUClBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T20:49:46.414174Z"},"content_sha256":"e1be981f7e97fafa3332a2bd48e05916e3eebdd3c498b76abaac2998ca1af10b","schema_version":"1.0","event_id":"sha256:e1be981f7e97fafa3332a2bd48e05916e3eebdd3c498b76abaac2998ca1af10b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q/bundle.json","state_url":"https://pith.science/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q/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":"2026-08-22T20:49:46Z","links":{"resolver":"https://pith.science/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q","bundle":"https://pith.science/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q/bundle.json","state":"https://pith.science/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q/state.json","well_known_bundle":"https://pith.science/.well-known/pith/C37Z7BMUVQS7EDN26RTWV5SG3Q/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:C37Z7BMUVQS7EDN26RTWV5SG3Q","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":"5236fab203f270bacddb57e2a0725383c4c0f214fbb38c24a108896173f0325c","cross_cats_sorted":["cs.CV","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-04-22T13:44:28Z","title_canon_sha256":"944032695a7cefad18e745cca34d4a4dfbf620a24906393394f6ad3df094f2c4"},"schema_version":"1.0","source":{"id":"2004.10568","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2004.10568","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"arxiv_version","alias_value":"2004.10568v2","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2004.10568","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"pith_short_12","alias_value":"C37Z7BMUVQS7","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"pith_short_16","alias_value":"C37Z7BMUVQS7EDN2","created_at":"2026-07-05T01:14:48Z"},{"alias_kind":"pith_short_8","alias_value":"C37Z7BMU","created_at":"2026-07-05T01:14:48Z"}],"graph_snapshots":[{"event_id":"sha256:e1be981f7e97fafa3332a2bd48e05916e3eebdd3c498b76abaac2998ca1af10b","target":"graph","created_at":"2026-07-05T01:14:48Z","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/2004.10568/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"When quantizing neural networks, assigning each floating-point weight to its nearest fixed-point value is the predominant approach. We find that, perhaps surprisingly, this is not the best we can do. In this paper, we propose AdaRound, a better weight-rounding mechanism for post-training quantization that adapts to the data and the task loss. AdaRound is fast, does not require fine-tuning of the network, and only uses a small amount of unlabelled data. We start by theoretically analyzing the rounding problem for a pre-trained neural network. By approximating the task loss with a Taylor series ","authors_text":"Christos Louizos, Markus Nagel, Mart van Baalen, Rana Ali Amjad, Tijmen Blankevoort","cross_cats":["cs.CV","stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-04-22T13:44:28Z","title":"Up or Down? Adaptive Rounding for Post-Training Quantization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2004.10568","kind":"arxiv","version":2},"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":"sha256:331c4be82e32570b31b0a5d7e2a1a71385ca45410956599b2400b071d4b792e8","target":"record","created_at":"2026-07-05T01:14:48Z","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":"5236fab203f270bacddb57e2a0725383c4c0f214fbb38c24a108896173f0325c","cross_cats_sorted":["cs.CV","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-04-22T13:44:28Z","title_canon_sha256":"944032695a7cefad18e745cca34d4a4dfbf620a24906393394f6ad3df094f2c4"},"schema_version":"1.0","source":{"id":"2004.10568","kind":"arxiv","version":2}},"canonical_sha256":"16ff9f8594ac25f20dbaf4676af646dc245692c6006b18c777eaab8d03b7b45f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"16ff9f8594ac25f20dbaf4676af646dc245692c6006b18c777eaab8d03b7b45f","first_computed_at":"2026-07-05T01:14:48.610871Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T01:14:48.610871Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"fli0smbvXzY8w8uMLEyH23YSS5ATH65RlQv88olhhBpy1puyDJtPQAY3R2xAoxayqLFLQgdPimw2j0ydMdxqAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T01:14:48.611288Z","signed_message":"canonical_sha256_bytes"},"source_id":"2004.10568","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:331c4be82e32570b31b0a5d7e2a1a71385ca45410956599b2400b071d4b792e8","sha256:e1be981f7e97fafa3332a2bd48e05916e3eebdd3c498b76abaac2998ca1af10b"],"state_sha256":"ffa34647225e32b8a55e0e94974eedee828724a0bb39f2eaf37165339a5be670"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Hrg7gPlc21sY7boUSH9znKx5fLuZh1TDA9S8BUSTf8ZvyhSb1kri2BtYbUwYQUp/E3APDGubmP4VGmOPz5SjAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-22T20:49:46.417889Z","bundle_sha256":"1f64ff51a358276d65b7616183ef1c7ad83e715930ba3b8ef2ce532c1ed1b9bf"}}