{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:NLG444PLBHBNVD67Y2ZOOK4U7I","short_pith_number":"pith:NLG444PL","schema_version":"1.0","canonical_sha256":"6acdce71eb09c2da8fdfc6b2e72b94fa3f5afd5959da45eee816ff64feaf43e9","source":{"kind":"arxiv","id":"2102.04503","version":1},"attestation_state":"computed","paper":{"title":"VS-Quant: Per-vector Scaled Quantization for Accurate Low-Precision Neural Network Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AR"],"primary_cat":"cs.LG","authors_text":"Brian Zimmer, Brucek Khailany, Haoxing Ren, Rangharajan Venkatesan, Steve Dai, William J. Dally","submitted_at":"2021-02-08T19:56:04Z","abstract_excerpt":"Quantization enables efficient acceleration of deep neural networks by reducing model memory footprint and exploiting low-cost integer math hardware units. Quantization maps floating-point weights and activations in a trained model to low-bitwidth integer values using scale factors. Excessive quantization, reducing precision too aggressively, results in accuracy degradation. When scale factors are shared at a coarse granularity across many dimensions of each tensor, effective precision of individual elements within the tensor are limited. To reduce quantization-related accuracy loss, we propos"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2102.04503","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-08T19:56:04Z","cross_cats_sorted":["cs.AR"],"title_canon_sha256":"2677b98a755acec972ead37e3bf8cf29356d488941672a3f5f45acb8df6b1db4","abstract_canon_sha256":"243e796fc8152984b30f1f87c75358508118daa1f0bc0b4b12cfd770d7152e47"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:14:14.449092Z","signature_b64":"W0r0oThig/+zJ9N924macgE0mDywdd/LLJCA4ChLyfbdU0/HoD3SQZfMc5HOt+bkHb3L4WnNnCDiz/J8epUpCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6acdce71eb09c2da8fdfc6b2e72b94fa3f5afd5959da45eee816ff64feaf43e9","last_reissued_at":"2026-07-05T02:14:14.448699Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:14:14.448699Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VS-Quant: Per-vector Scaled Quantization for Accurate Low-Precision Neural Network Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AR"],"primary_cat":"cs.LG","authors_text":"Brian Zimmer, Brucek Khailany, Haoxing Ren, Rangharajan Venkatesan, Steve Dai, William J. Dally","submitted_at":"2021-02-08T19:56:04Z","abstract_excerpt":"Quantization enables efficient acceleration of deep neural networks by reducing model memory footprint and exploiting low-cost integer math hardware units. Quantization maps floating-point weights and activations in a trained model to low-bitwidth integer values using scale factors. Excessive quantization, reducing precision too aggressively, results in accuracy degradation. When scale factors are shared at a coarse granularity across many dimensions of each tensor, effective precision of individual elements within the tensor are limited. To reduce quantization-related accuracy loss, we propos"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.04503","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/2102.04503/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2102.04503","created_at":"2026-07-05T02:14:14.448757+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.04503v1","created_at":"2026-07-05T02:14:14.448757+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.04503","created_at":"2026-07-05T02:14:14.448757+00:00"},{"alias_kind":"pith_short_12","alias_value":"NLG444PLBHBN","created_at":"2026-07-05T02:14:14.448757+00:00"},{"alias_kind":"pith_short_16","alias_value":"NLG444PLBHBNVD67","created_at":"2026-07-05T02:14:14.448757+00:00"},{"alias_kind":"pith_short_8","alias_value":"NLG444PL","created_at":"2026-07-05T02:14:14.448757+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2512.12930","citing_title":"SeVeDo: A Heterogeneous Transformer Accelerator for Low-Bit Inference via Hierarchical Group Quantization and SVD-Guided Mixed Precision","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I","json":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I.json","graph_json":"https://pith.science/api/pith-number/NLG444PLBHBNVD67Y2ZOOK4U7I/graph.json","events_json":"https://pith.science/api/pith-number/NLG444PLBHBNVD67Y2ZOOK4U7I/events.json","paper":"https://pith.science/paper/NLG444PL"},"agent_actions":{"view_html":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I","download_json":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I.json","view_paper":"https://pith.science/paper/NLG444PL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.04503&json=true","fetch_graph":"https://pith.science/api/pith-number/NLG444PLBHBNVD67Y2ZOOK4U7I/graph.json","fetch_events":"https://pith.science/api/pith-number/NLG444PLBHBNVD67Y2ZOOK4U7I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I/action/storage_attestation","attest_author":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I/action/author_attestation","sign_citation":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I/action/citation_signature","submit_replication":"https://pith.science/pith/NLG444PLBHBNVD67Y2ZOOK4U7I/action/replication_record"}},"created_at":"2026-07-05T02:14:14.448757+00:00","updated_at":"2026-07-05T02:14:14.448757+00:00"}