{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:GRDBRXJKNBHH5THYPSPQBSLN5G","short_pith_number":"pith:GRDBRXJK","schema_version":"1.0","canonical_sha256":"344618dd2a684e7eccf87c9f00c96de99039f25ab65c59fe53f9c49a34fa782c","source":{"kind":"arxiv","id":"2103.02800","version":1},"attestation_state":"computed","paper":{"title":"Hardware Acceleration of Fully Quantized BERT for Efficient Natural Language Processing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AR","authors_text":"Gang Li, Jian Cheng, Zejian Liu","submitted_at":"2021-03-04T02:49:16Z","abstract_excerpt":"BERT is the most recent Transformer-based model that achieves state-of-the-art performance in various NLP tasks. In this paper, we investigate the hardware acceleration of BERT on FPGA for edge computing. To tackle the issue of huge computational complexity and memory footprint, we propose to fully quantize the BERT (FQ-BERT), including weights, activations, softmax, layer normalization, and all the intermediate results. Experiments demonstrate that the FQ-BERT can achieve 7.94x compression for weights with negligible performance loss. We then propose an accelerator tailored for the FQ-BERT an"},"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":"2103.02800","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2021-03-04T02:49:16Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"9edd409f82d13af55cd56972b7454d2fbb73e851c645b61df5bbfc8156b7527e","abstract_canon_sha256":"adbc649b5d48b42bde955bdb04c6e1dc036f971e175208cd353cefaaf31faad9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:20:21.408813Z","signature_b64":"sZurnEnl0jkTBeEMyBTlIzCKcPRmZlG+3qOCda7RuiLJts8s/ALehQedIzO2qaTpe/9D8oxkIY2/Mc+VlbMtAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"344618dd2a684e7eccf87c9f00c96de99039f25ab65c59fe53f9c49a34fa782c","last_reissued_at":"2026-07-05T02:20:21.408306Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:20:21.408306Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hardware Acceleration of Fully Quantized BERT for Efficient Natural Language Processing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AR","authors_text":"Gang Li, Jian Cheng, Zejian Liu","submitted_at":"2021-03-04T02:49:16Z","abstract_excerpt":"BERT is the most recent Transformer-based model that achieves state-of-the-art performance in various NLP tasks. In this paper, we investigate the hardware acceleration of BERT on FPGA for edge computing. To tackle the issue of huge computational complexity and memory footprint, we propose to fully quantize the BERT (FQ-BERT), including weights, activations, softmax, layer normalization, and all the intermediate results. Experiments demonstrate that the FQ-BERT can achieve 7.94x compression for weights with negligible performance loss. We then propose an accelerator tailored for the FQ-BERT an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.02800","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/2103.02800/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":"2103.02800","created_at":"2026-07-05T02:20:21.408391+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.02800v1","created_at":"2026-07-05T02:20:21.408391+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.02800","created_at":"2026-07-05T02:20:21.408391+00:00"},{"alias_kind":"pith_short_12","alias_value":"GRDBRXJKNBHH","created_at":"2026-07-05T02:20:21.408391+00:00"},{"alias_kind":"pith_short_16","alias_value":"GRDBRXJKNBHH5THY","created_at":"2026-07-05T02:20:21.408391+00:00"},{"alias_kind":"pith_short_8","alias_value":"GRDBRXJK","created_at":"2026-07-05T02:20:21.408391+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G","json":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G.json","graph_json":"https://pith.science/api/pith-number/GRDBRXJKNBHH5THYPSPQBSLN5G/graph.json","events_json":"https://pith.science/api/pith-number/GRDBRXJKNBHH5THYPSPQBSLN5G/events.json","paper":"https://pith.science/paper/GRDBRXJK"},"agent_actions":{"view_html":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G","download_json":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G.json","view_paper":"https://pith.science/paper/GRDBRXJK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.02800&json=true","fetch_graph":"https://pith.science/api/pith-number/GRDBRXJKNBHH5THYPSPQBSLN5G/graph.json","fetch_events":"https://pith.science/api/pith-number/GRDBRXJKNBHH5THYPSPQBSLN5G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G/action/storage_attestation","attest_author":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G/action/author_attestation","sign_citation":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G/action/citation_signature","submit_replication":"https://pith.science/pith/GRDBRXJKNBHH5THYPSPQBSLN5G/action/replication_record"}},"created_at":"2026-07-05T02:20:21.408391+00:00","updated_at":"2026-07-05T02:20:21.408391+00:00"}