{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WDSHQWAWZCTCQAW5VRIOF2FAKF","short_pith_number":"pith:WDSHQWAW","schema_version":"1.0","canonical_sha256":"b0e4785816c8a62802ddac50e2e8a0514521260707ad8c829093fe67e331a102","source":{"kind":"arxiv","id":"2412.06321","version":1},"attestation_state":"computed","paper":{"title":"A Flexible Template for Edge Generative AI with High-Accuracy Accelerated Softmax & GELU","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Andrea Belano, Angelo Garofalo, Davide Rossi, Francesco Conti, Luca Benini, Yvan Tortorella","submitted_at":"2024-12-09T09:15:35Z","abstract_excerpt":"Transformer-based generative Artificial Intelligence (GenAI) models achieve remarkable results in a wide range of fields, including natural language processing, computer vision, and audio processing. However, this comes at the cost of increased complexity and the need of sophisticated non-linearities such as softmax and GELU. Even if Transformers are computationally dominated by matrix multiplications (MatMul), these non-linearities can become a performance bottleneck, especially if dedicated hardware is used to accelerate MatMul operators. In this work, we introduce a GenAI BFloat16 Transform"},"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":"2412.06321","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-12-09T09:15:35Z","cross_cats_sorted":[],"title_canon_sha256":"5f33455ec6156944d8a32cc18b92d3a7f3dd1fa33e8fe97d21c47345f40e4fb6","abstract_canon_sha256":"f9a65e75d84797a70f9063b3c9c48b51886fd3e52f7281fc7b796f48e12854ce"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:46:19.110026Z","signature_b64":"s0bsr7QCSHVpYHtEPn64jOoUA1YjVmHii7n4EXDql6GFTKqZeaofn2OJ4bkuJBZ9CA89h7MGSBP79D109ZDXDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b0e4785816c8a62802ddac50e2e8a0514521260707ad8c829093fe67e331a102","last_reissued_at":"2026-07-05T09:46:19.109551Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:46:19.109551Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Flexible Template for Edge Generative AI with High-Accuracy Accelerated Softmax & GELU","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Andrea Belano, Angelo Garofalo, Davide Rossi, Francesco Conti, Luca Benini, Yvan Tortorella","submitted_at":"2024-12-09T09:15:35Z","abstract_excerpt":"Transformer-based generative Artificial Intelligence (GenAI) models achieve remarkable results in a wide range of fields, including natural language processing, computer vision, and audio processing. However, this comes at the cost of increased complexity and the need of sophisticated non-linearities such as softmax and GELU. Even if Transformers are computationally dominated by matrix multiplications (MatMul), these non-linearities can become a performance bottleneck, especially if dedicated hardware is used to accelerate MatMul operators. In this work, we introduce a GenAI BFloat16 Transform"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.06321","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/2412.06321/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":"2412.06321","created_at":"2026-07-05T09:46:19.109606+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.06321v1","created_at":"2026-07-05T09:46:19.109606+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.06321","created_at":"2026-07-05T09:46:19.109606+00:00"},{"alias_kind":"pith_short_12","alias_value":"WDSHQWAWZCTC","created_at":"2026-07-05T09:46:19.109606+00:00"},{"alias_kind":"pith_short_16","alias_value":"WDSHQWAWZCTCQAW5","created_at":"2026-07-05T09:46:19.109606+00:00"},{"alias_kind":"pith_short_8","alias_value":"WDSHQWAW","created_at":"2026-07-05T09:46:19.109606+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.07935","citing_title":"The Hyperscale Lottery: How State-Space Models Have Sacrificed Edge Efficiency","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF","json":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF.json","graph_json":"https://pith.science/api/pith-number/WDSHQWAWZCTCQAW5VRIOF2FAKF/graph.json","events_json":"https://pith.science/api/pith-number/WDSHQWAWZCTCQAW5VRIOF2FAKF/events.json","paper":"https://pith.science/paper/WDSHQWAW"},"agent_actions":{"view_html":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF","download_json":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF.json","view_paper":"https://pith.science/paper/WDSHQWAW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.06321&json=true","fetch_graph":"https://pith.science/api/pith-number/WDSHQWAWZCTCQAW5VRIOF2FAKF/graph.json","fetch_events":"https://pith.science/api/pith-number/WDSHQWAWZCTCQAW5VRIOF2FAKF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF/action/storage_attestation","attest_author":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF/action/author_attestation","sign_citation":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF/action/citation_signature","submit_replication":"https://pith.science/pith/WDSHQWAWZCTCQAW5VRIOF2FAKF/action/replication_record"}},"created_at":"2026-07-05T09:46:19.109606+00:00","updated_at":"2026-07-05T09:46:19.109606+00:00"}