{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Z6CPUH35QXZ246BLFQLJL2D3AM","short_pith_number":"pith:Z6CPUH35","schema_version":"1.0","canonical_sha256":"cf84fa1f7d85f3ae782b2c1695e87b0318a4b2c5b75b56ee833b4064165e881f","source":{"kind":"arxiv","id":"2505.22194","version":2},"attestation_state":"computed","paper":{"title":"Refining Datapath for Microscaling ViTs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Aaron Zhao, Can Xiao, Jianyi Cheng","submitted_at":"2025-05-28T10:15:37Z","abstract_excerpt":"Vision Transformers (ViTs) leverage the transformer architecture to effectively capture global context, demonstrating strong performance in computer vision tasks. A major challenge in ViT hardware acceleration is that the model family contains complex arithmetic operations that are sensitive to model accuracy, such as the Softmax and LayerNorm operations, which cannot be mapped onto efficient hardware with low precision. Existing methods only exploit parallelism in the matrix multiplication operations of the model on hardware and keep these complex operations on the CPU. This results in subopt"},"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":"2505.22194","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2025-05-28T10:15:37Z","cross_cats_sorted":[],"title_canon_sha256":"7c50cca88ca9f48971213031ac4a39dcd40fc4d171dae3eb57ad3b863d3feab3","abstract_canon_sha256":"4a9782ba140c63681beac71d724c5068e0e3386b4cfc570ad07472be556a80b9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:47.338538Z","signature_b64":"Ca25p7sJIK9hHucMhHQisD4DcI4W3K3bFYCwjMuqsxoti0lokYjaajGnkSR3/iyIyBz9mjDPgQ6D6Vpl082gBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cf84fa1f7d85f3ae782b2c1695e87b0318a4b2c5b75b56ee833b4064165e881f","last_reissued_at":"2026-07-05T11:21:47.337990Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:47.337990Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Refining Datapath for Microscaling ViTs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Aaron Zhao, Can Xiao, Jianyi Cheng","submitted_at":"2025-05-28T10:15:37Z","abstract_excerpt":"Vision Transformers (ViTs) leverage the transformer architecture to effectively capture global context, demonstrating strong performance in computer vision tasks. A major challenge in ViT hardware acceleration is that the model family contains complex arithmetic operations that are sensitive to model accuracy, such as the Softmax and LayerNorm operations, which cannot be mapped onto efficient hardware with low precision. Existing methods only exploit parallelism in the matrix multiplication operations of the model on hardware and keep these complex operations on the CPU. This results in subopt"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.22194","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/2505.22194/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":"2505.22194","created_at":"2026-07-05T11:21:47.338052+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.22194v2","created_at":"2026-07-05T11:21:47.338052+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22194","created_at":"2026-07-05T11:21:47.338052+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z6CPUH35QXZ2","created_at":"2026-07-05T11:21:47.338052+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z6CPUH35QXZ246BL","created_at":"2026-07-05T11:21:47.338052+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z6CPUH35","created_at":"2026-07-05T11:21:47.338052+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/Z6CPUH35QXZ246BLFQLJL2D3AM","json":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM.json","graph_json":"https://pith.science/api/pith-number/Z6CPUH35QXZ246BLFQLJL2D3AM/graph.json","events_json":"https://pith.science/api/pith-number/Z6CPUH35QXZ246BLFQLJL2D3AM/events.json","paper":"https://pith.science/paper/Z6CPUH35"},"agent_actions":{"view_html":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM","download_json":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM.json","view_paper":"https://pith.science/paper/Z6CPUH35","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.22194&json=true","fetch_graph":"https://pith.science/api/pith-number/Z6CPUH35QXZ246BLFQLJL2D3AM/graph.json","fetch_events":"https://pith.science/api/pith-number/Z6CPUH35QXZ246BLFQLJL2D3AM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM/action/storage_attestation","attest_author":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM/action/author_attestation","sign_citation":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM/action/citation_signature","submit_replication":"https://pith.science/pith/Z6CPUH35QXZ246BLFQLJL2D3AM/action/replication_record"}},"created_at":"2026-07-05T11:21:47.338052+00:00","updated_at":"2026-07-05T11:21:47.338052+00:00"}