{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2Y35YDIXAYMZES2GIBCELXWZJG","short_pith_number":"pith:2Y35YDIX","schema_version":"1.0","canonical_sha256":"d637dc0d170619924b46404445ded949a63e9de0d2f2405a0b6f7fcced92d32d","source":{"kind":"arxiv","id":"2605.20295","version":1},"attestation_state":"computed","paper":{"title":"Quant.npu: Enabling Efficient Mobile NPU Inference for on-device LLMs via Fully Static Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chenghua Wang, Daliang Xu, Gang Huang, Jinghe Zhang, Mengwei Xu, Tao Qi, Weikai Xie, Yun Ma","submitted_at":"2026-05-19T10:48:35Z","abstract_excerpt":"Large language models (LLMs) are increasingly deployed on mobile devices, where Neural Processing Units (NPUs) necessitate fully static quantization for optimal inference efficiency. However, existing post-training quantization (PTQ) methods predominantly rely on dynamic activation quantization, rendering them incompatible with NPU hardware constraints. To bridge the gap between high-fidelity PTQ and NPU-constrained inference, we propose Quant.npu, a integer-only fully static quantization framework. It incorporates learnable quantization parameters and rotation matrices, enabling low-bit activ"},"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":"2605.20295","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T10:48:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1b523a981d156acb7b166ff90c25b02f0c0e655e7b3bddf54f6d061c05d30df9","abstract_canon_sha256":"c4c87543611bf809733577a59f79a47aeaee38ee22ec1757ec48ee6546ff0844"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T00:04:24.360161Z","signature_b64":"mHIiupu53H6Oo+Bh8dy5d8LnaDaKCHGloh6oXfQCmOjQPBC7rublL7eLLnM3ItOJ4ScwOlnnHvIcXz4uxlZ7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d637dc0d170619924b46404445ded949a63e9de0d2f2405a0b6f7fcced92d32d","last_reissued_at":"2026-05-21T00:04:24.359516Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T00:04:24.359516Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quant.npu: Enabling Efficient Mobile NPU Inference for on-device LLMs via Fully Static Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chenghua Wang, Daliang Xu, Gang Huang, Jinghe Zhang, Mengwei Xu, Tao Qi, Weikai Xie, Yun Ma","submitted_at":"2026-05-19T10:48:35Z","abstract_excerpt":"Large language models (LLMs) are increasingly deployed on mobile devices, where Neural Processing Units (NPUs) necessitate fully static quantization for optimal inference efficiency. However, existing post-training quantization (PTQ) methods predominantly rely on dynamic activation quantization, rendering them incompatible with NPU hardware constraints. To bridge the gap between high-fidelity PTQ and NPU-constrained inference, we propose Quant.npu, a integer-only fully static quantization framework. It incorporates learnable quantization parameters and rotation matrices, enabling low-bit activ"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20295","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/2605.20295/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":"2605.20295","created_at":"2026-05-21T00:04:24.359616+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20295v1","created_at":"2026-05-21T00:04:24.359616+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20295","created_at":"2026-05-21T00:04:24.359616+00:00"},{"alias_kind":"pith_short_12","alias_value":"2Y35YDIXAYMZ","created_at":"2026-05-21T00:04:24.359616+00:00"},{"alias_kind":"pith_short_16","alias_value":"2Y35YDIXAYMZES2G","created_at":"2026-05-21T00:04:24.359616+00:00"},{"alias_kind":"pith_short_8","alias_value":"2Y35YDIX","created_at":"2026-05-21T00:04:24.359616+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/2Y35YDIXAYMZES2GIBCELXWZJG","json":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG.json","graph_json":"https://pith.science/api/pith-number/2Y35YDIXAYMZES2GIBCELXWZJG/graph.json","events_json":"https://pith.science/api/pith-number/2Y35YDIXAYMZES2GIBCELXWZJG/events.json","paper":"https://pith.science/paper/2Y35YDIX"},"agent_actions":{"view_html":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG","download_json":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG.json","view_paper":"https://pith.science/paper/2Y35YDIX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20295&json=true","fetch_graph":"https://pith.science/api/pith-number/2Y35YDIXAYMZES2GIBCELXWZJG/graph.json","fetch_events":"https://pith.science/api/pith-number/2Y35YDIXAYMZES2GIBCELXWZJG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG/action/storage_attestation","attest_author":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG/action/author_attestation","sign_citation":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG/action/citation_signature","submit_replication":"https://pith.science/pith/2Y35YDIXAYMZES2GIBCELXWZJG/action/replication_record"}},"created_at":"2026-05-21T00:04:24.359616+00:00","updated_at":"2026-05-21T00:04:24.359616+00:00"}