{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WUQBOZO7JDILFIKIWLS4DT65UV","short_pith_number":"pith:WUQBOZO7","schema_version":"1.0","canonical_sha256":"b5201765df48d0b2a148b2e5c1cfdda55b96a634feaec3d365e08332ce41346b","source":{"kind":"arxiv","id":"2401.05031","version":1},"attestation_state":"computed","paper":{"title":"OTAS: An Elastic Transformer Serving System via Token Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Deze Zeng, Haozhao Wang, Jie Zhang, Jinyu Chen, Song Guo, Wenchao Xu, Zicong Hong","submitted_at":"2024-01-10T09:43:29Z","abstract_excerpt":"Transformer model empowered architectures have become a pillar of cloud services that keeps reshaping our society. However, the dynamic query loads and heterogeneous user requirements severely challenge current transformer serving systems, which rely on pre-training multiple variants of a foundation model, i.e., with different sizes, to accommodate varying service demands. Unfortunately, such a mechanism is unsuitable for large transformer models due to the additional training costs and excessive I/O delay. In this paper, we introduce OTAS, the first elastic serving system specially tailored f"},"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":"2401.05031","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2024-01-10T09:43:29Z","cross_cats_sorted":[],"title_canon_sha256":"b7f5a84058ad5513268fa2a80a3240b9f50f3ae3fd068f76c202ebac459560d1","abstract_canon_sha256":"e867960c41acde0682179849bbc2661eba165fde3e4267d0ef0bad894a262485"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:32:02.104939Z","signature_b64":"WBVe7jzFMR70cjkoD7WxaybUmfF0O1X5HGgzQzr7YThwMRrZpvEcZqSmik4mzJ5XrnTzeMSqJ1k4q576A/ksAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5201765df48d0b2a148b2e5c1cfdda55b96a634feaec3d365e08332ce41346b","last_reissued_at":"2026-07-05T07:32:02.104482Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:32:02.104482Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OTAS: An Elastic Transformer Serving System via Token Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Deze Zeng, Haozhao Wang, Jie Zhang, Jinyu Chen, Song Guo, Wenchao Xu, Zicong Hong","submitted_at":"2024-01-10T09:43:29Z","abstract_excerpt":"Transformer model empowered architectures have become a pillar of cloud services that keeps reshaping our society. However, the dynamic query loads and heterogeneous user requirements severely challenge current transformer serving systems, which rely on pre-training multiple variants of a foundation model, i.e., with different sizes, to accommodate varying service demands. Unfortunately, such a mechanism is unsuitable for large transformer models due to the additional training costs and excessive I/O delay. In this paper, we introduce OTAS, the first elastic serving system specially tailored f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.05031","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/2401.05031/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":"2401.05031","created_at":"2026-07-05T07:32:02.104543+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.05031v1","created_at":"2026-07-05T07:32:02.104543+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.05031","created_at":"2026-07-05T07:32:02.104543+00:00"},{"alias_kind":"pith_short_12","alias_value":"WUQBOZO7JDIL","created_at":"2026-07-05T07:32:02.104543+00:00"},{"alias_kind":"pith_short_16","alias_value":"WUQBOZO7JDILFIKI","created_at":"2026-07-05T07:32:02.104543+00:00"},{"alias_kind":"pith_short_8","alias_value":"WUQBOZO7","created_at":"2026-07-05T07:32:02.104543+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.16260","citing_title":"ToFe: Lagged Token Freezing and Reusing for Efficient Vision Transformer Inference","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV","json":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV.json","graph_json":"https://pith.science/api/pith-number/WUQBOZO7JDILFIKIWLS4DT65UV/graph.json","events_json":"https://pith.science/api/pith-number/WUQBOZO7JDILFIKIWLS4DT65UV/events.json","paper":"https://pith.science/paper/WUQBOZO7"},"agent_actions":{"view_html":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV","download_json":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV.json","view_paper":"https://pith.science/paper/WUQBOZO7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.05031&json=true","fetch_graph":"https://pith.science/api/pith-number/WUQBOZO7JDILFIKIWLS4DT65UV/graph.json","fetch_events":"https://pith.science/api/pith-number/WUQBOZO7JDILFIKIWLS4DT65UV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV/action/storage_attestation","attest_author":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV/action/author_attestation","sign_citation":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV/action/citation_signature","submit_replication":"https://pith.science/pith/WUQBOZO7JDILFIKIWLS4DT65UV/action/replication_record"}},"created_at":"2026-07-05T07:32:02.104543+00:00","updated_at":"2026-07-05T07:32:02.104543+00:00"}