{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RWFNAZ2MR4OOSUKLF36IO3YGDM","short_pith_number":"pith:RWFNAZ2M","schema_version":"1.0","canonical_sha256":"8d8ad0674c8f1ce9514b2efc876f061b16b60b8cbf03f50bb5d31ae670b5957b","source":{"kind":"arxiv","id":"2211.11152","version":2},"attestation_state":"computed","paper":{"title":"You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Caiwen Ding, Dongkuan Xu, Shengkun Tang, Tianchi Zhang, Yanzhi Wang, Yao Li, Yaqing Wang, Yi Liang, Zhenglun Kong","submitted_at":"2022-11-21T02:32:25Z","abstract_excerpt":"Large-scale Transformer models bring significant improvements for various downstream vision language tasks with a unified architecture. The performance improvements come with increasing model size, resulting in slow inference speed and increased cost for severing. While some certain predictions benefit from the full complexity of the large-scale model, not all of inputs need the same amount of computation to conduct, potentially leading to computation resource waste. To handle this challenge, early exiting is proposed to adaptively allocate computational power in term of input complexity to im"},"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":"2211.11152","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-11-21T02:32:25Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"76f285487d2f8e5ecaf78faf050a60a6ecd2ff4fa1751393f9bc14ed5e3d1419","abstract_canon_sha256":"9fac4a5fc672f9607c9bf790728f91ee81e3e6d961d5f976a9a34bcbd144a87a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:57:14.247341Z","signature_b64":"6zmAywetUev1ZRvKh0s4clscTLMdXumVPa4s0o84b4wMQFfKpz9xBM+vxEfwreGHY6W15LJjSkYrnmqIsjsRDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8d8ad0674c8f1ce9514b2efc876f061b16b60b8cbf03f50bb5d31ae670b5957b","last_reissued_at":"2026-07-05T05:57:14.246843Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:57:14.246843Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Caiwen Ding, Dongkuan Xu, Shengkun Tang, Tianchi Zhang, Yanzhi Wang, Yao Li, Yaqing Wang, Yi Liang, Zhenglun Kong","submitted_at":"2022-11-21T02:32:25Z","abstract_excerpt":"Large-scale Transformer models bring significant improvements for various downstream vision language tasks with a unified architecture. The performance improvements come with increasing model size, resulting in slow inference speed and increased cost for severing. While some certain predictions benefit from the full complexity of the large-scale model, not all of inputs need the same amount of computation to conduct, potentially leading to computation resource waste. To handle this challenge, early exiting is proposed to adaptively allocate computational power in term of input complexity to im"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.11152","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/2211.11152/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":"2211.11152","created_at":"2026-07-05T05:57:14.246910+00:00"},{"alias_kind":"arxiv_version","alias_value":"2211.11152v2","created_at":"2026-07-05T05:57:14.246910+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.11152","created_at":"2026-07-05T05:57:14.246910+00:00"},{"alias_kind":"pith_short_12","alias_value":"RWFNAZ2MR4OO","created_at":"2026-07-05T05:57:14.246910+00:00"},{"alias_kind":"pith_short_16","alias_value":"RWFNAZ2MR4OOSUKL","created_at":"2026-07-05T05:57:14.246910+00:00"},{"alias_kind":"pith_short_8","alias_value":"RWFNAZ2M","created_at":"2026-07-05T05:57:14.246910+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.05527","citing_title":"EdgeServing: Deadline-Aware Multi-DNN Serving at the Edge","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM","json":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM.json","graph_json":"https://pith.science/api/pith-number/RWFNAZ2MR4OOSUKLF36IO3YGDM/graph.json","events_json":"https://pith.science/api/pith-number/RWFNAZ2MR4OOSUKLF36IO3YGDM/events.json","paper":"https://pith.science/paper/RWFNAZ2M"},"agent_actions":{"view_html":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM","download_json":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM.json","view_paper":"https://pith.science/paper/RWFNAZ2M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2211.11152&json=true","fetch_graph":"https://pith.science/api/pith-number/RWFNAZ2MR4OOSUKLF36IO3YGDM/graph.json","fetch_events":"https://pith.science/api/pith-number/RWFNAZ2MR4OOSUKLF36IO3YGDM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM/action/storage_attestation","attest_author":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM/action/author_attestation","sign_citation":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM/action/citation_signature","submit_replication":"https://pith.science/pith/RWFNAZ2MR4OOSUKLF36IO3YGDM/action/replication_record"}},"created_at":"2026-07-05T05:57:14.246910+00:00","updated_at":"2026-07-05T05:57:14.246910+00:00"}