{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:26CC3WTXWPXR674G4WVNS2ELBH","short_pith_number":"pith:26CC3WTX","schema_version":"1.0","canonical_sha256":"d7842dda77b3ef1f7f86e5aad9688b09cc86487f26f4d74145c99a871396d99a","source":{"kind":"arxiv","id":"2010.07003","version":2},"attestation_state":"computed","paper":{"title":"Length-Adaptive Transformer: Train Once with Length Drop, Use Anytime with Search","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Gyuwan Kim, Kyunghyun Cho","submitted_at":"2020-10-14T12:28:08Z","abstract_excerpt":"Despite transformers' impressive accuracy, their computational cost is often prohibitive to use with limited computational resources. Most previous approaches to improve inference efficiency require a separate model for each possible computational budget. In this paper, we extend PoWER-BERT (Goyal et al., 2020) and propose Length-Adaptive Transformer that can be used for various inference scenarios after one-shot training. We train a transformer with LengthDrop, a structural variant of dropout, which stochastically determines a sequence length at each layer. We then conduct a multi-objective e"},"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":"2010.07003","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-14T12:28:08Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"47396999792f93b4642c62047567b767ee5a7e93931533194ffb78f8af21617e","abstract_canon_sha256":"9759c2bfa1a71080b1c3e0b07f0e7f8c838067779af297c1180856e8a901f9b6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:48:34.342520Z","signature_b64":"yhlxJkBbSlhW9bcTA6w4MNsyEcQzLptw1q23Igis8cWZQzIfFnf0kiwFek2mJrJo1sgbtOGWDoq1a3nvXo3LDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7842dda77b3ef1f7f86e5aad9688b09cc86487f26f4d74145c99a871396d99a","last_reissued_at":"2026-07-05T02:48:34.342001Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:48:34.342001Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Length-Adaptive Transformer: Train Once with Length Drop, Use Anytime with Search","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Gyuwan Kim, Kyunghyun Cho","submitted_at":"2020-10-14T12:28:08Z","abstract_excerpt":"Despite transformers' impressive accuracy, their computational cost is often prohibitive to use with limited computational resources. Most previous approaches to improve inference efficiency require a separate model for each possible computational budget. In this paper, we extend PoWER-BERT (Goyal et al., 2020) and propose Length-Adaptive Transformer that can be used for various inference scenarios after one-shot training. We train a transformer with LengthDrop, a structural variant of dropout, which stochastically determines a sequence length at each layer. We then conduct a multi-objective e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.07003","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/2010.07003/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":"2010.07003","created_at":"2026-07-05T02:48:34.342060+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.07003v2","created_at":"2026-07-05T02:48:34.342060+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.07003","created_at":"2026-07-05T02:48:34.342060+00:00"},{"alias_kind":"pith_short_12","alias_value":"26CC3WTXWPXR","created_at":"2026-07-05T02:48:34.342060+00:00"},{"alias_kind":"pith_short_16","alias_value":"26CC3WTXWPXR674G","created_at":"2026-07-05T02:48:34.342060+00:00"},{"alias_kind":"pith_short_8","alias_value":"26CC3WTX","created_at":"2026-07-05T02:48:34.342060+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29480","citing_title":"DTM-Codec: Dynamic Token Masking for VFR Speech Coding with Efficient Boundary Selection","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2210.09461","citing_title":"Token Merging: Your ViT But Faster","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH","json":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH.json","graph_json":"https://pith.science/api/pith-number/26CC3WTXWPXR674G4WVNS2ELBH/graph.json","events_json":"https://pith.science/api/pith-number/26CC3WTXWPXR674G4WVNS2ELBH/events.json","paper":"https://pith.science/paper/26CC3WTX"},"agent_actions":{"view_html":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH","download_json":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH.json","view_paper":"https://pith.science/paper/26CC3WTX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.07003&json=true","fetch_graph":"https://pith.science/api/pith-number/26CC3WTXWPXR674G4WVNS2ELBH/graph.json","fetch_events":"https://pith.science/api/pith-number/26CC3WTXWPXR674G4WVNS2ELBH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH/action/storage_attestation","attest_author":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH/action/author_attestation","sign_citation":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH/action/citation_signature","submit_replication":"https://pith.science/pith/26CC3WTXWPXR674G4WVNS2ELBH/action/replication_record"}},"created_at":"2026-07-05T02:48:34.342060+00:00","updated_at":"2026-07-05T02:48:34.342060+00:00"}