{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2FMSNSFE5VLDH6CXU76WSPBFG5","short_pith_number":"pith:2FMSNSFE","schema_version":"1.0","canonical_sha256":"d15926c8a4ed5633f857a7fd693c25377270701406567241563dd2bbd180eb60","source":{"kind":"arxiv","id":"2412.12496","version":4},"attestation_state":"computed","paper":{"title":"Faster Vision Mamba is Rebuilt in Minutes via Merged Token Re-training","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Kai Wang, Mingjia Shi, Ruiji Yu, Shanmukha Ramakrishna Vedantam, Wangbo Zhao, Xiaojiang Peng, Xuanlei Zhao, Yang You, Yuhao Zhou, Zekai Li, Zhiyuan Liang","submitted_at":"2024-12-17T02:56:35Z","abstract_excerpt":"Vision Mamba has shown close to state of the art performance on computer vision tasks, drawing much interest in increasing it's efficiency. A promising approach is token reduction (that has been successfully implemented in ViTs). Pruning informative tokens in Mamba leads to a high loss of key knowledge and degraded performance. An alternative, of merging tokens preserves more information than pruning, also suffers for large compression ratios. Our key insight is that a quick round of retraining after token merging yeilds robust results across various compression ratios. Empirically, pruned Vim"},"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":"2412.12496","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-17T02:56:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8ed734cfb767d0330673667f5a3f0f17c600ff9faeb7bd1ffb988c76b37b68e8","abstract_canon_sha256":"f19052d97487555a9a06dca785650ac939929a0270e3bbaee504444dbe2a824c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:48:22.994777Z","signature_b64":"LwxD4wfx22Q/piYt1L5geHh0Ys0iXXYaEEzsTJr0+Fleqdes69A/gI6YcJCTEPUHWtpLOVBHZiDmDK8oHyX1Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d15926c8a4ed5633f857a7fd693c25377270701406567241563dd2bbd180eb60","last_reissued_at":"2026-07-05T10:48:22.994272Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:48:22.994272Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Faster Vision Mamba is Rebuilt in Minutes via Merged Token Re-training","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Kai Wang, Mingjia Shi, Ruiji Yu, Shanmukha Ramakrishna Vedantam, Wangbo Zhao, Xiaojiang Peng, Xuanlei Zhao, Yang You, Yuhao Zhou, Zekai Li, Zhiyuan Liang","submitted_at":"2024-12-17T02:56:35Z","abstract_excerpt":"Vision Mamba has shown close to state of the art performance on computer vision tasks, drawing much interest in increasing it's efficiency. A promising approach is token reduction (that has been successfully implemented in ViTs). Pruning informative tokens in Mamba leads to a high loss of key knowledge and degraded performance. An alternative, of merging tokens preserves more information than pruning, also suffers for large compression ratios. Our key insight is that a quick round of retraining after token merging yeilds robust results across various compression ratios. Empirically, pruned Vim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.12496","kind":"arxiv","version":4},"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/2412.12496/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":"2412.12496","created_at":"2026-07-05T10:48:22.994333+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.12496v4","created_at":"2026-07-05T10:48:22.994333+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.12496","created_at":"2026-07-05T10:48:22.994333+00:00"},{"alias_kind":"pith_short_12","alias_value":"2FMSNSFE5VLD","created_at":"2026-07-05T10:48:22.994333+00:00"},{"alias_kind":"pith_short_16","alias_value":"2FMSNSFE5VLDH6CX","created_at":"2026-07-05T10:48:22.994333+00:00"},{"alias_kind":"pith_short_8","alias_value":"2FMSNSFE","created_at":"2026-07-05T10:48:22.994333+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19932","citing_title":"Spatial-Aware Reduction Framework: Towards Efficient and Faithful Visual State Space Models","ref_index":106,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5","json":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5.json","graph_json":"https://pith.science/api/pith-number/2FMSNSFE5VLDH6CXU76WSPBFG5/graph.json","events_json":"https://pith.science/api/pith-number/2FMSNSFE5VLDH6CXU76WSPBFG5/events.json","paper":"https://pith.science/paper/2FMSNSFE"},"agent_actions":{"view_html":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5","download_json":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5.json","view_paper":"https://pith.science/paper/2FMSNSFE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.12496&json=true","fetch_graph":"https://pith.science/api/pith-number/2FMSNSFE5VLDH6CXU76WSPBFG5/graph.json","fetch_events":"https://pith.science/api/pith-number/2FMSNSFE5VLDH6CXU76WSPBFG5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5/action/storage_attestation","attest_author":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5/action/author_attestation","sign_citation":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5/action/citation_signature","submit_replication":"https://pith.science/pith/2FMSNSFE5VLDH6CXU76WSPBFG5/action/replication_record"}},"created_at":"2026-07-05T10:48:22.994333+00:00","updated_at":"2026-07-05T10:48:22.994333+00:00"}