{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FMA7ZI5XJH3W6WXEXKEUDLCMC2","short_pith_number":"pith:FMA7ZI5X","schema_version":"1.0","canonical_sha256":"2b01fca3b749f76f5ae4ba8941ac4c16817ac6e25253df4445ef1c7295439f0d","source":{"kind":"arxiv","id":"2111.15667","version":3},"attestation_state":"computed","paper":{"title":"Adaptive Token Sampling For Efficient Vision Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Eric Sommerlade, Farnoush Rezaei Jafari, Hamed Pirsiavash, Hamid Reza Vaezi Joze, Juergen Gall, Mohsen Fayyaz, Soroush Abbasi Koohpayegani, Sunando Sengupta","submitted_at":"2021-11-30T18:56:57Z","abstract_excerpt":"While state-of-the-art vision transformer models achieve promising results in image classification, they are computationally expensive and require many GFLOPs. Although the GFLOPs of a vision transformer can be decreased by reducing the number of tokens in the network, there is no setting that is optimal for all input images. In this work, we therefore introduce a differentiable parameter-free Adaptive Token Sampler (ATS) module, which can be plugged into any existing vision transformer architecture. ATS empowers vision transformers by scoring and adaptively sampling significant tokens. As a r"},"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":"2111.15667","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-11-30T18:56:57Z","cross_cats_sorted":[],"title_canon_sha256":"d50b131f9a8a7a8e6c237b928f457cb386c261fe4d9b34c256fd38cbdd4918fe","abstract_canon_sha256":"8dbd07e53305d0c92dd8bbea77e2939e267506448e744ba921472c7ce8ceeaa7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:43:37.291957Z","signature_b64":"aZKwDqF9n0dP10T/ZgrRHCgsIMp/pWKNGQTAu3xmOjRoN2lCwVnyVNa46SJb8ddbZqY0Rx0L6v4B0iLKI9XIBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2b01fca3b749f76f5ae4ba8941ac4c16817ac6e25253df4445ef1c7295439f0d","last_reissued_at":"2026-07-05T04:43:37.291446Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:43:37.291446Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptive Token Sampling For Efficient Vision Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Eric Sommerlade, Farnoush Rezaei Jafari, Hamed Pirsiavash, Hamid Reza Vaezi Joze, Juergen Gall, Mohsen Fayyaz, Soroush Abbasi Koohpayegani, Sunando Sengupta","submitted_at":"2021-11-30T18:56:57Z","abstract_excerpt":"While state-of-the-art vision transformer models achieve promising results in image classification, they are computationally expensive and require many GFLOPs. Although the GFLOPs of a vision transformer can be decreased by reducing the number of tokens in the network, there is no setting that is optimal for all input images. In this work, we therefore introduce a differentiable parameter-free Adaptive Token Sampler (ATS) module, which can be plugged into any existing vision transformer architecture. ATS empowers vision transformers by scoring and adaptively sampling significant tokens. As a r"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2111.15667","kind":"arxiv","version":3},"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/2111.15667/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":"2111.15667","created_at":"2026-07-05T04:43:37.291529+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.15667v3","created_at":"2026-07-05T04:43:37.291529+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.15667","created_at":"2026-07-05T04:43:37.291529+00:00"},{"alias_kind":"pith_short_12","alias_value":"FMA7ZI5XJH3W","created_at":"2026-07-05T04:43:37.291529+00:00"},{"alias_kind":"pith_short_16","alias_value":"FMA7ZI5XJH3W6WXE","created_at":"2026-07-05T04:43:37.291529+00:00"},{"alias_kind":"pith_short_8","alias_value":"FMA7ZI5X","created_at":"2026-07-05T04:43:37.291529+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.14787","citing_title":"FOCUS: Fused Observation of Channels for Unveiling Spectra","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2","json":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2.json","graph_json":"https://pith.science/api/pith-number/FMA7ZI5XJH3W6WXEXKEUDLCMC2/graph.json","events_json":"https://pith.science/api/pith-number/FMA7ZI5XJH3W6WXEXKEUDLCMC2/events.json","paper":"https://pith.science/paper/FMA7ZI5X"},"agent_actions":{"view_html":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2","download_json":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2.json","view_paper":"https://pith.science/paper/FMA7ZI5X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.15667&json=true","fetch_graph":"https://pith.science/api/pith-number/FMA7ZI5XJH3W6WXEXKEUDLCMC2/graph.json","fetch_events":"https://pith.science/api/pith-number/FMA7ZI5XJH3W6WXEXKEUDLCMC2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2/action/storage_attestation","attest_author":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2/action/author_attestation","sign_citation":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2/action/citation_signature","submit_replication":"https://pith.science/pith/FMA7ZI5XJH3W6WXEXKEUDLCMC2/action/replication_record"}},"created_at":"2026-07-05T04:43:37.291529+00:00","updated_at":"2026-07-05T04:43:37.291529+00:00"}