{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:TGEXFIVUCXJILWUACX5FIOSQZN","short_pith_number":"pith:TGEXFIVU","schema_version":"1.0","canonical_sha256":"998972a2b415d285da8015fa543a50cb60019a14def6a252df43fef2edc4d62e","source":{"kind":"arxiv","id":"2111.13445","version":5},"attestation_state":"computed","paper":{"title":"How Well Do Sparse Imagenet Models Transfer?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexandra Peste, Dan Alistarh, Eugenia Iofinova, Mark Kurtz","submitted_at":"2021-11-26T11:58:51Z","abstract_excerpt":"Transfer learning is a classic paradigm by which models pretrained on large \"upstream\" datasets are adapted to yield good results on \"downstream\" specialized datasets. Generally, more accurate models on the \"upstream\" dataset tend to provide better transfer accuracy \"downstream\". In this work, we perform an in-depth investigation of this phenomenon in the context of convolutional neural networks (CNNs) trained on the ImageNet dataset, which have been pruned - that is, compressed by sparsifying their connections. We consider transfer using unstructured pruned models obtained by applying several"},"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":"2111.13445","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-11-26T11:58:51Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"97ea5fb59696766a26a88c9d0e97e6e38306ad73edd11a012942b3b7454690e8","abstract_canon_sha256":"e754b03a0bcef7fab173322889e1661f6245d4fd69c11aee4df5eb6e5cc8298a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:16:37.872357Z","signature_b64":"7K3YAbrJPGXSnMY4uFhVhvU7GOLT9WR3zxQXmBPXbMiLKllvSh4i8wZqqwasHWhS6MjZ+A1EmTztDW/r7S4mAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"998972a2b415d285da8015fa543a50cb60019a14def6a252df43fef2edc4d62e","last_reissued_at":"2026-07-05T04:16:37.871804Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:16:37.871804Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Well Do Sparse Imagenet Models Transfer?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexandra Peste, Dan Alistarh, Eugenia Iofinova, Mark Kurtz","submitted_at":"2021-11-26T11:58:51Z","abstract_excerpt":"Transfer learning is a classic paradigm by which models pretrained on large \"upstream\" datasets are adapted to yield good results on \"downstream\" specialized datasets. Generally, more accurate models on the \"upstream\" dataset tend to provide better transfer accuracy \"downstream\". In this work, we perform an in-depth investigation of this phenomenon in the context of convolutional neural networks (CNNs) trained on the ImageNet dataset, which have been pruned - that is, compressed by sparsifying their connections. We consider transfer using unstructured pruned models obtained by applying several"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2111.13445","kind":"arxiv","version":5},"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/2111.13445/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":"2111.13445","created_at":"2026-07-05T04:16:37.871862+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.13445v5","created_at":"2026-07-05T04:16:37.871862+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.13445","created_at":"2026-07-05T04:16:37.871862+00:00"},{"alias_kind":"pith_short_12","alias_value":"TGEXFIVUCXJI","created_at":"2026-07-05T04:16:37.871862+00:00"},{"alias_kind":"pith_short_16","alias_value":"TGEXFIVUCXJILWUA","created_at":"2026-07-05T04:16:37.871862+00:00"},{"alias_kind":"pith_short_8","alias_value":"TGEXFIVU","created_at":"2026-07-05T04:16:37.871862+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.00899","citing_title":"HASSLE-free: A unified Framework for Sparse plus Low-Rank Matrix Decomposition for LLMs","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN","json":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN.json","graph_json":"https://pith.science/api/pith-number/TGEXFIVUCXJILWUACX5FIOSQZN/graph.json","events_json":"https://pith.science/api/pith-number/TGEXFIVUCXJILWUACX5FIOSQZN/events.json","paper":"https://pith.science/paper/TGEXFIVU"},"agent_actions":{"view_html":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN","download_json":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN.json","view_paper":"https://pith.science/paper/TGEXFIVU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.13445&json=true","fetch_graph":"https://pith.science/api/pith-number/TGEXFIVUCXJILWUACX5FIOSQZN/graph.json","fetch_events":"https://pith.science/api/pith-number/TGEXFIVUCXJILWUACX5FIOSQZN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN/action/storage_attestation","attest_author":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN/action/author_attestation","sign_citation":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN/action/citation_signature","submit_replication":"https://pith.science/pith/TGEXFIVUCXJILWUACX5FIOSQZN/action/replication_record"}},"created_at":"2026-07-05T04:16:37.871862+00:00","updated_at":"2026-07-05T04:16:37.871862+00:00"}