{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3N3IZZAQOI3ZPVLZOKSM6WQLWR","short_pith_number":"pith:3N3IZZAQ","schema_version":"1.0","canonical_sha256":"db768ce410723797d57972a4cf5a0bb47e81b02155df01f4e7820c65d7330bb2","source":{"kind":"arxiv","id":"2310.16764","version":1},"attestation_state":"computed","paper":{"title":"ConvNets Match Vision Transformers at Scale","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.NE"],"primary_cat":"cs.CV","authors_text":"Andrew Brock, Leonard Berrada, Samuel L. Smith, Soham De","submitted_at":"2023-10-25T16:52:13Z","abstract_excerpt":"Many researchers believe that ConvNets perform well on small or moderately sized datasets, but are not competitive with Vision Transformers when given access to datasets on the web-scale. We challenge this belief by evaluating a performant ConvNet architecture pre-trained on JFT-4B, a large labelled dataset of images often used for training foundation models. We consider pre-training compute budgets between 0.4k and 110k TPU-v4 core compute hours, and train a series of networks of increasing depth and width from the NFNet model family. We observe a log-log scaling law between held out loss and"},"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":"2310.16764","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-10-25T16:52:13Z","cross_cats_sorted":["cs.LG","cs.NE"],"title_canon_sha256":"c738716e4fa747bcfe8e02fbfa0232653eae20b351e9dd66d4606b25dd69b3a3","abstract_canon_sha256":"bbd1f1d595b68d3d9ae5daf209a3e76f15548b2a0323ad7676914f59acb8a49d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:05:02.783933Z","signature_b64":"V5ybNeX8EmYhXdfJ3Vuil42PyChpJXVIWLzPvtqxLyXpAEqyn95bdVAdWY6RS4G5+/bp9C2JWo4DpFH4UvzvBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db768ce410723797d57972a4cf5a0bb47e81b02155df01f4e7820c65d7330bb2","last_reissued_at":"2026-07-05T07:05:02.783440Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:05:02.783440Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ConvNets Match Vision Transformers at Scale","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.NE"],"primary_cat":"cs.CV","authors_text":"Andrew Brock, Leonard Berrada, Samuel L. Smith, Soham De","submitted_at":"2023-10-25T16:52:13Z","abstract_excerpt":"Many researchers believe that ConvNets perform well on small or moderately sized datasets, but are not competitive with Vision Transformers when given access to datasets on the web-scale. We challenge this belief by evaluating a performant ConvNet architecture pre-trained on JFT-4B, a large labelled dataset of images often used for training foundation models. We consider pre-training compute budgets between 0.4k and 110k TPU-v4 core compute hours, and train a series of networks of increasing depth and width from the NFNet model family. We observe a log-log scaling law between held out loss and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.16764","kind":"arxiv","version":1},"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/2310.16764/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":"2310.16764","created_at":"2026-07-05T07:05:02.783504+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.16764v1","created_at":"2026-07-05T07:05:02.783504+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.16764","created_at":"2026-07-05T07:05:02.783504+00:00"},{"alias_kind":"pith_short_12","alias_value":"3N3IZZAQOI3Z","created_at":"2026-07-05T07:05:02.783504+00:00"},{"alias_kind":"pith_short_16","alias_value":"3N3IZZAQOI3ZPVLZ","created_at":"2026-07-05T07:05:02.783504+00:00"},{"alias_kind":"pith_short_8","alias_value":"3N3IZZAQ","created_at":"2026-07-05T07:05:02.783504+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.21691","citing_title":"There Will Be a Scientific Theory of Deep Learning","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR","json":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR.json","graph_json":"https://pith.science/api/pith-number/3N3IZZAQOI3ZPVLZOKSM6WQLWR/graph.json","events_json":"https://pith.science/api/pith-number/3N3IZZAQOI3ZPVLZOKSM6WQLWR/events.json","paper":"https://pith.science/paper/3N3IZZAQ"},"agent_actions":{"view_html":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR","download_json":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR.json","view_paper":"https://pith.science/paper/3N3IZZAQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.16764&json=true","fetch_graph":"https://pith.science/api/pith-number/3N3IZZAQOI3ZPVLZOKSM6WQLWR/graph.json","fetch_events":"https://pith.science/api/pith-number/3N3IZZAQOI3ZPVLZOKSM6WQLWR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR/action/storage_attestation","attest_author":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR/action/author_attestation","sign_citation":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR/action/citation_signature","submit_replication":"https://pith.science/pith/3N3IZZAQOI3ZPVLZOKSM6WQLWR/action/replication_record"}},"created_at":"2026-07-05T07:05:02.783504+00:00","updated_at":"2026-07-05T07:05:02.783504+00:00"}