{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:D7DKMJDMHILNAGPPPDC5LL6YRN","short_pith_number":"pith:D7DKMJDM","schema_version":"1.0","canonical_sha256":"1fc6a6246c3a16d019ef78c5d5afd88b6917d8d8296ed28aa250e7203ca7b2d3","source":{"kind":"arxiv","id":"2511.08704","version":2},"attestation_state":"computed","paper":{"title":"Rethinking Generative Image Pretraining: How Far Are We From Scaling Up Next-Pixel Prediction?","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Adams Wei Yu, Chen Liang, Lijun Yu, Quoc V. Le, Xinchen Yan, Yifeng Lu","submitted_at":"2025-11-11T19:11:02Z","abstract_excerpt":"This paper investigates the scaling properties of autoregressive next-pixel prediction, a simple, end-to-end yet under-explored framework for unified vision models. Starting with images at resolutions of 32x32, we train a family of Transformers using IsoFlops profiles across compute budgets up to 7e19 FLOPs and evaluate three distinct target metrics: next-pixel prediction objective, ImageNet classification accuracy, and generation-based completion measured by Fr'echet Distance. First, optimal scaling strategy is critically task-dependent. At a fixed resolution of 32x32 alone, the optimal scali"},"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":"2511.08704","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CV","submitted_at":"2025-11-11T19:11:02Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"41faded3d7fc7a7444438193eca805360594953d5cd6481a652a57e4d1535333","abstract_canon_sha256":"9fd09393b429989ac09499cc1dc73ea2cb6bc3c31fea9040bf9c8a2e95cab9d9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:17.165415Z","signature_b64":"Z/GyFMh77gFgKbCZgJnYWZl1zwDaTikcq2bJnow0hPfgRxu/PgJ5jg4GpbOtzRLrySQ+72ss+RuFXWw4o8wCDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1fc6a6246c3a16d019ef78c5d5afd88b6917d8d8296ed28aa250e7203ca7b2d3","last_reissued_at":"2026-05-20T00:04:17.164460Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:17.164460Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rethinking Generative Image Pretraining: How Far Are We From Scaling Up Next-Pixel Prediction?","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Adams Wei Yu, Chen Liang, Lijun Yu, Quoc V. Le, Xinchen Yan, Yifeng Lu","submitted_at":"2025-11-11T19:11:02Z","abstract_excerpt":"This paper investigates the scaling properties of autoregressive next-pixel prediction, a simple, end-to-end yet under-explored framework for unified vision models. Starting with images at resolutions of 32x32, we train a family of Transformers using IsoFlops profiles across compute budgets up to 7e19 FLOPs and evaluate three distinct target metrics: next-pixel prediction objective, ImageNet classification accuracy, and generation-based completion measured by Fr'echet Distance. First, optimal scaling strategy is critically task-dependent. At a fixed resolution of 32x32 alone, the optimal scali"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.08704","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/2511.08704/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":"2511.08704","created_at":"2026-05-20T00:04:17.164632+00:00"},{"alias_kind":"arxiv_version","alias_value":"2511.08704v2","created_at":"2026-05-20T00:04:17.164632+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.08704","created_at":"2026-05-20T00:04:17.164632+00:00"},{"alias_kind":"pith_short_12","alias_value":"D7DKMJDMHILN","created_at":"2026-05-20T00:04:17.164632+00:00"},{"alias_kind":"pith_short_16","alias_value":"D7DKMJDMHILNAGPP","created_at":"2026-05-20T00:04:17.164632+00:00"},{"alias_kind":"pith_short_8","alias_value":"D7DKMJDM","created_at":"2026-05-20T00:04:17.164632+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN","json":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN.json","graph_json":"https://pith.science/api/pith-number/D7DKMJDMHILNAGPPPDC5LL6YRN/graph.json","events_json":"https://pith.science/api/pith-number/D7DKMJDMHILNAGPPPDC5LL6YRN/events.json","paper":"https://pith.science/paper/D7DKMJDM"},"agent_actions":{"view_html":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN","download_json":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN.json","view_paper":"https://pith.science/paper/D7DKMJDM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2511.08704&json=true","fetch_graph":"https://pith.science/api/pith-number/D7DKMJDMHILNAGPPPDC5LL6YRN/graph.json","fetch_events":"https://pith.science/api/pith-number/D7DKMJDMHILNAGPPPDC5LL6YRN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN/action/storage_attestation","attest_author":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN/action/author_attestation","sign_citation":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN/action/citation_signature","submit_replication":"https://pith.science/pith/D7DKMJDMHILNAGPPPDC5LL6YRN/action/replication_record"}},"created_at":"2026-05-20T00:04:17.164632+00:00","updated_at":"2026-05-20T00:04:17.164632+00:00"}