{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:DC7TPP5KPV4L76QENPIUTZQUTD","short_pith_number":"pith:DC7TPP5K","schema_version":"1.0","canonical_sha256":"18bf37bfaa7d78bffa046bd149e61498f4a542e078b868fff1aa76a53f00a6ae","source":{"kind":"arxiv","id":"2203.01941","version":2},"attestation_state":"computed","paper":{"title":"Autoregressive Image Generation using Residual Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Chiheon Kim, Doyup Lee, Minsu Cho, Saehoon Kim, Wook-Shin Han","submitted_at":"2022-03-03T11:44:46Z","abstract_excerpt":"For autoregressive (AR) modeling of high-resolution images, vector quantization (VQ) represents an image as a sequence of discrete codes. A short sequence length is important for an AR model to reduce its computational costs to consider long-range interactions of codes. However, we postulate that previous VQ cannot shorten the code sequence and generate high-fidelity images together in terms of the rate-distortion trade-off. In this study, we propose the two-stage framework, which consists of Residual-Quantized VAE (RQ-VAE) and RQ-Transformer, to effectively generate high-resolution images. Gi"},"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":"2203.01941","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-03-03T11:44:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ec3776b17ac20cee8eeb6b79ee2e815640616e46bcb9759bfe1f9a927d309b54","abstract_canon_sha256":"755e07f02866b2ece6d57e1f33b45a855df9815f09db9ecea99c386c06f28a0f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:03:16.085345Z","signature_b64":"h/9oa3+0rTm1ulZy9DK5PZ6JNl7n+Dj5uWd20jZzkPzhXiWATaXhe0uABXZU0vnzL+HdtAfW8Ow8lDktOXBYDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18bf37bfaa7d78bffa046bd149e61498f4a542e078b868fff1aa76a53f00a6ae","last_reissued_at":"2026-07-05T04:03:16.084928Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:03:16.084928Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Autoregressive Image Generation using Residual Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Chiheon Kim, Doyup Lee, Minsu Cho, Saehoon Kim, Wook-Shin Han","submitted_at":"2022-03-03T11:44:46Z","abstract_excerpt":"For autoregressive (AR) modeling of high-resolution images, vector quantization (VQ) represents an image as a sequence of discrete codes. A short sequence length is important for an AR model to reduce its computational costs to consider long-range interactions of codes. However, we postulate that previous VQ cannot shorten the code sequence and generate high-fidelity images together in terms of the rate-distortion trade-off. In this study, we propose the two-stage framework, which consists of Residual-Quantized VAE (RQ-VAE) and RQ-Transformer, to effectively generate high-resolution images. Gi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.01941","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/2203.01941/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":"2203.01941","created_at":"2026-07-05T04:03:16.084983+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.01941v2","created_at":"2026-07-05T04:03:16.084983+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.01941","created_at":"2026-07-05T04:03:16.084983+00:00"},{"alias_kind":"pith_short_12","alias_value":"DC7TPP5KPV4L","created_at":"2026-07-05T04:03:16.084983+00:00"},{"alias_kind":"pith_short_16","alias_value":"DC7TPP5KPV4L76QE","created_at":"2026-07-05T04:03:16.084983+00:00"},{"alias_kind":"pith_short_8","alias_value":"DC7TPP5K","created_at":"2026-07-05T04:03:16.084983+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.02266","citing_title":"HERMES: A Multi-Granularity Labeling Substrate for Pre-training Data Mixtures","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11096","citing_title":"IDEAL: In-DEpth ALignment Makes A Discrete Representation AutoEncoder","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09605","citing_title":"Next-Token Prediction Learns Generalisable Representations of Sleep Physiology","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2209.15571","citing_title":"Building Normalizing Flows with Stochastic Interpolants","ref_index":104,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12965","citing_title":"Efficient Retrieval Scaling with Hierarchical Indexing for Large Scale Recommendation","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD","json":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD.json","graph_json":"https://pith.science/api/pith-number/DC7TPP5KPV4L76QENPIUTZQUTD/graph.json","events_json":"https://pith.science/api/pith-number/DC7TPP5KPV4L76QENPIUTZQUTD/events.json","paper":"https://pith.science/paper/DC7TPP5K"},"agent_actions":{"view_html":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD","download_json":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD.json","view_paper":"https://pith.science/paper/DC7TPP5K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.01941&json=true","fetch_graph":"https://pith.science/api/pith-number/DC7TPP5KPV4L76QENPIUTZQUTD/graph.json","fetch_events":"https://pith.science/api/pith-number/DC7TPP5KPV4L76QENPIUTZQUTD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD/action/storage_attestation","attest_author":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD/action/author_attestation","sign_citation":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD/action/citation_signature","submit_replication":"https://pith.science/pith/DC7TPP5KPV4L76QENPIUTZQUTD/action/replication_record"}},"created_at":"2026-07-05T04:03:16.084983+00:00","updated_at":"2026-07-05T04:03:16.084983+00:00"}