{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UD232AUREBYINFAA52VD7ANYTS","short_pith_number":"pith:UD232AUR","schema_version":"1.0","canonical_sha256":"a0f5bd02912070869400eeaa3f81b89cb02ffc41a348f08bbefa31bd6f36b485","source":{"kind":"arxiv","id":"2502.04056","version":1},"attestation_state":"computed","paper":{"title":"TQ-DiT: Efficient Time-Aware Quantization for Diffusion Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["eess.SP"],"primary_cat":"cs.LG","authors_text":"Hyojin Lee, Joonhyuk Kang, Younghye Hwang","submitted_at":"2025-02-06T13:14:52Z","abstract_excerpt":"Diffusion transformers (DiTs) combine transformer architectures with diffusion models. However, their computational complexity imposes significant limitations on real-time applications and sustainability of AI systems. In this study, we aim to enhance the computational efficiency through model quantization, which represents the weights and activation values with lower precision. Multi-region quantization (MRQ) is introduced to address the asymmetric distribution of network values in DiT blocks by allocating two scaling parameters to sub-regions. Additionally, time-grouping quantization (TGQ) i"},"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":"2502.04056","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-06T13:14:52Z","cross_cats_sorted":["eess.SP"],"title_canon_sha256":"075e8e53805124d267292ff07790b8bfff93d8c6b487fcf99bf87bc56dc09f19","abstract_canon_sha256":"623aa29bc4d53547b2a5534e4e988b4543a282fc7f0dc381cd4b5663ff1225a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:10:33.472718Z","signature_b64":"jE8Vo/9SqWhdovKwKTYoHEYZSsR0uoJFquhtp7nSZaSvaDTYksgFWUggxO9troZ7nVt6G37iwKxpYMnuDwaIAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a0f5bd02912070869400eeaa3f81b89cb02ffc41a348f08bbefa31bd6f36b485","last_reissued_at":"2026-07-05T10:10:33.472283Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:10:33.472283Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TQ-DiT: Efficient Time-Aware Quantization for Diffusion Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["eess.SP"],"primary_cat":"cs.LG","authors_text":"Hyojin Lee, Joonhyuk Kang, Younghye Hwang","submitted_at":"2025-02-06T13:14:52Z","abstract_excerpt":"Diffusion transformers (DiTs) combine transformer architectures with diffusion models. However, their computational complexity imposes significant limitations on real-time applications and sustainability of AI systems. In this study, we aim to enhance the computational efficiency through model quantization, which represents the weights and activation values with lower precision. Multi-region quantization (MRQ) is introduced to address the asymmetric distribution of network values in DiT blocks by allocating two scaling parameters to sub-regions. Additionally, time-grouping quantization (TGQ) i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.04056","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/2502.04056/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":"2502.04056","created_at":"2026-07-05T10:10:33.472338+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.04056v1","created_at":"2026-07-05T10:10:33.472338+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.04056","created_at":"2026-07-05T10:10:33.472338+00:00"},{"alias_kind":"pith_short_12","alias_value":"UD232AUREBYI","created_at":"2026-07-05T10:10:33.472338+00:00"},{"alias_kind":"pith_short_16","alias_value":"UD232AUREBYINFAA","created_at":"2026-07-05T10:10:33.472338+00:00"},{"alias_kind":"pith_short_8","alias_value":"UD232AUR","created_at":"2026-07-05T10:10:33.472338+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.00382","citing_title":"Vitality-Aware Compression for Efficient Image-to-Shape Diffusion Transformers","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2509.23582","citing_title":"RobuQ: Pushing DiTs to W1.58A2 via Robust Activation Quantization","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS","json":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS.json","graph_json":"https://pith.science/api/pith-number/UD232AUREBYINFAA52VD7ANYTS/graph.json","events_json":"https://pith.science/api/pith-number/UD232AUREBYINFAA52VD7ANYTS/events.json","paper":"https://pith.science/paper/UD232AUR"},"agent_actions":{"view_html":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS","download_json":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS.json","view_paper":"https://pith.science/paper/UD232AUR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.04056&json=true","fetch_graph":"https://pith.science/api/pith-number/UD232AUREBYINFAA52VD7ANYTS/graph.json","fetch_events":"https://pith.science/api/pith-number/UD232AUREBYINFAA52VD7ANYTS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS/action/storage_attestation","attest_author":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS/action/author_attestation","sign_citation":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS/action/citation_signature","submit_replication":"https://pith.science/pith/UD232AUREBYINFAA52VD7ANYTS/action/replication_record"}},"created_at":"2026-07-05T10:10:33.472338+00:00","updated_at":"2026-07-05T10:10:33.472338+00:00"}