{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ORSRO67RUOJBAKESA5EU6NJIHW","short_pith_number":"pith:ORSRO67R","schema_version":"1.0","canonical_sha256":"7465177bf1a39210289207494f35283db232009260822b6eea067d9e78e0a33d","source":{"kind":"arxiv","id":"2410.05317","version":4},"attestation_state":"computed","paper":{"title":"Accelerating Diffusion Transformers with Token-wise Feature Caching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Chang Zou, Linfeng Zhang, Siteng Huang, Ting Liu, Xuyang Liu","submitted_at":"2024-10-05T03:47:06Z","abstract_excerpt":"Diffusion transformers have shown significant effectiveness in both image and video synthesis at the expense of huge computation costs. To address this problem, feature caching methods have been introduced to accelerate diffusion transformers by caching the features in previous timesteps and reusing them in the following timesteps. However, previous caching methods ignore that different tokens exhibit different sensitivities to feature caching, and feature caching on some tokens may lead to 10$\\times$ more destruction to the overall generation quality compared with other tokens. In this paper,"},"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":"2410.05317","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-05T03:47:06Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"e89642604145feda75bcc2332c03af691c8f2172e486edc91a4e82e27f5e6180","abstract_canon_sha256":"f975d07947cda276f81a74360e0eff53bd01291110d587aa78123f519acbae7b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:38.402741Z","signature_b64":"OFsHbvzDJgW9QVC7aetyB1PEbpfJw7BN6KiugzT+1Rv2bHYtMEOTLfuEmf880dqKeQ+MBxWAEnaw3NkHSLRkCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7465177bf1a39210289207494f35283db232009260822b6eea067d9e78e0a33d","last_reissued_at":"2026-07-05T10:16:38.402188Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:38.402188Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Accelerating Diffusion Transformers with Token-wise Feature Caching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Chang Zou, Linfeng Zhang, Siteng Huang, Ting Liu, Xuyang Liu","submitted_at":"2024-10-05T03:47:06Z","abstract_excerpt":"Diffusion transformers have shown significant effectiveness in both image and video synthesis at the expense of huge computation costs. To address this problem, feature caching methods have been introduced to accelerate diffusion transformers by caching the features in previous timesteps and reusing them in the following timesteps. However, previous caching methods ignore that different tokens exhibit different sensitivities to feature caching, and feature caching on some tokens may lead to 10$\\times$ more destruction to the overall generation quality compared with other tokens. In this paper,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.05317","kind":"arxiv","version":4},"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/2410.05317/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":"2410.05317","created_at":"2026-07-05T10:16:38.402247+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.05317v4","created_at":"2026-07-05T10:16:38.402247+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.05317","created_at":"2026-07-05T10:16:38.402247+00:00"},{"alias_kind":"pith_short_12","alias_value":"ORSRO67RUOJB","created_at":"2026-07-05T10:16:38.402247+00:00"},{"alias_kind":"pith_short_16","alias_value":"ORSRO67RUOJBAKES","created_at":"2026-07-05T10:16:38.402247+00:00"},{"alias_kind":"pith_short_8","alias_value":"ORSRO67R","created_at":"2026-07-05T10:16:38.402247+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":19,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26769","citing_title":"ResilPhase: Plug-and-Play Phase Mapping and Noise-Resilient Macro-Trajectory Extrapolation for Diffusion Acceleration","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13496","citing_title":"Budget-Constrained Step-Level Diffusion Caching","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31026","citing_title":"OTCache: Optimal Transport for Geometry-Aware Caching in Diffusion Models","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2606.30557","citing_title":"EcoVideo: Entropy-Orchestrated Video Generation Paradigm in Cloud-Edge Dynamics","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25798","citing_title":"DiSC: Resolution-Scalable Acceleration of Diffusion Models by Exploiting Sparsity and Cached Token Reuse with Hash-based Distribution","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27075","citing_title":"BeCARE: Budgeted Cache Refresh for Diffusion Transformer Acceleration","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00310","citing_title":"Where to Refine, When to Stop: Rethinking Redundancy via Latent Discrepancy for Efficient Visual Autoregressive Generation","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22015","citing_title":"ORBIS: Output-Guided Token Reduction with Distribution-Aware Matching for Video Diffusion Acceleration","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18736","citing_title":"Spectral Progressive Diffusion for Efficient Image and Video Generation","ref_index":98,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18190","citing_title":"Dual-Rate Diffusion: Accelerating diffusion models with an interleaved heavy-light network","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18736","citing_title":"Spectral Progressive Diffusion for Efficient Image and Video Generation","ref_index":96,"is_internal_anchor":false},{"citing_arxiv_id":"2602.22575","citing_title":"S2O: Early Stopping for Sparse Attention via Online Permutation","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01725","citing_title":"Motion-Aware Caching for Efficient Autoregressive Video Generation","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14191","citing_title":"CoReDiT: Spatial Coherence-Guided Token Pruning and Reconstruction for Efficient Diffusion Transformers","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2604.28185","citing_title":"Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling","ref_index":105,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02152","citing_title":"SpecEdit: Training-Free Acceleration for Diffusion based Image Editing via Semantic Locking","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16492","citing_title":"LayerCache: Exploiting Layer-wise Velocity Heterogeneity for Efficient Flow Matching Inference","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01725","citing_title":"Motion-Aware Caching for Efficient Autoregressive Video Generation","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09227","citing_title":"Training-free, Perceptually Consistent Low-Resolution Previews with High-Resolution Image for Efficient Workflows of Diffusion Models","ref_index":56,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW","json":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW.json","graph_json":"https://pith.science/api/pith-number/ORSRO67RUOJBAKESA5EU6NJIHW/graph.json","events_json":"https://pith.science/api/pith-number/ORSRO67RUOJBAKESA5EU6NJIHW/events.json","paper":"https://pith.science/paper/ORSRO67R"},"agent_actions":{"view_html":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW","download_json":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW.json","view_paper":"https://pith.science/paper/ORSRO67R","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.05317&json=true","fetch_graph":"https://pith.science/api/pith-number/ORSRO67RUOJBAKESA5EU6NJIHW/graph.json","fetch_events":"https://pith.science/api/pith-number/ORSRO67RUOJBAKESA5EU6NJIHW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW/action/storage_attestation","attest_author":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW/action/author_attestation","sign_citation":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW/action/citation_signature","submit_replication":"https://pith.science/pith/ORSRO67RUOJBAKESA5EU6NJIHW/action/replication_record"}},"created_at":"2026-07-05T10:16:38.402247+00:00","updated_at":"2026-07-05T10:16:38.402247+00:00"}