{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NF37V4K3TSIP7ODREJCEJL5W5G","short_pith_number":"pith:NF37V4K3","schema_version":"1.0","canonical_sha256":"6977faf15b9c90ffb871224444afb6e9be27f646e5cfb0d240453e959eca0005","source":{"kind":"arxiv","id":"2506.14603","version":1},"attestation_state":"computed","paper":{"title":"Align Your Flow: Scaling Continuous-Time Flow Map Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Amirmojtaba Sabour, Karsten Kreis, Sanja Fidler","submitted_at":"2025-06-17T15:06:07Z","abstract_excerpt":"Diffusion- and flow-based models have emerged as state-of-the-art generative modeling approaches, but they require many sampling steps. Consistency models can distill these models into efficient one-step generators; however, unlike flow- and diffusion-based methods, their performance inevitably degrades when increasing the number of steps, which we show both analytically and empirically. Flow maps generalize these approaches by connecting any two noise levels in a single step and remain effective across all step counts. In this paper, we introduce two new continuous-time objectives for trainin"},"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":"2506.14603","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-06-17T15:06:07Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"261354830f11c5c738787f5f7d7dc3650f5e50a8b9b6fda72d5c051085f34b25","abstract_canon_sha256":"906d349f18234e57a3a63cfcf79dcdbaa477648e7d68d30e94f9216ef2f70aec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:22:59.604638Z","signature_b64":"MQkdOQmZ0KwRWAha0QodFbV47JobXRfQmZYMGkp1EGlJkuD12J6BcejmnaRJfFsYJiEwJjlCaOuKXH6ld3DTCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6977faf15b9c90ffb871224444afb6e9be27f646e5cfb0d240453e959eca0005","last_reissued_at":"2026-07-05T11:22:59.604102Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:22:59.604102Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Align Your Flow: Scaling Continuous-Time Flow Map Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Amirmojtaba Sabour, Karsten Kreis, Sanja Fidler","submitted_at":"2025-06-17T15:06:07Z","abstract_excerpt":"Diffusion- and flow-based models have emerged as state-of-the-art generative modeling approaches, but they require many sampling steps. Consistency models can distill these models into efficient one-step generators; however, unlike flow- and diffusion-based methods, their performance inevitably degrades when increasing the number of steps, which we show both analytically and empirically. Flow maps generalize these approaches by connecting any two noise levels in a single step and remain effective across all step counts. In this paper, we introduce two new continuous-time objectives for trainin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.14603","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/2506.14603/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":"2506.14603","created_at":"2026-07-05T11:22:59.604164+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.14603v1","created_at":"2026-07-05T11:22:59.604164+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.14603","created_at":"2026-07-05T11:22:59.604164+00:00"},{"alias_kind":"pith_short_12","alias_value":"NF37V4K3TSIP","created_at":"2026-07-05T11:22:59.604164+00:00"},{"alias_kind":"pith_short_16","alias_value":"NF37V4K3TSIP7ODR","created_at":"2026-07-05T11:22:59.604164+00:00"},{"alias_kind":"pith_short_8","alias_value":"NF37V4K3","created_at":"2026-07-05T11:22:59.604164+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":16,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25473","citing_title":"Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18478","citing_title":"Data-Forcing Distillation: Restoring Diversity and Fidelity in Few-Step Video Generation","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27147","citing_title":"How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01098","citing_title":"Implicit Drifting Policy: One-Step Action Generation via Conditional Expert Geometry","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21489","citing_title":"Variance Reduction for Expectations with Diffusion Teachers","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2602.16813","citing_title":"Flow Map Language Models: One-step Language Modeling via Continuous Denoising","ref_index":92,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27147","citing_title":"How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21489","citing_title":"Variance Reduction for Expectations with Diffusion Teachers","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17244","citing_title":"Drift Flow Matching","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2510.08431","citing_title":"Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2510.20512","citing_title":"Adversarial Concept Distillation for One-Step Diffusion Personalization","ref_index":70,"is_internal_anchor":false},{"citing_arxiv_id":"2602.10764","citing_title":"Dual-End Consistency Model","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2602.16813","citing_title":"Flow Map Language Models: One-step Language Modeling via Continuous Denoising","ref_index":92,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13724","citing_title":"AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillation","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27147","citing_title":"How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18168","citing_title":"Extending One-Step Image Generation from Class Labels to Text via Discriminative Text Representation","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G","json":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G.json","graph_json":"https://pith.science/api/pith-number/NF37V4K3TSIP7ODREJCEJL5W5G/graph.json","events_json":"https://pith.science/api/pith-number/NF37V4K3TSIP7ODREJCEJL5W5G/events.json","paper":"https://pith.science/paper/NF37V4K3"},"agent_actions":{"view_html":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G","download_json":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G.json","view_paper":"https://pith.science/paper/NF37V4K3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.14603&json=true","fetch_graph":"https://pith.science/api/pith-number/NF37V4K3TSIP7ODREJCEJL5W5G/graph.json","fetch_events":"https://pith.science/api/pith-number/NF37V4K3TSIP7ODREJCEJL5W5G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G/action/storage_attestation","attest_author":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G/action/author_attestation","sign_citation":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G/action/citation_signature","submit_replication":"https://pith.science/pith/NF37V4K3TSIP7ODREJCEJL5W5G/action/replication_record"}},"created_at":"2026-07-05T11:22:59.604164+00:00","updated_at":"2026-07-05T11:22:59.604164+00:00"}