{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6N7BSOIRTFSN723ZTB7XTAU2HT","short_pith_number":"pith:6N7BSOIR","schema_version":"1.0","canonical_sha256":"f37e1939119964dfeb79987f79829a3ced89e25f8aa8b2db553cca5a8d23f26e","source":{"kind":"arxiv","id":"2502.11128","version":2},"attestation_state":"computed","paper":{"title":"FELLE: Autoregressive Speech Synthesis with Token-Wise Coarse-to-Fine Flow Matching","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Haiyang Sun, Haoqin Sun, Hui Wang, Jiaming Zhou, Jinyu Li, Lingwei Meng, Shiwan Zhao, Shujie Liu, Yan Lu, Yanqing Liu, Yifan Yang, Yong Qin","submitted_at":"2025-02-16T13:54:32Z","abstract_excerpt":"To advance continuous-valued token modeling and temporal-coherence enforcement, we propose FELLE, an autoregressive model that integrates language modeling with token-wise flow matching. By leveraging the autoregressive nature of language models and the generative efficacy of flow matching, FELLE effectively predicts continuous-valued tokens (mel-spectrograms). For each continuous-valued token, FELLE modifies the general prior distribution in flow matching by incorporating information from the previous step, improving coherence and stability. Furthermore, to enhance synthesis quality, FELLE in"},"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":"2502.11128","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-16T13:54:32Z","cross_cats_sorted":["cs.SD","eess.AS"],"title_canon_sha256":"6cec18e5f0447120e79839f09d37de57cc949c4884aa22ae7a754a705ce1e7d2","abstract_canon_sha256":"75a241774cb76cb84e7f5a46662e15376a36ad02ce19900faa6fae0051d09ba5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:03:45.879578Z","signature_b64":"9iIhuIN5R7ABP5YcRA+AfnvOnK4BmNLdvkPsMPn9TiS3XG8cAf/m82lALTIRwObDfS+Bddgj0ADunLlvCqvRCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f37e1939119964dfeb79987f79829a3ced89e25f8aa8b2db553cca5a8d23f26e","last_reissued_at":"2026-07-05T12:03:45.878973Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:03:45.878973Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FELLE: Autoregressive Speech Synthesis with Token-Wise Coarse-to-Fine Flow Matching","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Haiyang Sun, Haoqin Sun, Hui Wang, Jiaming Zhou, Jinyu Li, Lingwei Meng, Shiwan Zhao, Shujie Liu, Yan Lu, Yanqing Liu, Yifan Yang, Yong Qin","submitted_at":"2025-02-16T13:54:32Z","abstract_excerpt":"To advance continuous-valued token modeling and temporal-coherence enforcement, we propose FELLE, an autoregressive model that integrates language modeling with token-wise flow matching. By leveraging the autoregressive nature of language models and the generative efficacy of flow matching, FELLE effectively predicts continuous-valued tokens (mel-spectrograms). For each continuous-valued token, FELLE modifies the general prior distribution in flow matching by incorporating information from the previous step, improving coherence and stability. Furthermore, to enhance synthesis quality, FELLE in"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.11128","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/2502.11128/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":"2502.11128","created_at":"2026-07-05T12:03:45.879046+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.11128v2","created_at":"2026-07-05T12:03:45.879046+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.11128","created_at":"2026-07-05T12:03:45.879046+00:00"},{"alias_kind":"pith_short_12","alias_value":"6N7BSOIRTFSN","created_at":"2026-07-05T12:03:45.879046+00:00"},{"alias_kind":"pith_short_16","alias_value":"6N7BSOIRTFSN723Z","created_at":"2026-07-05T12:03:45.879046+00:00"},{"alias_kind":"pith_short_8","alias_value":"6N7BSOIR","created_at":"2026-07-05T12:03:45.879046+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.02398","citing_title":"TTA-Bench: A Comprehensive Benchmark for Evaluating Text-to-Audio Models","ref_index":39,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT","json":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT.json","graph_json":"https://pith.science/api/pith-number/6N7BSOIRTFSN723ZTB7XTAU2HT/graph.json","events_json":"https://pith.science/api/pith-number/6N7BSOIRTFSN723ZTB7XTAU2HT/events.json","paper":"https://pith.science/paper/6N7BSOIR"},"agent_actions":{"view_html":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT","download_json":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT.json","view_paper":"https://pith.science/paper/6N7BSOIR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.11128&json=true","fetch_graph":"https://pith.science/api/pith-number/6N7BSOIRTFSN723ZTB7XTAU2HT/graph.json","fetch_events":"https://pith.science/api/pith-number/6N7BSOIRTFSN723ZTB7XTAU2HT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT/action/storage_attestation","attest_author":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT/action/author_attestation","sign_citation":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT/action/citation_signature","submit_replication":"https://pith.science/pith/6N7BSOIRTFSN723ZTB7XTAU2HT/action/replication_record"}},"created_at":"2026-07-05T12:03:45.879046+00:00","updated_at":"2026-07-05T12:03:45.879046+00:00"}