{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WQWWIB3PVHHXUVT7MWY6AHHGOH","short_pith_number":"pith:WQWWIB3P","schema_version":"1.0","canonical_sha256":"b42d64076fa9cf7a567f65b1e01ce671d1ce60cf1f6ed52b3970538e9bc6f323","source":{"kind":"arxiv","id":"2607.02119","version":1},"attestation_state":"computed","paper":{"title":"An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"eess.AS","authors_text":"Haoran Wang, Jinchuan Tian, Shinji Watanabe, Siddhant Arora","submitted_at":"2026-07-02T12:55:11Z","abstract_excerpt":"While Large Multimodal Models excel in comprehension, high-throughput inference engines lack native support for multimodal generation. This is severe in Speech Language Models, where generating multi-layered audio tokens via decoupled AR+NAR or synchronous Multi-Token Prediction (MTP) with delay-pattern interleaving conflicts with standard single-stream loops. We present a vLLM-based inference pipeline for unified speech understanding and generation. We extend autoregressive decoding to natively execute delay-pattern de-interleaving and coordinated multi-stream sampling, integrating an on-GPU "},"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":"2607.02119","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2026-07-02T12:55:11Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ec34913f3854443bfc2e5331b4c00e881bcd77e01a79c8afae9da21e0299fdbd","abstract_canon_sha256":"bd80f4dffd5bf22f00ed8430fba67df33f2224e1bff6d1f87c989ceb0090754c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:42.765068Z","signature_b64":"YNvVUKpKvb2WYKNI4U6YXZzQ/xKGBmBC/oSgVPFH71v1ZiRWMWdVzo6mtFRERZOkIorVdCBDq+GfQJbelC1/CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b42d64076fa9cf7a567f65b1e01ce671d1ce60cf1f6ed52b3970538e9bc6f323","last_reissued_at":"2026-07-03T01:17:42.764644Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:42.764644Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"eess.AS","authors_text":"Haoran Wang, Jinchuan Tian, Shinji Watanabe, Siddhant Arora","submitted_at":"2026-07-02T12:55:11Z","abstract_excerpt":"While Large Multimodal Models excel in comprehension, high-throughput inference engines lack native support for multimodal generation. This is severe in Speech Language Models, where generating multi-layered audio tokens via decoupled AR+NAR or synchronous Multi-Token Prediction (MTP) with delay-pattern interleaving conflicts with standard single-stream loops. We present a vLLM-based inference pipeline for unified speech understanding and generation. We extend autoregressive decoding to natively execute delay-pattern de-interleaving and coordinated multi-stream sampling, integrating an on-GPU "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.02119","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/2607.02119/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":"2607.02119","created_at":"2026-07-03T01:17:42.764701+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.02119v1","created_at":"2026-07-03T01:17:42.764701+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.02119","created_at":"2026-07-03T01:17:42.764701+00:00"},{"alias_kind":"pith_short_12","alias_value":"WQWWIB3PVHHX","created_at":"2026-07-03T01:17:42.764701+00:00"},{"alias_kind":"pith_short_16","alias_value":"WQWWIB3PVHHXUVT7","created_at":"2026-07-03T01:17:42.764701+00:00"},{"alias_kind":"pith_short_8","alias_value":"WQWWIB3P","created_at":"2026-07-03T01:17:42.764701+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.02119","citing_title":"An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH","json":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH.json","graph_json":"https://pith.science/api/pith-number/WQWWIB3PVHHXUVT7MWY6AHHGOH/graph.json","events_json":"https://pith.science/api/pith-number/WQWWIB3PVHHXUVT7MWY6AHHGOH/events.json","paper":"https://pith.science/paper/WQWWIB3P"},"agent_actions":{"view_html":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH","download_json":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH.json","view_paper":"https://pith.science/paper/WQWWIB3P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.02119&json=true","fetch_graph":"https://pith.science/api/pith-number/WQWWIB3PVHHXUVT7MWY6AHHGOH/graph.json","fetch_events":"https://pith.science/api/pith-number/WQWWIB3PVHHXUVT7MWY6AHHGOH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH/action/storage_attestation","attest_author":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH/action/author_attestation","sign_citation":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH/action/citation_signature","submit_replication":"https://pith.science/pith/WQWWIB3PVHHXUVT7MWY6AHHGOH/action/replication_record"}},"created_at":"2026-07-03T01:17:42.764701+00:00","updated_at":"2026-07-03T01:17:42.764701+00:00"}