{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SIZSHVBMDOOU6KIIYT6MOMKOJK","short_pith_number":"pith:SIZSHVBM","schema_version":"1.0","canonical_sha256":"923323d42c1b9d4f2908c4fcc7314e4ab1f98d965afda3dc713d55946011d686","source":{"kind":"arxiv","id":"2605.20199","version":1},"attestation_state":"computed","paper":{"title":"FlowLM: Few-Step Language Modeling via Diffusion-to-Flow Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Letian Chen, Peilin Zhao, Runzhe Zhang, Wenpeng Zhang, Zhouhan Lin","submitted_at":"2026-04-06T10:36:22Z","abstract_excerpt":"We present FlowLM, a flow matching language model transformed from pre-trained diffusion language models via efficient fine-tuning. By re-aligning the curved sampling trajectories of diffusion models into straight-line flows, FlowLM enables high quality few-step generation that rivals or even outperforms the quality of 2,000-step diffusion sampling with very few training epochs. Remarkably, finetuned FlowLM reaches performance saturation with only half as many training epochs as training from scratch, both approaches greatly outperforming the original diffusion model, thereby validating our me"},"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":"2605.20199","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-04-06T10:36:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"09712e776ac2829a222264b5497f9dd57bef92087f84f6f45d56a40ec4360e20","abstract_canon_sha256":"85f12aee19d3db0ef90ae998b6f4d80e6689e3636b1273dcfa1ada562f2978a8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T00:04:20.754907Z","signature_b64":"EG+bnT8HW0mCUsErsEmxwXzRqFVYl8gmxob1WoAOv8+xZaklnDFTGHPpqAz/jDrzGsMXA4EPtGg/C/wdv/myDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"923323d42c1b9d4f2908c4fcc7314e4ab1f98d965afda3dc713d55946011d686","last_reissued_at":"2026-05-21T00:04:20.754076Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T00:04:20.754076Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FlowLM: Few-Step Language Modeling via Diffusion-to-Flow Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Letian Chen, Peilin Zhao, Runzhe Zhang, Wenpeng Zhang, Zhouhan Lin","submitted_at":"2026-04-06T10:36:22Z","abstract_excerpt":"We present FlowLM, a flow matching language model transformed from pre-trained diffusion language models via efficient fine-tuning. By re-aligning the curved sampling trajectories of diffusion models into straight-line flows, FlowLM enables high quality few-step generation that rivals or even outperforms the quality of 2,000-step diffusion sampling with very few training epochs. Remarkably, finetuned FlowLM reaches performance saturation with only half as many training epochs as training from scratch, both approaches greatly outperforming the original diffusion model, thereby validating our me"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20199","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/2605.20199/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":"2605.20199","created_at":"2026-05-21T00:04:20.754174+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20199v1","created_at":"2026-05-21T00:04:20.754174+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20199","created_at":"2026-05-21T00:04:20.754174+00:00"},{"alias_kind":"pith_short_12","alias_value":"SIZSHVBMDOOU","created_at":"2026-05-21T00:04:20.754174+00:00"},{"alias_kind":"pith_short_16","alias_value":"SIZSHVBMDOOU6KII","created_at":"2026-05-21T00:04:20.754174+00:00"},{"alias_kind":"pith_short_8","alias_value":"SIZSHVBM","created_at":"2026-05-21T00:04:20.754174+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK","json":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK.json","graph_json":"https://pith.science/api/pith-number/SIZSHVBMDOOU6KIIYT6MOMKOJK/graph.json","events_json":"https://pith.science/api/pith-number/SIZSHVBMDOOU6KIIYT6MOMKOJK/events.json","paper":"https://pith.science/paper/SIZSHVBM"},"agent_actions":{"view_html":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK","download_json":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK.json","view_paper":"https://pith.science/paper/SIZSHVBM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20199&json=true","fetch_graph":"https://pith.science/api/pith-number/SIZSHVBMDOOU6KIIYT6MOMKOJK/graph.json","fetch_events":"https://pith.science/api/pith-number/SIZSHVBMDOOU6KIIYT6MOMKOJK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK/action/storage_attestation","attest_author":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK/action/author_attestation","sign_citation":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK/action/citation_signature","submit_replication":"https://pith.science/pith/SIZSHVBMDOOU6KIIYT6MOMKOJK/action/replication_record"}},"created_at":"2026-05-21T00:04:20.754174+00:00","updated_at":"2026-05-21T00:04:20.754174+00:00"}