{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:N2XFPXTJTVT7RPKYYGRLCSHNWE","short_pith_number":"pith:N2XFPXTJ","schema_version":"1.0","canonical_sha256":"6eae57de699d67f8bd58c1a2b148edb109dbdb6ce39d1b5cf778d190e7b6fea7","source":{"kind":"arxiv","id":"2309.01947","version":2},"attestation_state":"computed","paper":{"title":"TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-device ASR Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Ayushi Dalmia, Chunyang Wu, Danni Li, Dilin Wang, Haichuan Yang, Jay Mahadeokar, Junteng Jia, Mike Seltzer, Ozlem Kalinli, Raghuraman Krishnamoorthi, Vikas Chandra, Xin Lei, Yassir Fathullah, Yuan Shangguan","submitted_at":"2023-09-05T04:47:55Z","abstract_excerpt":"Automatic Speech Recognition (ASR) models need to be optimized for specific hardware before they can be deployed on devices. This can be done by tuning the model's hyperparameters or exploring variations in its architecture. Re-training and re-validating models after making these changes can be a resource-intensive task. This paper presents TODM (Train Once Deploy Many), a new approach to efficiently train many sizes of hardware-friendly on-device ASR models with comparable GPU-hours to that of a single training job. TODM leverages insights from prior work on Supernet, where Recurrent Neural N"},"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":"2309.01947","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-05T04:47:55Z","cross_cats_sorted":["cs.LG","cs.SD","eess.AS"],"title_canon_sha256":"be08833a05e291e2c94e34dcbb2460feb759c18f8721788bcfd2442783da182c","abstract_canon_sha256":"d33124389319605ac7db46adffed4c483bb272670ada56328800e672aa154ab7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:16:46.041865Z","signature_b64":"OpBhRwj1/YV/6nIQ8/UJQtuHFudB3hznqYCjErndpus0jQQeRs9/kZl/PNOJsSMU+scKS7yXk7Kc3CMJpbUFAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6eae57de699d67f8bd58c1a2b148edb109dbdb6ce39d1b5cf778d190e7b6fea7","last_reissued_at":"2026-07-05T07:16:46.041362Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:16:46.041362Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-device ASR Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Ayushi Dalmia, Chunyang Wu, Danni Li, Dilin Wang, Haichuan Yang, Jay Mahadeokar, Junteng Jia, Mike Seltzer, Ozlem Kalinli, Raghuraman Krishnamoorthi, Vikas Chandra, Xin Lei, Yassir Fathullah, Yuan Shangguan","submitted_at":"2023-09-05T04:47:55Z","abstract_excerpt":"Automatic Speech Recognition (ASR) models need to be optimized for specific hardware before they can be deployed on devices. This can be done by tuning the model's hyperparameters or exploring variations in its architecture. Re-training and re-validating models after making these changes can be a resource-intensive task. This paper presents TODM (Train Once Deploy Many), a new approach to efficiently train many sizes of hardware-friendly on-device ASR models with comparable GPU-hours to that of a single training job. TODM leverages insights from prior work on Supernet, where Recurrent Neural N"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.01947","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/2309.01947/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":"2309.01947","created_at":"2026-07-05T07:16:46.041430+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.01947v2","created_at":"2026-07-05T07:16:46.041430+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.01947","created_at":"2026-07-05T07:16:46.041430+00:00"},{"alias_kind":"pith_short_12","alias_value":"N2XFPXTJTVT7","created_at":"2026-07-05T07:16:46.041430+00:00"},{"alias_kind":"pith_short_16","alias_value":"N2XFPXTJTVT7RPKY","created_at":"2026-07-05T07:16:46.041430+00:00"},{"alias_kind":"pith_short_8","alias_value":"N2XFPXTJ","created_at":"2026-07-05T07:16:46.041430+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/N2XFPXTJTVT7RPKYYGRLCSHNWE","json":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE.json","graph_json":"https://pith.science/api/pith-number/N2XFPXTJTVT7RPKYYGRLCSHNWE/graph.json","events_json":"https://pith.science/api/pith-number/N2XFPXTJTVT7RPKYYGRLCSHNWE/events.json","paper":"https://pith.science/paper/N2XFPXTJ"},"agent_actions":{"view_html":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE","download_json":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE.json","view_paper":"https://pith.science/paper/N2XFPXTJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.01947&json=true","fetch_graph":"https://pith.science/api/pith-number/N2XFPXTJTVT7RPKYYGRLCSHNWE/graph.json","fetch_events":"https://pith.science/api/pith-number/N2XFPXTJTVT7RPKYYGRLCSHNWE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE/action/storage_attestation","attest_author":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE/action/author_attestation","sign_citation":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE/action/citation_signature","submit_replication":"https://pith.science/pith/N2XFPXTJTVT7RPKYYGRLCSHNWE/action/replication_record"}},"created_at":"2026-07-05T07:16:46.041430+00:00","updated_at":"2026-07-05T07:16:46.041430+00:00"}