{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:M5W3C4ZBXXQIQQ32QBV2OYTLKD","short_pith_number":"pith:M5W3C4ZB","schema_version":"1.0","canonical_sha256":"676db17321bde088437a806ba7626b50c6ca1467b61033c08476765c66ea42ac","source":{"kind":"arxiv","id":"2402.13723","version":1},"attestation_state":"computed","paper":{"title":"The Effect of Batch Size on Contrastive Self-Supervised Speech Representation Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"David A. van Leeuwen, Nik Vaessen","submitted_at":"2024-02-21T11:35:19Z","abstract_excerpt":"Foundation models in speech are often trained using many GPUs, which implicitly leads to large effective batch sizes. In this paper we study the effect of batch size on pre-training, both in terms of statistics that can be monitored during training, and in the effect on the performance of a downstream fine-tuning task. By using batch sizes varying from 87.5 seconds to 80 minutes of speech we show that, for a fixed amount of iterations, larger batch sizes result in better pre-trained models. However, there is lower limit for stability, and an upper limit for effectiveness. We then show that the"},"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":"2402.13723","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2024-02-21T11:35:19Z","cross_cats_sorted":["cs.LG","eess.AS"],"title_canon_sha256":"8b35a73408868b2ac64e554ded64f58e3087afb4596f523e6318216fe84dd774","abstract_canon_sha256":"0946c1fb7eb2c6b2c21f91fc7fee20d158495e8dd1d59bd3c744f0f2d4b69b18"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:47:48.065387Z","signature_b64":"qnElGGi6Tx1TztKvmSJnFUEV8+U/x54m2FzvEQez9NZfzfsn8x5UB1p+kxt7LlgF242sWeMuMp8hfTN/Qdb7BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"676db17321bde088437a806ba7626b50c6ca1467b61033c08476765c66ea42ac","last_reissued_at":"2026-07-05T07:47:48.064945Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:47:48.064945Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Effect of Batch Size on Contrastive Self-Supervised Speech Representation Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"David A. van Leeuwen, Nik Vaessen","submitted_at":"2024-02-21T11:35:19Z","abstract_excerpt":"Foundation models in speech are often trained using many GPUs, which implicitly leads to large effective batch sizes. In this paper we study the effect of batch size on pre-training, both in terms of statistics that can be monitored during training, and in the effect on the performance of a downstream fine-tuning task. By using batch sizes varying from 87.5 seconds to 80 minutes of speech we show that, for a fixed amount of iterations, larger batch sizes result in better pre-trained models. However, there is lower limit for stability, and an upper limit for effectiveness. We then show that the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.13723","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/2402.13723/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":"2402.13723","created_at":"2026-07-05T07:47:48.065003+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.13723v1","created_at":"2026-07-05T07:47:48.065003+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.13723","created_at":"2026-07-05T07:47:48.065003+00:00"},{"alias_kind":"pith_short_12","alias_value":"M5W3C4ZBXXQI","created_at":"2026-07-05T07:47:48.065003+00:00"},{"alias_kind":"pith_short_16","alias_value":"M5W3C4ZBXXQIQQ32","created_at":"2026-07-05T07:47:48.065003+00:00"},{"alias_kind":"pith_short_8","alias_value":"M5W3C4ZB","created_at":"2026-07-05T07:47:48.065003+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/M5W3C4ZBXXQIQQ32QBV2OYTLKD","json":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD.json","graph_json":"https://pith.science/api/pith-number/M5W3C4ZBXXQIQQ32QBV2OYTLKD/graph.json","events_json":"https://pith.science/api/pith-number/M5W3C4ZBXXQIQQ32QBV2OYTLKD/events.json","paper":"https://pith.science/paper/M5W3C4ZB"},"agent_actions":{"view_html":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD","download_json":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD.json","view_paper":"https://pith.science/paper/M5W3C4ZB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.13723&json=true","fetch_graph":"https://pith.science/api/pith-number/M5W3C4ZBXXQIQQ32QBV2OYTLKD/graph.json","fetch_events":"https://pith.science/api/pith-number/M5W3C4ZBXXQIQQ32QBV2OYTLKD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD/action/storage_attestation","attest_author":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD/action/author_attestation","sign_citation":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD/action/citation_signature","submit_replication":"https://pith.science/pith/M5W3C4ZBXXQIQQ32QBV2OYTLKD/action/replication_record"}},"created_at":"2026-07-05T07:47:48.065003+00:00","updated_at":"2026-07-05T07:47:48.065003+00:00"}