{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:EUQUH4WJBJ44TWTTJIAMVMGR7T","short_pith_number":"pith:EUQUH4WJ","schema_version":"1.0","canonical_sha256":"252143f2c90a79c9da734a00cab0d1fcc568eb023061c9f428de5aeb645d55c7","source":{"kind":"arxiv","id":"2011.09208","version":3},"attestation_state":"computed","paper":{"title":"Whale: Efficient Giant Model Training over Heterogeneous GPUs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Ang Wang, Jie Zhang, Langshi Chen, Le Jiang, Wei Lin, Wencong Xiao, Xianyan Jia, Xiaoyong Liu, Xinyuan Li, Yong Li, Zhen Zheng, Ziji Shi","submitted_at":"2020-11-18T10:54:31Z","abstract_excerpt":"The scaling up of deep neural networks has been demonstrated to be effective in improving model quality, but also encompasses several training challenges in terms of training efficiency, programmability, and resource adaptability. We present Whale, a general and efficient distributed training framework for giant models. To support various parallel strategies and their hybrids, Whale generalizes the programming interface by defining two new primitives in the form of model annotations, allowing for incorporating user hints. The Whale runtime utilizes those annotations and performs graph optimiza"},"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":"2011.09208","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2020-11-18T10:54:31Z","cross_cats_sorted":[],"title_canon_sha256":"237df2e29802145f7e02f11aeef09124cba8fc97b8ddb791133402b81b7162b2","abstract_canon_sha256":"4734f4b818526fd1a58651abd22b0bf2664d10e91bd0f4636ac333870615faab"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:29:07.169335Z","signature_b64":"qOgaKUJ0ygOcnIfB2e/Lpox9CT2PbbsFDpJHdvpFYbB+51FhZJ4uTf9YH8R/QEXtAUuwueVf0rm0gQtqTX5nDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"252143f2c90a79c9da734a00cab0d1fcc568eb023061c9f428de5aeb645d55c7","last_reissued_at":"2026-07-05T04:29:07.168861Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:29:07.168861Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Whale: Efficient Giant Model Training over Heterogeneous GPUs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Ang Wang, Jie Zhang, Langshi Chen, Le Jiang, Wei Lin, Wencong Xiao, Xianyan Jia, Xiaoyong Liu, Xinyuan Li, Yong Li, Zhen Zheng, Ziji Shi","submitted_at":"2020-11-18T10:54:31Z","abstract_excerpt":"The scaling up of deep neural networks has been demonstrated to be effective in improving model quality, but also encompasses several training challenges in terms of training efficiency, programmability, and resource adaptability. We present Whale, a general and efficient distributed training framework for giant models. To support various parallel strategies and their hybrids, Whale generalizes the programming interface by defining two new primitives in the form of model annotations, allowing for incorporating user hints. The Whale runtime utilizes those annotations and performs graph optimiza"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2011.09208","kind":"arxiv","version":3},"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/2011.09208/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":"2011.09208","created_at":"2026-07-05T04:29:07.168911+00:00"},{"alias_kind":"arxiv_version","alias_value":"2011.09208v3","created_at":"2026-07-05T04:29:07.168911+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2011.09208","created_at":"2026-07-05T04:29:07.168911+00:00"},{"alias_kind":"pith_short_12","alias_value":"EUQUH4WJBJ44","created_at":"2026-07-05T04:29:07.168911+00:00"},{"alias_kind":"pith_short_16","alias_value":"EUQUH4WJBJ44TWTT","created_at":"2026-07-05T04:29:07.168911+00:00"},{"alias_kind":"pith_short_8","alias_value":"EUQUH4WJ","created_at":"2026-07-05T04:29:07.168911+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10415","citing_title":"RATrain: A Resource-Aware Training Runtime for Large Language Models on Bandwidth-Constrained Heterogeneous Supercomputing Platforms","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T","json":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T.json","graph_json":"https://pith.science/api/pith-number/EUQUH4WJBJ44TWTTJIAMVMGR7T/graph.json","events_json":"https://pith.science/api/pith-number/EUQUH4WJBJ44TWTTJIAMVMGR7T/events.json","paper":"https://pith.science/paper/EUQUH4WJ"},"agent_actions":{"view_html":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T","download_json":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T.json","view_paper":"https://pith.science/paper/EUQUH4WJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2011.09208&json=true","fetch_graph":"https://pith.science/api/pith-number/EUQUH4WJBJ44TWTTJIAMVMGR7T/graph.json","fetch_events":"https://pith.science/api/pith-number/EUQUH4WJBJ44TWTTJIAMVMGR7T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T/action/storage_attestation","attest_author":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T/action/author_attestation","sign_citation":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T/action/citation_signature","submit_replication":"https://pith.science/pith/EUQUH4WJBJ44TWTTJIAMVMGR7T/action/replication_record"}},"created_at":"2026-07-05T04:29:07.168911+00:00","updated_at":"2026-07-05T04:29:07.168911+00:00"}