{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:OSL5OBPUDZEXQKDT3VOJ4TBZAT","short_pith_number":"pith:OSL5OBPU","canonical_record":{"source":{"id":"2212.05055","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-09T18:57:37Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"5c3f8e69669d60e011faf0a150ff78fca097efca7476bd997c0f57499acd6199","abstract_canon_sha256":"20b569b204e58e28c9ee2fc97213c31ddeaf4faf4917e670b8206b0920a766b3"},"schema_version":"1.0"},"canonical_sha256":"7497d705f41e49782873dd5c9e4c3904ea7f7e9e88b5b26de9e2110791bc957a","source":{"kind":"arxiv","id":"2212.05055","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2212.05055","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"arxiv_version","alias_value":"2212.05055v2","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.05055","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"pith_short_12","alias_value":"OSL5OBPUDZEX","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"pith_short_16","alias_value":"OSL5OBPUDZEXQKDT","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"pith_short_8","alias_value":"OSL5OBPU","created_at":"2026-07-05T05:43:03Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:OSL5OBPUDZEXQKDT3VOJ4TBZAT","target":"record","payload":{"canonical_record":{"source":{"id":"2212.05055","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-09T18:57:37Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"5c3f8e69669d60e011faf0a150ff78fca097efca7476bd997c0f57499acd6199","abstract_canon_sha256":"20b569b204e58e28c9ee2fc97213c31ddeaf4faf4917e670b8206b0920a766b3"},"schema_version":"1.0"},"canonical_sha256":"7497d705f41e49782873dd5c9e4c3904ea7f7e9e88b5b26de9e2110791bc957a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:43:03.849467Z","signature_b64":"gskquY+hxvdnjiF6YmqMQtBsSGtOsOGa/Xlw4w75Sdm3UFYmdjy/VKiK1aow4orF6cp8Zu82PEwx2zy20ovGBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7497d705f41e49782873dd5c9e4c3904ea7f7e9e88b5b26de9e2110791bc957a","last_reissued_at":"2026-07-05T05:43:03.849055Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:43:03.849055Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2212.05055","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:43:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Wlr6Cqk6U3EHmGDyZk3AGCSTynsasJLnjstfHgopExhWsLwq6r/SKvlHgjNFdp8P70QzRi+fpJRyZiOE4ilQDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-28T04:49:05.034566Z"},"content_sha256":"64ddf8fccd4442fc72fde3df1dd162d877ce800ddf62c4f7ddf2f185b37c02c6","schema_version":"1.0","event_id":"sha256:64ddf8fccd4442fc72fde3df1dd162d877ce800ddf62c4f7ddf2f185b37c02c6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:OSL5OBPUDZEXQKDT3VOJ4TBZAT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Aran Komatsuzaki, Basil Mustafa, Carlos Riquelme Ruiz, James Lee-Thorp, Joan Puigcerver, Joshua Ainslie, Mostafa Dehghani, Neil Houlsby, Yi Tay","submitted_at":"2022-12-09T18:57:37Z","abstract_excerpt":"Training large, deep neural networks to convergence can be prohibitively expensive. As a result, often only a small selection of popular, dense models are reused across different contexts and tasks. Increasingly, sparsely activated models, which seek to decouple model size from computation costs, are becoming an attractive alternative to dense models. Although more efficient in terms of quality and computation cost, sparse models remain data-hungry and costly to train from scratch in the large scale regime. In this work, we propose sparse upcycling -- a simple way to reuse sunk training costs "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.05055","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/2212.05055/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:43:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y0KnraSfWUPr7K4J9oOUPBbmojhIqoGAcIl8tSs59Cu9Su4jfEHW3KXwbXVZBh4iuMfFR4kiXx4HKPznlcyoBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-28T04:49:05.034952Z"},"content_sha256":"e74dc3855df68bf7d4f87b38b37d1e1462f30e6818c70d1c10fc13c8cd1cc7cc","schema_version":"1.0","event_id":"sha256:e74dc3855df68bf7d4f87b38b37d1e1462f30e6818c70d1c10fc13c8cd1cc7cc"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT/bundle.json","state_url":"https://pith.science/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-28T04:49:05Z","links":{"resolver":"https://pith.science/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT","bundle":"https://pith.science/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT/bundle.json","state":"https://pith.science/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/OSL5OBPUDZEXQKDT3VOJ4TBZAT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:OSL5OBPUDZEXQKDT3VOJ4TBZAT","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"20b569b204e58e28c9ee2fc97213c31ddeaf4faf4917e670b8206b0920a766b3","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-09T18:57:37Z","title_canon_sha256":"5c3f8e69669d60e011faf0a150ff78fca097efca7476bd997c0f57499acd6199"},"schema_version":"1.0","source":{"id":"2212.05055","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2212.05055","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"arxiv_version","alias_value":"2212.05055v2","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.05055","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"pith_short_12","alias_value":"OSL5OBPUDZEX","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"pith_short_16","alias_value":"OSL5OBPUDZEXQKDT","created_at":"2026-07-05T05:43:03Z"},{"alias_kind":"pith_short_8","alias_value":"OSL5OBPU","created_at":"2026-07-05T05:43:03Z"}],"graph_snapshots":[{"event_id":"sha256:e74dc3855df68bf7d4f87b38b37d1e1462f30e6818c70d1c10fc13c8cd1cc7cc","target":"graph","created_at":"2026-07-05T05:43:03Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2212.05055/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training large, deep neural networks to convergence can be prohibitively expensive. As a result, often only a small selection of popular, dense models are reused across different contexts and tasks. Increasingly, sparsely activated models, which seek to decouple model size from computation costs, are becoming an attractive alternative to dense models. Although more efficient in terms of quality and computation cost, sparse models remain data-hungry and costly to train from scratch in the large scale regime. In this work, we propose sparse upcycling -- a simple way to reuse sunk training costs ","authors_text":"Aran Komatsuzaki, Basil Mustafa, Carlos Riquelme Ruiz, James Lee-Thorp, Joan Puigcerver, Joshua Ainslie, Mostafa Dehghani, Neil Houlsby, Yi Tay","cross_cats":["cs.CL","cs.CV"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-09T18:57:37Z","title":"Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.05055","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:64ddf8fccd4442fc72fde3df1dd162d877ce800ddf62c4f7ddf2f185b37c02c6","target":"record","created_at":"2026-07-05T05:43:03Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"20b569b204e58e28c9ee2fc97213c31ddeaf4faf4917e670b8206b0920a766b3","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-12-09T18:57:37Z","title_canon_sha256":"5c3f8e69669d60e011faf0a150ff78fca097efca7476bd997c0f57499acd6199"},"schema_version":"1.0","source":{"id":"2212.05055","kind":"arxiv","version":2}},"canonical_sha256":"7497d705f41e49782873dd5c9e4c3904ea7f7e9e88b5b26de9e2110791bc957a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7497d705f41e49782873dd5c9e4c3904ea7f7e9e88b5b26de9e2110791bc957a","first_computed_at":"2026-07-05T05:43:03.849055Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:43:03.849055Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gskquY+hxvdnjiF6YmqMQtBsSGtOsOGa/Xlw4w75Sdm3UFYmdjy/VKiK1aow4orF6cp8Zu82PEwx2zy20ovGBA==","signature_status":"signed_v1","signed_at":"2026-07-05T05:43:03.849467Z","signed_message":"canonical_sha256_bytes"},"source_id":"2212.05055","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:64ddf8fccd4442fc72fde3df1dd162d877ce800ddf62c4f7ddf2f185b37c02c6","sha256:e74dc3855df68bf7d4f87b38b37d1e1462f30e6818c70d1c10fc13c8cd1cc7cc"],"state_sha256":"5009aa94dbcd06f117596ef0294c5a6bb4d9310926adde311974b7742e5cbdea"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LEnI990ZHTc1TpCG+wVK1P5lsLa5Pe4Z2G9IePT1ai/m84wXpzs/J+prB8ob8mqPuMApQaluwZn2yMd3Z/jqBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-28T04:49:05.037678Z","bundle_sha256":"b9171edb04fc89a6ce4ccdf68275fdbd86b3705c0b046b8b3cd5ad40c9a95d8d"}}