{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:YNRPOVXKQDJHZE2AMXWQQNQL6E","short_pith_number":"pith:YNRPOVXK","canonical_record":{"source":{"id":"2203.05482","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-03-10T17:03:49Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"51862f7d7f6becfeb6cdc9486965f7d936e4c3108bcb6fa4c70f9b29473872a0","abstract_canon_sha256":"ac88a086ddf3e7eda025b9c0225df91ca0fbcf234595736ffb5a82ec4ab2f60f"},"schema_version":"1.0"},"canonical_sha256":"c362f756ea80d27c934065ed08360bf11eb698975ccfb934d0523dc42e361a3e","source":{"kind":"arxiv","id":"2203.05482","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.05482","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"arxiv_version","alias_value":"2203.05482v3","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.05482","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"pith_short_12","alias_value":"YNRPOVXKQDJH","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"pith_short_16","alias_value":"YNRPOVXKQDJHZE2A","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"pith_short_8","alias_value":"YNRPOVXK","created_at":"2026-07-05T04:36:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:YNRPOVXKQDJHZE2AMXWQQNQL6E","target":"record","payload":{"canonical_record":{"source":{"id":"2203.05482","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-03-10T17:03:49Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"51862f7d7f6becfeb6cdc9486965f7d936e4c3108bcb6fa4c70f9b29473872a0","abstract_canon_sha256":"ac88a086ddf3e7eda025b9c0225df91ca0fbcf234595736ffb5a82ec4ab2f60f"},"schema_version":"1.0"},"canonical_sha256":"c362f756ea80d27c934065ed08360bf11eb698975ccfb934d0523dc42e361a3e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:36:47.860381Z","signature_b64":"udvpYZqmP54niAV2M0q1gqaNMbDfp0qT+wWgKL4YRo7errxpa2p7bRROBDrN6CIYaEIJQo3T5X6HXwF/mBA8BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c362f756ea80d27c934065ed08360bf11eb698975ccfb934d0523dc42e361a3e","last_reissued_at":"2026-07-05T04:36:47.859955Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:36:47.859955Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2203.05482","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T04:36:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9nhmf2vMt8Y8k2EXUupQC9Y3nUIQ7T5PwLFrQ/9Br2kCjnepyoS3pN6zDz33u9HTXQF544TvQIGfj07XgVCcBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T08:48:47.719085Z"},"content_sha256":"0aeab3ef0fa76acd203a68573c9dbbc1792f1ccfa4b598cd7730ecfec34c7a51","schema_version":"1.0","event_id":"sha256:0aeab3ef0fa76acd203a68573c9dbbc1792f1ccfa4b598cd7730ecfec34c7a51"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:YNRPOVXKQDJHZE2AMXWQQNQL6E","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Ali Farhadi, Ari S. Morcos, Gabriel Ilharco, Hongseok Namkoong, Ludwig Schmidt, Mitchell Wortsman, Raphael Gontijo-Lopes, Rebecca Roelofs, Samir Yitzhak Gadre, Simon Kornblith, Yair Carmon","submitted_at":"2022-03-10T17:03:49Z","abstract_excerpt":"The conventional recipe for maximizing model accuracy is to (1) train multiple models with various hyperparameters and (2) pick the individual model which performs best on a held-out validation set, discarding the remainder. In this paper, we revisit the second step of this procedure in the context of fine-tuning large pre-trained models, where fine-tuned models often appear to lie in a single low error basin. We show that averaging the weights of multiple models fine-tuned with different hyperparameter configurations often improves accuracy and robustness. Unlike a conventional ensemble, we m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.05482","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/2203.05482/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":"2026-07-05T04:36:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iYiTmdk6ZXIC40cfTN7ggwJt3iq+mYC73CFosf4dqK2z3WlKMZN+OH9shIWyWCj0aOYmgI64B5kNvVBbR2TUCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T08:48:47.719489Z"},"content_sha256":"6746bc822b652eac7f5d9da06c6c17e278d5e3cb94a220fa62be46d6ef215342","schema_version":"1.0","event_id":"sha256:6746bc822b652eac7f5d9da06c6c17e278d5e3cb94a220fa62be46d6ef215342"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E/bundle.json","state_url":"https://pith.science/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E/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":"2026-07-06T08:48:47Z","links":{"resolver":"https://pith.science/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E","bundle":"https://pith.science/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E/bundle.json","state":"https://pith.science/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YNRPOVXKQDJHZE2AMXWQQNQL6E/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:YNRPOVXKQDJHZE2AMXWQQNQL6E","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":"ac88a086ddf3e7eda025b9c0225df91ca0fbcf234595736ffb5a82ec4ab2f60f","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-03-10T17:03:49Z","title_canon_sha256":"51862f7d7f6becfeb6cdc9486965f7d936e4c3108bcb6fa4c70f9b29473872a0"},"schema_version":"1.0","source":{"id":"2203.05482","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.05482","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"arxiv_version","alias_value":"2203.05482v3","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.05482","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"pith_short_12","alias_value":"YNRPOVXKQDJH","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"pith_short_16","alias_value":"YNRPOVXKQDJHZE2A","created_at":"2026-07-05T04:36:47Z"},{"alias_kind":"pith_short_8","alias_value":"YNRPOVXK","created_at":"2026-07-05T04:36:47Z"}],"graph_snapshots":[{"event_id":"sha256:6746bc822b652eac7f5d9da06c6c17e278d5e3cb94a220fa62be46d6ef215342","target":"graph","created_at":"2026-07-05T04:36:47Z","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/2203.05482/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The conventional recipe for maximizing model accuracy is to (1) train multiple models with various hyperparameters and (2) pick the individual model which performs best on a held-out validation set, discarding the remainder. In this paper, we revisit the second step of this procedure in the context of fine-tuning large pre-trained models, where fine-tuned models often appear to lie in a single low error basin. We show that averaging the weights of multiple models fine-tuned with different hyperparameter configurations often improves accuracy and robustness. Unlike a conventional ensemble, we m","authors_text":"Ali Farhadi, Ari S. Morcos, Gabriel Ilharco, Hongseok Namkoong, Ludwig Schmidt, Mitchell Wortsman, Raphael Gontijo-Lopes, Rebecca Roelofs, Samir Yitzhak Gadre, Simon Kornblith, Yair Carmon","cross_cats":["cs.CL","cs.CV"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-03-10T17:03:49Z","title":"Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.05482","kind":"arxiv","version":3},"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":"sha256:0aeab3ef0fa76acd203a68573c9dbbc1792f1ccfa4b598cd7730ecfec34c7a51","target":"record","created_at":"2026-07-05T04:36:47Z","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":"ac88a086ddf3e7eda025b9c0225df91ca0fbcf234595736ffb5a82ec4ab2f60f","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-03-10T17:03:49Z","title_canon_sha256":"51862f7d7f6becfeb6cdc9486965f7d936e4c3108bcb6fa4c70f9b29473872a0"},"schema_version":"1.0","source":{"id":"2203.05482","kind":"arxiv","version":3}},"canonical_sha256":"c362f756ea80d27c934065ed08360bf11eb698975ccfb934d0523dc42e361a3e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c362f756ea80d27c934065ed08360bf11eb698975ccfb934d0523dc42e361a3e","first_computed_at":"2026-07-05T04:36:47.859955Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T04:36:47.859955Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"udvpYZqmP54niAV2M0q1gqaNMbDfp0qT+wWgKL4YRo7errxpa2p7bRROBDrN6CIYaEIJQo3T5X6HXwF/mBA8BA==","signature_status":"signed_v1","signed_at":"2026-07-05T04:36:47.860381Z","signed_message":"canonical_sha256_bytes"},"source_id":"2203.05482","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0aeab3ef0fa76acd203a68573c9dbbc1792f1ccfa4b598cd7730ecfec34c7a51","sha256:6746bc822b652eac7f5d9da06c6c17e278d5e3cb94a220fa62be46d6ef215342"],"state_sha256":"63f21abb354e811cb57e80f7c22868b36c5d6fe85f6f40a32820ee7efc036c0e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tSj9kKBLxpjwNbC6kOFCECvIUReyh+aSPwxGSfbI1AKFzHTgmfrmFoMSadRTbCIaUPLm8St7fsExvcouXB7bDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T08:48:47.721549Z","bundle_sha256":"8d022141766da3fc2e536ae9e58ebc81855d4387d48f260421f0e4d10b7f632e"}}