{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:ACWHURWLNTIRBZ4UQMPFQOTPO2","short_pith_number":"pith:ACWHURWL","canonical_record":{"source":{"id":"2204.03044","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-06T18:54:48Z","cross_cats_sorted":["cs.CV","cs.LG"],"title_canon_sha256":"d40e11d72e20e7a737370732ce729e66d8bac63394e22b1c8b537983345f5cba","abstract_canon_sha256":"10e9dd2345df6f81456d993ededcde7db5b14500e1c3399ac18342268093f9b8"},"schema_version":"1.0"},"canonical_sha256":"00ac7a46cb6cd110e794831e583a6f768f56d94d021517f9f111d79c9f5c9be7","source":{"kind":"arxiv","id":"2204.03044","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.03044","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"arxiv_version","alias_value":"2204.03044v1","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.03044","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"pith_short_12","alias_value":"ACWHURWLNTIR","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"pith_short_16","alias_value":"ACWHURWLNTIRBZ4U","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"pith_short_8","alias_value":"ACWHURWL","created_at":"2026-07-05T04:12:19Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:ACWHURWLNTIRBZ4UQMPFQOTPO2","target":"record","payload":{"canonical_record":{"source":{"id":"2204.03044","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-06T18:54:48Z","cross_cats_sorted":["cs.CV","cs.LG"],"title_canon_sha256":"d40e11d72e20e7a737370732ce729e66d8bac63394e22b1c8b537983345f5cba","abstract_canon_sha256":"10e9dd2345df6f81456d993ededcde7db5b14500e1c3399ac18342268093f9b8"},"schema_version":"1.0"},"canonical_sha256":"00ac7a46cb6cd110e794831e583a6f768f56d94d021517f9f111d79c9f5c9be7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:12:19.240238Z","signature_b64":"1fY1HKWYWX1U02utXYzyXWuWMqajSgNFqxrFg8Op5rrINGBnV+imLFjFAF2vnjtcXgBvG/k+ISdIJcqX1S2nBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"00ac7a46cb6cd110e794831e583a6f768f56d94d021517f9f111d79c9f5c9be7","last_reissued_at":"2026-07-05T04:12:19.239769Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:12:19.239769Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2204.03044","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T04:12:19Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TRGwR303O+TNR5VPmzY1urIffyEkLFjN8thQG8b/LvA38RqfZDSsdZLDNotYAAHBWE7vYMCfHWqeC5/jiiH6DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T11:45:48.603581Z"},"content_sha256":"dd1530ea1f42e6cdd16340dbf7f2f4dc2756043ba311e4f73b33f760394b424a","schema_version":"1.0","event_id":"sha256:dd1530ea1f42e6cdd16340dbf7f2f4dc2756043ba311e4f73b33f760394b424a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:ACWHURWLNTIRBZ4UQMPFQOTPO2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Fusing finetuned models for better pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Elad Venezian, Leshem Choshen, Noam Slonim, Yoav Katz","submitted_at":"2022-04-06T18:54:48Z","abstract_excerpt":"Pretrained models are the standard starting point for training. This approach consistently outperforms the use of a random initialization. However, pretraining is a costly endeavour that few can undertake.\n  In this paper, we create better base models at hardly any cost, by fusing multiple existing fine tuned models into one. Specifically, we fuse by averaging the weights of these models. We show that the fused model results surpass the pretrained model ones. We also show that fusing is often better than intertraining.\n  We find that fusing is less dependent on the target task. Furthermore, we"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.03044","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/2204.03044/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":"2026-07-05T04:12:19Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"l6jhcMTyeI5Kn5ttoVKlQ5p4Gp2hePSCHOexB9o0QLdN8LK35FeCvJflbRETmnizbIlc2wWH2Wk8etzA4/xmAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T11:45:48.604079Z"},"content_sha256":"409acafadf196fe9a187479d1f6b03e9c8963029b365962c7bf57c22bb2e81a4","schema_version":"1.0","event_id":"sha256:409acafadf196fe9a187479d1f6b03e9c8963029b365962c7bf57c22bb2e81a4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2/bundle.json","state_url":"https://pith.science/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2/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":"2026-08-05T11:45:48Z","links":{"resolver":"https://pith.science/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2","bundle":"https://pith.science/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2/bundle.json","state":"https://pith.science/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ACWHURWLNTIRBZ4UQMPFQOTPO2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:ACWHURWLNTIRBZ4UQMPFQOTPO2","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":"10e9dd2345df6f81456d993ededcde7db5b14500e1c3399ac18342268093f9b8","cross_cats_sorted":["cs.CV","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-06T18:54:48Z","title_canon_sha256":"d40e11d72e20e7a737370732ce729e66d8bac63394e22b1c8b537983345f5cba"},"schema_version":"1.0","source":{"id":"2204.03044","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.03044","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"arxiv_version","alias_value":"2204.03044v1","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.03044","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"pith_short_12","alias_value":"ACWHURWLNTIR","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"pith_short_16","alias_value":"ACWHURWLNTIRBZ4U","created_at":"2026-07-05T04:12:19Z"},{"alias_kind":"pith_short_8","alias_value":"ACWHURWL","created_at":"2026-07-05T04:12:19Z"}],"graph_snapshots":[{"event_id":"sha256:409acafadf196fe9a187479d1f6b03e9c8963029b365962c7bf57c22bb2e81a4","target":"graph","created_at":"2026-07-05T04:12:19Z","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/2204.03044/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pretrained models are the standard starting point for training. This approach consistently outperforms the use of a random initialization. However, pretraining is a costly endeavour that few can undertake.\n  In this paper, we create better base models at hardly any cost, by fusing multiple existing fine tuned models into one. Specifically, we fuse by averaging the weights of these models. We show that the fused model results surpass the pretrained model ones. We also show that fusing is often better than intertraining.\n  We find that fusing is less dependent on the target task. Furthermore, we","authors_text":"Elad Venezian, Leshem Choshen, Noam Slonim, Yoav Katz","cross_cats":["cs.CV","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-06T18:54:48Z","title":"Fusing finetuned models for better pretraining"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.03044","kind":"arxiv","version":1},"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:dd1530ea1f42e6cdd16340dbf7f2f4dc2756043ba311e4f73b33f760394b424a","target":"record","created_at":"2026-07-05T04:12:19Z","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":"10e9dd2345df6f81456d993ededcde7db5b14500e1c3399ac18342268093f9b8","cross_cats_sorted":["cs.CV","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-06T18:54:48Z","title_canon_sha256":"d40e11d72e20e7a737370732ce729e66d8bac63394e22b1c8b537983345f5cba"},"schema_version":"1.0","source":{"id":"2204.03044","kind":"arxiv","version":1}},"canonical_sha256":"00ac7a46cb6cd110e794831e583a6f768f56d94d021517f9f111d79c9f5c9be7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"00ac7a46cb6cd110e794831e583a6f768f56d94d021517f9f111d79c9f5c9be7","first_computed_at":"2026-07-05T04:12:19.239769Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T04:12:19.239769Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1fY1HKWYWX1U02utXYzyXWuWMqajSgNFqxrFg8Op5rrINGBnV+imLFjFAF2vnjtcXgBvG/k+ISdIJcqX1S2nBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T04:12:19.240238Z","signed_message":"canonical_sha256_bytes"},"source_id":"2204.03044","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:dd1530ea1f42e6cdd16340dbf7f2f4dc2756043ba311e4f73b33f760394b424a","sha256:409acafadf196fe9a187479d1f6b03e9c8963029b365962c7bf57c22bb2e81a4"],"state_sha256":"dfb8528fb49b5940a70447e6b69ea5fb0ea4621b848edf73eda5da59c44b4606"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"euvRXQRSbaTS7EVKLLGydiRFuJqSM1xqO+jWvm3CTVJpyeU3NglO1/0k848WiJE5l7cDJGpELKZgea+17gr7Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T11:45:48.608043Z","bundle_sha256":"072005a6ac46ad92062ac5b317cff7179eff03d1f8e96b704ee7fa1a36653412"}}