{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:4SOQHN6WTX5XVVLWJQWKF57RTI","short_pith_number":"pith:4SOQHN6W","canonical_record":{"source":{"id":"2308.01390","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-08-02T19:10:23Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"63c2f0843e42723e3598388b6146c56e7fbea819e120db01a141efebbe7c2583","abstract_canon_sha256":"1e893f201ce5c83c112cc1fd0d64167708aebb6e3f48e085afb530af8d87cfaf"},"schema_version":"1.0"},"canonical_sha256":"e49d03b7d69dfb7ad5764c2ca2f7f19a12eb8486bb5217b1d952e72685c51805","source":{"kind":"arxiv","id":"2308.01390","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2308.01390","created_at":"2026-05-18T02:47:22Z"},{"alias_kind":"arxiv_version","alias_value":"2308.01390v2","created_at":"2026-05-18T02:47:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.01390","created_at":"2026-05-18T02:47:22Z"},{"alias_kind":"pith_short_12","alias_value":"4SOQHN6WTX5X","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"4SOQHN6WTX5XVVLW","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"4SOQHN6W","created_at":"2026-05-18T12:33:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:4SOQHN6WTX5XVVLWJQWKF57RTI","target":"record","payload":{"canonical_record":{"source":{"id":"2308.01390","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-08-02T19:10:23Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"63c2f0843e42723e3598388b6146c56e7fbea819e120db01a141efebbe7c2583","abstract_canon_sha256":"1e893f201ce5c83c112cc1fd0d64167708aebb6e3f48e085afb530af8d87cfaf"},"schema_version":"1.0"},"canonical_sha256":"e49d03b7d69dfb7ad5764c2ca2f7f19a12eb8486bb5217b1d952e72685c51805","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:47:22.657533Z","signature_b64":"Axp2UZaECZg2EWekymHfHLE1IUa6FzHaIz3f7P5QerPw3Z4xpx3mj+3F8FcoFimd/SXqZovYil97F3+6ndq9BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e49d03b7d69dfb7ad5764c2ca2f7f19a12eb8486bb5217b1d952e72685c51805","last_reissued_at":"2026-05-18T02:47:22.656989Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:47:22.656989Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2308.01390","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":"2026-05-18T02:47:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EdgtynNObsXlTQlL0EDqUWDsL8911HtI58mGNjx3Do9tYueLEjgoxTWJ6mBDhwtaFLRnYnBgmGhWkEZxEVp8DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T22:10:08.234485Z"},"content_sha256":"c19297204fda17baf2c31b1ffd939260998cf2ce1bb2b7015cdd29b132e88828","schema_version":"1.0","event_id":"sha256:c19297204fda17baf2c31b1ffd939260998cf2ce1bb2b7015cdd29b132e88828"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:4SOQHN6WTX5XVVLWJQWKF57RTI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"OpenFlamingo: An Open-Source Framework for Training Large Autoregressive Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"OpenFlamingo delivers open-source vision-language models that reach 80-89 percent of Flamingo performance across seven datasets.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Anas Awadalla, Gabriel Ilharco, Irena Gao, Jack Hessel, Jenia Jitsev, Josh Gardner, Kalyani Marathe, Ludwig Schmidt, Mitchell Wortsman, Pang Wei Koh, Samir Gadre, Shiori Sagawa, Simon Kornblith, Wanrong Zhu, Yonatan Bitton, Yusuf Hanafy","submitted_at":"2023-08-02T19:10:23Z","abstract_excerpt":"We introduce OpenFlamingo, a family of autoregressive vision-language models ranging from 3B to 9B parameters. OpenFlamingo is an ongoing effort to produce an open-source replication of DeepMind's Flamingo models. On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance. This technical report describes our models, training data, hyperparameters, and evaluation suite. We share our models and code at https://github.com/mlfoundations/open_flamingo."},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the reported performance numbers were obtained under comparable evaluation conditions to the original Flamingo models and that the shared code and data suffice for independent reproduction of those numbers.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OpenFlamingo provides open-source autoregressive vision-language models that achieve 80-89% of Flamingo performance on seven vision-language datasets.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"OpenFlamingo delivers open-source vision-language models that reach 80-89 percent of Flamingo performance across seven datasets.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"417279ddea6a62d394a44c5530fe8cbfc631662c011447a41a1d28fad5eb7fcb"},"source":{"id":"2308.01390","kind":"arxiv","version":2},"verdict":{"id":"de2859ba-fef5-4cc6-8227-9b7988f6e3da","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T01:47:24.782890Z","strongest_claim":"On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance.","one_line_summary":"OpenFlamingo provides open-source autoregressive vision-language models that achieve 80-89% of Flamingo performance on seven vision-language datasets.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the reported performance numbers were obtained under comparable evaluation conditions to the original Flamingo models and that the shared code and data suffice for independent reproduction of those numbers.","pith_extraction_headline":"OpenFlamingo delivers open-source vision-language models that reach 80-89 percent of Flamingo performance across seven datasets."},"references":{"count":48,"sample":[{"doi":"","year":2022,"title":"Cm3: A causal masked multimodal model of the internet","work_id":"a4a6d3b6-13f5-437f-8081-765dd23198b9","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Lawrence Zitnick, Devi Parikh, and Dhruv Batra","work_id":"4fff8ab6-adfb-44b1-ab5a-541475d839a9","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Flamingo: a visual language model for few-shot learning","work_id":"268774f4-076c-46bf-b030-8ea92961a8fb","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Clip retrieval: Easily compute clip embeddings and build a clip re- trieval system with them","work_id":"f154ba7d-a04e-4fc1-845d-b60395a8be8e","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"On the Opportunities and Risks of Foundation Models","work_id":"a18039e9-928d-47c9-a836-32656a71bf71","ref_index":5,"cited_arxiv_id":"2108.07258","is_internal_anchor":true}],"resolved_work":48,"snapshot_sha256":"bbf5951173c887d91af0e512392928bb1900a328c995e7f058fec4299dd8a764","internal_anchors":13},"formal_canon":{"evidence_count":2,"snapshot_sha256":"8b677cd132d0519f4b07c8e460c2d268a550f5746f8e8c5e1dfa9dbf7094076e"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"de2859ba-fef5-4cc6-8227-9b7988f6e3da"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:47:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yGOvpI6knDUdCO7WHwMfoMjh1ZS9ZwnSLHMor5Lsi/raZzKCjRVM3SNtl6gY9nQU6KqRW+ioWqZqh7iSP034Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T22:10:08.235470Z"},"content_sha256":"7e9d3d95119660eb8c88a24dff250f9356d548998e4145c71786483e8503362a","schema_version":"1.0","event_id":"sha256:7e9d3d95119660eb8c88a24dff250f9356d548998e4145c71786483e8503362a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4SOQHN6WTX5XVVLWJQWKF57RTI/bundle.json","state_url":"https://pith.science/pith/4SOQHN6WTX5XVVLWJQWKF57RTI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4SOQHN6WTX5XVVLWJQWKF57RTI/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":"2026-05-25T22:10:08Z","links":{"resolver":"https://pith.science/pith/4SOQHN6WTX5XVVLWJQWKF57RTI","bundle":"https://pith.science/pith/4SOQHN6WTX5XVVLWJQWKF57RTI/bundle.json","state":"https://pith.science/pith/4SOQHN6WTX5XVVLWJQWKF57RTI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4SOQHN6WTX5XVVLWJQWKF57RTI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:4SOQHN6WTX5XVVLWJQWKF57RTI","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":"1e893f201ce5c83c112cc1fd0d64167708aebb6e3f48e085afb530af8d87cfaf","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-08-02T19:10:23Z","title_canon_sha256":"63c2f0843e42723e3598388b6146c56e7fbea819e120db01a141efebbe7c2583"},"schema_version":"1.0","source":{"id":"2308.01390","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2308.01390","created_at":"2026-05-18T02:47:22Z"},{"alias_kind":"arxiv_version","alias_value":"2308.01390v2","created_at":"2026-05-18T02:47:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.01390","created_at":"2026-05-18T02:47:22Z"},{"alias_kind":"pith_short_12","alias_value":"4SOQHN6WTX5X","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"4SOQHN6WTX5XVVLW","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"4SOQHN6W","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:7e9d3d95119660eb8c88a24dff250f9356d548998e4145c71786483e8503362a","target":"graph","created_at":"2026-05-18T02:47:22Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the reported performance numbers were obtained under comparable evaluation conditions to the original Flamingo models and that the shared code and data suffice for independent reproduction of those numbers."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"OpenFlamingo provides open-source autoregressive vision-language models that achieve 80-89% of Flamingo performance on seven vision-language datasets."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"OpenFlamingo delivers open-source vision-language models that reach 80-89 percent of Flamingo performance across seven datasets."}],"snapshot_sha256":"417279ddea6a62d394a44c5530fe8cbfc631662c011447a41a1d28fad5eb7fcb"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"8b677cd132d0519f4b07c8e460c2d268a550f5746f8e8c5e1dfa9dbf7094076e"},"paper":{"abstract_excerpt":"We introduce OpenFlamingo, a family of autoregressive vision-language models ranging from 3B to 9B parameters. OpenFlamingo is an ongoing effort to produce an open-source replication of DeepMind's Flamingo models. On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance. This technical report describes our models, training data, hyperparameters, and evaluation suite. We share our models and code at https://github.com/mlfoundations/open_flamingo.","authors_text":"Anas Awadalla, Gabriel Ilharco, Irena Gao, Jack Hessel, Jenia Jitsev, Josh Gardner, Kalyani Marathe, Ludwig Schmidt, Mitchell Wortsman, Pang Wei Koh, Samir Gadre, Shiori Sagawa, Simon Kornblith, Wanrong Zhu, Yonatan Bitton, Yusuf Hanafy","cross_cats":["cs.AI","cs.LG"],"headline":"OpenFlamingo delivers open-source vision-language models that reach 80-89 percent of Flamingo performance across seven datasets.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-08-02T19:10:23Z","title":"OpenFlamingo: An Open-Source Framework for Training Large Autoregressive Vision-Language Models"},"references":{"count":48,"internal_anchors":13,"resolved_work":48,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Cm3: A causal masked multimodal model of the internet","work_id":"a4a6d3b6-13f5-437f-8081-765dd23198b9","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Lawrence Zitnick, Devi Parikh, and Dhruv Batra","work_id":"4fff8ab6-adfb-44b1-ab5a-541475d839a9","year":2015},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Flamingo: a visual language model for few-shot learning","work_id":"268774f4-076c-46bf-b030-8ea92961a8fb","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Clip retrieval: Easily compute clip embeddings and build a clip re- trieval system with them","work_id":"f154ba7d-a04e-4fc1-845d-b60395a8be8e","year":2022},{"cited_arxiv_id":"2108.07258","doi":"","is_internal_anchor":true,"ref_index":5,"title":"On the Opportunities and Risks of Foundation Models","work_id":"a18039e9-928d-47c9-a836-32656a71bf71","year":2021}],"snapshot_sha256":"bbf5951173c887d91af0e512392928bb1900a328c995e7f058fec4299dd8a764"},"source":{"id":"2308.01390","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-14T01:47:24.782890Z","id":"de2859ba-fef5-4cc6-8227-9b7988f6e3da","model_set":{"reader":"grok-4.3"},"one_line_summary":"OpenFlamingo provides open-source autoregressive vision-language models that achieve 80-89% of Flamingo performance on seven vision-language datasets.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"OpenFlamingo delivers open-source vision-language models that reach 80-89 percent of Flamingo performance across seven datasets.","strongest_claim":"On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance.","weakest_assumption":"That the reported performance numbers were obtained under comparable evaluation conditions to the original Flamingo models and that the shared code and data suffice for independent reproduction of those numbers."}},"verdict_id":"de2859ba-fef5-4cc6-8227-9b7988f6e3da"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c19297204fda17baf2c31b1ffd939260998cf2ce1bb2b7015cdd29b132e88828","target":"record","created_at":"2026-05-18T02:47:22Z","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":"1e893f201ce5c83c112cc1fd0d64167708aebb6e3f48e085afb530af8d87cfaf","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-08-02T19:10:23Z","title_canon_sha256":"63c2f0843e42723e3598388b6146c56e7fbea819e120db01a141efebbe7c2583"},"schema_version":"1.0","source":{"id":"2308.01390","kind":"arxiv","version":2}},"canonical_sha256":"e49d03b7d69dfb7ad5764c2ca2f7f19a12eb8486bb5217b1d952e72685c51805","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e49d03b7d69dfb7ad5764c2ca2f7f19a12eb8486bb5217b1d952e72685c51805","first_computed_at":"2026-05-18T02:47:22.656989Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:47:22.656989Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Axp2UZaECZg2EWekymHfHLE1IUa6FzHaIz3f7P5QerPw3Z4xpx3mj+3F8FcoFimd/SXqZovYil97F3+6ndq9BA==","signature_status":"signed_v1","signed_at":"2026-05-18T02:47:22.657533Z","signed_message":"canonical_sha256_bytes"},"source_id":"2308.01390","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c19297204fda17baf2c31b1ffd939260998cf2ce1bb2b7015cdd29b132e88828","sha256:7e9d3d95119660eb8c88a24dff250f9356d548998e4145c71786483e8503362a"],"state_sha256":"7e0be3bc60470c9c906a526f05860d0e85e689d431082dc73ea69562b7a25860"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Wqizv3nd59+ojDZbn2gdQ7F096CJx9V6DXqDEK0xy933rAz2uMHvZoOsTfsvyXgYHEz47gMTejQeuO0mTozEBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T22:10:08.239603Z","bundle_sha256":"39fbf372ae47ba3c370f4549e5d6b452ffdb229182167d44a5385d5b360af2bf"}}