{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:A644DVVPXIPUTYXXSMHEHPJFVN","short_pith_number":"pith:A644DVVP","canonical_record":{"source":{"id":"2205.14100","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-05-27T17:03:38Z","cross_cats_sorted":[],"title_canon_sha256":"cf3073fe32bc581efcd673538e9c0024a428c84711e54694153947e82cf28c08","abstract_canon_sha256":"b38248447620917caee1d84f52027cab7bf16e7e0f3ce663de54ab3d12e2b386"},"schema_version":"1.0"},"canonical_sha256":"07b9c1d6afba1f49e2f7930e43bd25ab712b7f3f168b7448033b846371c8e7de","source":{"kind":"arxiv","id":"2205.14100","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2205.14100","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"arxiv_version","alias_value":"2205.14100v5","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.14100","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"pith_short_12","alias_value":"A644DVVPXIPU","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"A644DVVPXIPUTYXX","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"A644DVVP","created_at":"2026-05-18T12:33:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:A644DVVPXIPUTYXXSMHEHPJFVN","target":"record","payload":{"canonical_record":{"source":{"id":"2205.14100","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-05-27T17:03:38Z","cross_cats_sorted":[],"title_canon_sha256":"cf3073fe32bc581efcd673538e9c0024a428c84711e54694153947e82cf28c08","abstract_canon_sha256":"b38248447620917caee1d84f52027cab7bf16e7e0f3ce663de54ab3d12e2b386"},"schema_version":"1.0"},"canonical_sha256":"07b9c1d6afba1f49e2f7930e43bd25ab712b7f3f168b7448033b846371c8e7de","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:46.604370Z","signature_b64":"IkWkjcRKTMnpIgQNhXFPBYEyXW31yhw+b0LXYB7xOx74TsDgF0X7BgUkmYau1sqp/TevFG+TmneooZHaaGa9BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"07b9c1d6afba1f49e2f7930e43bd25ab712b7f3f168b7448033b846371c8e7de","last_reissued_at":"2026-05-17T23:38:46.603914Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:46.603914Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2205.14100","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bPdUx8Tv2wJCnsMqC7YGU9FjLWBRhSHx/0J+v7KC7o244yML0grYI9LwwfJceN6MbN7v0tm2Jfagiv1V8VIOAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T00:07:38.582299Z"},"content_sha256":"67f58464b396b8434301a716ade938f5849d218a0f3863be1ac2f914cc93626f","schema_version":"1.0","event_id":"sha256:67f58464b396b8434301a716ade938f5849d218a0f3863be1ac2f914cc93626f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:A644DVVPXIPUTYXXSMHEHPJFVN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"GIT: A Generative Image-to-text Transformer for Vision and Language","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A simplified generative image-to-text transformer unifies vision-language tasks and sets new state-of-the-art results on 12 benchmarks.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ce Liu, Jianfeng Wang, Kevin Lin, Lijuan Wang, Linjie Li, Xiaowei Hu, Zhe Gan, Zhengyuan Yang, Zicheng Liu","submitted_at":"2022-05-27T17:03:38Z","abstract_excerpt":"In this paper, we design and train a Generative Image-to-text Transformer, GIT, to unify vision-language tasks such as image/video captioning and question answering. While generative models provide a consistent network architecture between pre-training and fine-tuning, existing work typically contains complex structures (uni/multi-modal encoder/decoder) and depends on external modules such as object detectors/taggers and optical character recognition (OCR). In GIT, we simplify the architecture as one image encoder and one text decoder under a single language modeling task. We also scale up the"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"our GIT establishes new state of the arts on 12 challenging benchmarks with a large margin. For instance, our model surpasses the human performance for the first time on TextCaps (138.2 vs. 125.5 in CIDEr).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That scaling pre-training data volume and model size with the single language-modeling objective on a simplified encoder-decoder architecture is sufficient to unify and exceed prior performance on diverse vision-language tasks without external modules.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"GIT achieves new state-of-the-art results on 12 vision-language benchmarks, including surpassing human performance on TextCaps, via a simplified single-encoder single-decoder transformer scaled on large pre-training data.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A simplified generative image-to-text transformer unifies vision-language tasks and sets new state-of-the-art results on 12 benchmarks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"928d5208e46638de55a6531dd1641b40cfef3d89fe1f922df4f23b45a519faf9"},"source":{"id":"2205.14100","kind":"arxiv","version":5},"verdict":{"id":"5d678f26-068e-4457-bde3-0f1d15dbf73b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T20:51:12.001343Z","strongest_claim":"our GIT establishes new state of the arts on 12 challenging benchmarks with a large margin. For instance, our model surpasses the human performance for the first time on TextCaps (138.2 vs. 125.5 in CIDEr).","one_line_summary":"GIT achieves new state-of-the-art results on 12 vision-language benchmarks, including surpassing human performance on TextCaps, via a simplified single-encoder single-decoder transformer scaled on large pre-training data.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That scaling pre-training data volume and model size with the single language-modeling objective on a simplified encoder-decoder architecture is sufficient to unify and exceed prior performance on diverse vision-language tasks without external modules.","pith_extraction_headline":"A simplified generative image-to-text transformer unifies vision-language tasks and sets new state-of-the-art results on 12 benchmarks."},"references":{"count":36,"sample":[{"doi":"","year":null,"title":"Flamingo: a Visual Language Model for Few-Shot Learning","work_id":"a110f764-38dc-41b2-a802-53744ecea1fc","ref_index":1,"cited_arxiv_id":"2204.14198","is_internal_anchor":true},{"doi":"","year":2005,"title":"Language Models are Few-Shot Learners","work_id":"214732c0-2edd-44a0-af9e-28184a2b8279","ref_index":2,"cited_arxiv_id":"2005.14165","is_internal_anchor":true},{"doi":"","year":2020,"title":"98 on it. (13) GITTextcaps: The station logo. (14) GITTextcaps: A sign with the numbers 3, 642, 039, 055 on it. (15) GITMJST: republic GITMJST: peperoni GITMJST: jewellery GITMJST: promod GITTextcaps:","work_id":"07785aad-1e39-4af8-b2dd-53e659abb407","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1930,"title":"Bluetooth Beats buy beats by dr dre 3 wireless bluetooth headphones","work_id":"d0c29255-f705-4e89-9cef-cf4a0fd67f24","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","ref_index":5,"cited_arxiv_id":"2204.02311","is_internal_anchor":true}],"resolved_work":36,"snapshot_sha256":"54ee5e73979e1ae7ef082c7caab1f6bee77db1b5335e73dae8399a81da75a253","internal_anchors":11},"formal_canon":{"evidence_count":2,"snapshot_sha256":"5a872e9dd4e456f8ecc6df9f3d0afc8502760ce19393a01071a6a8655407811c"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"5d678f26-068e-4457-bde3-0f1d15dbf73b"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4RGMrMdstTyIJ5GrvrszZCnZdCZk/I11OnX1EJEAZM3OQwQ0ziiwK/zXh695zYQeG3g3SVmKJusGgKfsZGg3DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T00:07:38.582866Z"},"content_sha256":"5ddc903d0915e79d32928f430e1dd6c72b4c3fab62bd6c47196190229590d32e","schema_version":"1.0","event_id":"sha256:5ddc903d0915e79d32928f430e1dd6c72b4c3fab62bd6c47196190229590d32e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/A644DVVPXIPUTYXXSMHEHPJFVN/bundle.json","state_url":"https://pith.science/pith/A644DVVPXIPUTYXXSMHEHPJFVN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/A644DVVPXIPUTYXXSMHEHPJFVN/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":"2026-06-09T00:07:38Z","links":{"resolver":"https://pith.science/pith/A644DVVPXIPUTYXXSMHEHPJFVN","bundle":"https://pith.science/pith/A644DVVPXIPUTYXXSMHEHPJFVN/bundle.json","state":"https://pith.science/pith/A644DVVPXIPUTYXXSMHEHPJFVN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/A644DVVPXIPUTYXXSMHEHPJFVN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:A644DVVPXIPUTYXXSMHEHPJFVN","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":"b38248447620917caee1d84f52027cab7bf16e7e0f3ce663de54ab3d12e2b386","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-05-27T17:03:38Z","title_canon_sha256":"cf3073fe32bc581efcd673538e9c0024a428c84711e54694153947e82cf28c08"},"schema_version":"1.0","source":{"id":"2205.14100","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2205.14100","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"arxiv_version","alias_value":"2205.14100v5","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.14100","created_at":"2026-05-17T23:38:46Z"},{"alias_kind":"pith_short_12","alias_value":"A644DVVPXIPU","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"A644DVVPXIPUTYXX","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"A644DVVP","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:5ddc903d0915e79d32928f430e1dd6c72b4c3fab62bd6c47196190229590d32e","target":"graph","created_at":"2026-05-17T23:38:46Z","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":"our GIT establishes new state of the arts on 12 challenging benchmarks with a large margin. For instance, our model surpasses the human performance for the first time on TextCaps (138.2 vs. 125.5 in CIDEr)."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That scaling pre-training data volume and model size with the single language-modeling objective on a simplified encoder-decoder architecture is sufficient to unify and exceed prior performance on diverse vision-language tasks without external modules."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"GIT achieves new state-of-the-art results on 12 vision-language benchmarks, including surpassing human performance on TextCaps, via a simplified single-encoder single-decoder transformer scaled on large pre-training data."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A simplified generative image-to-text transformer unifies vision-language tasks and sets new state-of-the-art results on 12 benchmarks."}],"snapshot_sha256":"928d5208e46638de55a6531dd1641b40cfef3d89fe1f922df4f23b45a519faf9"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"5a872e9dd4e456f8ecc6df9f3d0afc8502760ce19393a01071a6a8655407811c"},"paper":{"abstract_excerpt":"In this paper, we design and train a Generative Image-to-text Transformer, GIT, to unify vision-language tasks such as image/video captioning and question answering. While generative models provide a consistent network architecture between pre-training and fine-tuning, existing work typically contains complex structures (uni/multi-modal encoder/decoder) and depends on external modules such as object detectors/taggers and optical character recognition (OCR). In GIT, we simplify the architecture as one image encoder and one text decoder under a single language modeling task. We also scale up the","authors_text":"Ce Liu, Jianfeng Wang, Kevin Lin, Lijuan Wang, Linjie Li, Xiaowei Hu, Zhe Gan, Zhengyuan Yang, Zicheng Liu","cross_cats":[],"headline":"A simplified generative image-to-text transformer unifies vision-language tasks and sets new state-of-the-art results on 12 benchmarks.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-05-27T17:03:38Z","title":"GIT: A Generative Image-to-text Transformer for Vision and Language"},"references":{"count":36,"internal_anchors":11,"resolved_work":36,"sample":[{"cited_arxiv_id":"2204.14198","doi":"","is_internal_anchor":true,"ref_index":1,"title":"Flamingo: a Visual Language Model for Few-Shot Learning","work_id":"a110f764-38dc-41b2-a802-53744ecea1fc","year":null},{"cited_arxiv_id":"2005.14165","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Language Models are Few-Shot Learners","work_id":"214732c0-2edd-44a0-af9e-28184a2b8279","year":2005},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"98 on it. (13) GITTextcaps: The station logo. (14) GITTextcaps: A sign with the numbers 3, 642, 039, 055 on it. (15) GITMJST: republic GITMJST: peperoni GITMJST: jewellery GITMJST: promod GITTextcaps:","work_id":"07785aad-1e39-4af8-b2dd-53e659abb407","year":2020},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Bluetooth Beats buy beats by dr dre 3 wireless bluetooth headphones","work_id":"d0c29255-f705-4e89-9cef-cf4a0fd67f24","year":1930},{"cited_arxiv_id":"2204.02311","doi":"","is_internal_anchor":true,"ref_index":5,"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","year":null}],"snapshot_sha256":"54ee5e73979e1ae7ef082c7caab1f6bee77db1b5335e73dae8399a81da75a253"},"source":{"id":"2205.14100","kind":"arxiv","version":5},"verdict":{"created_at":"2026-05-16T20:51:12.001343Z","id":"5d678f26-068e-4457-bde3-0f1d15dbf73b","model_set":{"reader":"grok-4.3"},"one_line_summary":"GIT achieves new state-of-the-art results on 12 vision-language benchmarks, including surpassing human performance on TextCaps, via a simplified single-encoder single-decoder transformer scaled on large pre-training data.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A simplified generative image-to-text transformer unifies vision-language tasks and sets new state-of-the-art results on 12 benchmarks.","strongest_claim":"our GIT establishes new state of the arts on 12 challenging benchmarks with a large margin. For instance, our model surpasses the human performance for the first time on TextCaps (138.2 vs. 125.5 in CIDEr).","weakest_assumption":"That scaling pre-training data volume and model size with the single language-modeling objective on a simplified encoder-decoder architecture is sufficient to unify and exceed prior performance on diverse vision-language tasks without external modules."}},"verdict_id":"5d678f26-068e-4457-bde3-0f1d15dbf73b"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:67f58464b396b8434301a716ade938f5849d218a0f3863be1ac2f914cc93626f","target":"record","created_at":"2026-05-17T23:38:46Z","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":"b38248447620917caee1d84f52027cab7bf16e7e0f3ce663de54ab3d12e2b386","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-05-27T17:03:38Z","title_canon_sha256":"cf3073fe32bc581efcd673538e9c0024a428c84711e54694153947e82cf28c08"},"schema_version":"1.0","source":{"id":"2205.14100","kind":"arxiv","version":5}},"canonical_sha256":"07b9c1d6afba1f49e2f7930e43bd25ab712b7f3f168b7448033b846371c8e7de","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"07b9c1d6afba1f49e2f7930e43bd25ab712b7f3f168b7448033b846371c8e7de","first_computed_at":"2026-05-17T23:38:46.603914Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:46.603914Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"IkWkjcRKTMnpIgQNhXFPBYEyXW31yhw+b0LXYB7xOx74TsDgF0X7BgUkmYau1sqp/TevFG+TmneooZHaaGa9BA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:46.604370Z","signed_message":"canonical_sha256_bytes"},"source_id":"2205.14100","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:67f58464b396b8434301a716ade938f5849d218a0f3863be1ac2f914cc93626f","sha256:5ddc903d0915e79d32928f430e1dd6c72b4c3fab62bd6c47196190229590d32e"],"state_sha256":"98c8fb3a67ef3d7a344de068a904c0e41eae4f4b081ec4acedde6229ec60cea8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vT9N9xLYpsKPQ6rxy1sDrH4uu0jQqlCKj7UpmIa18BSwHZ9f09Wl8i+GFCqdEWNa+1BHlHKJiM3XHp8y0uCCCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-09T00:07:38.585442Z","bundle_sha256":"3b25c030d572d03b53ffe66f8a4f99ef453954c375bea1cf7f1e5fe1859a0e71"}}