{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2017:DJZTI7JZ3NHRONNFBRALVVK22W","short_pith_number":"pith:DJZTI7JZ","canonical_record":{"source":{"id":"1704.04861","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2017-04-17T03:57:34Z","cross_cats_sorted":[],"title_canon_sha256":"9433d332ad8cd595cec802d044a8d6e834dc3ccd7a579d82cb7500f82afe65ae","abstract_canon_sha256":"4cf0e09b9fc6a52e5cf29774cb99ef621cc601b47d46670d6189c24c70bb9cfc"},"schema_version":"1.0"},"canonical_sha256":"1a73347d39db4f1735a50c40bad55ad5b0ef9f800f0624227bf8636789982b2d","source":{"kind":"arxiv","id":"1704.04861","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1704.04861","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"arxiv_version","alias_value":"1704.04861v1","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1704.04861","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"pith_short_12","alias_value":"DJZTI7JZ3NHR","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"pith_short_16","alias_value":"DJZTI7JZ3NHRONNF","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"pith_short_8","alias_value":"DJZTI7JZ","created_at":"2026-07-04T21:53:30Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2017:DJZTI7JZ3NHRONNFBRALVVK22W","target":"record","payload":{"canonical_record":{"source":{"id":"1704.04861","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2017-04-17T03:57:34Z","cross_cats_sorted":[],"title_canon_sha256":"9433d332ad8cd595cec802d044a8d6e834dc3ccd7a579d82cb7500f82afe65ae","abstract_canon_sha256":"4cf0e09b9fc6a52e5cf29774cb99ef621cc601b47d46670d6189c24c70bb9cfc"},"schema_version":"1.0"},"canonical_sha256":"1a73347d39db4f1735a50c40bad55ad5b0ef9f800f0624227bf8636789982b2d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-04T21:53:30.683675Z","signature_b64":"CDZNwiPpHwmy0hL9o1OPaGqPpoLkJ6u3nF8q8H6Gzr6JXIC5tLYpuN2xRW9ec7XrlvvyTIWvLdA7alnn8ITNAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a73347d39db4f1735a50c40bad55ad5b0ef9f800f0624227bf8636789982b2d","last_reissued_at":"2026-07-04T21:53:30.683169Z","signature_status":"signed_v1","first_computed_at":"2026-07-04T21:53:30.683169Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1704.04861","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":"2026-07-04T21:53:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PvWJ24+ewxXId4iOL82Cvw0M8f+VcwqdXJcnHFD6TzeHNqT8+EviIYlAm+s2EWVcFcMqE7bKJyfC4FJOXYoRDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T15:19:51.375886Z"},"content_sha256":"85a8eca9150fdb19e5a7628df2dd9c46970c020a08e49edaf93df802f05eda95","schema_version":"1.0","event_id":"sha256:85a8eca9150fdb19e5a7628df2dd9c46970c020a08e49edaf93df802f05eda95"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2017:DJZTI7JZ3NHRONNFBRALVVK22W","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MobileNets use depth-wise separable convolutions and two global scaling hyperparameters to build lightweight networks that trade latency for accuracy on mobile devices.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Andrew G. Howard, Bo Chen, Dmitry Kalenichenko, Hartwig Adam, Marco Andreetto, Menglong Zhu, Tobias Weyand, Weijun Wang","submitted_at":"2017-04-17T03:57:34Z","abstract_excerpt":"We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy. These hyper-parameters allow the model builder to choose the right sized model for their application based on the constraints of the problem. We present extensive experiments on resource and accuracy tradeoffs and show strong performance compared to othe"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That depthwise separable convolutions plus uniform width and resolution scaling preserve enough representational power across the full range of target tasks and hardware constraints without needing task-specific redesign.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MobileNets introduce depthwise separable convolutions plus width and resolution multipliers to produce efficient CNNs that trade off latency and accuracy for mobile and embedded vision applications.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MobileNets use depth-wise separable convolutions and two global scaling hyperparameters to build lightweight networks that trade latency for accuracy on mobile devices.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"515e59a5a903991d85c381d26f2cf1bada12eeb68dae08175c763d8b0641ad9e"},"source":{"id":"1704.04861","kind":"arxiv","version":1},"verdict":{"id":"f6904dc0-61b8-4ab2-874a-d10a1986bd2a","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-11T02:45:24.717356Z","strongest_claim":"We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy.","one_line_summary":"MobileNets introduce depthwise separable convolutions plus width and resolution multipliers to produce efficient CNNs that trade off latency and accuracy for mobile and embedded vision applications.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That depthwise separable convolutions plus uniform width and resolution scaling preserve enough representational power across the full range of target tasks and hardware constraints without needing task-specific redesign.","pith_extraction_headline":"MobileNets use depth-wise separable convolutions and two global scaling hyperparameters to build lightweight networks that trade latency for accuracy on mobile devices."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1704.04861/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":37,"sample":[{"doi":"","year":2015,"title":"M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro, G. S. Corrado, A. Davis, J. Dean, M. Devin, et al. Tensorflow: Large-scale machine learning on heterogeneous systems, 2015. Software avail","work_id":"edc51558-2aac-4595-b42a-5fb545be7fb3","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Caffe model zoo","work_id":"837929ac-c926-488b-8638-8b896885021c","ref_index":2,"cited_arxiv_id":"1504.04788","is_internal_anchor":false},{"doi":"","year":2016,"title":"Xception: Deep Learning with Depthwise Separable Convolutions","work_id":"5e38a2cc-82da-4330-875e-a77799981f0b","ref_index":3,"cited_arxiv_id":"1610.02357","is_internal_anchor":false},{"doi":"","year":2014,"title":"Training deep neural networks with low precision multiplications","work_id":"8c21a31d-1633-44c7-a302-8f9242620270","ref_index":4,"cited_arxiv_id":"1412.7024","is_internal_anchor":false},{"doi":"","year":2015,"title":"Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding","work_id":"d79f6a26-2d92-4f7c-aea0-8aabf3b668c8","ref_index":5,"cited_arxiv_id":"1510.00149","is_internal_anchor":true}],"resolved_work":37,"snapshot_sha256":"0ad3edcef7680bb56c47748d30c6fefdbf322c5a4bd35cd3cab438c378261178","internal_anchors":8},"formal_canon":{"evidence_count":2,"snapshot_sha256":"de683b7f5afdf6d8bfce39d95b79582aded465bbba9a205281ca1e1ca0433380"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"f6904dc0-61b8-4ab2-874a-d10a1986bd2a"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-04T21:53:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"woozZLFBzWgj7y2MLFSqWCBcHuBLzxwKnnEzyWa+qw32N6SuYHdyFxfmBPgk3q0sycRlikNk7bppIRv/OIXZDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T15:19:51.377747Z"},"content_sha256":"9934f20c70ef0c250bdb83d9141c226d98615ce1e5e6e0bdbddf4f1d5c5b4113","schema_version":"1.0","event_id":"sha256:9934f20c70ef0c250bdb83d9141c226d98615ce1e5e6e0bdbddf4f1d5c5b4113"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DJZTI7JZ3NHRONNFBRALVVK22W/bundle.json","state_url":"https://pith.science/pith/DJZTI7JZ3NHRONNFBRALVVK22W/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DJZTI7JZ3NHRONNFBRALVVK22W/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":"2026-08-14T15:19:51Z","links":{"resolver":"https://pith.science/pith/DJZTI7JZ3NHRONNFBRALVVK22W","bundle":"https://pith.science/pith/DJZTI7JZ3NHRONNFBRALVVK22W/bundle.json","state":"https://pith.science/pith/DJZTI7JZ3NHRONNFBRALVVK22W/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DJZTI7JZ3NHRONNFBRALVVK22W/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2017:DJZTI7JZ3NHRONNFBRALVVK22W","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":"4cf0e09b9fc6a52e5cf29774cb99ef621cc601b47d46670d6189c24c70bb9cfc","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2017-04-17T03:57:34Z","title_canon_sha256":"9433d332ad8cd595cec802d044a8d6e834dc3ccd7a579d82cb7500f82afe65ae"},"schema_version":"1.0","source":{"id":"1704.04861","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1704.04861","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"arxiv_version","alias_value":"1704.04861v1","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1704.04861","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"pith_short_12","alias_value":"DJZTI7JZ3NHR","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"pith_short_16","alias_value":"DJZTI7JZ3NHRONNF","created_at":"2026-07-04T21:53:30Z"},{"alias_kind":"pith_short_8","alias_value":"DJZTI7JZ","created_at":"2026-07-04T21:53:30Z"}],"graph_snapshots":[{"event_id":"sha256:9934f20c70ef0c250bdb83d9141c226d98615ce1e5e6e0bdbddf4f1d5c5b4113","target":"graph","created_at":"2026-07-04T21:53:30Z","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":"We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That depthwise separable convolutions plus uniform width and resolution scaling preserve enough representational power across the full range of target tasks and hardware constraints without needing task-specific redesign."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"MobileNets introduce depthwise separable convolutions plus width and resolution multipliers to produce efficient CNNs that trade off latency and accuracy for mobile and embedded vision applications."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"MobileNets use depth-wise separable convolutions and two global scaling hyperparameters to build lightweight networks that trade latency for accuracy on mobile devices."}],"snapshot_sha256":"515e59a5a903991d85c381d26f2cf1bada12eeb68dae08175c763d8b0641ad9e"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"de683b7f5afdf6d8bfce39d95b79582aded465bbba9a205281ca1e1ca0433380"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/1704.04861/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy. These hyper-parameters allow the model builder to choose the right sized model for their application based on the constraints of the problem. We present extensive experiments on resource and accuracy tradeoffs and show strong performance compared to othe","authors_text":"Andrew G. Howard, Bo Chen, Dmitry Kalenichenko, Hartwig Adam, Marco Andreetto, Menglong Zhu, Tobias Weyand, Weijun Wang","cross_cats":[],"headline":"MobileNets use depth-wise separable convolutions and two global scaling hyperparameters to build lightweight networks that trade latency for accuracy on mobile devices.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2017-04-17T03:57:34Z","title":"MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications"},"references":{"count":37,"internal_anchors":8,"resolved_work":37,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro, G. S. Corrado, A. Davis, J. Dean, M. Devin, et al. Tensorflow: Large-scale machine learning on heterogeneous systems, 2015. Software avail","work_id":"edc51558-2aac-4595-b42a-5fb545be7fb3","year":2015},{"cited_arxiv_id":"1504.04788","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Caffe model zoo","work_id":"837929ac-c926-488b-8638-8b896885021c","year":2015},{"cited_arxiv_id":"1610.02357","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Xception: Deep Learning with Depthwise Separable Convolutions","work_id":"5e38a2cc-82da-4330-875e-a77799981f0b","year":2016},{"cited_arxiv_id":"1412.7024","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Training deep neural networks with low precision multiplications","work_id":"8c21a31d-1633-44c7-a302-8f9242620270","year":2014},{"cited_arxiv_id":"1510.00149","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding","work_id":"d79f6a26-2d92-4f7c-aea0-8aabf3b668c8","year":2015}],"snapshot_sha256":"0ad3edcef7680bb56c47748d30c6fefdbf322c5a4bd35cd3cab438c378261178"},"source":{"id":"1704.04861","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-11T02:45:24.717356Z","id":"f6904dc0-61b8-4ab2-874a-d10a1986bd2a","model_set":{"reader":"grok-4.3"},"one_line_summary":"MobileNets introduce depthwise separable convolutions plus width and resolution multipliers to produce efficient CNNs that trade off latency and accuracy for mobile and embedded vision applications.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"MobileNets use depth-wise separable convolutions and two global scaling hyperparameters to build lightweight networks that trade latency for accuracy on mobile devices.","strongest_claim":"We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy.","weakest_assumption":"That depthwise separable convolutions plus uniform width and resolution scaling preserve enough representational power across the full range of target tasks and hardware constraints without needing task-specific redesign."}},"verdict_id":"f6904dc0-61b8-4ab2-874a-d10a1986bd2a"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:85a8eca9150fdb19e5a7628df2dd9c46970c020a08e49edaf93df802f05eda95","target":"record","created_at":"2026-07-04T21:53:30Z","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":"4cf0e09b9fc6a52e5cf29774cb99ef621cc601b47d46670d6189c24c70bb9cfc","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2017-04-17T03:57:34Z","title_canon_sha256":"9433d332ad8cd595cec802d044a8d6e834dc3ccd7a579d82cb7500f82afe65ae"},"schema_version":"1.0","source":{"id":"1704.04861","kind":"arxiv","version":1}},"canonical_sha256":"1a73347d39db4f1735a50c40bad55ad5b0ef9f800f0624227bf8636789982b2d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1a73347d39db4f1735a50c40bad55ad5b0ef9f800f0624227bf8636789982b2d","first_computed_at":"2026-07-04T21:53:30.683169Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-04T21:53:30.683169Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"CDZNwiPpHwmy0hL9o1OPaGqPpoLkJ6u3nF8q8H6Gzr6JXIC5tLYpuN2xRW9ec7XrlvvyTIWvLdA7alnn8ITNAA==","signature_status":"signed_v1","signed_at":"2026-07-04T21:53:30.683675Z","signed_message":"canonical_sha256_bytes"},"source_id":"1704.04861","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:85a8eca9150fdb19e5a7628df2dd9c46970c020a08e49edaf93df802f05eda95","sha256:9934f20c70ef0c250bdb83d9141c226d98615ce1e5e6e0bdbddf4f1d5c5b4113"],"state_sha256":"2d8670d4ed0e0ebb76b25ce70598919db9b877506f4cd086dc2c5e30675a3eec"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lKiZ8MPjmygXEGCUR0FN98EXbDpImitLpAWo+bqz06n6GI/t27q2jB/ehckwflKGhob6Zp9GSj4MgZWShlhZDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-14T15:19:51.385207Z","bundle_sha256":"a5ae5cae7338465ad6aedfc27f7b50bb168b5550ad952bae651b078b6c9df550"}}