V2.5 annotations, provenance and grouped partition

.gitattributes

@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+case-manifest.jsonl filter=lfs diff=lfs merge=lfs -text
+annotations/train.json filter=lfs diff=lfs merge=lfs -text

annotations/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43715e316fb2aee75c1fdfb518b6b1d5bad30c5046f43591cd039c2b4f985db7
+size 14052715

audit/exclusions.json

@@ -0,0 +1 @@
+{"conflicts": [{"pixel_sha256": "2f82b50a7887f71dd37f23bf1cdefd556516087dc3e2e7edef3f44f77db4cbd4", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "bc1ec76b23954648905731a21f310aaab0d40646e9437c012a043363f311d7b4"}, {"pixel_sha256": "2f82b50a7887f71dd37f23bf1cdefd556516087dc3e2e7edef3f44f77db4cbd4", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "049dbb25c2d430dd45b8845a454d319e3315807cdbb91923d49fccc11b79b156"}, {"pixel_sha256": "ae7c4727c5ccfe87d9bf7007e215bd285fbdadd365c2cc6c5c9387e988d5ed50", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "474ffa86db8f2e50a17624c0f62e7742ec42eac9aabc35ed0800d9eaeb93a9fc"}, {"pixel_sha256": "ae7c4727c5ccfe87d9bf7007e215bd285fbdadd365c2cc6c5c9387e988d5ed50", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "636ffdb054150ec4bcbec219d79cd93a18eb5583afcdb7d036478d5444b3cbab"}, {"pixel_sha256": "bda23e352c9c3550adbdf91d6331d6d9ee8a75d5736dc27d6d5d638442a45917", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "5bb5a81180cab2ede89a23021c028c650c0ea0a58acbdb6fa4972370f5d37684"}, {"pixel_sha256": "bda23e352c9c3550adbdf91d6331d6d9ee8a75d5736dc27d6d5d638442a45917", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "7c096ad50e4c153944f20fb30ef146981b6d424474ce22ac1b94e2bf967169c4"}, {"pixel_sha256": "e0aa186c20be7a698e85dd3c3a54bf9da4698ac360333c8319e997cabd04e00c", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "5c02ab3210cf131de49fd70b19e24118467fa93e85aba60ca451d427805c8440"}, {"pixel_sha256": "e0aa186c20be7a698e85dd3c3a54bf9da4698ac360333c8319e997cabd04e00c", "reason": "identical_decoded_pixels_conflicting_final_annotations", "sha256": "a9ec5ee667c28301a4976f65db3ceb02bfd45fc8caecccdbee01950f76c77fc5"}], "duplicates": [{"kept": "2cad30c1bbdca0512c5397f11014e85183f56085ffdbe0c8abd5a4064950db76", "reason": "identical_decoded_rgb", "removed": "c3ab8e15f4ad79706bc77bafc20fa3675b62179146718a9d9e04570f44c5ee8e"}]}

case-manifest.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:584291355b628b0e606e56f011b1bf9ed22a5d4d124d73fb51f8fa8eb4d9561d
+size 98348666

categories.json

@@ -0,0 +1 @@
+[{"id": 1, "name": "smoke_visible", "supercategory": "fire_smoke"}, {"id": 2, "name": "flame_visible", "supercategory": "fire_smoke"}]

manifest.json

@@ -0,0 +1 @@
+{"annotations": "One final COCO xywh list; negatives are empty. Source points remain provenance, not detection targets.", "case_manifest_sha256": "584291355b628b0e606e56f011b1bf9ed22a5d4d124d73fb51f8fa8eb4d9561d", "categories": [{"id": 1, "name": "smoke_visible", "supercategory": "fire_smoke"}, {"id": 2, "name": "flame_visible", "supercategory": "fire_smoke"}], "created_at": 1791016084.433933, "images": 38965, "objects": 50591, "pixel_policy": "Original encoded bytes; raw decoded pixels without EXIF orientation.", "ready": false, "schema": "fireviewer-ready-to-train-v2.5", "shards": [{"bytes": 100433314, "path": "data/train-00000.parquet", "rows": 224, "sha256": "968fb42d7ba118a7164cea24d6623baf468f5e396261e4f7e21c6f09517a143d", "split": "train"}, {"bytes": 100081103, "path": "data/train-00001.parquet", "rows": 144, "sha256": "7ce0a52558d0b3a48937a8e0edcdf8c3310eab7699812a5c5f3130fd192e2077", "split": "train"}, {"bytes": 94147676, "path": "data/train-00002.parquet", "rows": 156, "sha256": "1c19f4d4010e5b2d5a2ecc761188aa69bbde02de6e07da7375d279a0ecce15d3", "split": "train"}, {"bytes": 97888185, "path": "data/train-00003.parquet", "rows": 111, "sha256": "f41d1fbb9b90c8225a6eb36a90fd12206bede960916d982875044a6914622cce", "split": "train"}, {"bytes": 100300288, "path": "data/train-00004.parquet", "rows": 168, "sha256": "096b96dca27304bbe52898c24e8f16ea95cb309c029ac538627dbb76534bb101", "split": "train"}, {"bytes": 100021114, "path": "data/train-00005.parquet", "rows": 210, "sha256": "0fbb0e8e00b972699cee4863b3a2214122b765ce96915b759f94b1802c60e38c", "split": "train"}, {"bytes": 100337935, "path": "data/train-00006.parquet", "rows": 134, "sha256": "88a78bfedeeff28feb7986853c9aec95512691a8ea11f758e9a7332a8d3fc614", "split": "train"}, {"bytes": 98897830, "path": "data/train-00007.parquet", "rows": 224, "sha256": "544f25789e51654c691729301d7ae2cdc5f708621d562b8cff77b81471c208d5", "split": "train"}, {"bytes": 98001341, "path": "data/train-00008.parquet", "rows": 138, "sha256": "15712ae66516ee2a587d443dc85aa22365466faae09edc82883326824c79c4d4", "split": "train"}, {"bytes": 99773121, "path": "data/train-00009.parquet", "rows": 130, "sha256": "3042dee0bcb02f784076b3848eff0e384d269b91da9f1ae088a9736e9e2adccd", "split": "train"}, {"bytes": 98423019, "path": "data/train-00010.parquet", "rows": 173, "sha256": "9674aa0830bd3ef2fb3da160fa3100d1c333d3a1624db55cd0edd75bbb4294a7", "split": "train"}, {"bytes": 97661189, "path": "data/train-00011.parquet", "rows": 205, "sha256": "85e5ae1f3400b1e13bba232eed7fe91ac15c98dadf3fe6babfd851dcd04b5917", "split": "train"}, {"bytes": 86735920, "path": "data/train-00012.parquet", "rows": 114, "sha256": "2f8a3aa7d004a858121cdc609ba269c16462a2e9254ac5972e59887f249b18cc", "split": "train"}, {"bytes": 77902364, "path": "data/train-00013.parquet", "rows": 94, "sha256": "664cf4265ec777becca0530da90faac9f38de6c68fa3abd2f017af9e9dc7190c", "split": "train"}, {"bytes": 94566444, "path": "data/train-00014.parquet", "rows": 75, "sha256": "4621e728745f08c3c8c2d057fa0f95b7e71710f24d177988f94c2351b098147d", "split": "train"}, {"bytes": 98911606, "path": "data/train-00015.parquet", "rows": 218, "sha256": "9b12dfa282aa6cff323e8c950c2724287ab291b09b092e6ab75923322b8410ed", "split": "train"}, {"bytes": 99408642, "path": "data/train-00016.parquet", "rows": 176, "sha256": "0f63083eb0abb43c3baabb2beb05a3a22939a8607b5e8b535ae1066da09891ee", "split": "train"}, {"bytes": 97481872, "path": "data/train-00017.parquet", "rows": 124, "sha256": "30975f235b08660b85caf214bebd35f92a3ba29592d5b9b76d08588d41e36ca2", "split": "train"}, {"bytes": 99672024, "path": "data/train-00018.parquet", "rows": 211, "sha256": "24a68dd9a00b76d332bb40b4cbda45f8c46c751173587fc953e776ff93bb923b", "split": "train"}, {"bytes": 99335424, "path": "data/train-00019.parquet", "rows": 148, "sha256": "1d00b77d8f8c993f54ea691a88976aaf4a3d55d560f8383d2959b1964534068c", "split": "train"}, {"bytes": 99481678, "path": "data/train-00020.parquet", "rows": 143, "sha256": "e09b3166203b98cb15b89d6e6816db49cc7475b0f3b7304959d286f3baa6ce75", "split": "train"}, {"bytes": 99454089, "path": "data/train-00021.parquet", "rows": 237, "sha256": "1096fac8365196a244d1eefe647412e0940c5a397a2d8c5b63a9ae06cfabd6e7", "split": "train"}, {"bytes": 99479840, "path": "data/train-00022.parquet", "rows": 134, "sha256": "f1a67c2c2b1cac3d8d098843fea25584eeb64b42e7190b24a49ba8361a2fe62d", "split": "train"}, {"bytes": 95133886, "path": "data/train-00023.parquet", "rows": 143, "sha256": "e4d76f2b1168994210fa7db58f56a55893bcdd20c0caab6604e8d5b5141330c0", "split": "train"}, {"bytes": 100242119, "path": "data/train-00024.parquet", "rows": 89, "sha256": "32789b9d99c7100d9425440a94520dca6fa8d5386f7757946d63536de7d991f9", "split": "train"}, {"bytes": 82139979, "path": "data/train-00025.parquet", "rows": 124, "sha256": "1bb5ba0abcc444603347ac48e4d76bf1f16bf69614256226cb28393109d7276b", "split": "train"}, {"bytes": 83150546, "path": "data/train-00026.parquet", "rows": 126, "sha256": "d7fa47369962ad41740063b94489be1596f4249e5b8dadd9a3f3c9102808a7fd", "split": "train"}, {"bytes": 95410832, "path": "data/train-00027.parquet", "rows": 110, "sha256": "28cc2fc205ed0f719e12c51762461d38ffb6781cdc40569162d6b1066719da8e", "split": "train"}, {"bytes": 99708851, "path": "data/train-00028.parquet", "rows": 158, "sha256": "08405ad88a1b9fe799c41277bd8b5972b7d69ea21428085f4ed4f0bd1b691ff3", "split": "train"}, {"bytes": 96279635, "path": "data/train-00029.parquet", "rows": 159, "sha256": "a78527128d2a8049b075e3c3d0eb96e95bee33bbd880328efe53283de10f3ca8", "split": "train"}, {"bytes": 99532274, "path": "data/train-00030.parquet", "rows": 82, "sha256": "b457e9f5474ec5462ad267128a09676fa559e0866e266182b238b0c5aa31cbe0", "split": "train"}, {"bytes": 99211374, "path": "data/train-00031.parquet", "rows": 141, "sha256": "24bbdc43709c61ed28387668ff3a13cc5b73b1c8bf05b6c1f16df3e6dac3fe3e", "split": "train"}, {"bytes": 100086638, "path": "data/train-00032.parquet", "rows": 155, "sha256": "70d0287835461568ea72d722ca767bf2a497b2499cb9c9165cde5df4fadf2c8a", "split": "train"}, {"bytes": 90498333, "path": "data/train-00033.parquet", "rows": 101, "sha256": "5475b2c3d54e51dfa2dd842e20c676f21af1ab99d55f89d159018784fe8e58ee", "split": "train"}, {"bytes": 96969390, "path": "data/train-00034.parquet", "rows": 73, "sha256": "8d707de6a4f72df869225d46ffacb6076bb0331388ad1952574c04184cd9eecd", "split": "train"}, {"bytes": 85705412, "path": "data/train-00035.parquet", "rows": 110, "sha256": "f318ded521cbc666271de92d867e180757c1a0bde8465c2c2088d8da70ea0215", "split": "train"}, {"bytes": 99566656, "path": "data/train-00036.parquet", "rows": 44, "sha256": "28318decf2a72093e0dc886ec820d9193cc3c02503603d87faf94e8312a77fc8", "split": "train"}, {"bytes": 89836639, "path": "data/train-00037.parquet", "rows": 129, "sha256": "5df86724bfccdec479d508e3f19d5d3eca64fda2a47122bdb165b27e0960ea30", "split": "train"}, {"bytes": 98166184, "path": "data/train-00038.parquet", "rows": 222, "sha256": "5abfc6f53fa6656d29c162c9a3a4b830de2af5ee8a0695233b81d5be243f05f8", "split": "train"}, {"bytes": 82811978, "path": "data/train-00039.parquet", "rows": 112, "sha256": "b538186f6af61a59cb4fc71511a6dfb8e733baab0ded5a3a01d6e406f9fbec9f", "split": "train"}, {"bytes": 90137659, "path": "data/train-00040.parquet", "rows": 96, "sha256": "91e67cb2565a8abd21be9df849025d2d3e8a73e3ddf6558bdfb66f93a2a1f440", "split": "train"}, {"bytes": 92704178, "path": "data/train-00041.parquet", "rows": 103, "sha256": "fdd0dc7c323aaecc17399c88da90df04a904e8da465ff728436db126239ce34b", "split": "train"}, {"bytes": 91761271, "path": "data/train-00042.parquet", "rows": 140, "sha256": "fe1d900f571bd1699f8d140f09d60f61f76e563d69cb975fe23d8a2e0141d97b", "split": "train"}, {"bytes": 90497835, "path": "data/train-00043.parquet", "rows": 160, "sha256": "d841c8f2bec05106a7eb574f68375835ea42d8ce268b7d6b9fda7186a0d7b318", "split": "train"}, {"bytes": 71432820, "path": "data/train-00044.parquet", "rows": 104, "sha256": "e0f4223d079b52799d34fbb6ab76dfe9380cdd9c49eac3e9bc5cfb304a1bd747", "split": "train"}, {"bytes": 99893633, "path": "data/train-00045.parquet", "rows": 100, "sha256": "84306250268d1984539a404fce6f6080b59285b58bd2fa4c035ff13be4947f13", "split": "train"}, {"bytes": 71248303, "path": "data/train-00046.parquet", "rows": 102, "sha256": "9dc89911d5cbf6da46208935ebe2e97e4572b1eb22ebb71c9e479f17e295c44c", "split": "train"}, {"bytes": 84852890, "path": "data/train-00047.parquet", "rows": 136, "sha256": "88036331a4bd87b819f52c051ea8b5fe23462bc377c094def636de342480b666", "split": "train"}, {"bytes": 91129246, "path": "data/train-00048.parquet", "rows": 181, "sha256": "465ba0bd033d25201f100845f503dde59e8d0ae173334ae1e5c4f3f9de791073", "split": "train"}, {"bytes": 97157205, "path": "data/train-00049.parquet", "rows": 114, "sha256": "d863ed360ae637d3559be7db6440060623d26b613ac1c54c099dfee567cddfa0", "split": "train"}, {"bytes": 98255129, "path": "data/train-00050.parquet", "rows": 109, "sha256": "833bd58aa9c02828e12aa6b5bb4ce3035bb3f9185c948b2718ac241a0cc1dcd1", "split": "train"}, {"bytes": 99205111, "path": "data/train-00051.parquet", "rows": 84, "sha256": "276a6a9859889ea39059fad1a23a96a7a75090c227f50845a629c85fcac10a62", "split": "train"}, {"bytes": 92780876, "path": "data/train-00052.parquet", "rows": 120, "sha256": "ef68311c3b298ed55f92e9fe189c257017daec8b49abfd91f9f5f9e2fbe4ea88", "split": "train"}, {"bytes": 99768287, "path": "data/train-00053.parquet", "rows": 177, "sha256": "a83d082fd543281408bdf2b06b9618447407d677bcd1b9c3b6fcaf5ec1d4cc9a", "split": "train"}, {"bytes": 77971195, "path": "data/train-00054.parquet", "rows": 109, "sha256": "167eaa6e06e4cc596d8ba3b0ba180766f71daa94308794a4dbef16c798a1190c", "split": "train"}, {"bytes": 99612757, "path": "data/train-00055.parquet", "rows": 130, "sha256": "2b3339f271079cd69dc6f81c98750d168f75be6eaca4631a08d1e7c87689c460", "split": "train"}, {"bytes": 98551212, "path": "data/train-00056.parquet", "rows": 221, "sha256": "d3089f069dc3a557589d346f0fba6a09a5128a496a3277db2f7d9133de485640", "split": "train"}, {"bytes": 98882039, "path": "data/train-00057.parquet", "rows": 160, "sha256": "7c56c3a59c5dcc2b1a257e03e373e8aad12e3b6b04ba36712d03defd036f6554", "split": "train"}, {"bytes": 76178224, "path": "data/train-00058.parquet", "rows": 91, "sha256": "eee3c6f0675a1b7a6928e6531edd7dccb27830c1a420e56947fa9b865a9a68cf", "split": "train"}, {"bytes": 90987852, "path": "data/train-00059.parquet", "rows": 81, "sha256": "3cf54040ac5f6d8f99c6721b8aac6efa3191268ecad60134d3b40003e9114441", "split": "train"}, {"bytes": 82307459, "path": "data/train-00060.parquet", "rows": 131, "sha256": "149c89fa31e603c01d6055ccc4be94238310f83c4c0bfa6dd2cf80a304c8efd4", "split": "train"}, {"bytes": 99932911, "path": "data/train-00061.parquet", "rows": 78, "sha256": "67e41b34d89fa741d6d9d5a61d8b8d20adb1675bd90065be1394955ca479b3a4", "split": "train"}, {"bytes": 97929309, "path": "data/train-00062.parquet", "rows": 160, "sha256": "a253b377108e5bf3288d44a69c7f6e4e155b2d4c65c87144b7a32e2ac8dcd009", "split": "train"}, {"bytes": 82357461, "path": "data/train-00063.parquet", "rows": 73, "sha256": "1f577eeda6e6de93ea99b9d894d42e218ff737f0da18afe93cb5cbb9fa872799", "split": "train"}, {"bytes": 84433714, "path": "data/train-00064.parquet", "rows": 46, "sha256": "badb9ef30d1051fb6579b3d1764772fa507a5ec88cf203ef9f6688370daf7603", "split": "train"}, {"bytes": 100210392, "path": "data/train-00065.parquet", "rows": 96, "sha256": "9947308cfb8caee88ee5862ba65a8a53c58e87ca44705543317f541c88d75e41", "split": "train"}, {"bytes": 97253160, "path": "data/train-00066.parquet", "rows": 123, "sha256": "7e09f981b251bd09166af357b98d01895ff9adabcd4e75477fa95c3246d70c5c", "split": "train"}, {"bytes": 95539112, "path": "data/train-00067.parquet", "rows": 143, "sha256": "1007f3bd707ba45d05583606fa0944bc904126426474bb7a0b741e4ac2f0c57f", "split": "train"}, {"bytes": 86200170, "path": "data/train-00068.parquet", "rows": 156, "sha256": "7d2b95e236f8a856380461510bd083e59974b8ee9c25916b00469d5a2cf7ae89", "split": "train"}, {"bytes": 93241042, "path": "data/train-00069.parquet", "rows": 91, "sha256": "abd8e4e99fc423929517570dd8b3b20eb707d55f16216f1385a1969c8cb03a7e", "split": "train"}, {"bytes": 100290692, "path": "data/train-00070.parquet", "rows": 161, "sha256": "9a570d66d5939013380baea891b4ed572130a6dc5a205db90a8212ac2cc19c1e", "split": "train"}, {"bytes": 99385271, "path": "data/train-00071.parquet", "rows": 158, "sha256": "36f625f6953ebe26295a2ffa383b6d7c2e2a10c063376ed454c347a99cd98aed", "split": "train"}, {"bytes": 96888050, "path": "data/train-00072.parquet", "rows": 78, "sha256": "63b188c31b72758a11a328a915d30881206ea6d15110e474dcd20de2372d7318", "split": "train"}, {"bytes": 99773743, "path": "data/train-00073.parquet", "rows": 206, "sha256": "0ef7cc3334d0e066c358a938f014ba12bb25b337ec2a0bc18b188fbc13d0796c", "split": "train"}, {"bytes": 99645845, "path": "data/train-00074.parquet", "rows": 199, "sha256": "a9fec4dcce0384872fa5c16343d8903bb4e71570e8ce554947eadcac7c99842d", "split": "train"}, {"bytes": 99903994, "path": "data/train-00075.parquet", "rows": 163, "sha256": "ec72494e7f65504e343438dcaa06429392a9c5395628fc461b38599b77ab5f20", "split": "train"}, {"bytes": 93579171, "path": "data/train-00076.parquet", "rows": 100, "sha256": "ed3406d6c08bdf5a453840181b02a3597479e6fb28c98dfa3358e44e1b2a9135", "split": "train"}, {"bytes": 98996648, "path": "data/train-00077.parquet", "rows": 136, "sha256": "a742b184cc8c8273cdae09d0ed6469ea2678ac4b3eb52ce24318bbc7b246b1cb", "split": "train"}, {"bytes": 99562779, "path": "data/train-00078.parquet", "rows": 130, "sha256": "71dc9fa1310dc1fc39472fa82a07bcded998197a3789463efca0099bfa046002", "split": "train"}, {"bytes": 61708811, "path": "data/train-00079.parquet", "rows": 115, "sha256": "d9d4dfe73a0db3b3938ac633b0aff97980c09f877bb2247a9b106d0ea9855511", "split": "train"}, {"bytes": 100484606, "path": "data/train-00080.parquet", "rows": 113, "sha256": "14216621d087cfe92d5b011a5107d10d178974c5d247678d1b8144091bbb4788", "split": "train"}, {"bytes": 87433913, "path": "data/train-00081.parquet", "rows": 88, "sha256": "d2cafee05f889145a278989f54e27e191d3e7d1a6c4ee8cd871323ef7c4227a1", "split": "train"}, {"bytes": 97032201, "path": "data/train-00082.parquet", "rows": 184, "sha256": "6c6755607493c6dc9fbf92b5c7d3c7303bd8652964cd15f2f3666e013d7d38eb", "split": "train"}, {"bytes": 85793073, "path": "data/train-00083.parquet", "rows": 178, "sha256": "fba89d3d216ebddd7edc5f58efd4ba0d823c19ff978c1bb2f534e9f3642c00c0", "split": "train"}, {"bytes": 93845549, "path": "data/train-00084.parquet", "rows": 75, "sha256": "814a8b521dd445ef7ffa464445bce1b1e78d006be2dcbeed48fd5217af44eb86", "split": "train"}, {"bytes": 99384261, "path": "data/train-00085.parquet", "rows": 61, "sha256": "b9f8bb8ca1b190f4a9e1553196d71897756f44a2425f8fe58e39ffe8189afdc0", "split": "train"}, {"bytes": 99922663, "path": "data/train-00086.parquet", "rows": 169, "sha256": "36e27561b9012e5830bc79af1513f33fd05a4c9f4c7fa6e98b5e8aa1fa153dbc", "split": "train"}, {"bytes": 96390192, "path": "data/train-00087.parquet", "rows": 106, "sha256": "0783ff60c17e6fa5fff61b25b5bac6dc3bb8a2be257d0c8f10ae45d140873630", "split": "train"}, {"bytes": 98893445, "path": "data/train-00088.parquet", "rows": 142, "sha256": "70e45dcde25b3a54d4e1b7a130402db0e98e835216dc4d3d2ef2f87937c335d6", "split": "train"}, {"bytes": 100361148, "path": "data/train-00089.parquet", "rows": 156, "sha256": "99390bc35139e98b271b01bacac2437cbf678abb9f1ccb2a0202c7a57620e345", "split": "train"}, {"bytes": 95390411, "path": "data/train-00090.parquet", "rows": 176, "sha256": "deee3b9c08e10af5d2f841ce51e7b40667aa3e3c1610cdb311f150974864482b", "split": "train"}, {"bytes": 100193444, "path": "data/train-00091.parquet", "rows": 176, "sha256": "4023ef87f61b488b5fd3fa2bf33f8efca6b41657c273b2c954ce4b26ce39cca3", "split": "train"}, {"bytes": 94341583, "path": "data/train-00092.parquet", "rows": 161, "sha256": "4a4406a666404f524be43ae1292cef118e1b333e80c0b0ae20e3152cd7ddd8bb", "split": "train"}, {"bytes": 96508486, "path": "data/train-00093.parquet", "rows": 100, "sha256": "4115ded6f71dac7153de049dc81aa78d3b8d49b71606c212ed78bb7cce2e7ea8", "split": "train"}, {"bytes": 88809656, "path": "data/train-00094.parquet", "rows": 137, "sha256": "cfe5c2e8f36e30aed5481e44353dce00c6b79654fc93411a168ca8b47c46fbfa", "split": "train"}, {"bytes": 96931453, "path": "data/train-00095.parquet", "rows": 66, "sha256": "ec6330608b872d63806a752d92c6659a937ede78e6716d1480408aa5a8e469fa", "split": "train"}, {"bytes": 99901880, "path": "data/train-00096.parquet", "rows": 114, "sha256": "60b34bf41a044cdf8b60b4c36be5ae5e54aea629c715e4cc139c8c0f44eb04bc", "split": "train"}, {"bytes": 89117411, "path": "data/train-00097.parquet", "rows": 169, "sha256": "c55007333d4a81434a450235aa8864e73512e60ed480a4381a2821f2ceff26f5", "split": "train"}, {"bytes": 100040028, "path": "data/train-00098.parquet", "rows": 180, "sha256": "ca5f17281a1bd6916cc00868de829da4a28b71ac57f26f6f0cc902b846e716c6", "split": "train"}, {"bytes": 99933369, "path": "data/train-00099.parquet", "rows": 158, "sha256": "5d7b172b287a69eb48dbac26d7fa170d5c2566ffd91c0f104938ce10252ec56a", "split": "train"}, {"bytes": 98580236, "path": "data/train-00100.parquet", "rows": 206, "sha256": "3269ca835fdab9de5828e314c9990b22a87a2d00f009c6f5482e61f5634976da", "split": "train"}, {"bytes": 96604421, "path": "data/train-00101.parquet", "rows": 98, "sha256": "09ccb4d68603660edfc82abcd15560ee980a9aaf09bc1baba2fe6342d1668ff5", "split": "train"}, {"bytes": 96379852, "path": "data/train-00102.parquet", "rows": 144, "sha256": "384ceaf9ec3647ea351c834d1fd913df232560c9648cb60101bbd2edc393c097", "split": "train"}, {"bytes": 96931429, "path": "data/train-00103.parquet", "rows": 198, "sha256": "1a87505bf7fbf07cad64e512a4290e93d6d0a69daf5e34fad96a24c3c26b0b4c", "split": "train"}, {"bytes": 98875645, "path": "data/train-00104.parquet", "rows": 167, "sha256": "279b896af9c04c923ae410bbc54b84039aadcdfad1a51a88e25caa1add117b42", "split": "train"}, {"bytes": 99281473, "path": "data/train-00105.parquet", "rows": 165, "sha256": "9d29b0ed845c918525e7a43367035f969bfd6664db1ec1eadc5c915c106e141e", "split": "train"}, {"bytes": 98457984, "path": "data/train-00106.parquet", "rows": 130, "sha256": "88596994b1d7e0ab2262105c6ca49466483eec289283f7d84696497f34f77c42", "split": "train"}, {"bytes": 99608681, "path": "data/train-00107.parquet", "rows": 162, "sha256": "26d188040b2684b7849dd04ba7f1c9adc9117ce647b3ebcd99a7814bf1c91db4", "split": "train"}, {"bytes": 96465040, "path": "data/train-00108.parquet", "rows": 81, "sha256": "a0ec63674c326a5b59c8e769d36c6e6423cdc6446704913df09e751f407fd95a", "split": "train"}, {"bytes": 81900788, "path": "data/train-00109.parquet", "rows": 139, "sha256": "2272ed9dce56f1aa3b7f94e6385012c33a8fb53cca9e5a15c935ee1fae21aaac", "split": "train"}, {"bytes": 89614191, "path": "data/train-00110.parquet", "rows": 133, "sha256": "3e4de781251ead139a7abc127abda1cbbf6cea2025fa30942f34a1ee94a0d3c6", "split": "train"}, {"bytes": 99631554, "path": "data/train-00111.parquet", "rows": 156, "sha256": "b20825e1c1526fbe9cc7d61f070c6559ca95148fb954d50a07f992f6b98efb18", "split": "train"}, {"bytes": 96595648, "path": "data/train-00112.parquet", "rows": 112, "sha256": "60f2b7e79ccca15fffaa68432414ac727f6435086abe8a4923243f2ac7609a23", "split": "train"}, {"bytes": 84600613, "path": "data/train-00113.parquet", "rows": 124, "sha256": "9d81fb11c28b817690ea2540b94e20df43d24a086f2c5f69991bfa719e56804f", "split": "train"}, {"bytes": 99095367, "path": "data/train-00114.parquet", "rows": 177, "sha256": "db37521fade29d68fbdf69de242f379318ff1ca7f67237a05de84d7349b094d8", "split": "train"}, {"bytes": 96075786, "path": "data/train-00115.parquet", "rows": 217, "sha256": "bbd74bc1ad7736b58bdfc862a1ff5b34b3ef7a79933a6b46fa9f8c55105b2f44", "split": "train"}, {"bytes": 98640740, "path": "data/train-00116.parquet", "rows": 126, "sha256": "f7d07f960a6e37e3a97507d34e40eb9f7796b21fb80fef3c8b17ac285cbc56a8", "split": "train"}, {"bytes": 96044557, "path": "data/train-00117.parquet", "rows": 115, "sha256": "54d65672b5611f17e305f917b6f2c84c9d22dd2a31cf8ed5ea180b9e2dc4e865", "split": "train"}, {"bytes": 98190145, "path": "data/train-00118.parquet", "rows": 148, "sha256": "c3037e3c1ea6f846fa416651e4eedad07c125e895dc04fb3da4a60f47855af9d", "split": "train"}, {"bytes": 91737670, "path": "data/train-00119.parquet", "rows": 171, "sha256": "a397fb4f63a59eac23b341ae803066d793000756786ec667510324d429434e1b", "split": "train"}, {"bytes": 100201733, "path": "data/train-00120.parquet", "rows": 114, "sha256": "61bd1966cbae98d4c5524a2e1241a495bacf133151d9695460bbddf9c7cc1a37", "split": "train"}, {"bytes": 98602672, "path": "data/train-00121.parquet", "rows": 159, "sha256": "2946a81e9a2caaa80030af690a5d092379fef1327d92eb651fba0ea2bd18812c", "split": "train"}, {"bytes": 99961250, "path": "data/train-00122.parquet", "rows": 70, "sha256": "5b2f2ee0b24c8b4463316c808bddcef2f7608879990eb2461ab86e80dcbde9db", "split": "train"}, {"bytes": 98562418, "path": "data/train-00123.parquet", "rows": 163, "sha256": "fb9f4dcc7c69cf0b2006888d06d83d522f131d5acf98cc0e274d465f7d106279", "split": "train"}, {"bytes": 87386759, "path": "data/train-00124.parquet", "rows": 67, "sha256": "0f4c54ae19403295e16c21e5312f64a322f5624baf9f6e2b6df2d37774453cbe", "split": "train"}, {"bytes": 100096879, "path": "data/train-00125.parquet", "rows": 175, "sha256": "fa603cc0d66fe1eba4980676fbc9d3d304f61115aecd658bb3fa3821dac5d94a", "split": "train"}, {"bytes": 90004300, "path": "data/train-00126.parquet", "rows": 117, "sha256": "57e7cafc1ce3345e8cc02d0a635bd1696ad00bab69239d864f76ac5550510921", "split": "train"}, {"bytes": 93407362, "path": "data/train-00127.parquet", "rows": 178, "sha256": "c6e015258c8a8b20a6feb1bb1faa15ccb954dcda5b5cc126e6f5e2755652d723", "split": "train"}, {"bytes": 99761779, "path": "data/train-00128.parquet", "rows": 144, "sha256": "6800714e7f58c9ab54cb06929c89fdb51cf4d17c221a94bdbe25ac1906327b39", "split": "train"}, {"bytes": 98125976, "path": "data/train-00129.parquet", "rows": 177, "sha256": "19b62493be490390a0641bc7496121f18bcbb23653fe3aa540ae1f74651d1f13", "split": "train"}, {"bytes": 99316893, "path": "data/train-00130.parquet", "rows": 104, "sha256": "0c6080e6f977edfd674e3a79672fa02a7ad2180bd7c9ade4e802feb372936baa", "split": "train"}, {"bytes": 99602433, "path": "data/train-00131.parquet", "rows": 175, "sha256": "51e7884a0db72eba617fa687bf2f11339a291438a18f2cc0f28c250d86cb3d3b", "split": "train"}, {"bytes": 97805640, "path": "data/train-00132.parquet", "rows": 190, "sha256": "2104606f9c6e950b19cfd36dde68d70dbf9454d17e3c6541ed4ace3247e31384", "split": "train"}, {"bytes": 89500361, "path": "data/train-00133.parquet", "rows": 194, "sha256": "ac281dc6679cc4e2840d53d41972b4a4d10adccd5ab98f18a58cbc343e9d1611", "split": "train"}, {"bytes": 94167108, "path": "data/train-00134.parquet", "rows": 125, "sha256": "9a69027a13c6f9904d366bc1bac31ab6b8055abc9b0cdd72ce9e9b08843d761a", "split": "train"}, {"bytes": 80790791, "path": "data/train-00135.parquet", "rows": 102, "sha256": "961a19f757d732ef25c245795e623bdc7ba3c021e88a19759672b0dabe9bca92", "split": "train"}, {"bytes": 89181912, "path": "data/train-00136.parquet", "rows": 54, "sha256": "8bf17ced64ba1627f952fe1b3835bfdd7076477b11e725ba7bedf2b5eae8650c", "split": "train"}, {"bytes": 98734720, "path": "data/train-00137.parquet", "rows": 180, "sha256": "50ca5a4aab1075a190a451d0795b7a4d6d6202ae74ba590c3d4415f3b7498e09", "split": "train"}, {"bytes": 97056742, "path": "data/train-00138.parquet", "rows": 146, "sha256": "ce234b5bb8fcdd84156d89890a911e48495dfed0f3b406cfccad287f1a0674d9", "split": "train"}, {"bytes": 99534430, "path": "data/train-00139.parquet", "rows": 143, "sha256": "f358bc189e9cadf694fba113e229d02de608a6e52176b22ea17e42b9c5155cdc", "split": "train"}, {"bytes": 92696590, "path": "data/train-00140.parquet", "rows": 190, "sha256": "e38af8fe29ecd7106584544de79c468a15dc4c261fc6c33359600fc2e59d951a", "split": "train"}, {"bytes": 100127329, "path": "data/train-00141.parquet", "rows": 141, "sha256": "d122708a5505e03b03cc4e361b1d2a0ca8a4667d882c7f056ff854166483e77c", "split": "train"}, {"bytes": 83371303, "path": "data/train-00142.parquet", "rows": 173, "sha256": "326fb706206d0ef0e27f6517326769864d0d3528a50c01b157d53940e26bdd5d", "split": "train"}, {"bytes": 100189840, "path": "data/train-00143.parquet", "rows": 72, "sha256": "7f5f1fa2932044d7361d7e0cc8028e0cf613b492361cf915a9d149a82c41dc71", "split": "train"}, {"bytes": 100200102, "path": "data/train-00144.parquet", "rows": 97, "sha256": "51fbcf48e52d912d3fca9bf2c56b78c0e499e2087823db35c366b1c53584b76a", "split": "train"}, {"bytes": 94544975, "path": "data/train-00145.parquet", "rows": 173, "sha256": "beb8558991bf389cfa47b46bb4ccb7ba3fa9347f3848f8589895a3ceed5f9c7f", "split": "train"}, {"bytes": 95188719, "path": "data/train-00146.parquet", "rows": 174, "sha256": "eafe1611849295ccd623942664ff32ed2cc726fc9c6f4c7c05eb4c92b10ce39e", "split": "train"}, {"bytes": 94534836, "path": "data/train-00147.parquet", "rows": 202, "sha256": "d42c669e369664ca7dd366273739a0e1175947b0c672cd129d7c5fdf2e638b93", "split": "train"}, {"bytes": 91672778, "path": "data/train-00148.parquet", "rows": 139, "sha256": "c3f876df453d8783f9d455ede836b463a2447770b41697b87be069b6b87d0171", "split": "train"}, {"bytes": 98057631, "path": "data/train-00149.parquet", "rows": 188, "sha256": "4122745a67dc1bb0aa1961b5c3fa27f55b9551c22a52e610b8586d0935a62539", "split": "train"}, {"bytes": 81472129, "path": "data/train-00150.parquet", "rows": 75, "sha256": "e3a443f72736e50a936d88435aac0dc217d109340d5281427ce3ba2ba79d0b88", "split": "train"}, {"bytes": 99196733, "path": "data/train-00151.parquet", "rows": 146, "sha256": "7353d47de3d1f90e0b6872b38fe5a94161e64267fe47ada4eea8de786ab3724b", "split": "train"}, {"bytes": 91652659, "path": "data/train-00152.parquet", "rows": 108, "sha256": "156c38b656ce2d88e5fec46373114df55098d7216db4ed79d65084e605cdec29", "split": "train"}, {"bytes": 97995030, "path": "data/train-00153.parquet", "rows": 112, "sha256": "da553b6cf1cbfd8e2d8dfe056a7bf9bf9b2d5650488914ce7cdc70ee6fd0e314", "split": "train"}, {"bytes": 97253034, "path": "data/train-00154.parquet", "rows": 192, "sha256": "dc0d66452ed4713592321d0b445c2fe1e33eb5e3f87cb752b160ebd5e533da42", "split": "train"}, {"bytes": 97943820, "path": "data/train-00155.parquet", "rows": 136, "sha256": "251a6acd7b4d74e68594a8d06708be4aa72a06ffbfde13f51a2c889ffd7c03f0", "split": "train"}, {"bytes": 90276505, "path": "data/train-00156.parquet", "rows": 208, "sha256": "4ed78c06f094d01eaca399355d4e27eb2cdb9494f22f0e29ce65c4b7f1442950", "split": "train"}, {"bytes": 97960658, "path": "data/train-00157.parquet", "rows": 148, "sha256": "268199252aeae706e77604721f028cd9b624637916b895e9c59508a9b3c1271e", "split": "train"}, {"bytes": 99605115, "path": "data/train-00158.parquet", "rows": 174, "sha256": "b4538184fa006b73d1b4827783121307e8de0d2c720dead6b6623626a3566398", "split": "train"}, {"bytes": 98748738, "path": "data/train-00159.parquet", "rows": 115, "sha256": "754d2ad18a821ae7bd0b304d80bc60848b707cae3041fd815c3f0ac43b6572ce", "split": "train"}, {"bytes": 98657894, "path": "data/train-00160.parquet", "rows": 141, "sha256": "2063979534bd7f80beeeef9efa7f2c35316fc7c39483782b80fe1f68108dd4bc", "split": "train"}, {"bytes": 99814211, "path": "data/train-00161.parquet", "rows": 127, "sha256": "667d8e4cfef9b7000cb746296d6d51045c9c901aa39448149edd68818ee3afd7", "split": "train"}, {"bytes": 94773280, "path": "data/train-00162.parquet", "rows": 158, "sha256": "af31ad999b0eee24d6a4f880f4d1cda8a7f5389d4a006f6b334afbf6bcbafc06", "split": "train"}, {"bytes": 99940951, "path": "data/train-00163.parquet", "rows": 162, "sha256": "2edc29692290ea672e58a19a0c0e93cf7e7756e45324006970845fa83a64c196", "split": "train"}, {"bytes": 97215132, "path": "data/train-00164.parquet", "rows": 114, "sha256": "c0f4480aa46af0e9604eae56d4f36a173c21bbafad1718552af70649212a26a8", "split": "train"}, {"bytes": 100151208, "path": "data/train-00165.parquet", "rows": 146, "sha256": "ef43644c8dd8af628444a05f5d774111e7794321c2fbaf11386f7fbcdd08b72f", "split": "train"}, {"bytes": 99611433, "path": "data/train-00166.parquet", "rows": 152, "sha256": "0111ba971d8f64790d4c5cafc9933d43c8a0ecc4a1590f8ca3573cc1595f289b", "split": "train"}, {"bytes": 100103676, "path": "data/train-00167.parquet", "rows": 116, "sha256": "9c33c85e9b580dfe50109733c5f59977aa75d2f326885ae54c58fcf81eb38ae6", "split": "train"}, {"bytes": 98914296, "path": "data/train-00168.parquet", "rows": 121, "sha256": "13f98d8e95389561b6cb56fdcaeb27178c03a8b9ef7516a8e3cda84b16fa47e1", "split": "train"}, {"bytes": 98229514, "path": "data/train-00169.parquet", "rows": 181, "sha256": "0d6d5478ef3967bc9ed5e8b9126a77671d5483361900707ac46e941a852c6892", "split": "train"}, {"bytes": 99573997, "path": "data/train-00170.parquet", "rows": 154, "sha256": "fec380ec27cdcd89c4b37e1dab472bc634e1980286e4bc14ec6244f443cf230a", "split": "train"}, {"bytes": 95777515, "path": "data/train-00171.parquet", "rows": 100, "sha256": "14cdc21c5db13c7ccad3c725e68a74342958bd73bf5c92810978e548ee021e28", "split": "train"}, {"bytes": 97893580, "path": "data/train-00172.parquet", "rows": 81, "sha256": "50e967b4b79ace7c7419548af39b9f7e40bdb0f6fa5689d7fb06643c1e9b7b26", "split": "train"}, {"bytes": 100354038, "path": "data/train-00173.parquet", "rows": 122, "sha256": "8f33255c734a4af94b5ab3b0c1dd906ae590d8854618356d77c02a52350d7cb0", "split": "train"}, {"bytes": 84825560, "path": "data/train-00174.parquet", "rows": 142, "sha256": "6b54a6fef5fe0525511d80b5bfd1af734b113e73a37b27caf2630754353dce90", "split": "train"}, {"bytes": 85339835, "path": "data/train-00175.parquet", "rows": 131, "sha256": "2a9c1e60ae36a576dc8c597d98910db8e1a64bf877ac5059a10d71b2a831e4f7", "split": "train"}, {"bytes": 95416330, "path": "data/train-00176.parquet", "rows": 85, "sha256": "dfbfc83955c2761d1292607283ed3012f0c9b06a47bf325d20cd4b3970362e41", "split": "train"}, {"bytes": 96648628, "path": "data/train-00177.parquet", "rows": 156, "sha256": "8d43b3a7ecc9b8e12af871f9a315a7757c4827c057b0281bc1db36b5dfc0d4c8", "split": "train"}, {"bytes": 100000200, "path": "data/train-00178.parquet", "rows": 168, "sha256": "398a02b6be5260822e2c068bc7746ebb427762af1eb1c97362ed5be39d3e07c9", "split": "train"}, {"bytes": 95689863, "path": "data/train-00179.parquet", "rows": 187, "sha256": "ff97212f53906c2da0254d0252e3993a15f4ac4350c063b9abe0547d5b50530a", "split": "train"}, {"bytes": 100279322, "path": "data/train-00180.parquet", "rows": 72, "sha256": "64c7cf9080c46b104d63293cd6aa2b4265c0e2b09d2b32b34b5025cdc9f3ac13", "split": "train"}, {"bytes": 81500439, "path": "data/train-00181.parquet", "rows": 151, "sha256": "0d42cc28bcb972ff6639392dff9032565b86d61aead1125bef866f4bfce33f19", "split": "train"}, {"bytes": 96006962, "path": "data/train-00182.parquet", "rows": 100, "sha256": "9b3f55485cf0edae7dab43398533bbb1e5824efaf9bb6ca0e7d9a7194c885915", "split": "train"}, {"bytes": 69113834, "path": "data/train-00183.parquet", "rows": 94, "sha256": "6cc6f6e199cb5f3fc904386bd66a7055a3e4c07a7e068f393a48903dcfd6528c", "split": "train"}, {"bytes": 83159458, "path": "data/train-00184.parquet", "rows": 107, "sha256": "0280e7261542b19822aea83e9323404c974ca0daea86b29cdcfb27b8c301877d", "split": "train"}, {"bytes": 100446436, "path": "data/train-00185.parquet", "rows": 106, "sha256": "aa0b962e01d1a38b8e8ef7c63687822de84438e1a55e54d2befa210046afff43", "split": "train"}, {"bytes": 87484304, "path": "data/train-00186.parquet", "rows": 139, "sha256": "5d439b55c2059362a37c81b558256714abe5944201ca2f6fcf09f9ffc3e927d1", "split": "train"}, {"bytes": 96688017, "path": "data/train-00187.parquet", "rows": 111, "sha256": "d9ae40eb8980131cb8ba0df8905ead5e08d7f757d0eff18345d62d232008fd56", "split": "train"}, {"bytes": 98341246, "path": "data/train-00188.parquet", "rows": 118, "sha256": "f9e10772bfe2a692ed5d9c2d0be94ea8ae1bac4d8136db3dd8756b424fdf87f3", "split": "train"}, {"bytes": 95659125, "path": "data/train-00189.parquet", "rows": 125, "sha256": "684ae0adc3de82e61d9a146477a8c6f04b80bec0934005386205e35313d0f4bd", "split": "train"}, {"bytes": 96706897, "path": "data/train-00190.parquet", "rows": 92, "sha256": "23b51214763322844dca6dacb5b7df8bbe5c8248c3c6ac8899fe3a9b933cbf16", "split": "train"}, {"bytes": 99805320, "path": "data/train-00191.parquet", "rows": 149, "sha256": "80ba3b3673026893afd10358ab7da6ba77b415ef4850767b9b43a7422b9803dc", "split": "train"}, {"bytes": 99359943, "path": "data/train-00192.parquet", "rows": 170, "sha256": "89857046ed6cb6a0b552a65f8c4146f806134da0155050493ddb024aec1ab20b", "split": "train"}, {"bytes": 99947763, "path": "data/train-00193.parquet", "rows": 177, "sha256": "ca5a5fae9802e8d20663026c2b094fde1e63934a4256dd806369be29f6c7196b", "split": "train"}, {"bytes": 87546064, "path": "data/train-00194.parquet", "rows": 101, "sha256": "1845b509f5e31584d5e1a5b172b4868e1af8f2e85efe54a5eef3b44fa014b544", "split": "train"}, {"bytes": 95654716, "path": "data/train-00195.parquet", "rows": 179, "sha256": "c0b9b2e8142e7e24ef1c8d0da764bbc5e0d290367bafff300681fc7c4395f710", "split": "train"}, {"bytes": 100312501, "path": "data/train-00196.parquet", "rows": 90, "sha256": "65f174b69c6b28e9e6fbd0a92558d689543dcc015998394e713cb1ca87a81440", "split": "train"}, {"bytes": 98445589, "path": "data/train-00197.parquet", "rows": 181, "sha256": "7ce88715321e1c474610ed7611469386d2594ad86a38ecf9646283e567e9e2e1", "split": "train"}, {"bytes": 100055806, "path": "data/train-00198.parquet", "rows": 103, "sha256": "66efb8ae5769688ea1ee1fd53eaecdea914f8d27df0c95a07449d3b198c9928c", "split": "train"}, {"bytes": 99426199, "path": "data/train-00199.parquet", "rows": 135, "sha256": "b7fe388153cfb76e0c3ff05ae7fa6e766c6b47eaddf72094888ee1b2df51f4ea", "split": "train"}, {"bytes": 99253296, "path": "data/train-00200.parquet", "rows": 181, "sha256": "f892c1bc13bf357d929f3e98cf58cc55b4962e9950f40dbe882dc8d5dd57a1b8", "split": "train"}, {"bytes": 94919499, "path": "data/train-00201.parquet", "rows": 151, "sha256": "fc42cae853ffd27abc48a27c9dcbbf368fc924e3de8996caf79e362049e260bb", "split": "train"}, {"bytes": 99810591, "path": "data/train-00202.parquet", "rows": 171, "sha256": "c6303c3bce6c27f9e4752c5cb04bcf19746ad812eb84ab55dc55f47e07801b27", "split": "train"}, {"bytes": 97954282, "path": "data/train-00203.parquet", "rows": 185, "sha256": "52c08e63a7f024240d1ca8f7b209ec78edb4566a542dcf124d245781541b27bb", "split": "train"}, {"bytes": 91177259, "path": "data/train-00204.parquet", "rows": 94, "sha256": "35aec26acb7c27f016c975dae7cca832b22b31954b174ea7f31bcd7e2b598980", "split": "train"}, {"bytes": 98712618, "path": "data/train-00205.parquet", "rows": 120, "sha256": "4398f2323937a01c47dbc1cecb1c46879f7900af241f6721d62e091f2e61db05", "split": "train"}, {"bytes": 86318604, "path": "data/train-00206.parquet", "rows": 167, "sha256": "a3c7e551e72e183c1b88e064d96e7bfe48d464b712c5730b7be6e1e15a8e9e81", "split": "train"}, {"bytes": 97942619, "path": "data/train-00207.parquet", "rows": 148, "sha256": "03062b8e14c4640a7759b4dddffc9725c881967c87b539c2d24e9f97b1dd1854", "split": "train"}, {"bytes": 97142653, "path": "data/train-00208.parquet", "rows": 127, "sha256": "7b36644590116b360d36020767721d9628e885da65c77759894b46a532c99e84", "split": "train"}, {"bytes": 85935145, "path": "data/train-00209.parquet", "rows": 107, "sha256": "7fe94b4351f3d597e7c988b63a710a1c2cd6e12663a28378a73fa421569a2072", "split": "train"}, {"bytes": 99518892, "path": "data/train-00210.parquet", "rows": 121, "sha256": "bbd8494a895882f3e24d1dc226dc45b6a183a55876e17abd274066f613cca5fb", "split": "train"}, {"bytes": 73386858, "path": "data/train-00211.parquet", "rows": 175, "sha256": "a1797b19a399f9ff82c20648e2a6a51838bb4ae26afe5a9b12881f9def0cb807", "split": "train"}, {"bytes": 98001515, "path": "data/train-00212.parquet", "rows": 70, "sha256": "1d4017d1ac44f8a857b329d9520dacf308934e58a55a4bb0c6732ed300fea4dd", "split": "train"}, {"bytes": 100075013, "path": "data/train-00213.parquet", "rows": 183, "sha256": "e2e9d6bce44e98da054df5cb94e77a804b62410de9821400d7f8307ad0eb86af", "split": "train"}, {"bytes": 97962091, "path": "data/train-00214.parquet", "rows": 161, "sha256": "799a50a4720c76c8baec15d6de84db1c38bcade61ff8fb529833c0e2ab503897", "split": "train"}, {"bytes": 99236541, "path": "data/train-00215.parquet", "rows": 182, "sha256": "d9eb104d4c5ef5e7e086109e4303f5435c464e71c93dad369ba88bc48e37ec4d", "split": "train"}, {"bytes": 99714164, "path": "data/train-00216.parquet", "rows": 224, "sha256": "1b6d124b435b5ce60ad08d0445b8a16e47d380484d376fad1c3a9f1998b61383", "split": "train"}, {"bytes": 94645580, "path": "data/train-00217.parquet", "rows": 142, "sha256": "40bb135d3619bc1a772e5767689cbb32f28c54a5390a54d58429eabdee0cdb63", "split": "train"}, {"bytes": 98762905, "path": "data/train-00218.parquet", "rows": 69, "sha256": "c3c3eb7b02165c06620f1252049b08753b1127ea976f13236a387c525e401ecb", "split": "train"}, {"bytes": 94951836, "path": "data/train-00219.parquet", "rows": 124, "sha256": "9a267a7bad58dc8c1509306bca01e1a6bcf8125552e02db3ede4b538d535fceb", "split": "train"}, {"bytes": 98336299, "path": "data/train-00220.parquet", "rows": 226, "sha256": "1c409bfb488255bce62a97ec34b9680dbfff6a61f2cc5f5631f00ccfb8a9161c", "split": "train"}, {"bytes": 98101017, "path": "data/train-00221.parquet", "rows": 184, "sha256": "e27064003f551d014c54e64a489b5a8d4b19f8c56daae8a5080f404c582d3b93", "split": "train"}, {"bytes": 99495734, "path": "data/train-00222.parquet", "rows": 199, "sha256": "89d03a7b337b43e18fd5d29bf8f9433593bb5cfd594b4a717740e74955960406", "split": "train"}, {"bytes": 91182075, "path": "data/train-00223.parquet", "rows": 155, "sha256": "d46ffba8ba52ddd0ac5b152d3175b15ec9debd12fa09dd25c19e660037da40d1", "split": "train"}, {"bytes": 97275465, "path": "data/train-00224.parquet", "rows": 108, "sha256": "472324b541d862be9758e8a227b3371ee5f0b2e816f473477bb29ed83c170d6c", "split": "train"}, {"bytes": 98792635, "path": "data/train-00225.parquet", "rows": 122, "sha256": "fafeb01b3b34db419479ce33e1102e19a7bdf1cea636297558f0cecee871698e", "split": "train"}, {"bytes": 97754019, "path": "data/train-00226.parquet", "rows": 96, "sha256": "e7538d1822fe4df468cee96b059032eb29dd770b3909db29a54771a45fc362f0", "split": "train"}, {"bytes": 66928488, "path": "data/train-00227.parquet", "rows": 104, "sha256": "491bdebe6879a5cbd7665585dd765389fc149b563338392f031df2d2fc4cea54", "split": "train"}, {"bytes": 86106434, "path": "data/train-00228.parquet", "rows": 69, "sha256": "aaa6b481c4a50803b3879f73a972f8c5c3ecdb978f80b28b5035e2a21ae36f4f", "split": "train"}, {"bytes": 98846155, "path": "data/validation-00000.parquet", "rows": 244, "sha256": "cccf220875bc93679ec7d689e56a1a87026949561142ba0abaa1e59f7a431059", "split": "validation"}, {"bytes": 47190778, "path": "data/validation-00001.parquet", "rows": 196, "sha256": "b4db330ff89e317f828d65dff78eb7b5e1d616fdcb964f89bf35577274686f81", "split": "validation"}, {"bytes": 99446700, "path": "data/validation-00002.parquet", "rows": 34, "sha256": "56f09551835005a11d3fc3ffa70d792c9afc6149526550376e3c3e6652aadf83", "split": "validation"}, {"bytes": 69896326, "path": "data/validation-00003.parquet", "rows": 256, "sha256": "5a6c68b625435ca29a0cd8232a9730c2a7fe8bbd265a363a13fee0ae4352d0fc", "split": "validation"}, {"bytes": 65359702, "path": "data/validation-00004.parquet", "rows": 256, "sha256": "e0e91090fcd4d2838f68311cff3a949f6bc5c3cb1ed1e72fb6602e48d530d3cc", "split": "validation"}, {"bytes": 80907793, "path": "data/validation-00005.parquet", "rows": 256, "sha256": "3f418915f3e1b0818f306b663e6594d2e586ec7c9b8bd4a00fdc8efa1fb342b3", "split": "validation"}, {"bytes": 99357225, "path": "data/validation-00006.parquet", "rows": 157, "sha256": "c441847d2a968e18be0902e4a0190a3d26be6bcf3ed41c7b2c2d8c2915a9035a", "split": "validation"}, {"bytes": 65838488, "path": "data/validation-00007.parquet", "rows": 256, "sha256": "cff6716a69cb3b4e20fcac02dbbed15b9055a9190af39ea49fb73cda0bd113d5", "split": "validation"}, {"bytes": 84190208, "path": "data/validation-00008.parquet", "rows": 256, "sha256": "e2f7b37607016eabce3fefcc240a7a03efc66626bc55dc9bceabbc89db4e7b84", "split": "validation"}, {"bytes": 75232664, "path": "data/validation-00009.parquet", "rows": 256, "sha256": "76544c5a56b8d8c9797e2706dd9a33ac775929687e4a6e12a5c46b63cdbce743", "split": "validation"}, {"bytes": 92675192, "path": "data/validation-00010.parquet", "rows": 256, "sha256": "f78665674b7fa93fa71a20306be9a31ed80c4f117780dfccca1ea011fb5e7e79", "split": "validation"}, {"bytes": 98445254, "path": "data/validation-00011.parquet", "rows": 184, "sha256": "e23f9bb57cd4278b584f854c137e8d80eaea93fe3532195874c3c8131b9939a6", "split": "validation"}, {"bytes": 83961613, "path": "data/validation-00012.parquet", "rows": 256, "sha256": "37307f914ef0311041fd95164d09d6141f167cfb6deafae0e3c54cec0078b59a", "split": "validation"}, {"bytes": 84651406, "path": "data/validation-00013.parquet", "rows": 256, "sha256": "995ed935f1dc8586319fbc27fbe53f9a6bf88b809f44adba7e619ae4eb6a09e0", "split": "validation"}, {"bytes": 99104972, "path": "data/validation-00014.parquet", "rows": 213, "sha256": "c150c612e75382fadef8d2a5d5255fb9d60179223f5430c2b1a5a73ec063f544", "split": "validation"}, {"bytes": 95648614, "path": "data/validation-00015.parquet", "rows": 256, "sha256": "4e29c15bbee4e815835f0d8fa7098a24aff3834bb1711621f42ccc2408aba16b", "split": "validation"}, {"bytes": 6600119, "path": "data/validation-00016.parquet", "rows": 31, "sha256": "c9123ff1a52592feebbc1e6acb601d671fab84b0b29a9e18e3f26f52330369cf", "split": "validation"}, {"bytes": 82403838, "path": "data/test-00000.parquet", "rows": 256, "sha256": "fbd4fd2e9ec7c119f20245c45e41c5f53d2c933647396ea40c43a6a4aa0058df", "split": "test"}, {"bytes": 92669959, "path": "data/test-00001.parquet", "rows": 256, "sha256": "3cc3aa35668515dd53c2f7238f81a22c52a1f71cd4fd4ae9fd93a56485e95f10", "split": "test"}, {"bytes": 94209336, "path": "data/test-00002.parquet", "rows": 256, "sha256": "d5f3f11eee9fb666ad362357f0b11acb1c1db66ad9ea88fb5383c153c44b1d99", "split": "test"}, {"bytes": 60084827, "path": "data/test-00003.parquet", "rows": 193, "sha256": "3edd9e02b3696a04ac17ebc154e06a74d0494a94a96f221a9eff5109d6615fa8", "split": "test"}, {"bytes": 99574615, "path": "data/test-00004.parquet", "rows": 151, "sha256": "13b5f12af2968a305af0ec544457826dda1d30452f59ce406d65dfac203b6646", "split": "test"}, {"bytes": 89986531, "path": "data/test-00005.parquet", "rows": 256, "sha256": "3ec90974691042680c88b80a1e3ff1ce00d702558f526dcfa489979dbc3f4b40", "split": "test"}, {"bytes": 99394990, "path": "data/test-00006.parquet", "rows": 198, "sha256": "0976d12b00642659095ff8be25696312b3adf78153d48c942b5d3417052d39e6", "split": "test"}, {"bytes": 98587865, "path": "data/test-00007.parquet", "rows": 234, "sha256": "801853760a7dfdf69b025bc996ea4c846fb3a6ff7b6c9a6018695dadf86d2bdd", "split": "test"}, {"bytes": 88158167, "path": "data/test-00008.parquet", "rows": 256, "sha256": "99aec80a349aeaddd122aa2b1eb0bc610b68becd4a30bbd73dd561d88eb496fc", "split": "test"}, {"bytes": 99369325, "path": "data/test-00009.parquet", "rows": 111, "sha256": "1b782dcacca42ffe511a84d91fcc8e6dc974f181ee049985d457bbfc46a64cc0", "split": "test"}, {"bytes": 99541986, "path": "data/test-00010.parquet", "rows": 216, "sha256": "09c4c9a3160f1915b5ff750d4090d2b7e5ceb7b4ec30db2a86a9c3f5e69d1389", "split": "test"}, {"bytes": 98375797, "path": "data/test-00011.parquet", "rows": 217, "sha256": "dd66394f3688ce22aac3a9eec4873683437c5c99448f03e92646ff5668700787", "split": "test"}, {"bytes": 72059949, "path": "data/test-00012.parquet", "rows": 169, "sha256": "8f09d6fffd036e7ef4cfe1775cd70e4d48b158387d7230deefd5840113ca7b34", "split": "test"}, {"bytes": 99370046, "path": "data/test-00013.parquet", "rows": 119, "sha256": "1ebb74556fb8b265e14fd0856da7e08ac64a07dd6518f994af79f2cfcb6ca759", "split": "test"}, {"bytes": 88972625, "path": "data/test-00014.parquet", "rows": 256, "sha256": "129af38f47dd71c208d7d1b7c37b9fcc975cd3a4883fa09d9a19a914f14b207e", "split": "test"}, {"bytes": 81517716, "path": "data/test-00015.parquet", "rows": 256, "sha256": "2864207f65372bfa6d320f1e97a3783a0d048832a3446408801e195720fc98c1", "split": "test"}, {"bytes": 89889552, "path": "data/test-00016.parquet", "rows": 256, "sha256": "c015a6f6db30e165d160b668ded898300025f4fa21b7cb8034a3ec786548b156", "split": "test"}, {"bytes": 8956896, "path": "data/test-00017.parquet", "rows": 33, "sha256": "cf2acddae1d252015dc7daa7d6ef3d6bd99e19d42375402fa75ba870823bfa75", "split": "test"}], "sources": {"fireviewer/fire-smoke-detection-corpus-v2": {"license_scope": "mixed; consult original source provenance", "revision": "c1336415e1f9e7f67f332a6ec212345fdf3ffdb2"}, "fireviewer/fire_and_smoke_detection_very_hard_negative": {"license_scope": "mixed; consult original source provenance", "revision": "04ed0590954245b524dd265a6da63462b2e50f62"}}, "split_receipt": {"conflicting_images_quarantined": 8, "cross_split_detected_near_duplicate_edges": 0, "cross_split_groups": 0, "decoded_pixel_duplicates": 1, "grouping": "native RGB hashes, normalized source origins, known capture groups, conservative FLAME sequence grouping, pHash Hamming<=8 AND dHash<=12 AND aspect-ratio log distance<=0.2", "historical_images_forced_train": 10086, "input_images": 38974, "limitation": "Automated near-duplicate and known-origin separation; no claim of exhaustive incident identity or individual semantic label revalidation.", "metadata_sha256": "bf2e9771e1ce90971b15b119f1f1f74632c9987477afd79c8e87eb5a18fed99d", "near_duplicate_edges": 223857, "schema": "fireviewer-v25-splits-v1", "seed": 42, "splits": {"test": {"groups": 1632, "historical": 0, "images": 3689, "negative_kinds": {"native_negative": 1229, "very_hard_negative": 21}, "objects": {"flame_visible": 2749, "smoke_visible": 2062}, "small_objects": {"flame_visible": 592, "smoke_visible": 63}}, "train": {"groups": 19575, "historical": 10086, "images": 31657, "negative_kinds": {"native_negative": 10165, "v2_negative": 504, "very_hard_negative": 153}, "objects": {"flame_visible": 23939, "smoke_visible": 16894}, "small_objects": {"flame_visible": 5599, "smoke_visible": 507}}, "validation": {"groups": 1695, "historical": 0, "images": 3619, "negative_kinds": {"native_negative": 1259, "very_hard_negative": 21}, "objects": {"flame_visible": 2808, "smoke_visible": 2139}, "small_objects": {"flame_visible": 533, "smoke_visible": 62}}}, "target_fractions": {"test": 0.1, "train": 0.8, "validation": 0.1}, "unique_kept": 38965}, "splits": {"test": 3689, "train": 31657, "validation": 3619}}

provenance/negative/anamibnjafar0--flamevision.json

@@ -0,0 +1,110 @@
+{
+  "authority_sha256": "6fe912334677393474ed297e669ef41f41551930eea46e17f3dcb91ebed40dac",
+  "classification_evidence": {
+    "classification_audit": {
+      "allow_detection_mapping": true,
+      "boxes_usable": false,
+      "cases_to_correct": [
+        {
+          "issue": "Visible smoke missing from fire-only detection annotations",
+          "sample": 7
+        },
+        {
+          "issue": "Visible smoke missing from fire-only detection annotations",
+          "sample": 8
+        },
+        {
+          "issue": "Visible smoke missing from fire-only detection annotations",
+          "sample": 9
+        }
+      ],
+      "classification_allowed_labels": [
+        "fire",
+        "nofire"
+      ],
+      "classification_usable": true,
+      "decision": "needs_correction",
+      "evidence": "audit/samples/anamibnjafar0--flamevision/sample.json and sheet-01/02/03.jpg",
+      "findings": "All nine sampled classification labels match visible incident versus no incident. Three detection examples (7, 8, 9) have plausible fire boxes but intentionally omit substantial visible smoke. Suitable as source fire-only annotation; incomplete for visible_fire + visible_smoke corpus. Preserve aerial views and classification/detection subsets separately. Closely related controlled-burn scenes appear across original splits; rebuild by event/sequence.",
+      "presence_semantics": "wildfire-event-presence-including-smoke",
+      "review_kind": "root_visual_sample_review",
+      "reviewed_at": "2026-10-02T00:28:00.669072+00:00",
+      "reviewed_samples": [
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12
+      ],
+      "source_labels": [
+        "fire",
+        "nofire"
+      ],
+      "target_schema": "classification-fire-nofire; detection-fire-only requiring smoke completion",
+      "training_split_requires_rebuild": true,
+      "validation_scope": "12-image visual sample only; three boxed images; not exhaustive"
+    },
+    "path": "/content/drive/MyDrive/FireViewer/datasets-20261002/collected/a-annoter-ou-corriger/anamibnjafar0--flamevision/v1/inventory.jsonl",
+    "sha256": "11bdfc70a96f68228a45d8e52305d3f423d9c222717c64ca57b5b025f7d4c857",
+    "source_metadata_sha256": "ba10cd69a5d1ac3ce559b4c3540419b76bfeae7944b9de512f556750d6309e94"
+  },
+  "individual_visual_validation_claimed": false,
+  "manual_review_enabled": false,
+  "original_labels_and_splits_preserved": true,
+  "root_repository_license_does_not_replace_source_licenses": true,
+  "source": "anamibnjafar0/flamevision",
+  "source_metadata": {
+    "frozen_download_version": 1,
+    "group": "aerial",
+    "kaggle_metadata": {
+      "creatorName": "Anam Ibn Jafar",
+      "creatorNameNullable": "Anam Ibn Jafar",
+      "currentVersionNumber": 1,
+      "currentVersionNumberNullable": 1,
+      "description": "The FlameVision dataset is a comprehensive aerial image dataset designed specifically for detecting and classifying wildfires. It consists of a total of 8600 high-resolution images, with 5000 images depicting fire and the remaining 3600 images depicting non-fire scenes. The images are provided in PNG format for classification tasks and JPG format for detection tasks. The dataset is organized into two primary folders, one for detection and the other for classification, with further subdivisions into train, validation, and test sets for each folder. To facilitate accurate object detection, the dataset also includes 4500 image annotation files. These annotation files contain manual annotations in XML format, which specify the exact positions of objects and their corresponding labels within the images. The annotations were performed using Roboflow, ensuring high quality and consistency across the dataset. One of the notable features of the FlameVision dataset is its compatibility with various convolutional neural network (CNN) architectures, including EfficientNet, DenseNet, VGG-16, ResNet50, YOLO, and R-CNN. This makes it a versatile and valuable resource for researchers and practitioners in the field of wildfire detection and classification, enabling the development and evaluation of sophisticated ML models.",
+      "descriptionNullable": "The FlameVision dataset is a comprehensive aerial image dataset designed specifically for detecting and classifying wildfires. It consists of a total of 8600 high-resolution images, with 5000 images depicting fire and the remaining 3600 images depicting non-fire scenes. The images are provided in PNG format for classification tasks and JPG format for detection tasks. The dataset is organized into two primary folders, one for detection and the other for classification, with further subdivisions into train, validation, and test sets for each folder. To facilitate accurate object detection, the dataset also includes 4500 image annotation files. These annotation files contain manual annotations in XML format, which specify the exact positions of objects and their corresponding labels within the images. The annotations were performed using Roboflow, ensuring high quality and consistency across the dataset. One of the notable features of the FlameVision dataset is its compatibility with various convolutional neural network (CNN) architectures, including EfficientNet, DenseNet, VGG-16, ResNet50, YOLO, and R-CNN. This makes it a versatile and valuable resource for researchers and practitioners in the field of wildfire detection and classification, enabling the development and evaluation of sophisticated ML models.",
+      "hasCreatorName": true,
+      "hasCurrentVersionNumber": true,
+      "hasDescription": true,
+      "hasLicenseName": true,
+      "hasSubtitle": true,
+      "hasTitle": true,
+      "lastUpdated": "2023-07-18T10:30:34.863Z",
+      "licenseName": "CC0: Public Domain",
+      "licenseNameNullable": "CC0: Public Domain",
+      "subtitle": "A new dataset for wildfire classification and detection using aerial imagery",
+      "subtitleNullable": "A new dataset for wildfire classification and detection using aerial imagery",
+      "title": "FlameVision",
+      "titleNullable": "FlameVision",
+      "versions": [
+        {
+          "creationDate": "2023-07-18T10:30:34.863Z",
+          "creatorName": "Anam Ibn Jafar",
+          "creatorNameNullable": "Anam Ibn Jafar",
+          "creatorRef": "flamevision",
+          "creatorRefNullable": "flamevision",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Initial release",
+          "versionNotesNullable": "Initial release",
+          "versionNumber": 1
+        }
+      ]
+    },
+    "license_status": "clear",
+    "retrieved_at": "2026-10-01T23:22:28.369945+00:00",
+    "source_url": "https://www.kaggle.com/datasets/anamibnjafar0/flamevision"
+  },
+  "source_metadata_sha256": "ba10cd69a5d1ac3ce559b4c3540419b76bfeae7944b9de512f556750d6309e94"
+}

provenance/negative/arfintanim--fire-detection.json

@@ -0,0 +1,180 @@
+{
+  "authority_sha256": "6fe912334677393474ed297e669ef41f41551930eea46e17f3dcb91ebed40dac",
+  "classification_evidence": {
+    "classification_audit": {
+      "allow_detection_mapping": false,
+      "boxes_usable": false,
+      "cases_to_correct": [
+        {
+          "issue": "Resolve local class namespace before canonical class adaptation",
+          "sample": 3
+        },
+        {
+          "issue": "Visible smoke without source box",
+          "sample": 2
+        },
+        {
+          "issue": "Visible flame/smoke without source boxes",
+          "sample": 4
+        },
+        {
+          "issue": "Fire-only aerial annotations omit smoke",
+          "sample": 5
+        }
+      ],
+      "classification_allowed_labels": [
+        "fire",
+        "nofire"
+      ],
+      "classification_usable": true,
+      "decision": "needs_correction",
+      "evidence": "audit/samples/arfintanim--fire-detection/sample.json and sheet-01/02/03.jpg",
+      "findings": "Mixed source archive. Sample 3 tiny distant smoke box is spatially plausible but category is displayed only as unresolved ID 0: mapping must be established from that exact annotation namespace. Samples 2 and 4 have visible targets with no associated boxes. Satellite example 5 has flame boxes but no smoke coverage; keep aerial domain separate. Unboxed fire/nofire classification examples 7-10 have consistent labels, and must not be treated as annotated detection examples. Rain example 6 and ordinary skyline/landscape 1/11 are hard-negative candidates. Bright night-edge 12 remains ambiguous.",
+      "review_kind": "root_visual_sample_review",
+      "reviewed_at": "2026-10-02T00:34:18.749816+00:00",
+      "reviewed_samples": [
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12
+      ],
+      "source_labels": [
+        "Fire",
+        "Smoke",
+        "fire",
+        "nofire",
+        "unresolved-id-0"
+      ],
+      "target_schema": "per-source namespaces; classification-fire-nofire and detection subsets separated",
+      "training_split_requires_rebuild": true,
+      "validation_scope": "12-image stratified visual sample across distinct source tasks; not exhaustive"
+    },
+    "path": "/content/drive/MyDrive/FireViewer/datasets-20261002/collected/a-annoter-ou-corriger/arfintanim--fire-detection/v5/inventory.jsonl",
+    "sha256": "853f12a6a730d21d90a930f4cd1ca0147a864b97ce7065d6233d9c41506e53c2",
+    "source_metadata_sha256": "83717375b0d206b9770d136b5edd6c3ed817353d1874b07b1e9993c8d6c31c9a"
+  },
+  "individual_visual_validation_claimed": false,
+  "manual_review_enabled": false,
+  "original_labels_and_splits_preserved": true,
+  "root_repository_license_does_not_replace_source_licenses": true,
+  "source": "arfintanim/fire-detection",
+  "source_metadata": {
+    "frozen_download_version": 5,
+    "group": "mixed-real-rgb",
+    "kaggle_metadata": {
+      "creatorName": "Sharia Arfin Tanim",
+      "creatorNameNullable": "Sharia Arfin Tanim",
+      "currentVersionNumber": 5,
+      "currentVersionNumberNullable": 5,
+      "description": "# **Combined Forest Fire Datasets**\n\nThis dataset is a compilation of forest fire-related datasets sourced from various repositories and research platforms. It combines data from multiple sources to provide a comprehensive collection for analysis and research purposes. The datasets included in this compilation are:\n\n1. **Dataset 1 (gjmr63rz2r-1)**: Retrieved from [Mendeley](https://data.mendeley.com/datasets/gjmr63rz2r/1)\n2. **Dataset 2 (archive)**: Retrieved from [Kaggle](https://www.kaggle.com/datasets/alik05/forest-fire-dataset)\n3. **Dataset 3 (ForestFireDataset(Object Detection))**: Retrieved from [Mendeley](https://data.mendeley.com/datasets/fcsjwd9gr6/1)\n4. **Dataset 4 (D-Fire)**: Retrieved from [GitHub](https://github.com/gaiasd/DFireDataset?tab=readme-ov-file)\n5. **Dataset 5 (ForestFireDataset(Classifications))**: Retrieved from [Mendeley](https://data.mendeley.com/datasets/fcsjwd9gr6/1)\n6. **Dataset 6 (mivia-fire)**: Retrieved from [Mivia](https://mivia.unisa.it/)\n7. **Dataset 7 (mivia-smoke)**: Retrieved from [Mivia](https://mivia.unisa.it/)\n\n\nThe datasets cover various aspects of forest fires, including object detection, classifications, and related imagery. By combining these datasets, researchers and analysts can access a diverse range of data for studying forest fires, developing predictive models, and exploring mitigation strategies.\n\n# **Copyright Notice:**\n\nThe individual datasets included in this compilation retain their respective copyrights. Users are encouraged to refer to the original sources for specific usage terms and citation requirements. This combined dataset is provided for research and educational purposes only.\n",
+      "descriptionNullable": "# **Combined Forest Fire Datasets**\n\nThis dataset is a compilation of forest fire-related datasets sourced from various repositories and research platforms. It combines data from multiple sources to provide a comprehensive collection for analysis and research purposes. The datasets included in this compilation are:\n\n1. **Dataset 1 (gjmr63rz2r-1)**: Retrieved from [Mendeley](https://data.mendeley.com/datasets/gjmr63rz2r/1)\n2. **Dataset 2 (archive)**: Retrieved from [Kaggle](https://www.kaggle.com/datasets/alik05/forest-fire-dataset)\n3. **Dataset 3 (ForestFireDataset(Object Detection))**: Retrieved from [Mendeley](https://data.mendeley.com/datasets/fcsjwd9gr6/1)\n4. **Dataset 4 (D-Fire)**: Retrieved from [GitHub](https://github.com/gaiasd/DFireDataset?tab=readme-ov-file)\n5. **Dataset 5 (ForestFireDataset(Classifications))**: Retrieved from [Mendeley](https://data.mendeley.com/datasets/fcsjwd9gr6/1)\n6. **Dataset 6 (mivia-fire)**: Retrieved from [Mivia](https://mivia.unisa.it/)\n7. **Dataset 7 (mivia-smoke)**: Retrieved from [Mivia](https://mivia.unisa.it/)\n\n\nThe datasets cover various aspects of forest fires, including object detection, classifications, and related imagery. By combining these datasets, researchers and analysts can access a diverse range of data for studying forest fires, developing predictive models, and exploring mitigation strategies.\n\n# **Copyright Notice:**\n\nThe individual datasets included in this compilation retain their respective copyrights. Users are encouraged to refer to the original sources for specific usage terms and citation requirements. This combined dataset is provided for research and educational purposes only.\n",
+      "hasCreatorName": true,
+      "hasCurrentVersionNumber": true,
+      "hasDescription": true,
+      "hasLicenseName": true,
+      "hasSubtitle": true,
+      "hasTitle": true,
+      "lastUpdated": "2024-06-03T19:18:21.307Z",
+      "licenseName": "CC BY-SA 4.0",
+      "licenseNameNullable": "CC BY-SA 4.0",
+      "subtitle": "Learnify Research Lab",
+      "subtitleNullable": "Learnify Research Lab",
+      "title": "Fire Detection",
+      "titleNullable": "Fire Detection",
+      "versions": [
+        {
+          "creationDate": "2024-06-03T19:18:21.307Z",
+          "creatorName": "Sharia Arfin Tanim",
+          "creatorNameNullable": "Sharia Arfin Tanim",
+          "creatorRef": "fire-detection",
+          "creatorRefNullable": "fire-detection",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Update 2024-06-03",
+          "versionNotesNullable": "Update 2024-06-03",
+          "versionNumber": 5
+        },
+        {
+          "creationDate": "2024-06-02T17:09:44.043Z",
+          "creatorName": "Sharia Arfin Tanim",
+          "creatorNameNullable": "Sharia Arfin Tanim",
+          "creatorRef": "fire-detection",
+          "creatorRefNullable": "fire-detection",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Update 2024-06-02",
+          "versionNotesNullable": "Update 2024-06-02",
+          "versionNumber": 4
+        },
+        {
+          "creationDate": "2024-06-02T16:17:42.9Z",
+          "creatorName": "Sharia Arfin Tanim",
+          "creatorNameNullable": "Sharia Arfin Tanim",
+          "creatorRef": "fire-detection",
+          "creatorRefNullable": "fire-detection",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Update 2024-06-02",
+          "versionNotesNullable": "Update 2024-06-02",
+          "versionNumber": 3
+        },
+        {
+          "creationDate": "2024-06-02T16:10:21.567Z",
+          "creatorName": "Sharia Arfin Tanim",
+          "creatorNameNullable": "Sharia Arfin Tanim",
+          "creatorRef": "fire-detection",
+          "creatorRefNullable": "fire-detection",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Version 2",
+          "versionNotesNullable": "Version 2",
+          "versionNumber": 2
+        },
+        {
+          "creationDate": "2024-06-02T15:47:41.943Z",
+          "creatorName": "Sharia Arfin Tanim",
+          "creatorNameNullable": "Sharia Arfin Tanim",
+          "creatorRef": "fire-detection",
+          "creatorRefNullable": "fire-detection",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Initial release",
+          "versionNotesNullable": "Initial release",
+          "versionNumber": 1
+        }
+      ]
+    },
+    "license_status": "mixed_upstream_licenses",
+    "retrieved_at": "2026-10-01T23:36:20.024956+00:00",
+    "source_url": "https://www.kaggle.com/datasets/arfintanim/fire-detection"
+  },
+  "source_metadata_sha256": "83717375b0d206b9770d136b5edd6c3ed817353d1874b07b1e9993c8d6c31c9a"
+}

provenance/negative/brycehopkins--flame-3-computer-vision-subset-sycan-marsh.json

@@ -0,0 +1,101 @@
+{
+  "authority_sha256": "6fe912334677393474ed297e669ef41f41551930eea46e17f3dcb91ebed40dac",
+  "classification_evidence": {
+    "classification_audit": {
+      "allow_detection_mapping": false,
+      "boxes_usable": false,
+      "classification_allowed_labels": [
+        "fire",
+        "nofire"
+      ],
+      "classification_allowed_modalities": [
+        "rgb"
+      ],
+      "classification_usable": true,
+      "decision": "needs_correction",
+      "detection_action": "Add dedicated smoke/fire boxes before detection integration; retain thermal native bytes/calibration",
+      "evidence": "audit/samples/brycehopkins--flame-3-computer-vision-subset-sycan-marsh/sample.json and sheet-01/02/03.jpg",
+      "findings": "Six RGB samples (1,3,5,7,9,11) match fire-event versus no-event: positive images often show smoke without visible flame. No boxes. The six thermal samples must remain a distinct modality; hot regions do not alone establish visible flame and grayscale TIFF previews need radiometric calibration/normalization for reliable review. Thermal sample 12 labeled No Fire still shows strong colored heat contrast; do not treat it as a combustion-free negative without temporal/source definition. Multiple raw/corrected-FOV and thermal representations of one acquisition must share the same split group.",
+      "presence_semantics": "wildfire-event-presence-including-smoke",
+      "review_kind": "root_visual_sample_review",
+      "reviewed_at": "2026-10-02T00:37:02.055414+00:00",
+      "reviewed_samples": [
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12
+      ],
+      "source_labels": [
+        "Fire",
+        "No Fire"
+      ],
+      "target_schema": "separate aerial RGB and radiometric/thermal classification",
+      "thermal_visual_label_validation_completed": false,
+      "training_split_requires_rebuild": true,
+      "validation_scope": "12-image visual sample across RGB and thermal; only RGB classification approved by sample"
+    },
+    "path": "/content/drive/MyDrive/FireViewer/datasets-20261002/collected/a-annoter-ou-corriger/brycehopkins--flame-3-computer-vision-subset-sycan-marsh/v1/inventory.jsonl",
+    "sha256": "486a00ca61e5b01fb2b6ee394d0deb1964735617475413e5fdab2fa55761db6d",
+    "source_metadata_sha256": "88ea0ed8ba48b609fe3b52b662460d31fa935f83f318102d7ce9d6e9ac9c78fa"
+  },
+  "individual_visual_validation_claimed": false,
+  "manual_review_enabled": false,
+  "original_labels_and_splits_preserved": true,
+  "root_repository_license_does_not_replace_source_licenses": true,
+  "source": "brycehopkins/flame-3-computer-vision-subset-sycan-marsh",
+  "source_metadata": {
+    "frozen_download_version": 1,
+    "group": "aerial-thermal",
+    "kaggle_metadata": {
+      "creatorName": "Bryce Hopkins",
+      "creatorNameNullable": "Bryce Hopkins",
+      "currentVersionNumber": 1,
+      "currentVersionNumberNullable": 1,
+      "description": "FLAME 3 is the third dataset in the FLAME series of aerial UAV-collected side-by-side multi-spectral wildlands fire imagery (see [FLAME 1](https://ieee-dataport.org/open-access/flame-dataset-aerial-imagery-pile-burn-detection-using-drones-uavs) and [FLAME 2](https://ieee-dataport.org/open-access/flame-2-fire-detection-and-modeling-aerial-multi-spectral-image-dataset)). This set is a single-burn subset of the Larger FLAME 3 dataset focusing specifically on Computer Vision tasks such as fire detection and segmentation. Included are 622 image quartets labeled Fire and 116 image quartets labeled No Fire. The No Fire images are of the surrounding forestry of the prescribed burn plot. Each image quartet is composed of four images - a raw RGB image, a raw thermal image, a corrected FOV RGB image, and a thermal TIFF. Each of the four data types are detailed in the below Table 1. More information on data collection methods, data processing procedures, and data labeling can be found in https://arxiv.org/abs/2412.02831. \n\nThis dataset is based on the research conducted in the paper: FLAME 3 Dataset: Unleashing the Power of Radiometric Thermal UAV Imagery for Wildfire Management. It provides detailed insights and analysis related to forest fire monitoring and modeling.\n\nIf you use this dataset in your research or projects, please cite the original paper as follows:\nAPA: \nHopkins, B., ONeill, L., Marinaccio, M., Rowell, E., Parsons, R., Flanary, S., Nazim I, Seielstad C, Afghah, F. (2024). FLAME 3 Dataset: Unleashing the Power of Radiometric Thermal UAV Imagery for Wildfire Management. arXiv preprint arXiv:2412.02831.\n\nBibTeX:\n@misc{hopkins2024flame3datasetunleashing,\n      title={FLAME 3 Dataset: Unleashing the Power of Radiometric Thermal UAV Imagery for Wildfire Management}, \n      author={Bryce Hopkins and Leo ONeill and Michael Marinaccio and Eric Rowell and Russell Parsons and Sarah Flanary and Irtija Nazim and Carl Seielstad and Fatemeh Afghah},\n      year={2024},\n      eprint={2412.02831},\n      archivePrefix={arXiv},\n      primaryClass={cs.CV},\n      url={https://arxiv.org/abs/2412.02831}, \n}\n\n\n**Table 1: FLAME 3 Datatypes**\n| Data Type | Resolution | File Extension | Description |\n| --- | --- | --- | --- |\n| Raw RGB Image | 4000 x 3000 | .JPG | Unprocessed visual spectrum image collected simultaneously with the Raw Thermal Image. |\n| Raw Thermal Image | 640 x 512 | .JPG | Thermal image collected simultaneously with the Raw RGB Image. Regenerated from the Thermal TIFF to ensure thermal color palette consistency across dataset (preserving JPG meta-data). |\n| Corrected FOV RGB Image | 640 x 512 | .JPG | Copy of Raw RGB Image that has been translated, scaled, and cropped such that the image's pixels align with the Raw Thermal Image. |\n| Thermal TIFF | 640 x 512 | .TIFF | Raster of per-pixel Celsius temperature values. Maximum temperature saturates at approximately 500 degrees Celsius. |\n\nAll of the images contained in this FLAME 3 subset were collected from a single prescribed fire at Sycan Marsh, Oregon between 10/25/23 and 10/27/23. All images were collected with a DJI M30T UAV.",
+      "descriptionNullable": "FLAME 3 is the third dataset in the FLAME series of aerial UAV-collected side-by-side multi-spectral wildlands fire imagery (see [FLAME 1](https://ieee-dataport.org/open-access/flame-dataset-aerial-imagery-pile-burn-detection-using-drones-uavs) and [FLAME 2](https://ieee-dataport.org/open-access/flame-2-fire-detection-and-modeling-aerial-multi-spectral-image-dataset)). This set is a single-burn subset of the Larger FLAME 3 dataset focusing specifically on Computer Vision tasks such as fire detection and segmentation. Included are 622 image quartets labeled Fire and 116 image quartets labeled No Fire. The No Fire images are of the surrounding forestry of the prescribed burn plot. Each image quartet is composed of four images - a raw RGB image, a raw thermal image, a corrected FOV RGB image, and a thermal TIFF. Each of the four data types are detailed in the below Table 1. More information on data collection methods, data processing procedures, and data labeling can be found in https://arxiv.org/abs/2412.02831. \n\nThis dataset is based on the research conducted in the paper: FLAME 3 Dataset: Unleashing the Power of Radiometric Thermal UAV Imagery for Wildfire Management. It provides detailed insights and analysis related to forest fire monitoring and modeling.\n\nIf you use this dataset in your research or projects, please cite the original paper as follows:\nAPA: \nHopkins, B., ONeill, L., Marinaccio, M., Rowell, E., Parsons, R., Flanary, S., Nazim I, Seielstad C, Afghah, F. (2024). FLAME 3 Dataset: Unleashing the Power of Radiometric Thermal UAV Imagery for Wildfire Management. arXiv preprint arXiv:2412.02831.\n\nBibTeX:\n@misc{hopkins2024flame3datasetunleashing,\n      title={FLAME 3 Dataset: Unleashing the Power of Radiometric Thermal UAV Imagery for Wildfire Management}, \n      author={Bryce Hopkins and Leo ONeill and Michael Marinaccio and Eric Rowell and Russell Parsons and Sarah Flanary and Irtija Nazim and Carl Seielstad and Fatemeh Afghah},\n      year={2024},\n      eprint={2412.02831},\n      archivePrefix={arXiv},\n      primaryClass={cs.CV},\n      url={https://arxiv.org/abs/2412.02831}, \n}\n\n\n**Table 1: FLAME 3 Datatypes**\n| Data Type | Resolution | File Extension | Description |\n| --- | --- | --- | --- |\n| Raw RGB Image | 4000 x 3000 | .JPG | Unprocessed visual spectrum image collected simultaneously with the Raw Thermal Image. |\n| Raw Thermal Image | 640 x 512 | .JPG | Thermal image collected simultaneously with the Raw RGB Image. Regenerated from the Thermal TIFF to ensure thermal color palette consistency across dataset (preserving JPG meta-data). |\n| Corrected FOV RGB Image | 640 x 512 | .JPG | Copy of Raw RGB Image that has been translated, scaled, and cropped such that the image's pixels align with the Raw Thermal Image. |\n| Thermal TIFF | 640 x 512 | .TIFF | Raster of per-pixel Celsius temperature values. Maximum temperature saturates at approximately 500 degrees Celsius. |\n\nAll of the images contained in this FLAME 3 subset were collected from a single prescribed fire at Sycan Marsh, Oregon between 10/25/23 and 10/27/23. All images were collected with a DJI M30T UAV.",
+      "hasCreatorName": true,
+      "hasCurrentVersionNumber": true,
+      "hasDescription": true,
+      "hasLicenseName": true,
+      "hasSubtitle": true,
+      "hasTitle": true,
+      "lastUpdated": "2024-05-21T21:56:10.413Z",
+      "licenseName": "MIT",
+      "licenseNameNullable": "MIT",
+      "subtitle": " UAV-Collected Radiometric Thermal and Visual Spectrum Wildfire Imagery",
+      "subtitleNullable": " UAV-Collected Radiometric Thermal and Visual Spectrum Wildfire Imagery",
+      "title": "FLAME 3 Computer Vision Subset (Sycan Marsh)",
+      "titleNullable": "FLAME 3 Computer Vision Subset (Sycan Marsh)",
+      "versions": [
+        {
+          "creationDate": "2024-05-21T21:56:10.413Z",
+          "creatorName": "Bryce Hopkins",
+          "creatorNameNullable": "Bryce Hopkins",
+          "creatorRef": "flame-3-computer-vision-subset-sycan-marsh",
+          "creatorRefNullable": "flame-3-computer-vision-subset-sycan-marsh",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Initial release",
+          "versionNotesNullable": "Initial release",
+          "versionNumber": 1
+        }
+      ]
+    },
+    "license_status": "clear",
+    "retrieved_at": "2026-10-01T23:25:13.895144+00:00",
+    "source_url": "https://www.kaggle.com/datasets/brycehopkins/flame-3-computer-vision-subset-sycan-marsh"
+  },
+  "source_metadata_sha256": "88ea0ed8ba48b609fe3b52b662460d31fa935f83f318102d7ce9d6e9ac9c78fa"
+}

provenance/negative/busier--wildfire.json

@@ -0,0 +1,95 @@
+{
+  "authority_sha256": "6fe912334677393474ed297e669ef41f41551930eea46e17f3dcb91ebed40dac",
+  "classification_evidence": {
+    "classification_audit": {
+      "allow_detection_mapping": false,
+      "boxes_usable": false,
+      "classification_allowed_labels": [
+        "fire",
+        "nofire"
+      ],
+      "classification_usable": true,
+      "decision": "usable",
+      "detection_action": "Boxes must be newly annotated before detection use",
+      "evidence": "audit/samples/busier--wildfire/sample.json and sheet-01/02/03.jpg",
+      "findings": "The twelve binary labels appear consistent. Positive samples include forest fire, urban fire and cooking flame; negatives include forests, fog, animals and flowers. No boxes. General fire classification, not a wildfire-only or smoke-visible detection corpus. Some images are low resolution and watermarked; retain them unchanged and retain provenance.",
+      "review_kind": "root_visual_sample_review",
+      "reviewed_samples": [
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12
+      ],
+      "source_labels": [
+        "Fire",
+        "Nofire",
+        "NoFire"
+      ],
+      "target_schema": "classification-fire-nofire",
+      "validation_scope": "12-image sample only; no claim that all source labels are correct"
+    },
+    "path": "/content/drive/MyDrive/FireViewer/datasets-20261002/collected/exploitable-controle-par-lot/busier--wildfire/v1/inventory.jsonl",
+    "sha256": "7a5476ed293b60d24a756d1cc7ff110f85da852ed606b0a2327a4e3b1897fb9a",
+    "source_metadata_sha256": "1a2f7d5714534f9730e7da30ef0dc89f8937adee10c7382e0184ca48ff29afbc"
+  },
+  "individual_visual_validation_claimed": false,
+  "manual_review_enabled": false,
+  "original_labels_and_splits_preserved": true,
+  "root_repository_license_does_not_replace_source_licenses": true,
+  "source": "busier/wildfire",
+  "source_metadata": {
+    "frozen_download_version": 1,
+    "group": "real-rgb-unclassified",
+    "kaggle_metadata": {
+      "creatorName": "Rafael Amaro",
+      "creatorNameNullable": "Rafael Amaro",
+      "currentVersionNumber": 1,
+      "currentVersionNumberNullable": 1,
+      "description": "This dataset was designed to train machine learning models aimed at detecting forest fires, within the classification problem category. The related project and implementation code can be found at: https://github.com/rafaelolimpioamaro/Deep-Learning-for-Forest-Fire-Detection.git.",
+      "descriptionNullable": "This dataset was designed to train machine learning models aimed at detecting forest fires, within the classification problem category. The related project and implementation code can be found at: https://github.com/rafaelolimpioamaro/Deep-Learning-for-Forest-Fire-Detection.git.",
+      "hasCreatorName": true,
+      "hasCurrentVersionNumber": true,
+      "hasDescription": true,
+      "hasLicenseName": true,
+      "hasSubtitle": true,
+      "hasTitle": true,
+      "lastUpdated": "2025-04-25T17:27:08.033Z",
+      "licenseName": "CC0: Public Domain",
+      "licenseNameNullable": "CC0: Public Domain",
+      "subtitle": "A dataset for training and evaluating forest fire detection models.",
+      "subtitleNullable": "A dataset for training and evaluating forest fire detection models.",
+      "title": "WildFire",
+      "titleNullable": "WildFire",
+      "versions": [
+        {
+          "creationDate": "2025-04-25T17:27:08.033Z",
+          "creatorName": "Rafael Amaro",
+          "creatorNameNullable": "Rafael Amaro",
+          "creatorRef": "wildfire",
+          "creatorRefNullable": "wildfire",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Initial release",
+          "versionNotesNullable": "Initial release",
+          "versionNumber": 1
+        }
+      ]
+    },
+    "license_status": "clear",
+    "retrieved_at": "2026-10-01T23:11:01.017330+00:00",
+    "source_url": "https://www.kaggle.com/datasets/busier/wildfire"
+  },
+  "source_metadata_sha256": "1a2f7d5714534f9730e7da30ef0dc89f8937adee10c7382e0184ca48ff29afbc"
+}

provenance/negative/elmadafri--the-wildfire-dataset.json

@@ -0,0 +1,126 @@
+{
+  "authority_sha256": "6fe912334677393474ed297e669ef41f41551930eea46e17f3dcb91ebed40dac",
+  "classification_evidence": {
+    "classification_audit": {
+      "allow_detection_mapping": false,
+      "boxes_usable": false,
+      "classification_allowed_labels": [
+        "fire",
+        "nofire"
+      ],
+      "classification_usable": true,
+      "decision": "usable",
+      "evidence": "audit/samples/elmadafri--the-wildfire-dataset/sample.json and sheet-01/02/03.jpg",
+      "findings": "Six positive samples depict wildfire flames, rising smoke or incident haze; sample 7 shows an orange smoky canyon but no discrete flame, so the source is event-presence classification. Six nofire samples contain normal forest, sunlight, fog or sunset without a visible incident. No box annotations. Sampling includes distant ground smoke suitable for future detection annotation; preserve original photographic provenance.",
+      "presence_semantics": "wildfire-event-presence-including-smoke",
+      "review_kind": "root_visual_sample_review",
+      "reviewed_samples": [
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12
+      ],
+      "source_labels": [
+        "fire",
+        "nofire"
+      ],
+      "target_schema": "classification-fire-nofire",
+      "validation_scope": "12-image sample only; not exhaustive"
+    },
+    "path": "/content/drive/MyDrive/FireViewer/datasets-20261002/collected/exploitable-controle-par-lot/elmadafri--the-wildfire-dataset/v3/inventory.jsonl",
+    "sha256": "d3624be0c26c3ef66a321f74898903ea7700c2b3669882a108683e27ffeb55be",
+    "source_metadata_sha256": "dbb67e08eae23ca06c130622009555402ed247a5c96459af4d4481cc330906b5"
+  },
+  "individual_visual_validation_claimed": false,
+  "manual_review_enabled": false,
+  "original_labels_and_splits_preserved": true,
+  "root_repository_license_does_not_replace_source_licenses": true,
+  "source": "elmadafri/the-wildfire-dataset",
+  "source_metadata": {
+    "frozen_download_version": 3,
+    "group": "mixed-ground-aerial",
+    "kaggle_metadata": {
+      "creatorName": "Ismail El madafri",
+      "creatorNameNullable": "Ismail El madafri",
+      "currentVersionNumber": 3,
+      "currentVersionNumberNullable": 3,
+      "description": "Thank you for your interest in and use of the Wildfire dataset.  If you find this dataset valuable in your research, we kindly ask that you cite **the original article** where it was first introduced, rather than the URL to this Kaggle page. \n\n**Please cite:** El-Madafri I, Peña M, Olmedo-Torre N. The Wildfire Dataset: Enhancing Deep Learning-Based Forest Fire Detection with a Diverse Evolving Open-Source Dataset Focused on Data Representativeness and a Novel Multi-Task Learning Approach. Forests. 2023; 14(9):1697. https://doi.org/10.3390/f14091697 \n\nThe wildfire dataset symbolizes a concerted effort to explore the capabilities of RGB imagery in the realm of forest fire detection through machine learning methodologies. Spanning 2,700 aerial and ground-based images, it has been curated from a diverse array of online platforms such as government databases, Flickr, and Unsplash. By capturing a wide spectrum of environmental scenarios, forest variants, geographical locations, and the intricate dynamics of forest ecosystems and fire events, the dataset stands as a thoughtful benchmark for research in forest fire detection.\n\nEvery image within this dataset is sourced from the Public Domain, emphasizing the dataset's commitment to transparency. Users can delve into the specifics, accessing detailed information about the URL origins and resolutions for each image. A defining feature of our research is the innovative Multi-Task Learning framework that integrates multi-class confounding elements, crafted specifically to refine forest fire detection. This methodology aims to bolster the model's accuracy and reduce false alarms. Its merits become particularly evident when benchmarked against traditional classification techniques.\n\n**Dataset Highlights:**\n\n**Resolution Insights:**\n\n- **Average:** 4057×3155 pixels\n- **Minimum:** 153×206 pixels\n- **Maximum:** 19699×8974 pixels\n- **Standard Deviation (Width):** 1867.47 pixels\n- **Standard Deviation (Height):** 1388.60 pixels\n\nThese specifications underline the dataset's commitment to providing detailed and quality-oriented imagery apt for in-depth machine learning analysis.\n\n- **Scale Variability:** The collection mirrors the heterogeneity of its sources, showcasing images of diverse real-world scales.\n- **Dataset Evolvement:** Recognizing the need for adaptability, this dataset is dynamic. Continuous enrichment with images, videos, and diverse data types is anticipated, depending on user feedback and requirements.\n\nAs a gesture of fostering open research while acknowledging intellectual contributions, this dataset is licensed under the CC BY 4.0 License. We kindly request that users, when leveraging this dataset in their work or publications, extend the courtesy of citing the associated research paper. \n\n**Citation:** El-Madafri I, Peña M, Olmedo-Torre N. The Wildfire Dataset: Enhancing Deep Learning-Based Forest Fire Detection with a Diverse Evolving Open-Source Dataset Focused on Data Representativeness and a Novel Multi-Task Learning Approach. Forests. 2023; 14(9):1697. \nhttps://doi.org/10.3390/f14091697 \n\n### Authors and Affiliations\n\nThis dataset has been created by:\n\n**Ismail El-Madafri**\n Universitat Politècnica de Catalunya - BarcelonaTech (UPC)\n\n**Marta Peña**\n Universitat Politècnica de Catalunya - BarcelonaTech (UPC)\n\n**Noelia Olmedo-Torre**\n Universitat Politècnica de Catalunya - BarcelonaTech (UPC)\n\nPlease cite this work appropriately in any publications or projects where you use this dataset.",
+      "descriptionNullable": "Thank you for your interest in and use of the Wildfire dataset.  If you find this dataset valuable in your research, we kindly ask that you cite **the original article** where it was first introduced, rather than the URL to this Kaggle page. \n\n**Please cite:** El-Madafri I, Peña M, Olmedo-Torre N. The Wildfire Dataset: Enhancing Deep Learning-Based Forest Fire Detection with a Diverse Evolving Open-Source Dataset Focused on Data Representativeness and a Novel Multi-Task Learning Approach. Forests. 2023; 14(9):1697. https://doi.org/10.3390/f14091697 \n\nThe wildfire dataset symbolizes a concerted effort to explore the capabilities of RGB imagery in the realm of forest fire detection through machine learning methodologies. Spanning 2,700 aerial and ground-based images, it has been curated from a diverse array of online platforms such as government databases, Flickr, and Unsplash. By capturing a wide spectrum of environmental scenarios, forest variants, geographical locations, and the intricate dynamics of forest ecosystems and fire events, the dataset stands as a thoughtful benchmark for research in forest fire detection.\n\nEvery image within this dataset is sourced from the Public Domain, emphasizing the dataset's commitment to transparency. Users can delve into the specifics, accessing detailed information about the URL origins and resolutions for each image. A defining feature of our research is the innovative Multi-Task Learning framework that integrates multi-class confounding elements, crafted specifically to refine forest fire detection. This methodology aims to bolster the model's accuracy and reduce false alarms. Its merits become particularly evident when benchmarked against traditional classification techniques.\n\n**Dataset Highlights:**\n\n**Resolution Insights:**\n\n- **Average:** 4057×3155 pixels\n- **Minimum:** 153×206 pixels\n- **Maximum:** 19699×8974 pixels\n- **Standard Deviation (Width):** 1867.47 pixels\n- **Standard Deviation (Height):** 1388.60 pixels\n\nThese specifications underline the dataset's commitment to providing detailed and quality-oriented imagery apt for in-depth machine learning analysis.\n\n- **Scale Variability:** The collection mirrors the heterogeneity of its sources, showcasing images of diverse real-world scales.\n- **Dataset Evolvement:** Recognizing the need for adaptability, this dataset is dynamic. Continuous enrichment with images, videos, and diverse data types is anticipated, depending on user feedback and requirements.\n\nAs a gesture of fostering open research while acknowledging intellectual contributions, this dataset is licensed under the CC BY 4.0 License. We kindly request that users, when leveraging this dataset in their work or publications, extend the courtesy of citing the associated research paper. \n\n**Citation:** El-Madafri I, Peña M, Olmedo-Torre N. The Wildfire Dataset: Enhancing Deep Learning-Based Forest Fire Detection with a Diverse Evolving Open-Source Dataset Focused on Data Representativeness and a Novel Multi-Task Learning Approach. Forests. 2023; 14(9):1697. \nhttps://doi.org/10.3390/f14091697 \n\n### Authors and Affiliations\n\nThis dataset has been created by:\n\n**Ismail El-Madafri**\n Universitat Politècnica de Catalunya - BarcelonaTech (UPC)\n\n**Marta Peña**\n Universitat Politècnica de Catalunya - BarcelonaTech (UPC)\n\n**Noelia Olmedo-Torre**\n Universitat Politècnica de Catalunya - BarcelonaTech (UPC)\n\nPlease cite this work appropriately in any publications or projects where you use this dataset.",
+      "hasCreatorName": true,
+      "hasCurrentVersionNumber": true,
+      "hasDescription": true,
+      "hasLicenseName": true,
+      "hasSubtitle": true,
+      "hasTitle": true,
+      "lastUpdated": "2024-03-17T00:12:31.783Z",
+      "licenseName": "Attribution 4.0 International (CC BY 4.0)",
+      "licenseNameNullable": "Attribution 4.0 International (CC BY 4.0)",
+      "subtitle": "Enhancing Deep Learning-Based Forest Fire Detection with the Wildfire Dataset",
+      "subtitleNullable": "Enhancing Deep Learning-Based Forest Fire Detection with the Wildfire Dataset",
+      "title": "The wildfire dataset",
+      "titleNullable": "The wildfire dataset",
+      "versions": [
+        {
+          "creationDate": "2024-03-17T00:12:31.783Z",
+          "creatorName": "Ismail El madafri",
+          "creatorNameNullable": "Ismail El madafri",
+          "creatorRef": "the-wildfire-dataset",
+          "creatorRefNullable": "the-wildfire-dataset",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Version 3",
+          "versionNotesNullable": "Version 3",
+          "versionNumber": 3
+        },
+        {
+          "creationDate": "2024-03-15T22:04:29.12Z",
+          "creatorName": "Ismail El madafri",
+          "creatorNameNullable": "Ismail El madafri",
+          "creatorRef": "the-wildfire-dataset",
+          "creatorRefNullable": "the-wildfire-dataset",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Update_1: Simplified File Structure for Enhanced Accessibility",
+          "versionNotesNullable": "Update_1: Simplified File Structure for Enhanced Accessibility",
+          "versionNumber": 2
+        },
+        {
+          "creationDate": "2023-08-22T21:09:18.133Z",
+          "creatorName": "Ismail El madafri",
+          "creatorNameNullable": "Ismail El madafri",
+          "creatorRef": "the-wildfire-dataset",
+          "creatorRefNullable": "the-wildfire-dataset",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Initial release",
+          "versionNotesNullable": "Initial release",
+          "versionNumber": 1
+        }
+      ]
+    },
+    "license_status": "clear",
+    "retrieved_at": "2026-10-01T23:12:34.963895+00:00",
+    "source_url": "https://www.kaggle.com/datasets/elmadafri/the-wildfire-dataset"
+  },
+  "source_metadata_sha256": "dbb67e08eae23ca06c130622009555402ed247a5c96459af4d4481cc330906b5"
+}

provenance/negative/johanjohnthomas--corrected-wildfires.json

@@ -0,0 +1,110 @@
+{
+  "authority_sha256": "6fe912334677393474ed297e669ef41f41551930eea46e17f3dcb91ebed40dac",
+  "classification_evidence": {
+    "classification_audit": {
+      "allow_detection_mapping": false,
+      "boxes_usable": false,
+      "classification_allowed_labels": [
+        "fire",
+        "nofire"
+      ],
+      "classification_usable": true,
+      "decision": "usable",
+      "evidence": "audit/samples/johanjohnthomas--corrected-wildfires/sample.json and sheet-01/02/03.jpg",
+      "findings": "The six fire samples show wildfire flame or smoke; sample 1 is an aerial smoke scene, 5 and 7 show smoke without clearly visible flames. Six nofire scenes show sunset, forest, mountains or water without a visible incident. Labels are consistent for wildfire-event presence, not flame-only detection. No box annotations exist. Aerial and ground views must stay explicit.",
+      "presence_semantics": "wildfire-event-presence-including-smoke",
+      "review_kind": "root_visual_sample_review",
+      "reviewed_samples": [
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12
+      ],
+      "source_labels": [
+        "fire",
+        "nofire"
+      ],
+      "target_schema": "classification-fire-nofire",
+      "validation_scope": "12-image sample only; no exhaustive claim"
+    },
+    "path": "/content/drive/MyDrive/FireViewer/datasets-20261002/collected/exploitable-controle-par-lot/johanjohnthomas--corrected-wildfires/v2/inventory.jsonl",
+    "sha256": "777c08ea51743c0bbc29e5910740285c6ba527bb932941381dce5b2f4a5788e8",
+    "source_metadata_sha256": "bb339402173426e0b5e72f98a656eba2e1afcf8fb8ad507b2ee8dfecf03b1db1"
+  },
+  "individual_visual_validation_claimed": false,
+  "manual_review_enabled": false,
+  "original_labels_and_splits_preserved": true,
+  "root_repository_license_does_not_replace_source_licenses": true,
+  "source": "johanjohnthomas/corrected-wildfires",
+  "source_metadata": {
+    "frozen_download_version": 2,
+    "group": "mixed-ground-aerial",
+    "kaggle_metadata": {
+      "creatorName": "Johan John Thomas",
+      "creatorNameNullable": "Johan John Thomas",
+      "currentVersionNumber": 2,
+      "currentVersionNumberNullable": 2,
+      "description": "This data set is taken from this [Kaggle Page](https://www.kaggle.com/datasets/elmadafri/the-wildfire-dataset/data)\nUpon using the above dataset, I've noticed that there were a few misclassified labels, and also a few images which were very unclear, impossible to distinguish. \nThus I manually cleaned it, and removed/reclassified entries which were unclear or mislabelled.\n\nThe original repository has mentioned to cite the paper introducing the dataset, so I will cite it below:\n>El-Madafri I, Peña M, Olmedo-Torre N. The Wildfire Dataset: Enhancing Deep Learning-Based Forest Fire Detection with a Diverse Evolving Open-Source Dataset Focused on Data Representativeness and a Novel Multi-Task Learning Approach. Forests. 2023; 14(9):1697. https://doi.org/10.3390/f14091697\n\n",
+      "descriptionNullable": "This data set is taken from this [Kaggle Page](https://www.kaggle.com/datasets/elmadafri/the-wildfire-dataset/data)\nUpon using the above dataset, I've noticed that there were a few misclassified labels, and also a few images which were very unclear, impossible to distinguish. \nThus I manually cleaned it, and removed/reclassified entries which were unclear or mislabelled.\n\nThe original repository has mentioned to cite the paper introducing the dataset, so I will cite it below:\n>El-Madafri I, Peña M, Olmedo-Torre N. The Wildfire Dataset: Enhancing Deep Learning-Based Forest Fire Detection with a Diverse Evolving Open-Source Dataset Focused on Data Representativeness and a Novel Multi-Task Learning Approach. Forests. 2023; 14(9):1697. https://doi.org/10.3390/f14091697\n\n",
+      "hasCreatorName": true,
+      "hasCurrentVersionNumber": true,
+      "hasDescription": true,
+      "hasLicenseName": true,
+      "hasSubtitle": true,
+      "hasTitle": true,
+      "lastUpdated": "2024-11-01T13:33:26.317Z",
+      "licenseName": "Attribution 4.0 International (CC BY 4.0)",
+      "licenseNameNullable": "Attribution 4.0 International (CC BY 4.0)",
+      "subtitle": "For DeepLearning Based Wildfire Detection with Confounder Images",
+      "subtitleNullable": "For DeepLearning Based Wildfire Detection with Confounder Images",
+      "title": "Wildfires",
+      "titleNullable": "Wildfires",
+      "versions": [
+        {
+          "creationDate": "2024-11-01T13:33:26.317Z",
+          "creatorName": "Johan John Thomas",
+          "creatorNameNullable": "Johan John Thomas",
+          "creatorRef": "corrected-wildfires",
+          "creatorRefNullable": "corrected-wildfires",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Updated Dataset with more cleaned entries",
+          "versionNotesNullable": "Updated Dataset with more cleaned entries",
+          "versionNumber": 2
+        },
+        {
+          "creationDate": "2024-11-01T11:12:07.267Z",
+          "creatorName": "Johan John Thomas",
+          "creatorNameNullable": "Johan John Thomas",
+          "creatorRef": "corrected-wildfires",
+          "creatorRefNullable": "corrected-wildfires",
+          "hasCreatorName": true,
+          "hasCreatorRef": true,
+          "hasStatus": true,
+          "hasVersionNotes": true,
+          "status": "Ready",
+          "statusNullable": "Ready",
+          "versionNotes": "Initial release",
+          "versionNotesNullable": "Initial release",
+          "versionNumber": 1
+        }
+      ]
+    },
+    "license_status": "clear",
+    "retrieved_at": "2026-10-01T23:12:18.707750+00:00",
+    "source_url": "https://www.kaggle.com/datasets/johanjohnthomas/corrected-wildfires"
+  },
+  "source_metadata_sha256": "bb339402173426e0b5e72f98a656eba2e1afcf8fb8ad507b2ee8dfecf03b1db1"
+}

provenance/negative/original-dataset-card.md

@@ -0,0 +1,3 @@
+---
+license: apache-2.0
+---

provenance/v2-source-card.md

@@ -0,0 +1,328 @@
+---
+license: other
+pretty_name: FireViewer corrected fire and smoke corpus V2
+task_categories:
+- object-detection
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*.parquet
+  - split: validation
+    path: data/validation-*.parquet
+  - split: test
+    path: data/test-*.parquet
+---
+
+# FireViewer V2 — corrected corpus
+
+26126 unique native images. Splits: {'test': 332, 'train': 24806, 'validation': 983}.
+
+One final `objects` list per image: COCO xywh in native pixels, category1 smoke_visible / category2 flame_visible, paired source_base points in native pixels or null coordinates when hidden. Original encoded image bytes are embedded. No old predictions or alternative annotation versions are included.
+
+Capture groups remain disjoint across splits. Existing group assignments and the 651 final manual archive versions are preserved. Negative images are limited to 5% of each split. Context fields large_view and wildfire remain available.
+
+The frozen 19,377-image source audit remains in progress. The original corrected cohort contains completed direct visual versions with inspected final renders; native source imports are documented separately below; formerly ambiguous cases require effective principal review. This is not a claim that all originals passed review. Incremental batches contain 1,000 added or updated eligible versions; exclusions and cap withdrawals are recorded separately. Small remainders require explicit closeout. See manifest.json, case-manifest.jsonl and incremental-batches/*/receipt.json.
+
+
+Direct correction batch 5f32ebfe8e3ad7edde3299ab: 5 new images / 10 final objects. Native images and existing annotations were reviewed directly on grouped contact sheets; ambiguous cases are held. No paid model call. Authority: assistant_direct_visual_correction; human acceptance not claimed. Known scene version limit: 5; the global historic near-scene audit remains incomplete.
+
+## Native smoke source import — 2026-10-02
+
+737 native images from `ahemateja19bec1025/wildfiresmokedatasetyolo` (Kaggle revision 2). 0 exact image duplicates skipped. The twelve-image source sample was inspected with native boxes. All added encoded images, hashes, dimensions and annotation geometry were checked; this is not individual visual validation of every image. Source annotations are retained as `native_source_annotations_sample_audited`. This smoke-only source has no located source-base points: coordinates are null with uncertain visibility. All new camera scenes are assigned to train, keeping the existing validation and test benchmark unchanged. Original source splits and provenance remain on Drive. See `source-imports/999d945ed0b84853833c38eb/receipt.json`.
+
+## Individually inspected native source cases — 2026-10-02
+
+28 additional images had their displayed native annotations checked individually. Only the accepted sample ordinals were imported; failed, incomplete and ambiguous cases remain on Drive. This is no claim that the source datasets have been checked in full. All new native bytes, SHA, dimensions, final objects and Parquet roundtrips are verified. Base points are unannotated and remain null/uncertain. Existing validation/test and human versions are unchanged. See `source-imports/e886d5c52a15cb6b596e2dbf/receipt.json` and `visual-evidence.json`. Source splits and provenance remain on Drive. Exact image deduplication is complete for this batch; a global perceptual scene-leakage audit is not claimed.
+
+## Individually inspected native source cases — 2026-10-02
+
+24 additional images had their displayed native annotations checked individually. Only the accepted sample ordinals were imported; failed, incomplete and ambiguous cases remain on Drive. This is no claim that the source datasets have been checked in full. All new native bytes, SHA, dimensions, final objects and Parquet roundtrips are verified. Base points are unannotated and remain null/uncertain. Existing validation/test and human versions are unchanged. See `source-imports/8e5b4714a061e35b3e0c95e3/receipt.json` and `visual-evidence.json`. Source splits and provenance remain on Drive. Exact image deduplication is complete for this batch; a global perceptual scene-leakage audit is not claimed.
+
+## Individually inspected native source cases — 2026-10-02
+
+16 additional images had their displayed native annotations checked individually. Only the accepted sample ordinals were imported; failed, incomplete and ambiguous cases remain on Drive. This is no claim that the source datasets have been checked in full. All new native bytes, SHA, dimensions, final objects and Parquet roundtrips are verified. Base points are unannotated and remain null/uncertain. Existing validation/test and human versions are unchanged. See `source-imports/e4a1068046b1ea63aaf76611/receipt.json` and `visual-evidence.json`. Source splits and provenance remain on Drive. Exact image deduplication is complete for this batch; a global perceptual scene-leakage audit is not claimed.
+
+## Individually inspected native source cases — 2026-10-02
+
+15 additional images had their displayed native annotations checked individually. Only the accepted sample ordinals were imported; failed, incomplete and ambiguous cases remain on Drive. This is no claim that the source datasets have been checked in full. All new native bytes, SHA, dimensions, final objects and Parquet roundtrips are verified. Base points are unannotated and remain null/uncertain. Existing validation/test and human versions are unchanged. See `source-imports/5c06d474c188f296203b1d80/receipt.json` and `visual-evidence.json`. Source splits and provenance remain on Drive. Exact image deduplication is complete for this batch; a global perceptual scene-leakage audit is not claimed.
+
+
+## Native subset with visual sample audit — 2026-10-02
+
+361 native images imported after a coherent subset sample audit. The listed sample ordinals were inspected visually; unseen members were checked for source namespace, one final annotation version, native SHA, decoded dimensions and geometry. This is not individual visual validation of every imported image. Extra variants of the same original frame and known held versions were excluded in the cohort selection. Existing human versions and validation/test remain unchanged. Source-base points remain null/uncertain. Global perceptual leakage audit is not claimed. See source-imports/39a32a05d3912e3975ca52f9/sample-reviews.json and receipt.json.
+
+
+## Native subset with visual sample audit — 2026-10-02
+
+97 native images imported after a coherent subset sample audit. The listed sample ordinals were inspected visually; unseen members were checked for source namespace, one final annotation version, native SHA, decoded dimensions and geometry. This is not individual visual validation of every imported image. Extra variants of the same original frame and known held versions were excluded in the cohort selection. Existing human versions and validation/test remain unchanged. Source-base points remain null/uncertain. Global perceptual leakage audit is not claimed. See source-imports/0ec2a068360f19d351ac8606/sample-reviews.json and receipt.json.
+
+
+## Native subset with visual sample audit — 2026-10-02
+
+563 native images imported after a coherent subset sample audit. The listed sample ordinals were inspected visually; unseen members were checked for source namespace, one final annotation version, native SHA, decoded dimensions and geometry. This is not individual visual validation of every imported image. Extra variants of the same original frame and known held versions were excluded in the cohort selection. Existing human versions and validation/test remain unchanged. Source-base points remain null/uncertain. Global perceptual leakage audit is not claimed. See source-imports/51d2f432cb4c384222fe7dc5/sample-reviews.json and receipt.json.
+
+
+## Native subset with visual sample audit — 2026-10-02
+
+16 native images imported after a coherent subset sample audit. The listed sample ordinals were inspected visually; unseen members were checked for source namespace, one final annotation version, native SHA, decoded dimensions and geometry. This is not individual visual validation of every imported image. Extra variants of the same original frame and known held versions were excluded in the cohort selection. Existing human versions and validation/test remain unchanged. Source-base points remain null/uncertain. Global perceptual leakage audit is not claimed. See source-imports/cc629d90ff57b7d153deda14/sample-reviews.json and receipt.json.
+
+
+## Native subset with visual sample audit — 2026-10-02
+
+66 native images imported after a coherent subset sample audit. The listed sample ordinals were inspected visually; unseen members were checked for source namespace, one final annotation version, native SHA, decoded dimensions and geometry. This is not individual visual validation of every imported image. Extra variants of the same original frame and known held versions were excluded in the cohort selection. Existing human versions and validation/test remain unchanged. Source-base points remain null/uncertain. Global perceptual leakage audit is not claimed. See source-imports/cb368887ce36cbaf1d9e6100/sample-reviews.json and receipt.json.
+
+
+## Native subset with visual sample audit — 2026-10-02
+
+133 native images imported after a coherent subset sample audit. The listed sample ordinals were inspected visually; unseen members were checked for source namespace, one final annotation version, native SHA, decoded dimensions and geometry. This is not individual visual validation of every imported image. Extra variants of the same original frame and known held versions were excluded in the cohort selection. Existing human versions and validation/test remain unchanged. Source-base points remain null/uncertain. Global perceptual leakage audit is not claimed. See source-imports/a962f09eb0a5bca54faec33d/sample-reviews.json and receipt.json.
+
+
+### Human slideshow decisions c07a0bb537065c22ee5f3135
+
+7 native images were individually accepted by the user with their native boxes displayed. All previous V2 cases, human versions, validation and test are preserved. See source-imports/c07a0bb537065c22ee5f3135/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 625852e32245e5f598a333a5
+
+32 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/625852e32245e5f598a333a5/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions ecc98766e3ba23d1754a14ec
+
+32 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/ecc98766e3ba23d1754a14ec/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions cd4bb5b7bd358420bc47a98c
+
+32 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/cd4bb5b7bd358420bc47a98c/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 144353cbdc669811a64076f1
+
+32 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/144353cbdc669811a64076f1/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions eff5c5aafd27e8e2c9eb0d8a
+
+32 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/eff5c5aafd27e8e2c9eb0d8a/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions aefa73775192cfa2e5f83118
+
+32 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/aefa73775192cfa2e5f83118/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 5aecb949fd190090b6a23ee8
+
+25 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/5aecb949fd190090b6a23ee8/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 37e252df1645dac6cec7fce8
+
+2 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/37e252df1645dac6cec7fce8/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions ab58a298a9147699accfccf8
+
+3 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/ab58a298a9147699accfccf8/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 746728275bed835e647b14ff
+
+1 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/746728275bed835e647b14ff/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 3c88fe083bbbd485b999e031
+
+5 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/3c88fe083bbbd485b999e031/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions ce011db222fbdd5c5067f7b2
+
+4 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/ce011db222fbdd5c5067f7b2/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 0dd074ea94fd6a516eb50f1b
+
+5 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/0dd074ea94fd6a516eb50f1b/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 854a8c09c47acd5fe2b1c2bd
+
+5 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/854a8c09c47acd5fe2b1c2bd/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 7a18d7c947e3df905b6c8728
+
+3 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/7a18d7c947e3df905b6c8728/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 81c3c8e3b5b8255a2fcd7f5a
+
+8 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/81c3c8e3b5b8255a2fcd7f5a/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 2c87f7feeb77e154388be4e9
+
+11 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/2c87f7feeb77e154388be4e9/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions e91c86c884dd2247b1a68276
+
+3 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/e91c86c884dd2247b1a68276/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 9a398ea6f0eb8c15b5d71433
+
+3 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/9a398ea6f0eb8c15b5d71433/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 0b00a601d167db9696b78684
+
+3 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/0b00a601d167db9696b78684/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 2f04a74dc62a282b65f0be96
+
+1 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/2f04a74dc62a282b65f0be96/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions a9bc321b99148d1b626a9fb6
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/a9bc321b99148d1b626a9fb6/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 9575fbb9d954be92551b9cf4
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/9575fbb9d954be92551b9cf4/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions c5922907d0cb245b189ac9ff
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/c5922907d0cb245b189ac9ff/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions ed05a1f7193cff05e3a68021
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/ed05a1f7193cff05e3a68021/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 18aaa06b23f2579227a1b584
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/18aaa06b23f2579227a1b584/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions b4293b3eaf42502dee048a4b
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/b4293b3eaf42502dee048a4b/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 6bb994e58e501b7c086e16d6
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/6bb994e58e501b7c086e16d6/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions c169070947a83b478a0430ec
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/c169070947a83b478a0430ec/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions ac93aaf80cb81d2e42d9864e
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/ac93aaf80cb81d2e42d9864e/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions fc8b670d7dd7f136fa6b5619
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/fc8b670d7dd7f136fa6b5619/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 36680c6a7102d0f918ef981c
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/36680c6a7102d0f918ef981c/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions ff201bb57368c099294e4562
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/ff201bb57368c099294e4562/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions f85cf061e471468c4a43e4f6
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/f85cf061e471468c4a43e4f6/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions bee57ae5f8dee4ef6966d147
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/bee57ae5f8dee4ef6966d147/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions dfe1226de47e7d072d9fcd65
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/dfe1226de47e7d072d9fcd65/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions a96e57fc4488206f46b9e9d4
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/a96e57fc4488206f46b9e9d4/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 872fbbd5fb0cd4ac8db3d99a
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/872fbbd5fb0cd4ac8db3d99a/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 6c9be4edbad771e4fe4fab41
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/6c9be4edbad771e4fe4fab41/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 60006ee704b7449e2a0665d7
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/60006ee704b7449e2a0665d7/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 6e49ab198505fe07b19d872b
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/6e49ab198505fe07b19d872b/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions cd53f23ac14be0902f8902b7
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/cd53f23ac14be0902f8902b7/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 97466fcfa0ef3eabae0d2988
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/97466fcfa0ef3eabae0d2988/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 037d25b952ed3e3eab222ad8
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/037d25b952ed3e3eab222ad8/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions c87a58add09b7ccb2ec16877
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/c87a58add09b7ccb2ec16877/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 1632df1dd32aff48aa1f6d5a
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/1632df1dd32aff48aa1f6d5a/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions e4038c5c6ff3ce99d9fabfe9
+
+512 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/e4038c5c6ff3ce99d9fabfe9/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+### Human slideshow decisions 0d5e8eb9a7a7e977d5183dd6
+
+372 native images were accepted by the user, individually or under the explicit batch approval recorded in the evidence. All previous V2 cases, human versions, validation and test are preserved. See source-imports/0d5e8eb9a7a7e977d5183dd6/human-review-evidence.json for all native versions, source licences (including upstream conflicts), native splits and exact namespace proofs. New cases are grouped by whole source in train. Missing source points remain null/uncertain. A global near-scene leakage audit is incomplete.
+
+
+## MiMo proposals with direct assistant correction — aab63e4a9c539b8883a6bfb8
+
+12 previously unpublished native SHA cases were reviewed as a batch. Five cases (14 objects) have one final assistant-corrected annotation list; seven materially uncertain cases remain outside final objects. MiMo supplied proposals only. Native annotations, provenance and original license statuses are preserved in source-imports/aab63e4a9c539b8883a6bfb8/assistant-final-review.json. No human acceptance is claimed. The 26,116 pre-existing cases and native split assignments are preserved. Unknown source points remain null/uncertain. The global near-scene leakage audit remains incomplete.

split-receipt.json

@@ -0,0 +1 @@
+{"conflicting_images_quarantined": 8, "cross_split_detected_near_duplicate_edges": 0, "cross_split_groups": 0, "decoded_pixel_duplicates": 1, "grouping": "native RGB hashes, normalized source origins, known capture groups, conservative FLAME sequence grouping, pHash Hamming<=8 AND dHash<=12 AND aspect-ratio log distance<=0.2", "historical_images_forced_train": 10086, "input_images": 38974, "limitation": "Automated near-duplicate and known-origin separation; no claim of exhaustive incident identity or individual semantic label revalidation.", "metadata_sha256": "bf2e9771e1ce90971b15b119f1f1f74632c9987477afd79c8e87eb5a18fed99d", "near_duplicate_edges": 223857, "schema": "fireviewer-v25-splits-v1", "seed": 42, "splits": {"test": {"groups": 1632, "historical": 0, "images": 3689, "negative_kinds": {"native_negative": 1229, "very_hard_negative": 21}, "objects": {"flame_visible": 2749, "smoke_visible": 2062}, "small_objects": {"flame_visible": 592, "smoke_visible": 63}}, "train": {"groups": 19575, "historical": 10086, "images": 31657, "negative_kinds": {"native_negative": 10165, "v2_negative": 504, "very_hard_negative": 153}, "objects": {"flame_visible": 23939, "smoke_visible": 16894}, "small_objects": {"flame_visible": 5599, "smoke_visible": 507}}, "validation": {"groups": 1695, "historical": 0, "images": 3619, "negative_kinds": {"native_negative": 1259, "very_hard_negative": 21}, "objects": {"flame_visible": 2808, "smoke_visible": 2139}, "small_objects": {"flame_visible": 533, "smoke_visible": 62}}}, "target_fractions": {"test": 0.1, "train": 0.8, "validation": 0.1}, "unique_kept": 38965}