{ "args": { "arm": "rich-span", "corpus": "v3", "out": "runs/parser-t4/e5-s2", "seed": 2, "epochs": 24, "batch": 64, "threads": 4, "device": "cuda", "extra_real": [ "data/label-v2/dev-review-500.jsonl", "data/label-v2/public-dev-spans.jsonl", "data/label-v2/hard-shapes-reviewed.jsonl", "data/label-v2/scene-dotted-reviewed.jsonl", "data/label-v2/path-echo.jsonl", "data/label-v2/path-echo2.jsonl", "data/label-v2/hard-titles.jsonl", "data/label-v2/r2-cjk.jsonl", "data/label-v2/r1b.reviewed.jsonl", "data/label-v2/r1c.reviewed.jsonl", "data/label-v2/r3-misc.jsonl", "data/label-v2/r4-resid.jsonl", "data/label-v2/r5-tail.jsonl", "data/label-v2/r6-stick.jsonl", "data/label-v2/r7-last.jsonl", "data/label-v2/r8-final.jsonl", "data/label-v2/r9-end.jsonl", "data/label-v2/al-v1.jsonl", "data/label-v2/al-v1.jsonl" ], "real_repeat": 5, "encoder_dir": null, "smoke": false, "exposures": 9568, "schedule": "cosine", "warmup": 0.06, "ema": 0.999, "no_export": true, "amp": true, "bucket": true, "char_classes": true, "identity_dropout": [ 0.25, 0.05, 0.5 ], "augment_real": 0.5, "select_real": "data/label-v2/fresh-204-gold.jsonl", "semicrf": 0.5, "uniform_classes": true, "teacher": null, "kd_weight": 0.0, "kd_temperature": 2.0, "extra_real_rows": 2291, "extra_real_rows_per_file": { "data/label-v2/dev-review-500.jsonl": 500, "data/label-v2/public-dev-spans.jsonl": 146, "data/label-v2/hard-shapes-reviewed.jsonl": 111, "data/label-v2/scene-dotted-reviewed.jsonl": 62, "data/label-v2/path-echo.jsonl": 30, "data/label-v2/path-echo2.jsonl": 60, "data/label-v2/hard-titles.jsonl": 94, "data/label-v2/r2-cjk.jsonl": 142, "data/label-v2/r1b.reviewed.jsonl": 50, "data/label-v2/r1c.reviewed.jsonl": 80, "data/label-v2/r3-misc.jsonl": 132, "data/label-v2/r4-resid.jsonl": 115, "data/label-v2/r5-tail.jsonl": 85, "data/label-v2/r6-stick.jsonl": 123, "data/label-v2/r7-last.jsonl": 98, "data/label-v2/r8-final.jsonl": 80, "data/label-v2/r9-end.jsonl": 125, "data/label-v2/al-v1.jsonl": 129 }, "extra_real_sha256": { "data/label-v2/dev-review-500.jsonl": "87dd284d9ade8758f94ce260603f032ea0bdb3b95370c65b6cc1e61d17ab913e", "data/label-v2/public-dev-spans.jsonl": "3406b81d405cc6d38eaa01eb6072805307f18b8f939f3818f78255be4d7a1673", "data/label-v2/hard-shapes-reviewed.jsonl": "b4a7b5f49afb9e63d33689ae6677dc4f579f7bda65fbb996c07f9d69503a325f", "data/label-v2/scene-dotted-reviewed.jsonl": "b69409cd92afed85f56b3a9acaac4ac17abed55a44df765931c1edaf6fc65aac", "data/label-v2/path-echo.jsonl": "111da563ebbb60848d5890f2761eae719b49494df74a4ab186aaf95d5bcf07ea", "data/label-v2/path-echo2.jsonl": "19cff6261543fa666a4fb852455d83809272116a7a71d2eacac9d71c50d4f04d", "data/label-v2/hard-titles.jsonl": "0a830ca99fff9ea932c1f911787139487e907c88753d3fda9991bb95ba0a9f4a", "data/label-v2/r2-cjk.jsonl": "1f516b667d7f5ac32343f6a90cf0655e8427bf5ab1e6505f26d2242fdf559817", "data/label-v2/r1b.reviewed.jsonl": "725101d9d544dfd77b2608885f298ca988cf6d7686e7309cfee76f9b0c9ccd8f", "data/label-v2/r1c.reviewed.jsonl": "f474112deb753220730eaafa2ecf18a886097e09fa5ff43c6ac7592506380b7b", "data/label-v2/r3-misc.jsonl": "ca84ef3f5070b113a50aae9ab961c186c95bfe839412c6f69bf4e02d4956da93", "data/label-v2/r4-resid.jsonl": "76a13288c5728bede229ebb101ed2a4c25877140f98753507073b20580eeb276", "data/label-v2/r5-tail.jsonl": "162dec17ea44a25cc4b349a895c4d32b2e6ef69ee1c2afb82b1c3318b3b8d317", "data/label-v2/r6-stick.jsonl": "a6c6a3e1654a913b03c342a08d46b14d4fb4b5f8a13c07bc31adabadc0e9b2a0", "data/label-v2/r7-last.jsonl": "695223fd7ded82b6a7e11554a5072602b50ad26d431e0b0ab7b95a3a251fcd8a", "data/label-v2/r8-final.jsonl": "265ede3d516000774c70eb6bac3355f8796a7ef98fadff20293aba75c7619f03", "data/label-v2/r9-end.jsonl": "db641b57cd4ac4b59bb914a98f67813887d3d38322f9cdca5f8282a4ccdfdc72", "data/label-v2/al-v1.jsonl": "2bd984716f0d3fd23419a0f2c539eb234f589bdd86d84d1e0d457d2e3a7a2781" }, "training_skipped_over_limit": 0, "validation_skipped_over_limit": 0 }, "representation": "unicode", "data": "rich", "head": "span", "primary": "span", "exposures_per_epoch": 4784, "training_pool_rows": 40455, "validation_rows": 1450, "claim": "synthetic development selection only", "versions": { "torch": "2.11.0+cu128", "transformers": "5.16.1", "numpy": "2.1.3" }, "python": "3.13.15", "device": "cuda", "cuda": "12.8", "gpu": "Tesla T4", "source_hashes": { "fp/parser3.py": "6c26270acc5234f625378aa2cd5935fa434507aa3ffa4ebd208959489ae58176", "fp/parser3_model.py": "be0ef67cf753a730b7fa86538d36fe50cfddea53c71b70ac7a29dee1cbaf39ee", "fp/corpus3.py": "f8ca257fe716b6d933e3d324fdbcc65a9b399c80f461ae0a8250278a3de684de", "fp/corpus2.py": "a1953d116fa81f0a3103d25d5a20ba7216adc7405aab337888165f7b0a636893", "fp/normalize.py": "7e5d096f9f36f2d3d289bf3d75b6d680bbfa17339431c154a00dc7fe8268ee01", "fp/parser2.py": "89abf39364c34d1e159d56553f1044cce06040d23898c6b8c5fa849b4dae7efb", "docs/PARSER-V3.md": "43909edbfe8b27295c039c8f7c69ffbaa65e6ad91f5e3342479505de78100f3a", "schema.json": "b42b525b313e922b2e6c6d85bcc28573f89502d9ee2ebb9a63d1b475ae8ab8a1" }, "snapshot": "/content/anifilebert-dapt512-long", "snapshot_hashes": { "anifilebert/__init__.py": "56c1ffe657abefbdf36b1282e86b654c0e198ed1518ec4063ad36ef9ce9061ec", "anifilebert/config.py": "9761b6b8b298da8484196e938e4bf7347f337fd68ca4f490cafd649481a1073b", "anifilebert/dataset.py": "9359204fbd4e055b38d9c712735331137d3a8170fb70ebf88badaaa0e47b2b2b", "anifilebert/inference.py": "5a0684569478a26054e858d78713c981364dc2728090b23f68c8bd0e0edb8586", "anifilebert/label_repairs.py": "9859860f37af5535429337dbf72edc2beedbca634c2d17090b28084e226bd42b", "anifilebert/labels.py": "8c456f499baa0328a135059c0846c297cd39b88abcd578ebad552d74d4d29fd7", "anifilebert/model.py": "d66715295cefe93b3689f30becba072161380b29c4ba43ea163473feafc3afca", "anifilebert/tokenizer.py": "5d745c9925ce83af1c9c6c743439aad8adb6c4ac38bea65068bdd7fffa79e879", "anifilebert/train.py": "018f3c06b6a47a67a8afef98bac753c63c6b4bdc7e95d48e408bc52f42a94b0b", "anifilebert/virtual_dataset.py": "3b705f73925bce46dcd622d25823682f608c20fb6ed178d076cd8e2d1f48e0ec", "config.json": "11f764a6f118d237daa8bb39667f8f8557585a8b1847a33ed3cd5e6dc38b6d32", "dapt.json": "ecb05ec5fc6cf72835270f787efbb78de276ed75c065cf2e3ae7f69878835ade", "label_schema.json": "4cea27b41ffab31a8bec446e5baf5e9c5fc9d72236b0359fdf7b9cff1f7221a3", "model.safetensors": "509125db5cc4bee9f3949b44e1c17f86b89406c8efee6a7376c39de8b3db7a0d", "tokenizer_config.json": "3219e20b10e78496c5acdc39e21e8df5e7c56864bf461041a58721903e2650f8", "vocab.char.json": "cd41768013f3077ca63cf5e6171f65c79a44e384e6a36a07cbb930aaea6a32b6", "vocab.json": "e2d7a5a42fb7e10ba7436a3d4fb568f3011bd8a3e61db0d1124e4c2b1507a0ce" }, "encoder_dir": null, "corpus_manifest_sha256": "63f6aa162b2ccc0ed441b705e577311c4a5f7f5bb782681d4ae7a3f8c76c8c12", "corpus": "v3", "corpus2_manifest_sha256": "e175466af0d756ed7edad1df6a622a70ee1299af5772ddbd312872be6cd7a8a6", "parameters": 28964388, "labels": [ "O", "B-TITLE", "I-TITLE", "B-EPISODE_TITLE", "I-EPISODE_TITLE", "B-GROUP", "I-GROUP", "B-SEASON", "I-SEASON", "B-EPISODE", "I-EPISODE", "B-RANGE", "I-RANGE", "B-LIST", "I-LIST", "B-YEAR", "I-YEAR", "B-RESOLUTION", "I-RESOLUTION", "B-DIMENSIONS", "I-DIMENSIONS", "B-DISCARD", "I-DISCARD" ], "unknown_input_train": { "pt": { "rows": 4669, "characters": 337822, "unknown": 0, "byte_fallback": 3111 }, "ru": { "rows": 4669, "characters": 316628, "unknown": 0, "byte_fallback": 30705 }, "es": { "rows": 5002, "characters": 349497, "unknown": 0, "byte_fallback": 2862 }, "ja": { "rows": 4995, "characters": 256590, "unknown": 0, "byte_fallback": 2872 }, "zh": { "rows": 4662, "characters": 243002, "unknown": 0, "byte_fallback": 3148 }, "en": { "rows": 5003, "characters": 343908, "unknown": 0, "byte_fallback": 2886 }, "real": { "rows": 11455, "characters": 632680, "unknown": 0, "byte_fallback": 3740 } }, "unknown_input_validation": { "pt": { "rows": 238, "characters": 18138, "unknown": 0, "byte_fallback": 148 }, "zh": { "rows": 231, "characters": 12464, "unknown": 0, "byte_fallback": 107 }, "en": { "rows": 255, "characters": 17351, "unknown": 0, "byte_fallback": 114 }, "ja": { "rows": 248, "characters": 12935, "unknown": 0, "byte_fallback": 108 }, "es": { "rows": 247, "characters": 17607, "unknown": 0, "byte_fallback": 110 }, "ru": { "rows": 231, "characters": 16171, "unknown": 0, "byte_fallback": 1404 } }, "char_classes": true, "selection": "earliest highest exact-record on --select-real, synthetic validation exact as tie-break" }