{
  "schema_version": 2,
  "created_at_utc": "2026-08-21T19:56:58.229241+00:00",
  "provider_first_built_on": "Lightning AI",
  "shared_data_directory": "/teamspace/studios/this_studio/apertus_shared/swiss-retrieval-data",
  "recipe": {
    "schema_version": 2,
    "smoke_test": false,
    "seed": 42,
    "wiki_date": "20231101",
    "swiss_mono_targets": {
      "de": 6000,
      "en": 4000,
      "fr": 5000,
      "it": 4000,
      "rm": 2500,
      "als": 2500
    },
    "mono_eval_per_lang": 80,
    "voting_train_per_direction": 500,
    "diamond_eval_per_direction": 30,
    "swissgov_train_per_direction": 150,
    "swissgov_eval_per_direction": 20,
    "preservation_targets": {
      "es": 200,
      "pt": 200,
      "nl": 150,
      "pl": 150,
      "tr": 150,
      "ar": 150,
      "hi": 150,
      "zh": 150,
      "sw": 150
    },
    "min_doc_chars": 220,
    "max_doc_chars": 1800,
    "min_parallel_chars": 60,
    "max_parallel_chars": 1600
  },
  "recipe_signature": "fd8f7d87d50416d4d732c68dfffd2ce3de5a58aac879ccb677185d06186c400e",
  "sources": {
    "wikimedia/wikipedia": {
      "url": "https://huggingface.co/datasets/wikimedia/wikipedia",
      "revision": "b04c8d1ceb2f5cd4588862100d08de323dccfbaa"
    },
    "ZurichNLP/SwissGov-RSD": {
      "url": "https://huggingface.co/datasets/ZurichNLP/SwissGov-RSD",
      "revision": "f16aa0536811b8c0cf975fede21eec5996e4b3c1"
    },
    "eljuanina/VotingBooklets-v1": {
      "url": "https://huggingface.co/datasets/eljuanina/VotingBooklets-v1",
      "revision": "8db11c674d6993aee084fe65ed3ab6d9ee102fd7"
    },
    "eljuanina/VotingBooklets-Diamond-v1": {
      "url": "https://huggingface.co/datasets/eljuanina/VotingBooklets-Diamond-v1",
      "revision": "476c96e74a72b65122cf6fc01943be125f7f3324"
    }
  },
  "processed_file_sha256": {
    "train.jsonl": "d84c3dc14244c746deed98cee9cd5332e16f483b3cab039a53fe7c8c994dfcaf",
    "train_moe.jsonl": "287bee32718f74d333f48e94c38e5150ed70921f47e15871940c83ce7e13d22a",
    "eval_mono.jsonl": "b974eebcc3b3c68b443eec04315ab96ff3d5c4cf7fc0a54785fc50e1597ecc27",
    "eval_cross.jsonl": "82ba1f01c9099a39b30bec0b33dbc48b5bf6998973f45387d389a8a951144f19",
    "data_stats.json": "ee05572b590907e92ba753799994a8914d53dbb1962e16482ab3c42b78dd2e30"
  },
  "train_triplets_identical_between_dense_and_moe": true
}