reports/encoding.json

{
  "model": "Qwen/Qwen3.5-4B",
  "revision": "851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a",
  "truncated_rows": 0,
  "splits": {
    "train": {
      "rows": 87621,
      "tokens": 98928078,
      "max_tokens": 11673
    },
    "validation": {
      "rows": 11278,
      "tokens": 13647382,
      "max_tokens": 8543
    },
    "calibration": {
      "rows": 2968,
      "tokens": 3282800,
      "max_tokens": 7911
    },
    "test": {
      "rows": 7446,
      "tokens": 4300500,
      "max_tokens": 9872
    }
  }
}