{
  "format": "mlx_lm text jsonl",
  "target_tool_format": "native_lfm",
  "base_templates": 19,
  "copies": 24,
  "splits": {
    "train": 392,
    "valid": 32,
    "test": 32
  }
}