test/dataset.json

{
  "pad_id": 0,
  "ignore_label_id": -100,
  "blank_identifier_id": 0,
  "vocab_size": 389,
  "seq_len": 512,
  "num_puzzle_identifiers": 256,
  "total_groups": 7772,
  "mean_puzzle_examples": 41.17382913021101,
  "total_puzzles": 320003,
  "sets": [
    "all"
  ]
}