tokenizer_training_metadata.json

{
  "timestamp_utc": "2026-02-16T20:50:18.532387+00:00",
  "language": "pol_Latn",
  "track": "paper_spm32k",
  "training_data": "/netscratch/nrauscher/projects/BA-hydra/cross_lingual_transfer_multilingual/data/languages/pol_Latn/splits/sub/sub_charcap43gb_seed42/train/docs.parquet",
  "training_data_rows": 121766,
  "subsplit": "sub_charcap43gb_seed42",
  "sentencepiece": {
    "model_type": "bpe",
    "vocab_size": 32000,
    "character_coverage": 1.0,
    "byte_fallback": true,
    "hard_vocab_limit": false,
    "input_sentence_size": 0,
    "shuffle_input_sentence": true,
    "num_threads": 32,
    "pad_id": 0,
    "eos_id": 1,
    "unk_id": 2,
    "bos_id": -1
  },
  "t5": {
    "extra_ids": 100,
    "tokenizer_length": 32100,
    "fast_tokenizer_export_ok": false
  },
  "duration_sec": 63.32048845291138
}