{
  "actual_vocab_size": 49152,
  "adapter_kind": "hf_json",
  "benchmark_config_sha256": "e87516aca76ba2ceb43e82d1a2e55705c6218a4b1d347e59f3f1d0881edc94f8",
  "candidate": "GollemTokenizer",
  "implementation": {
    "algorithm": "GollemTokenizer byte-level BPE",
    "byte_fallback": true,
    "decoder": "ByteLevel",
    "pretokenizer": "ByteLevel(add_prefix_space=False)",
    "recipe": "candidates/external/GollemTokenizer/train_tokenizer.py"
  },
  "input": {
    "dataset": "SlayerLab/minimal-en-corpus-5b",
    "document_order": "manifest shard order, then Parquet row order",
    "documents": 275265,
    "manifest_sha256": "315152f9ae75541459f1dc568250b082944ee086747f8e3814d3c566779e9a71",
    "max_documents": null,
    "normalization": "identity",
    "revision": "870948fee236859fd9f72f7057f39252dceb697f",
    "split": "train",
    "text_field": "text",
    "training_segmentation": {
      "boundary_preference": [
        "newline",
        "ASCII space",
        "UTF-8 code point"
      ],
      "lossless_within_document": true,
      "max_observed_utf8_bytes": 4096,
      "max_utf8_bytes": 4096,
      "segments": 500377,
      "sha256": "00af3de0ef9b72a2f82949d10321eaa1a80f49b19303419857f78a51ad5aa50d"
    },
    "utf8_bytes": 1355918775
  },
  "mode": "trained",
  "model_files": [
    "tokenizer.json"
  ],
  "model_sha256": {
    "tokenizer.json": "cfa7178d8f85e00460aeaf0bd9901cf5f2e2b6da8bebd08ea805f99c059593ca"
  },
  "requested_vocab_size": 49152,
  "schema_version": 1,
  "source_revision": "ca4fdafbe2ee5049361fe7bacbd9bfd8c331f9e2",
  "special_tokens": {
    "<|endoftext|>": 0,
    "<|im_end|>": 2,
    "<|im_start|>": 1
  },
  "training_resources": {
    "cpu_system_seconds": 320.707631,
    "cpu_user_seconds": 717.464117,
    "peak_rss_bytes": 5778391040,
    "utf8_bytes_per_second": 23138712.459609956,
    "wall_seconds": 58.599577542045154
  },
  "training_seconds": 58.599577542045154
}