{
  "source_dataset": "HuggingFaceFW/fineweb-2",
  "revision": "af9c13333eb981300149d5ca60a8e9d659b276b9",
  "model_name": "CausalNLP/gpt2-ar-zh-ko-ja-120k",
  "languages": [
    "arb_Arab",
    "cmn_Hani",
    "jpn_Jpan",
    "kor_Hang"
  ],
  "skip_tokens_per_language": 200000000,
  "kept_tokens_per_language": 1000000,
  "context_size": 16,
  "total_tokens": 4000000,
  "language_manifests": [
    {
      "source_dataset": "HuggingFaceFW/fineweb-2",
      "revision": "af9c13333eb981300149d5ca60a8e9d659b276b9",
      "model_name": "CausalNLP/gpt2-ar-zh-ko-ja-120k",
      "language": "arb_Arab",
      "skip_tokens": 200000000,
      "kept_tokens": 1000000,
      "context_size": 16,
      "sequences": 62500,
      "raw_tokens_seen": 201008964,
      "documents_seen": 230293,
      "documents_contributing": 439,
      "parquet_file": "arb_Arab.parquet"
    },
    {
      "source_dataset": "HuggingFaceFW/fineweb-2",
      "revision": "af9c13333eb981300149d5ca60a8e9d659b276b9",
      "model_name": "CausalNLP/gpt2-ar-zh-ko-ja-120k",
      "language": "cmn_Hani",
      "skip_tokens": 200000000,
      "kept_tokens": 1000000,
      "context_size": 16,
      "sequences": 62500,
      "raw_tokens_seen": 201010759,
      "documents_seen": 235091,
      "documents_contributing": 1148,
      "parquet_file": "cmn_Hani.parquet"
    },
    {
      "source_dataset": "HuggingFaceFW/fineweb-2",
      "revision": "af9c13333eb981300149d5ca60a8e9d659b276b9",
      "model_name": "CausalNLP/gpt2-ar-zh-ko-ja-120k",
      "language": "jpn_Jpan",
      "skip_tokens": 200000000,
      "kept_tokens": 1000000,
      "context_size": 16,
      "sequences": 62500,
      "raw_tokens_seen": 201012199,
      "documents_seen": 305332,
      "documents_contributing": 1616,
      "parquet_file": "jpn_Jpan.parquet"
    },
    {
      "source_dataset": "HuggingFaceFW/fineweb-2",
      "revision": "af9c13333eb981300149d5ca60a8e9d659b276b9",
      "model_name": "CausalNLP/gpt2-ar-zh-ko-ja-120k",
      "language": "kor_Hang",
      "skip_tokens": 200000000,
      "kept_tokens": 1000000,
      "context_size": 16,
      "sequences": 62500,
      "raw_tokens_seen": 201011803,
      "documents_seen": 285489,
      "documents_contributing": 1455,
      "parquet_file": "kor_Hang.parquet"
    }
  ]
}