data/manifest.json

{
  "total_tokens": 4000008192,
  "total_bytes": 8000016384,
  "format": "uint16",
  "tokenizer": "akpsahan/s8",
  "vocab_size": 3200,
  "datasets": {
    "math": {
      "repo": "open-web-math/open-web-math",
      "tokens": 4000000000,
      "budget": 4000000000,
      "coverage": 100.0
    },
    "ot": {
      "repo": "akpsahan/ot",
      "tokens": 8192,
      "budget": 8192,
      "coverage": 100.0
    }
  }
}