{
  "schema_version": 1,
  "corpus": "primer-pretext-270k",
  "package": "text_only",
  "layout": "tasks",
  "rows": {
    "jigsaw/train.jsonl": 10035,
    "jigsaw/val.jsonl": 231
  },
  "tasks": {
    "jigsaw": {
      "train": 10035,
      "val": 231
    }
  },
  "image_records": 55017,
  "images_by_upstream": {
    "Matterport3D": 55017
  },
  "image_bytes_included": 0,
  "images_without_md5": 0,
  "note": "Rows whose images come from Matterport3D or Ego-Exo4D. Users obtain those datasets under their own terms; images.jsonl names every image (paths relative to this folder) and gives the size and MD5 of the file used in training.",
  "files": [
    {
      "path": "jigsaw/train.jsonl",
      "bytes": 32951881,
      "sha256": "c902d0296e38caf0b5a09c086aef5f3f1f9279ddac595b742cd8f221d66f4a39"
    },
    {
      "path": "jigsaw/val.jsonl",
      "bytes": 762653,
      "sha256": "979fcccf98085937264e5d4d7b2813d7650311053b6621b0569912e4bc7e0aba"
    },
    {
      "path": "images.jsonl",
      "bytes": 32656800,
      "sha256": "7f9c11b1bb21d6c6f842d5e0ef8c05ac19aa16af5e02cc00146e050584220c12"
    },
    {
      "path": "provenance.jsonl",
      "bytes": 3873401,
      "sha256": "d344d1f89e1ca7b9a4d33e5c8576aea4e3bff58bdcab66ca0d500485173ee1a2"
    }
  ]
}