tether.json

{
  "config_path": "dataset.json",
  "dataset": {
    "splits": {
      "test": {
        "column_names": [
          "head",
          "relation",
          "tail"
        ],
        "num_rows": 5133,
        "type": "Dataset"
      },
      "train": {
        "column_names": [
          "head",
          "relation",
          "tail"
        ],
        "num_rows": 20614279,
        "type": "Dataset"
      },
      "validation": {
        "column_names": [
          "head",
          "relation",
          "tail"
        ],
        "num_rows": 5163,
        "type": "Dataset"
      }
    },
    "type": "DatasetDict"
  },
  "dataset_card": "README.md",
  "dataset_id": "wikidata5m",
  "exported_paths": [
    "train.parquet",
    "validation.parquet",
    "test.parquet"
  ],
  "format": "parquet",
  "manifest": "tether.json",
  "max_shard_size": "1GB",
  "output_dir": ".",
  "profile": {
    "citation": "Wang et al. (2021), KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation.",
    "dataset_id": "wikidata5m",
    "default_repo_id": null,
    "description": "Knowledge graph triples with transductive and inductive splits.",
    "homepage": "https://deepgraphlearning.github.io/project/wikidata5m",
    "license": "cc0-1.0 (Wikidata)",
    "tags": [
      "knowledge-graph",
      "wikidata",
      "triples"
    ],
    "title": "Wikidata5M"
  },
  "repo_id": "ysginc/wikidata5m",
  "report": {
    "dataset_id": "wikidata5m",
    "dataset_type": "DatasetDict",
    "entity_file": "/home/jschultz/source/personal/llm-datasets/datasets/wikidata5m/data/wikidata5m_entity.txt",
    "extensions": [
      "request_fingerprint",
      "required_splits",
      "metadata_report"
    ],
    "name": null,
    "num_proc": null,
    "num_rows": {
      "test": 5133,
      "train": 20614279,
      "validation": 5163
    },
    "profile": {
      "citation": "Wang et al. (2021), KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation.",
      "dataset_id": "wikidata5m",
      "default_repo_id": null,
      "description": "Knowledge graph triples with transductive and inductive splits.",
      "homepage": "https://deepgraphlearning.github.io/project/wikidata5m",
      "license": "cc0-1.0 (Wikidata)",
      "tags": [
        "knowledge-graph",
        "wikidata",
        "triples"
      ],
      "title": "Wikidata5M"
    },
    "relation_file": "/home/jschultz/source/personal/llm-datasets/datasets/wikidata5m/data/wikidata5m_relation.txt",
    "request_fingerprint": "741467886f874c13e0b2a7aa59590c65755f3272e830c265c33ff63a056dae16",
    "revision": null,
    "source": "csv",
    "split": null,
    "streaming": false,
    "variant": "transductive"
  },
  "request": {
    "column_names": [
      "head",
      "relation",
      "tail"
    ],
    "data_files": {
      "test": "/home/jschultz/source/personal/llm-datasets/datasets/wikidata5m/data/wikidata5m_transductive_test.txt",
      "train": "/home/jschultz/source/personal/llm-datasets/datasets/wikidata5m/data/wikidata5m_transductive_train.txt",
      "validation": "/home/jschultz/source/personal/llm-datasets/datasets/wikidata5m/data/wikidata5m_transductive_valid.txt"
    },
    "delimiter": "\t",
    "path": "csv",
    "streaming": false
  },
  "standardization": {
    "rename_map": {
      "head": "subject",
      "relation": "predicate",
      "tail": "object"
    },
    "selected_columns": [],
    "split_aliases": {
      "dev": "validation",
      "val": "validation",
      "valid": "validation"
    }
  }
}