宝石

参考文献:

mlsum_de

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/mlsum_de')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_covid' 5058
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 10695
'train' 220748
'validation' 11392
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "topic": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "date": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

mlsum_es

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/mlsum_es')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_covid' 1938年
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 13366
'train' 259888
'validation' 9977
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "topic": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "date": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_es_en_v0

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_es_en_v0')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 19797
'train' 79515
'validation' 8835
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_ru_en_v0

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_ru_en_v0')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 9094
'train' 36898
'validation' 4100
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_tr_en_v0

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_tr_en_v0')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 808
'train' 3193
'validation' 355
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_vi_en_v0

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_vi_en_v0')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 2167
'train' 9206
'validation' 1023
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_arabic_ar

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_arabic_ar')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 5841
'train' 20441
'validation' 2919
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "ar",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "ar",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_chinese_zh

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_chinese_zh')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 3775
'train' 13211
'validation' 1886年
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "zh",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "zh",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_czech_cs

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_czech_cs')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 1438
'train' 5033
'validation' 718
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "cs",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "cs",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_dutch_nl

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_dutch_nl')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 6248
'train' 21866
'validation' 3123
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "nl",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "nl",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_english_en

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_english_en')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 28614
'train' 99020
'validation' 13823
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "en",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "en",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_french_fr

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_french_fr')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 12731
'train' 44556
'validation' 6364
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "fr",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "fr",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_german_de

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_german_de')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 11669
'train' 40839
'validation' 5833
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "de",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "de",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_hindi_hi

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_hindi_hi')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 1984年
'train' 6942
'validation' 991
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "hi",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "hi",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_indonesian_id

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_indonesian_id')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 9497
'train' 33237
'validation' 4747
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "id",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "id",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_italian_it

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_italian_it')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 10189
'train' 35661
'validation' 5093
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "it",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "it",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_japanese_ja

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_japanese_ja')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 2530
'train' 8853
'validation' 1264
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "ja",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "ja",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_korean_ko

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_korean_ko')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 2436
'train' 8524
'validation' 1216
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "ko",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "ko",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_portuguese_pt

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_portuguese_pt')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 16331
'train' 57159
'validation' 8165
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "pt",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "pt",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_russian_ru

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_russian_ru')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 10580
'train' 37028
'validation' 5288
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "ru",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "ru",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_spanish_es

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_spanish_es')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 22632
'train' 79212
'validation' 11316
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "es",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "es",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_thai_th

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_thai_th')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 2950
'train' 10325
'validation' 1475年
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "th",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "th",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_turkish_tr

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_turkish_tr')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 900
'train' 3148
'validation' 449
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "tr",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "tr",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

wiki_lingua_vietnamese_vi

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_lingua_vietnamese_vi')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 3917
'train' 13707
'validation' 1957年
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source_aligned": {
        "languages": [
            "vi",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "target_aligned": {
        "languages": [
            "vi",
            "en"
        ],
        "id": null,
        "_type": "Translation"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

xsum

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/xsum')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_backtranslation' 500
'challenge_test_bfp_02' 500
'challenge_test_bfp_05' 500
'challenge_test_covid' 401
'challenge_test_nopunc' 500
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 1166
'train' 23206
'validation' 1117
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "xsum_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "document": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

共通世代

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/common_gen')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_scramble' 500
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 1497年
'train' 67389
'validation' 993
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "concept_set_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "concepts": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

cs_レストラン

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/cs_restaurants')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_scramble' 500
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 842
'train' 3569
'validation' 781
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "dialog_act": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "dialog_act_delexicalized": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target_delexicalized": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

ダーツ

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/dart')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 5097
'train' 62659
'validation' 2768
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "dart_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "tripleset": [
        [
            {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        ]
    ],
    "subtree_was_extended": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "target_sources": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

e2e_nlg

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/e2e_nlg')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_scramble' 500
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 4693
'train' 33525
'validation' 4299
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "meaning_representation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

トット

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/totto')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_scramble' 500
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 7700
'train' 121153
'validation' 7700
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "totto_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "table_page_title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "table_webpage_url": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "table_section_title": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "table_section_text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "table": [
        [
            {
                "column_span": {
                    "dtype": "int32",
                    "id": null,
                    "_type": "Value"
                },
                "is_header": {
                    "dtype": "bool",
                    "id": null,
                    "_type": "Value"
                },
                "row_span": {
                    "dtype": "int32",
                    "id": null,
                    "_type": "Value"
                },
                "value": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                }
            }
        ]
    ],
    "highlighted_cells": [
        [
            {
                "dtype": "int32",
                "id": null,
                "_type": "Value"
            }
        ]
    ],
    "example_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_annotations": [
        {
            "original_sentence": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "sentence_after_deletion": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "sentence_after_ambiguity": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "final_sentence": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        }
    ],
    "overlap_subset": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

web_nlg_ja

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/web_nlg_en')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_numbers' 500
'challenge_test_scramble' 500
'challenge_train_sample' 502
'challenge_validation_sample' 499
'test' 1779年
'train' 35426
'validation' 1667年
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "input": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "category": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "webnlg_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

web_nlg_ru

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/web_nlg_ru')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_scramble' 500
'challenge_train_sample' 501
'challenge_validation_sample' 500
'test' 1102
'train' 14630
'validation' 790
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "input": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "category": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "webnlg_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

wiki_auto_asset_turk

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/wiki_auto_asset_turk')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_asset_backtranslation' 359
'challenge_test_asset_bfp02' 359
'challenge_test_asset_bfp05' 359
'challenge_test_asset_nopunc' 359
'challenge_test_turk_backtranslation' 359
'challenge_test_turk_bfp02' 359
'challenge_test_turk_bfp05' 359
'challenge_test_turk_nopunc' 359
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test_asset' 359
'test_turk' 359
'train' 483801
'validation' 20000
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "source": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

スキーマ_ガイド付きダイアログ

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:gem/schema_guided_dialog')
  • 説明
GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation,
both through human annotations and automated Metrics.

GEM aims to:
- measure NLG progress across 13 datasets spanning many NLG tasks and languages.
- provide an in-depth analysis of data and models presented via data statements and challenge sets.
- develop standards for evaluation of generated text using both automated and human metrics.

It is our goal to regularly update GEM and to encourage toward more inclusive practices in dataset development
by extending existing data or developing datasets for additional languages.
  • ライセンス: CC-BY-SA-4.0
  • バージョン: 1.1.0
  • 分割:
スプリット
'challenge_test_backtranslation' 500
'challenge_test_bfp02' 500
'challenge_test_bfp05' 500
'challenge_test_nopunc' 500
'challenge_test_scramble' 500
'challenge_train_sample' 500
'challenge_validation_sample' 500
'test' 10000
'train' 164982
'validation' 10000
  • 特徴
{
    "gem_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "gem_parent_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "dialog_acts": [
        {
            "act": {
                "num_classes": 18,
                "names": [
                    "AFFIRM",
                    "AFFIRM_INTENT",
                    "CONFIRM",
                    "GOODBYE",
                    "INFORM",
                    "INFORM_COUNT",
                    "INFORM_INTENT",
                    "NEGATE",
                    "NEGATE_INTENT",
                    "NOTIFY_FAILURE",
                    "NOTIFY_SUCCESS",
                    "OFFER",
                    "OFFER_INTENT",
                    "REQUEST",
                    "REQUEST_ALTS",
                    "REQ_MORE",
                    "SELECT",
                    "THANK_YOU"
                ],
                "id": null,
                "_type": "ClassLabel"
            },
            "slot": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "values": [
                {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                }
            ]
        }
    ],
    "context": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ],
    "dialog_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "service": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "turn_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "prompt": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "target": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "references": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}