落とす
コレクションでコンテンツを整理
必要に応じて、コンテンツの保存と分類を行います。
既存の読解力ベンチマークでのシステム パフォーマンスが人間のパフォーマンスに近いかそれを上回っているため、テキストの段落を実際に読むシステムの能力を向上させる新しいハード データセットが必要です。 DROP は、クラウドソーシングされ、敵対的に作成された 96k の質問のベンチマークであり、システムは質問内の参照をおそらく複数の入力位置に解決し、それらに対して個別の操作 (加算、カウント、並べ替えなど) を実行する必要があります。これらの操作では、以前のデータセットで必要だったよりも、段落の内容をより包括的に理解する必要があります。
スプリット | 例 |
---|
'dev' | 9,536 |
'train' | 77,409 |
FeaturesDict({
'answer': Text(shape=(), dtype=string),
'passage': Text(shape=(), dtype=string),
'query_id': Text(shape=(), dtype=string),
'question': Text(shape=(), dtype=string),
'validated_answers': Sequence(Text(shape=(), dtype=string)),
})
特徴 | クラス | 形 | Dtype | 説明 |
---|
| 特徴辞書 | | | |
答え | 文章 | | ストリング | |
通路 | 文章 | | ストリング | |
query_id | 文章 | | ストリング | |
質問 | 文章 | | ストリング | |
validated_answers | シーケンス(テキスト) | (なし、) | ストリング | |
@inproceedings{Dua2019DROP,
author={Dheeru Dua and Yizhong Wang and Pradeep Dasigi and Gabriel Stanovsky and Sameer Singh and Matt Gardner},
title={ {DROP}: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs},
booktitle={Proc. of NAACL},
year={2019}
}
特に記載のない限り、このページのコンテンツはクリエイティブ・コモンズの表示 4.0 ライセンスにより使用許諾されます。コードサンプルは Apache 2.0 ライセンスにより使用許諾されます。詳しくは、Google Developers サイトのポリシーをご覧ください。Java は Oracle および関連会社の登録商標です。
最終更新日 2022-12-06 UTC。
[null,null,["最終更新日 2022-12-06 UTC。"],[],[],null,["# drop\n\n\u003cbr /\u003e\n\n- **Description**:\n\nWith system performance on existing reading comprehension benchmarks nearing or\nsurpassing human performance, we need a new, hard dataset that improves systems'\ncapabilities to actually read paragraphs of text. DROP is a crowdsourced,\nadversarially-created, 96k-question benchmark, in which a system must resolve\nreferences in a question, perhaps to multiple input positions, and perform\ndiscrete operations over them (such as addition, counting, or sorting). These\noperations require a much more comprehensive understanding of the content of\nparagraphs than what was necessary for prior datasets.\n\n- **Additional Documentation** :\n [Explore on Papers With Code\n north_east](https://paperswithcode.com/dataset/drop)\n\n- **Homepage** : \u003chttps://allennlp.org/drop\u003e\n\n- **Source code** :\n [`tfds.text.drop.Drop`](https://github.com/tensorflow/datasets/tree/master/tensorflow_datasets/text/drop/drop.py)\n\n- **Versions**:\n\n - `1.0.0`: Initial release.\n - **`2.0.0`** (default): Add all options for the answers.\n- **Download size** : `7.92 MiB`\n\n- **Dataset size** : `116.24 MiB`\n\n- **Auto-cached**\n ([documentation](https://www.tensorflow.org/datasets/performances#auto-caching)):\n Yes\n\n- **Splits**:\n\n| Split | Examples |\n|-----------|----------|\n| `'dev'` | 9,536 |\n| `'train'` | 77,409 |\n\n- **Feature structure**:\n\n FeaturesDict({\n 'answer': Text(shape=(), dtype=string),\n 'passage': Text(shape=(), dtype=string),\n 'query_id': Text(shape=(), dtype=string),\n 'question': Text(shape=(), dtype=string),\n 'validated_answers': Sequence(Text(shape=(), dtype=string)),\n })\n\n- **Feature documentation**:\n\n| Feature | Class | Shape | Dtype | Description |\n|-------------------|----------------|---------|--------|-------------|\n| | FeaturesDict | | | |\n| answer | Text | | string | |\n| passage | Text | | string | |\n| query_id | Text | | string | |\n| question | Text | | string | |\n| validated_answers | Sequence(Text) | (None,) | string | |\n\n- **Supervised keys** (See\n [`as_supervised` doc](https://www.tensorflow.org/datasets/api_docs/python/tfds/load#args)):\n `None`\n\n- **Figure**\n ([tfds.show_examples](https://www.tensorflow.org/datasets/api_docs/python/tfds/visualization/show_examples)):\n Not supported.\n\n- **Examples**\n ([tfds.as_dataframe](https://www.tensorflow.org/datasets/api_docs/python/tfds/as_dataframe)):\n\nDisplay examples... \n\n- **Citation**:\n\n @inproceedings{Dua2019DROP,\n author={Dheeru Dua and Yizhong Wang and Pradeep Dasigi and Gabriel Stanovsky and Sameer Singh and Matt Gardner},\n title={ {DROP}: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs},\n booktitle={Proc. of NAACL},\n year={2019}\n }"]]