--- license: cc-by-sa-4.0 pretty_name: tasksource/mtop (mirror) source_datasets: - tasksource/mtop tags: - laya - laya-finetune - mirror - domain-general - use-conditional - genre-other - label-text - label-multiclass - annot-unknown - lang-multi task_categories: - text-classification --- # tasksource/mtop(複製) ## データの内容 MTOP(Li et al., 2020、多言語のタスク指向の意味解析のベンチマーク)に対応するデータ。1 行は質問文(`question`)と意味表現(`logical_form`)の組で、`intent`、`domain`、`lang`、`spans`、`tokenized_question` の列を持つ。元のカードは論文の引用だけで、この説明は列名と引用から書いた。 ## データの概要 | 項目 | 値 | |---|---| | 行数 | 104,445(1 構成・3 分割の合計) | | 列数 | 8 | | 容量 | 15.77 MB(ファイル 5 本) | | ファイルの形式 | parquet 3 | | 公開 | public | ### 構成と分割 | 構成(config) | 分割(split)ごとの行数 | 列数 | |---|---|---:| | `default` | `train` 73,928、`validation` 10,837、`test` 19,680 | 8 | ### 列(構成 `default`・分割 `train`) | 列 | 型 | 先頭の行の値 | |---|---|---| | `idx` | `string` | `3232353834303431` | | `intent` | `string` | `IN:SEND_MESSAGE` | | `spans` | `string` | `25:40:SL:GROUP` | | `question` | `string` | `Antworte im Chat mit der Kirchenführung` | | `domain` | `string` | `messaging` | | `lang` | `string` | `de_XX` | | `logical_form` | `string` | `[IN:SEND_MESSAGE [SL:GROUP Kirchenführung ] ]` | | `tokenized_question` | `string` | `{"tokens":["Antworte","im","Chat","mit","der","K…` | 行数と列は HF のデータセットビューアの集計(元のデータセットの現在の版で数えたもの)。全データセットの一覧は [HYUKI5/dataset-catalog](https://huggingface.co/datasets/HYUKI5/dataset-catalog) にある。 ## 複製について [`tasksource/mtop`](https://huggingface.co/datasets/tasksource/mtop) をリビジョン `9fa95557ba2809ad05943880eee5231b4f2463fb` の時点でそのまま複製したもの。中身は変えていない。判断モデル Laya(`convaiinnovations/laya`)の微調整に使う候補として、調査メモ(`Laya 学習用データセット調査`)に載せたデータを集めている。 ## 出典 - 元のデータセット: https://huggingface.co/datasets/tasksource/mtop - 複製したリビジョン: `9fa95557ba2809ad05943880eee5231b4f2463fb`(2026-09-23 10:37:35+00:00) - 元のライセンス表記: `cc-by-sa-4.0` ## 論文 - https://arxiv.org/abs/2008.09335 ## アノテーションの方法 元のカードにアノテーションの節が無かった。下の「調査メモの記述」と「元のカード」を参照。 ## 微調整に使うときの注意 - 調査メモの台帳での区分: B(継承条件・gated・モデル由来の条件・上流の規約などが付く)。区分は記載の文言で機械的に付けたもので、法的な判定ではない。学習に入れる前に元のカードとライセンスの原文で確かめる。 - Laya の入力は 2,048 token まで(ModernBERT-large / mmBERT)。長い本文は先頭から黙って切られるので、切れる行は事前に除くか、要約・切り出しをしてから入れる。 - 分割は記事・文書単位で行い、同じ本文から作った別の問いを学習側と較正側に分けない。温度と閾値は学習に使っていない分割で別々に当てる。 ## 調査メモの記述 | メモ | 節 | ライセンス(記載) | 確認 | 注記 | |---|---|---|---|---| | `jev_sources_tasksource_bekko.csv` | CSV(tasksource-jev-typed-decisions) | cc-by-sa-4.0 / license_use=commercial | 確認済 | license_use=commercial / [Jev 元データ source] source_annotation_converted(deterministic recast; no model r | | `jev_sources_tasksource_bekko.csv` | CSV(bekko-system-one-dataset-v0) | CC BY-SA 4.0 | 確認済 | [Jev 元データ source] source_annotation_converted(Choice/Noul/Score recast; soft t | ## 元のカード(日本語訳) 以下は元のデータセットのカードの本文を日本語に訳したもの(機械翻訳)。訳に疑問があるときは、末尾の原文を確かめる。 --- https://arxiv.org/pdf/2008.09335.pdf ``` @article{li2020mtop, title={MTOP: A comprehensive multilingual task-oriented semantic parsing benchmark}, author={Li, Haoran and Arora, Abhinav and Chen, Shuohui and Gupta, Anchit and Gupta, Sonal and Mehdad, Yashar}, journal={arXiv preprint arXiv:2008.09335}, year={2020} } ```
原文(英語) https://arxiv.org/pdf/2008.09335.pdf ``` @article{li2020mtop, title={MTOP: A comprehensive multilingual task-oriented semantic parsing benchmark}, author={Li, Haoran and Arora, Abhinav and Chen, Shuohui and Gupta, Anchit and Gupta, Sonal and Mehdad, Yashar}, journal={arXiv preprint arXiv:2008.09335}, year={2020} } ```