babi_qa

参考文献:

• コード
• ハグフェイス

en-qa1

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:babi_qa/en-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa2

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:babi_qa/en-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa3

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:babi_qa/en-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa4

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:babi_qa/en-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa5

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:babi_qa/en-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa6

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:babi_qa/en-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa7

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:babi_qa/en-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa8

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:babi_qa/en-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa9

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:babi_qa/en-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa10

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:babi_qa/en-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa11

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:babi_qa/en-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa12

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:babi_qa/en-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa13

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:babi_qa/en-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa14

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:babi_qa/en-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa15

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:babi_qa/en-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa16

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:babi_qa/en-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa17

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:babi_qa/en-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	125

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa18

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:babi_qa/en-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	199
'train'	198

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa19

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:babi_qa/en-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-qa20

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:babi_qa/en-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	93
'train'	94

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa1

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:babi_qa/hn-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa2

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:babi_qa/hn-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa3

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:babi_qa/hn-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	167
'train'	167

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa4

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:babi_qa/hn-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa5

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:babi_qa/hn-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa6

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:babi_qa/hn-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa7

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:babi_qa/hn-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa8

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:babi_qa/hn-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa9

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:babi_qa/hn-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa10

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:babi_qa/hn-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa11

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:babi_qa/hn-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa12

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:babi_qa/hn-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa13

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:babi_qa/hn-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	125

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa14

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:babi_qa/hn-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa15

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:babi_qa/hn-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa16

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:babi_qa/hn-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa17

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:babi_qa/hn-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	125

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa18

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:babi_qa/hn-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	198
'train'	198

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa19

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:babi_qa/hn-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-qa20

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:babi_qa/hn-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	94
'train'	93

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa1

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:babi_qa/en-10k-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa2

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:babi_qa/en-10k-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa3

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:babi_qa/en-10k-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa4

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:babi_qa/en-10k-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa5

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:babi_qa/en-10k-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa6

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:babi_qa/en-10k-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa7

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:babi_qa/en-10k-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa8

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:babi_qa/en-10k-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa9

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:babi_qa/en-10k-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa10

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:babi_qa/en-10k-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa11

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:babi_qa/en-10k-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa12

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:babi_qa/en-10k-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa13

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:babi_qa/en-10k-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa14

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:babi_qa/en-10k-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa15

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:babi_qa/en-10k-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	2500

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa16

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:babi_qa/en-10k-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa17

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:babi_qa/en-10k-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	1250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa18

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:babi_qa/en-10k-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	199
'train'	1978年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa19

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:babi_qa/en-10k-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-10k-qa20

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:babi_qa/en-10k-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	93
'train'	933

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効なqa1

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:babi_qa/en-valid-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa2

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:babi_qa/en-valid-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa3

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:babi_qa/en-valid-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa4

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:babi_qa/en-valid-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	900
'validation'	100

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa5

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:babi_qa/en-valid-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa6

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:babi_qa/en-valid-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa7

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:babi_qa/en-valid-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa8

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:babi_qa/en-valid-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa9

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:babi_qa/en-valid-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa10

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:babi_qa/en-valid-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa11

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:babi_qa/en-valid-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa12

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:babi_qa/en-valid-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa13

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:babi_qa/en-valid-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa14

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:babi_qa/en-valid-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	180
'validation'	20

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa15

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:babi_qa/en-valid-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	225
'validation'	25

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa16

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:babi_qa/en-valid-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	900
'validation'	100

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa17

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:babi_qa/en-valid-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	113
'validation'	12

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa18

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:babi_qa/en-valid-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	199
'train'	179
'validation'	19

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa19

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:babi_qa/en-valid-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	900
'validation'	100

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-有効-qa20

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:babi_qa/en-valid-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	93
'train'	85
'validation'	9

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa1

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:babi_qa/en-valid-10k-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa2

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:babi_qa/en-valid-10k-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa3

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:babi_qa/en-valid-10k-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa4

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:babi_qa/en-valid-10k-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	9000
'validation'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa5

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:babi_qa/en-valid-10k-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa6

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:babi_qa/en-valid-10k-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa7

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:babi_qa/en-valid-10k-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa8

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:babi_qa/en-valid-10k-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa9

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:babi_qa/en-valid-10k-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa10

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:babi_qa/en-valid-10k-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa11

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:babi_qa/en-valid-10k-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa12

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:babi_qa/en-valid-10k-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa13

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:babi_qa/en-valid-10k-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa14

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:babi_qa/en-valid-10k-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	1800
'validation'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa15

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:babi_qa/en-valid-10k-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	2250
'validation'	250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa16

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:babi_qa/en-valid-10k-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	9000
'validation'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa17

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:babi_qa/en-valid-10k-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	1125
'validation'	125

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa18

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:babi_qa/en-valid-10k-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	199
'train'	1781年
'validation'	197

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa19

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:babi_qa/en-valid-10k-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	9000
'validation'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

en-valid-10k-qa20

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:babi_qa/en-valid-10k-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	93
'train'	840
'validation'	93

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa1

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:babi_qa/hn-10k-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa2

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:babi_qa/hn-10k-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa3

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:babi_qa/hn-10k-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	167
'train'	1667年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa4

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:babi_qa/hn-10k-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa5

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:babi_qa/hn-10k-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa6

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:babi_qa/hn-10k-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa7

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:babi_qa/hn-10k-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa8

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:babi_qa/hn-10k-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa9

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:babi_qa/hn-10k-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa10

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:babi_qa/hn-10k-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa11

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:babi_qa/hn-10k-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa12

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:babi_qa/hn-10k-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa13

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:babi_qa/hn-10k-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	1250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa14

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:babi_qa/hn-10k-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa15

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:babi_qa/hn-10k-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	2500

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa16

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:babi_qa/hn-10k-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa17

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:babi_qa/hn-10k-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	1250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa18

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:babi_qa/hn-10k-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	198
'train'	1977年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa19

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:babi_qa/hn-10k-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

hn-10k-qa20

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:babi_qa/hn-10k-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	94
'train'	934

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-QA1

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:babi_qa/shuffled-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフルQA2

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:babi_qa/shuffled-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa3

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:babi_qa/shuffled-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa4

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:babi_qa/shuffled-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa5

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:babi_qa/shuffled-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa6

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:babi_qa/shuffled-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa7

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:babi_qa/shuffled-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa8

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:babi_qa/shuffled-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-QA9

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:babi_qa/shuffled-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa10

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:babi_qa/shuffled-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa11

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:babi_qa/shuffled-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa12

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:babi_qa/shuffled-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa13

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:babi_qa/shuffled-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa14

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:babi_qa/shuffled-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	200

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa15

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:babi_qa/shuffled-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa16

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:babi_qa/shuffled-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa17

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:babi_qa/shuffled-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	125

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa18

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:babi_qa/shuffled-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	199
'train'	198

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa19

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:babi_qa/shuffled-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	1000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-qa20

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:babi_qa/shuffled-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	93
'train'	94

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa1

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:babi_qa/shuffled-10k-qa1')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa2

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:babi_qa/shuffled-10k-qa2')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa3

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:babi_qa/shuffled-10k-qa3')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa4

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:babi_qa/shuffled-10k-qa4')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa5

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:babi_qa/shuffled-10k-qa5')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa6

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:babi_qa/shuffled-10k-qa6')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa7

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:babi_qa/shuffled-10k-qa7')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa8

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:babi_qa/shuffled-10k-qa8')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa9

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:babi_qa/shuffled-10k-qa9')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa10

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:babi_qa/shuffled-10k-qa10')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa11

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:babi_qa/shuffled-10k-qa11')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa12

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:babi_qa/shuffled-10k-qa12')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa13

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:babi_qa/shuffled-10k-qa13')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa14

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:babi_qa/shuffled-10k-qa14')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	200
'train'	2000年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa15

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:babi_qa/shuffled-10k-qa15')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	250
'train'	2500

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa16

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:babi_qa/shuffled-10k-qa16')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa17

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:babi_qa/shuffled-10k-qa17')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	125
'train'	1250

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa18

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:babi_qa/shuffled-10k-qa18')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	199
'train'	1978年

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa19

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:babi_qa/shuffled-10k-qa19')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	1000
'train'	10000

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

シャッフル-10k-qa20

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:babi_qa/shuffled-10k-qa20')

• 説明：

The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.

• ライセンス: クリエイティブ・コモンズ表示 3.0 ライセンス
• バージョン: 1.2.0
• 分割:

スプリット	例
'test'	93
'train'	933

• 特徴：

{
    "story": {
        "feature": {
            "id": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "type": {
                "num_classes": 2,
                "names": [
                    "context",
                    "question"
                ],
                "names_file": null,
                "id": null,
                "_type": "ClassLabel"
            },
            "text": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            },
            "supporting_ids": {
                "feature": {
                    "dtype": "string",
                    "id": null,
                    "_type": "Value"
                },
                "length": -1,
                "id": null,
                "_type": "Sequence"
            },
            "answer": {
                "dtype": "string",
                "id": null,
                "_type": "Value"
            }
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}