qangaroo

Tài liệu tham khảo:

y khoa

Sử dụng lệnh sau để tải tập dữ liệu này trong TFDS:

ds = tfds.load('huggingface:qangaroo/medhop')
  • Sự miêu tả :
We have created two new Reading Comprehension datasets focussing on multi-hop (alias multi-step) inference.

Several pieces of information often jointly imply another fact. In multi-hop inference, a new fact is derived by combining facts via a chain of multiple steps.

Our aim is to build Reading Comprehension methods that perform multi-hop inference on text, where individual facts are spread out across different documents.

The two QAngaroo datasets provide a training and evaluation resource for such methods.
  • Giấy phép : Không có giấy phép được biết đến
  • Phiên bản : 1.0.0
  • Chia tách :
Tách ra Ví dụ
'train' 1620
'validation' 342
  • Đặc trưng :
{
    "query": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "supports": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "candidates": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "answer": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

đeo mặt nạ_medhop

Sử dụng lệnh sau để tải tập dữ liệu này trong TFDS:

ds = tfds.load('huggingface:qangaroo/masked_medhop')
  • Sự miêu tả :
We have created two new Reading Comprehension datasets focussing on multi-hop (alias multi-step) inference.

Several pieces of information often jointly imply another fact. In multi-hop inference, a new fact is derived by combining facts via a chain of multiple steps.

Our aim is to build Reading Comprehension methods that perform multi-hop inference on text, where individual facts are spread out across different documents.

The two QAngaroo datasets provide a training and evaluation resource for such methods.
  • Giấy phép : Không có giấy phép được biết đến
  • Phiên bản : 1.0.0
  • Chia tách :
Tách ra Ví dụ
'train' 1620
'validation' 342
  • Đặc trưng :
{
    "query": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "supports": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "candidates": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "answer": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

wikihop

Sử dụng lệnh sau để tải tập dữ liệu này trong TFDS:

ds = tfds.load('huggingface:qangaroo/wikihop')
  • Sự miêu tả :
We have created two new Reading Comprehension datasets focussing on multi-hop (alias multi-step) inference.

Several pieces of information often jointly imply another fact. In multi-hop inference, a new fact is derived by combining facts via a chain of multiple steps.

Our aim is to build Reading Comprehension methods that perform multi-hop inference on text, where individual facts are spread out across different documents.

The two QAngaroo datasets provide a training and evaluation resource for such methods.
  • Giấy phép : Không có giấy phép được biết đến
  • Phiên bản : 1.0.0
  • Chia tách :
Tách ra Ví dụ
'train' 43738
'validation' 5129
  • Đặc trưng :
{
    "query": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "supports": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "candidates": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "answer": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

đeo mặt nạ_wikihop

Sử dụng lệnh sau để tải tập dữ liệu này trong TFDS:

ds = tfds.load('huggingface:qangaroo/masked_wikihop')
  • Sự miêu tả :
We have created two new Reading Comprehension datasets focussing on multi-hop (alias multi-step) inference.

Several pieces of information often jointly imply another fact. In multi-hop inference, a new fact is derived by combining facts via a chain of multiple steps.

Our aim is to build Reading Comprehension methods that perform multi-hop inference on text, where individual facts are spread out across different documents.

The two QAngaroo datasets provide a training and evaluation resource for such methods.
  • Giấy phép : Không có giấy phép được biết đến
  • Phiên bản : 1.0.0
  • Chia tách :
Tách ra Ví dụ
'train' 43738
'validation' 5129
  • Đặc trưng :
{
    "query": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "supports": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "candidates": {
        "feature": {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    },
    "answer": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}