nsmc

Tài liệu tham khảo:

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

ds = tfds.load('huggingface:nsmc')
  • Sự miêu tả :
This is a movie review dataset in the Korean language. Reviews were scraped from Naver movies. The dataset construction is based on the method noted in Large movie review dataset from Maas et al., 2011.
  • Giấy phép : CC0 1.0 Universal (CC0 1.0)
  • Phiên bản : 1.1.0
  • Chia tách :
Tách ra Ví dụ
'test' 50000
'train' 150000
  • Đặc trưng :
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "document": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 2,
        "names": [
            "negative",
            "positive"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}