sinh khối

Tài liệu tham khảo:

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

ds = tfds.load('huggingface:biosses')
  • Sự miêu tả :
BIOSSES is a benchmark dataset for biomedical sentence similarity estimation. The dataset comprises 100 sentence pairs, in which each sentence was selected from the TAC (Text Analysis Conference) Biomedical Summarization Track Training Dataset containing articles from the biomedical domain. The sentence pairs were evaluated by five different human experts that judged their similarity and gave scores ranging from 0 (no relation) to 4 (equivalent).
  • Giấy phép : BIOSSES được cung cấp theo các điều khoản của Giấy phép Công cộng Chung GNU v.3.0.

  • Phiên bản : 0.0.0

  • Chia tách :

Tách ra Ví dụ
'train' 100
  • Đặc trưng :
{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}